4 5805688099241461709 PDF

Accessing the E-book edition
Using the VitalSource® ebook DOWNLOAD AND READ OFFLINE

Access to the VitalBookTM ebook accompanying this book is To use your ebook offline, download BookShelf to your PC,
via VitalSource® Bookshelf – an ebook reader which allows Mac, iOS device, Android device or Kindle Fire, and log in to
you to make and share notes and highlights on your ebooks your Bookshelf account to access your ebook:
and search across all of the ebooks that you hold on your
On your PC/Mac
VitalSource Bookshelf. You can access the ebook online or
Go to https://support.vitalsource.com/hc/en-us and
offline on your smartphone, tablet or PC/Mac and your notes
follow the instructions to download the free VitalSource
and highlights will automatically stay in sync no matter where
Bookshelf app to your PC or Mac and log into your
you make them.
Bookshelf account.
1. Create a VitalSource Bookshelf account at On your iPhone/iPod Touch/iPad

https://online.vitalsource.com/user/new or log into Download the free VitalSource Bookshelf App available
your existing account if you already have one. via the iTunes App Store and log into your Bookshelf
account. You can find more information at https://support.
2. Redeem the code provided in the panel vitalsource.com/hc/en-us/categories/200134217-
below to get online access to the ebook. Bookshelf-for-iOS
Log in to Bookshelf and select Redeem at the top right
of the screen. Enter the redemption code shown on the On your Android™ smartphone or tablet
scratch-off panel below in the Redeem Code pop-up and Download the free VitalSource Bookshelf App available
press Redeem. Once the code has been redeemed your via Google Play and log into your Bookshelf account. You can
ebook will download and appear in your library. find more information at https://support.vitalsource.com/
hc/en-us/categories/200139976-Bookshelf-for-Android-
and-Kindle-Fire
On your Kindle Fire

Download the free VitalSource Bookshelf App available
from Amazon and log into your Bookshelf account. You can
find more information at https://support.vitalsource.com/
hc/en-us/categories/200139976-Bookshelf-for-Android-
and-Kindle-Fire
N.B. The code in the scratch-off panel can only be used once.
When you have created a Bookshelf account and redeemed
No returns if this code has been revealed. the code you will be able to access the ebook online or offline
on your smartphone, tablet or PC/Mac.
SUPPORT
If you have any questions about downloading Bookshelf,
creating your account, or accessing and using your ebook
edition, please visit http://support.vitalsource.com/
SERIES IN MATERNAL-FETAL MEDICINE
Published in association with the
Journal of Maternal-Fetal & Neonatal Medicine
Edited by
Gian Carlo Di Renzo and Dev Maulik
Howard Carp, Recurrent Pregnancy Loss, ISBN 9780415421300

Vincenzo Berghella, Obstetric Evidence Based Guidelines,
ISBN 9780415701884
Vincenzo Berghella, Maternal-Fetal Evidence Based Guidelines,
ISBN 9780415432818
Moshe Hod, Lois Jovanovic, Gian Carlo Di Renzo, Alberto de Leiva,
Oded Langer, Textbook of Diabetes and Pregnancy, Second Edition,
ISBN 9780415426206
Simcha Yagel, Norman H. Silverman, Ulrich Gembruch,
Fetal Cardiology, Second Edition, ISBN 9780415432658
Fabio Facchinetti, Gustaaf A. Dekker, Dante Baronciani,
George Saade, Stillbirth: Understanding and Management,
ISBN 9780415473903
Vincenzo Berghella, Maternal–Fetal Evidence Based Guidelines,
Second Edition, ISBN 9781841848228
Vincenzo Berghella, Obstetric Evidence Based Guidelines, Second Edition,
ISBN 9781841848242
Howard Carp, Recurrent Pregnancy Loss: Causes, Controversies, and

Treatment, Second Edition, ISBN 9781482216141
Moshe Hod, Lois G. Jovanovic, Gian Carlo Di Renzo, Alberto De Leiva,
Oded Langer, Textbook of Diabetes and Pregnancy, Third Edition,
ISBN 9781482213607
Edited by
Moshe Hod MD
Director, Division of Maternal Fetal Medicine
Rabin Medical Center
Sackler Faculty of Medicine, Tel-Aviv University
Petah-Tiqva, Israel
Lois G. Jovanovic MD
Clinical Professor of Medicine, University of Southern California
Keck School of Medicine
Adjunct Professor of Biomolecular Science and Engineering
University of California at Santa Barbara
CEO and Chief Scientific Officer
Sansum Diabetes Research Institute, Santa Barbara, CA, USA
Gian Carlo Di Renzo MD PhD
Professor and Chairman
Department of Obstetrics and Gynecology
Director, Perinatal and Reproductive Medicine Center and Midwifery School, University Hospital
Perugia, Italy
Director, Permanent International and European School of Perinatal and Reproductive Medicine (PREIS)
Florence, Italy
Alberto de Leiva MD PhD
Professor of Medicine, Universitat Autònoma de Barcelona
Director, Department of Endocrinology, Diabetes and Nutrition
Hospital de la Santa Creu i Sant Pau
Principal Investigator, EDUAB-HSP, CIBER-BBN, ISCIII
Vice President and Scientific Director, Fundación DIABEM
Barcelona, Spain
Oded Langer MD PhD

Former Babcock Professor and Chairman
St. Luke’s–Roosevelt Hospital Center
University Hospital for Columbia University
New York, NY, USA
MATLAB® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks does not warrant the accuracy of the text or exercises in
this book. This book’s use or discussion of MATLAB® software or related products does not constitute endorsement or sponsorship by The MathWorks
of a particular pedagogical approach or particular use of the MATLAB® software.
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2016 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works

Version Date: 20160414
International Standard Book Number-13: 978-1-4822-1362-1 (eBook - PDF)
This book contains information obtained from authentic and highly regarded sources. While all reasonable efforts have been made to publish reliable
data and information, neither the author[s] nor the publisher can accept any legal responsibility or liability for any errors or omissions that may be made.
The publishers wish to make clear that any views or opinions expressed in this book by individual editors, authors or contributors are personal to them
and do not necessarily reflect the views/opinions of the publishers. The information or guidance contained in this book is intended for use by medical,
scientific or health-care professionals and is provided strictly as a supplement to the medical or other professional’s own judgement, their knowledge of
the patient’s medical history, relevant manufacturer’s instructions and the appropriate best practice guidelines. Because of the rapid advances in medi-
cal science, any information or advice on dosages, procedures or diagnoses should be independently verified. The reader is strongly urged to consult
the relevant national drug formulary and the drug companies’ and device or material manufacturers’ printed instructions, and their websites, before
administering or utilizing any of the drugs, devices or materials mentioned in this book. This book does not indicate whether a particular treatment is
appropriate or suitable for a particular individual. Ultimately it is the sole responsibility of the medical professional to make his or her own professional
judgements, so as to advise and treat patients appropriately. The authors and publishers have also attempted to trace the copyright holders of all material
reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material
has not been acknowledged please write and let us know so we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic,
mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or
retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact
the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides
licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment
has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation
without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
To the most important people in my life
My wife Zipi; my sons Roy, Elad, and Yotam; my parents Esther and Michael; my grandchildren Dan,
Guy, Noa, Carmel, and Dor; and their mothers Maya and Timi
For their tolerance, patience, and love—they made it all possible

Moshe Hod
This page intentionally left blank
Contents
Preface xi
Contributors xvii
1. Introduction: Merging the legacies and hypotheses—Maternal medicine meets fetal medicine 1
Moshe Hod, Kypros Nicolaides, Hamutal Meiri, and Nicky Lieberman
2. History of diabetic pregnancy 11
David R. Hadden
3. Metabolism in normal pregnancy 17
Emilio Herrera and Henar Ortega-Senovilla
4. Intermediary metabolism in pregnancies complicated by gestational diabetes 28
Bartolomé Bonet, María Bonet-Alavés, and Isabel Sánchez-Vera
5. Nutrient delivery and metabolism in the fetus 34
William W. Hay, Jr., Paul J. Rozance, Stephanie R. Wesolowski, and Laura D. Brown
6. Pathogenesis of gestational diabetes mellitus 49
Yariv Yogev
7. Autoimmunity in gestational diabetes mellitus 57
Alberto de Leiva, Dídac Mauricio, and Rosa Corcoy
8. Epidemiology of gestational diabetes mellitus 69
Yariv Yogev, Avi Ben Haroush, Moshe Hod, and Jeremy Oats
9. Genetics of diabetic pregnancy 78
Komal Bajaj and Susan J. Gross
10. Animal models in diabetes and pregnancy research 84
Catherine Yzydorczyk, Delphine Mitanchez, and Umberto Simeoni
11. Pathologic abnormalities of placental structure and function in diabetes 91
Rhonda Bentley-Lewis, Maria Rosaria Raspollini, and Drucilla Roberts
12. The great obstetric syndromes: The roots of disease 97
Rinat Gabbay-Benziv and Ahmet A. Baschat
13. Placental origins of diabesity and the origin of preeclampsia 100
Gernot Desoye and Berthold Huppertz
14. Diagnosis of gestational diabetes mellitus 110
Donald R. Coustan and Boyd E. Metzger
15. Cost-effectiveness of screening and management programs for gestational diabetes mellitus 119
Louise K. Weile, James G. Kahn, Elliot Marseille, and Nicolai Lohse
16. Changing health policy: From study to national policy 131
Ofra Kalter-Leibovici, Nicky Lieberman, Ronni Gamzu, and Moshe Hod
17. Ideal weight gain in diabetic pregnancy 136
Gerard H.A. Visser and Harold W. de Valk
18. Medical nutritional therapy for gestational diabetes mellitus 138
Lois Jovanovic
19. Pharmacologic treatment of gestational diabetes mellitus: When to start and what agent to use 147
Celeste P. Durnwald and Mark B. Landon
20. Gestational diabetes mellitus: The consequences of not treating 157
Oded Langer
vii
viii Contents
21. Gestational diabetes mellitus in multiple pregnancies 169

Matteo Andrea Bonomo and Angela Napoli
22. Glycemic goals in diabetic pregnancy and defining “good control”: Maternal and fetal perspective 179
Liran Hiersch and Yariv Yogev
23. Insulin therapy in pregnancy 187
Lois Jovanovic and John L. Kitzmiller
24. Use of oral hypoglycemic agents in pregnancy 200
Oded Langer
25. The drug dilemma of oral antidiabetic agents in pregnancy: Metformin 211
Yoel Toledano, Moshe Zloczower, and Nicky Lieberman
26. Facing noncommunicable diseases’ global epidemic: The battle of prevention starts
in utero—The FIGO challenge 219
Luis Cabero and Sabaratnam Arulkumaran
27. Links between maternal health and noncommunicable diseases 226
Anil Kapur
28. Diabetic pregnancy in the developing world 234
Eran Hadar, Eran Ashwal, and Moshe Hod
29. Managing diabetic pregnancy in China 242
Huixia Yang, Weiwei Zhu, and Rina Su
30. Gestational diabetes mellitus, obesity, and pregnancy outcomes 246
Harold David McIntyre, Marloes Dekker-Nitert, Helen Lorraine Graham Barrett,
and Leonie Kaye Callaway
31. Obesity versus glycemic control: Which contributes more to adverse pregnancy outcome? 253
Amir Aviram and Yariv Yogev
32. Pharmacological treatment for the obese gestational diabetes mellitus patient 259
Fiona C. Denison and Rebecca M. Reynolds
33. Role of exercise in reducing the risks of gestational diabetes mellitus in obese women 266
Raul Artal
34. Role of bariatric surgery in obese women planning pregnancy 273
Ron Charach and Eyal Sheiner
35. Fetal lung maturity 287
Gian Carlo Di Renzo, Giulia Babucci, and Graziano Clerici
36. Monitoring during the later stage of pregnancy and during labor: Glycemic considerations 299
Harold W. de Valk and Gerard H.A. Visser
37. Timing and mode of delivery 305
Salvatore Alberico and Gianpaolo Maso
38. Management of the macrosomic fetus 312
Federico Mecacci, Marianna Pina Rambaldi, and Giorgio Mello
39. Congenital malformations in diabetic pregnancy: Prevalence and types 315
Paul Merlob
40. Diabetic embryopathy in the preimplantation embryo 321
Asher Ornoy and Noa Bischitz
41. Postimplantation diabetic embryopathy 329
Ulf J. Eriksson and Parri Wentzel
42. Fetal malformations detected with magnetic resonance imaging in the diabetic mother 351
Tuangsit Wataganara
43. Continuous glucose monitoring in pregnancy 362
Marlon Pragnell and Aaron Kowalski
Contents ix
44. Insulin infusion pumps in pregnancy 368

Ilana Jaye Halperin and Denice S. Feig
45. Closed-loop insulin delivery in type 1 diabetes pregnancy 373
Zoe A. Stewart and Helen R. Murphy
46. Noninvasive glucose monitoring 381
Itai Ben-David and Pierre Singer
47. Reproduction and its impact on health and disease 391
Sara Ornaghi and Michael J. Paidas
48. Diabetes, pregnancy, and the developmental origins of health and disease 403
Gerard H.A. Visser and Mark A. Hanson
49. Interventions to improve pregnancy outcome in obese pregnancy: Implications for mother and child 408
Rahat Maitland and Lucilla Poston
50. Lifestyle interventions to reduce risk of diabetes among high-risk pregnant and postpartum women 415
Lisa Chasan-Taber
51. Can fetal macrosomia be predicted and prevented? 425
Maria Farren and Michael Turner
52. Hypoglycemia in diabetic pregnancy 432
Graziano Di Cianni, Cristina Lencioni, Emilia Lacaria, and Laura Russo
53. Hypertensive disorders and diabetic pregnancy 441
Jacob Bar, Moshe Hod, and Michal Kovo
54. Diabetic retinopathy 453
Nir Melamed and Moshe Hod
55. Diabetic nephropathy 466
Elisabeth R. Mathiesen, Lene Ringholm, and Peter Damm
56. Diabetic ketoacidosis 473
Annunziata Lapolla and Maria Grazia Dalfrà
57. Thyroid disease in pregnancy 479
Yoel Toledano and Gabriella Solomon
58. Quality of care for the woman with diabetes at pregnancy 489
Alberto de Leiva, Rosa Corcoy, Alejandra de Leiva-Pérez, and Eulàlia Brugués
59. Early pregnancy loss and perinatal mortality 502
Kinneret Tenenbaum-Gavish, Anat Shmuely, and Moshe Hod
60. Short-term implications of gestational diabetes mellitus: The neonate 512
Delphine Mitanchez, Catherine Yzydorczyk, and Umberto Simeoni
61. Long-term outcomes after gestational diabetes mellitus exposure in the offspring 519
62. Metabolomics and diabetic pregnancy 524
Angelica Dessì, Roberta Carboni, and Vassilios Fanos
63. Fetal growth restriction: Evidence-based clinical management 529
Eduard Gratacós and Francesc Figueras
Index 535
Preface
In 2014, the International Federation of Gynecology and the cost-effectiveness of the universal diagnosis approach.
Obstetrics (FIGO) embarked on a new gestational diabetes A summary of the main areas of focus is provided in the fol-
mellitus (GDM) initiative with the ambitious objectives of lowing, although we strongly suggest reading the original
(1) raising awareness of the links between hyperglycemia document, which is open access and includes references and
and poor maternal and fetal outcomes and to the future can be found at www.figo.org/figo-project-publications.
health risks to mother and offspring, and demanding a
clearly defined global health agenda to tackle this issue,
and (2) creating a consensus document that provides guid- Gestational diabetes mellitus
ance for testing, management, and care of women with
GDM regardless of resource setting and disseminating and Hyperglycemia is one of the most common medical condi-
encouraging its use. In order to develop such international tions women encounter during pregnancy, with an estimated
guidance, FIGO brought together a group of experts (Moshe one in six live births (16.8%) to women with some form of
Hod, Anil Kapur, David A. Sacks, Eran Hadar, Mukesh hyperglycemia in pregnancy. While 16% of these cases may
Agarwal, Gian Carlo Di Renzo, Luis Cabero Roura, Harold be due to diabetes in pregnancy (either preexisting diabetes—
David McIntyre, Jessica L. Morris, and Hema Divakar) to type 1 or type 2—which antedates pregnancy or is first iden-
develop a document to frame the issues around gestational tified during testing in the index pregnancy), the majority
diabetes and suggest key actions to address the health bur- (84%) is due to gestational diabetes mellitus (GDM).
den posed by it. The result—“The International Federation of The occurrence of GDM parallels the prevalence of
Gynecology and Obstetrics (FIGO) Initiative on gestational impaired glucose tolerance (IGT), obesity, and type 2 diabe-
diabetes mellitus: A pragmatic guide for diagnosis, manage- tes mellitus (T2DM) in a given population. These conditions
ment, and care”—was published in the International Journal are on the rise globally. Moreover, the age of onset of diabetes
of Gynecology and Obstetrics 131 (S3) (2015) S173–S211 and and pre-diabetes is declining while the age of childbearing is
launched at the FIGO World Congress in October 2015 in increasing. There is also an increase in the rate of overweight
Vancouver. and obese women of reproductive age; thus, more women
Despite challenges of providing guidance given the lim- entering pregnancy have risk factors that make them vulner-
ited high-quality evidence available, this guide outlines able to hyperglycemia during pregnancy.
current global standards for the testing, management, and GDM is associated with a higher incidence of maternal
care of women with GDM and provides pragmatic recom- morbidity, including cesarean deliveries, shoulder dystocia,
mendations, which, because of their level of acceptability, birth trauma, hypertensive disorders of pregnancy (includ-
feasibility, and ease of implementation, have the potential to ing pre-eclampsia), and subsequent development of T2DM.
produce a significant impact. Suggestions are provided for Perinatal and neonatal morbidities also increase; the latter
a variety of different regional and resource settings based include macrosomia, birth injury, hypoglycemia, polycy-
on their financial, human, and infrastructure resources, themia, and hyperbilirubinemia. Long-term sequelae in
as well as for research priorities to bridge the gap between offspring with in utero exposure to maternal hyperglyce-
current knowledge and evidence. In assessing the quality mia may include higher risks for obesity and diabetes later
of evidence and grading of the strength of recommenda- in life.
tions, the guide follows the terminology proposed by the In most parts of low- and middle-income countries
Grading of Recommendations, Assessment, Development (LMICs) (which contribute to over 85% of the annual global
and Evaluation (GRADE) Working Group in which strong deliveries), most women are either not screened or improp-
recommendations are numbered as 1 and conditional/weak erly screened for diabetes during pregnancy—despite these
recommendations are numbered 2 with the quality of sup- countries accounting for 80% of the global diabetes burden
porting evidence labeled from very low quality to high qual- and for 90% of all cases of maternal and perinatal deaths
ity of evidence. and poor pregnancy outcomes. In particular, eight LMICs—
The guidelines were extremely well received globally but India, China, Nigeria, Pakistan, Indonesia, Bangladesh,
it is the next phase that will be even more challenging for Brazil, and Mexico—account for 55% of the global live births
FIGO—the implementation of this extensive document via (70 million live births annually) and 55% of the global bur-
capacity building, education, and advocacy, as well through den of diabetes (209.5 million) and should be key targets
establishing a research network which will be able to provide for any focused strategy on addressing the global burden of
evidence on operational and clinical implementation of the GDM pregnancies. These countries have been identified as
guidelines and provide health economics evidence to support priority countries for all future GDM interventions.
xi
xii Preface
Given the interaction between hyperglycemia and poor be at high risk, and universal testing will have to be strictly
pregnancy outcomes, the role of in utero imprinting in implemented and measured to ensure that all women are
increasing the risk of diabetes and cardiometabolic disorders offered the test.
in the offspring of mothers with hyperglycemia in pregnancy, The diagnosis of diabetes in pregnancy as defined by the
and increasing maternal vulnerability to future diabetes and WHO criteria and the diagnosis of GDM should be made
cardiovascular disorders, there needs to be a greater global using a single-step 75 g OGTT as per the recommendation of
focus on preventing, screening, diagnosing, and managing the IADPSG (2010) and WHO (2013). FIGO suggests various
hyperglycemia in pregnancy. The relevance of GDM as a pri- alternatives based on resource settings in Table P.1.
ority for maternal health and its impact on the future burden
of noncommunicable diseases is no longer in doubt, but how
best to deal with the issue remains contentious, as there are Glucose measurement: Technical
many gaps in knowledge on how to prevent, diagnose, and
manage GDM to optimize care and outcome. These must be considerations in laboratory and
addressed through future research. point-of-care testing
Most glucose measurements in laboratories are performed
on serum or plasma. Faster laboratory turnaround time is
Diagnosing GDM one reason that plasma has become the gold standard for
glucose measurement. However, in most laboratory panels
Global healthcare organizations and professional bodies have
(i.e., the comprehensive metabolic panel), serum is the most
advocated a plethora of diverse algorithms for screening and
suitable sample for all other laboratory tests performed, and
diagnosis of GDM that have been criticized for lacking valida-
so a “panel” glucose is usually a serum glucose.
tion, inasmuch as they were developed based on tenuous data,
Ideally, for diagnosis of GDM, reliable test results should
the biased result of expert opinions, which were based on eco-
be based on venous plasma samples properly collected and
nomic considerations or were convenience oriented, thereby
transported prior to laboratory testing by an accredited labo-
creating confusion and uncertainty among care providers.
ratory. However, this ideal situation may not be present in
One underlying yet fundamental problem, as shown consis-
many primary care settings, particularly in the developing
tently by several studies, including the Hyperglycemia and
world where proper facility for collection, transport, storage,
Adverse Pregnancy Outcomes (HAPO) study, is that the risk
or testing may not exist. In this situation, FIGO recommends
of poor pregnancy outcomes associated with hyperglycemia
that it is acceptable to use a plasma calibrated handheld glu-
is continuous, with no clear inflection points.
cometer with properly stored test strips to measure plasma
It is therefore clear that any set of criteria for the diagno-
glucose. Regular calibration should be undertaken with
sis of GDM proposed will need to evolve from a consensus
standard test solutions (usually supplied by the glucose
approach, balancing risks and benefits in particular social,
meter manufacturer).
economic, and clinical contexts. In addition to different cut-
off values, the lack of consensus among different professional
bodies for an algorithm for screening and diagnosis of GDM
is perhaps an even larger problem. Management of hyperglycemia
Selective testing based on clinical risk factors for GDM during pregnancy
evolved from the view that in populations with a low risk
of GDM, subjecting all pregnant women to a laboratory test Fetal and maternal outcomes are directly correlated with
was not considered cost-effective. Variations in risk factors the degree of maternal glycemic control. The primary goal
have resulted in different approaches, generally with poor of treatment for pregnancies complicated by diabetes is to
sensitivity and specificity. The major problem of risk factor– ensure as close to normal outcome as possible for the mother
based screening is its high demand on healthcare providers and offspring by controlling maternal hyperglycemia. Since
with more complex protocols for testing, which result in fetal macrosomia is the most frequent complication of dia-
lower compliance by both patients and healthcare providers. betes, special effort should be directed toward its diagno-
Given the high rates of hyperglycemia in pregnancy in sis and prevention. Fetal assessment can be achieved by a
most populations and that selective testing based on known fetal kick count, biophysical profile, and cardiotocography
risk factors has poor sensitivity for detection of GDM in a (nonstress test).
given population, it seems appropriate to recommend uni- Maternal hyperglycemia and macrosomia are associated
versal rather than risk factor–based testing. This approach with increased risk of intrauterine fetal death and other
is strongly recommended by FIGO and is particularly rele- adverse outcomes. Therefore, induction of labor may be con-
vant to LMICs where 90% of all cases of GDM are found and sidered at 38−39 weeks, although there is no good-quality
ascertainment of risk factors is poor owing to low levels of evidence to support such an approach. Thus, some guide-
education and awareness and poor record keeping. In many lines suggest that a pregnancy with good glycemic control
of these countries, there is little justification for selective and a seemingly appropriate gestational-age fetus ought to
testing, as they also have ethnic populations considered to continue until 40−41 weeks. Given the significantly greater
Preface xiii
Table P.1 Options for diagnosis of GDM based on resource settings
Strategy
Who to test and
Setting when Diagnostic test Interpretationa Grade
Fully resourced settings All women at Measure FPG, RBG, or 1|⊕⊕⊕O
booking/first HbA1c to detect
trimester diabetes in pregnancy
24−28 weeks If negative: perform 75 g
2-hour OGTT
Fully resourced settings All women at Perform 75 g 2-hour 2|⊕OOO
serving ethnic booking/first OGTT to detect
populations at high trimester diabetes in pregnancy
riskb 24−28 weeks If negative: repeat 75 g
2-hour OGTT
Any setting (basic); All women Perform 75 g 2-hour 1|⊕⊕⊕O
particularly medium- between 24 OGTT
to low-resource and 28 weeks
settings serving
ethnic populations
at risk
Alternative strategies as currently used in specified countries
China: Medium- to All women at Measure FPG to detect >7.0 mmol/L or 2|⊕OOO
low-resource settings booking/first diabetes in pregnancy >126 mg/dL.
serving populations trimester FPG values between 5.6
at high risk and 6.9 mmol/L (100
and 125 mg/dL)
consider as GDM
24−28 weeks If negative: perform 75 g Value >5.1 mmol/L or 1|⊕⊕⊕O
2-hour OGTT >92 mg/dL diagnostic
Or of GDM
To reduce number of 2|⊕OOO
OGTTs measure FPG.
Only in women with
values between 4.5 and
5.0 mmol/L (81 and 90
mg/dL) perform 75 g
2-hour OGTT
Indian subcontinent: All women at Measure fasting or Reading between 7.8 and 2|⊕OOO
Medium- to booking/first nonfasting 2-hour value 11.0 mmol/L or 140
low-resource settings trimester after 75 g OGTT and 199 mg/dL
serving rural/ indicates GDM
semi-urban/urban 24−28 weeks If negative: repeat test
ethnic populations
at high risk
Latin America: All women at Measure FPG to detect >7.0 mmol/L or 2|⊕OOO
Medium- to booking/first diabetes in pregnancy >126 mg/dL.
low-resource settings trimester FPG values between 5.6
and 6.9 mmol/L (100
and 125 mg/dL)
consider as GDM
24−28 weeks If negative: perform 75 g 75 g 2-hour glucose value
2-hour OGTT >7.8 mmol/L or >140
mg/dL is diagnostic of
GDMd
(Continued)
xiv Preface
Table P.1 (Continued) Options for diagnosis of GDM based on resource settings
Strategy
Who to test and
Setting when Diagnostic test Interpretationa Grade
United Kingdom: All Selected women Perform 75 g 2-hour FPG of 100 mg/dL or
settings at booking/as OGTT 5.6 mmol/L or above or
soon as 2-hour plasma glucose
possiblee of 140 mg/dL or
7.8 mmol/L or above
24−28 weeks If negative: perform 75 g
is diagnosticg
2-hour OGTT
Offered also to
other women
with risk
factors for
GDMf
Abbreviations: FPG, fasting plasma glucose; RBG, random blood glucose; HbA1c, glycosylated hemoglobin; GDM, gestational
diabetes mellitus; OGTT, oral glucose tolerance test.
a Interpret as per IADPSG/WHO/IDF guidelines unless stated otherwise.
b Asians are at high risk of hyperglycemia during pregnancy, which may include previously undiagnosed diabetes. The propor-
tion of previously undiagnosed diabetes is highest in the youngest age group particularly among women. In Asian populations,
FPG and HbA1c have much lower sensitivity to diagnose diabetes than the 2-hour post-glucose value. In a study of 11 Asian
cohorts, more than half of the diabetic subjects had isolated postchallenge hyperglycemia. In a study in China, 46.6% of the
participants with undiagnosed diabetes (44.1% of the men and 50.2% of the women) had isolated increased 2-hour plasma
glucose levels after an OGTT. Therefore, the need to identify postprandial hyperglycemia seems especially relevant in Asian
populations.
c Diabetes in Pregnancy Study Group in India (DIPSI) Guideline.
d Latin America Study Group.
e Women with a past history of GDM or women with glycosuria of 2+ or above on one occasion or of 1+ or above on two or more
occasions (as detected by reagent strip testing during routine prenatal care in the current pregnancy).
f BMI above 30 (calculated as weight in kilograms divided by height in meters squared), previous macrosomic baby weighing
4.5 kg or above, family history of diabetes, first-degree relative with diabetes, minority ethnic family origin with a high prevalence
of diabetes.
g National Institute for Health and Care Excellence (NICE) Diabetes in pregnancy: management of diabetes and its complications
from preconception to the postnatal period. NICE guidelines [NG3]. Published February 2015. http://www.nice.org.uk/guidance/
ng3/evidence.
risk of shoulder dystocia at any birthweight above 3750 g for diabetes, hypertensive complications, stillbirth, and cesar-
babies of women with diabetes, consideration may be given ean delivery. Recommendations are given for weight gain
to elective cesarean delivery when the best estimate of fetal during pregnancy for women with diabetes. Nutritional
weight exceeds 4000 g. Recommendations are provided for therapy includes an individualized food plan to optimize
prenatal supervision, fetal growth assessment, fetal well- glycemic control. It should be based on personal and cul-
being surveillance, and timing and mode of delivery. tural eating habits, physical activity, blood glucose measure-
Blood glucose control can be evaluated in one of three ments, and the expected physiological effects of pregnancy
ways: glycosylated hemoglobin (HbA1c), self-monitoring of on the mother and her fetus. Nutritional intervention for
blood glucose, and continuous glucose monitoring. The rec- diabetes, specifically pregnancy complicated with diabetes,
ommendations for glucose monitoring in women with GDM is consistently considered a fundamental treatment modal-
are given in the document. Attempts must be made to achieve ity and is the first-line therapy for all women diagnosed with
glucose levels as close as possible to those seen in normal GDM. Women with GDM and DIP must receive practical
pregnancy. Elevated glucose values, specifically postpran- education that empowers them to choose the right quantity
dial glucose levels, are associated with adverse pregnancy and quality of food. Recommendations are given for nutri-
outcomes in patients with hyperglycemia in pregnancy. Data tion therapy in women with GDM, and for physical activity,
suggest that postprandial glucose levels are more closely which has been shown to have benefits.
associated with macrosomia than fasting glucose levels. No Management using pharmacological treatment may also
controlled study has, as yet, established the optimal plasma be required. In the short term, for women with GDM requir-
glucose level(s) to prevent increased fetal risk. Glycemic tar- ing drug treatment, glyburide is inferior to both insulin and
gets for women with GDM are provided. metformin, while metformin (plus insulin when required)
Overweight and obese women before pregnancy are at performs slightly better than insulin. Recommendations
an increased risk for pregnancy complications including for pharmacological treatment in women with GDM are
Preface xv
given. It is important to note that there is no long-term Preconception care

evidence on the safety of oral antidiabetic drugs (OADs).
The following insulins may be considered safe and effective Preconception care is a set of assessment measures and inter-
treatment during pregnancy: regular insulin, NPH, lispro, ventions undertaken prior to conception. These are aimed at
aspart, and detemir. identifying and modifying medical, behavioral, and social
risks to women’s health during pregnancy, which may pre-
vent or mitigate adverse pregnancy outcomes.
Postpartum management Pregnancies should be planned and maternal assessment
with possible interventions should occur prior to conception
The postpartum period is crucial, not only in terms of to improve pregnancy outcome and maternal health. This
addressing the immediate perinatal problems but also in the may not only improve immediate maternal, perinatal, and
long term for establishing the basis for early preventive health neonatal outcomes, but possibly may have long-term benefi-
for both mother and child, who are at a heightened risk for cial effects on both the mother and her baby, lasting well into
future obesity, metabolic syndrome, diabetes, hypertension, adulthood and impacting next-generation offspring, through
and cardiovascular disorders. epigenetic changes and intrauterine fetal programming. It is
Mothers with GDM and diabetes in pregnancy need to estimated that 30%−90% of women have at least one condi-
be encouraged and supported in initiating and maintaining tion or risk factor, such as anemia, undernutrition, obesity,
breastfeeding. Breastfeeding has been shown to be protec- diabetes, hypertension, and thyroid disorders, that may benefit
tive against the occurrence of infant and maternal compli- from an appropriate preconception intervention; however,
cations, including reduction in childhood obesity, T2DM, only 30%−50% of pregnancies are planned and receive proper
and even T1DM. Moreover, breastfeeding helps postpar- preconception care. The key challenges are increasing aware-
tum weight loss. Treatment with insulin or commonly used ness and acceptance of the concept of preconception counsel-
OADs, such as glyburide and metformin, is not a contrain- ing and increasing affordability and access to preconception
dication to breastfeeding as levels of OAD medications in services to women of reproductive age.
breast milk are negligible and do not cause hypoglycemia Universal preconception care, as a concept, is still a chal-
in the baby. lenge in most parts of the world, where a significant propor-
For all women diagnosed with hyperglycemia for the first tion of women do not have access to prenatal care or receive
time during pregnancy (GDM and DIP), the glycemic status only one or two prenatal visits, the concept of preconception
should be reevaluated with a 75 g oral OGTT at 6−12 weeks care is a far-off goal but envisaged as an intervention that
after delivery with diagnosis based on the WHO criteria for could dramatically change maternal and neonatal health
diabetes, impaired fasting glucose (IFG), and impaired glu- and outcomes. Screening for conditions such as malnutri-
cose tolerance (IGT) in the nonpregnant state. Women who tion, anemia, overweight and obesity, hypertension, diabetes,
do not have diabetes or pre-diabetes, according to these defi- and thyroid dysfunction may have a significant impact. For
nitions, are still at risk of progression to diabetes and other women with diabetes, preconception care is also cost-saving
cardiovascular problems and require ongoing surveillance and yet only half the women with diabetes undergo appropri-
according to local protocol. ate preconception glycemic control.
Irrespective of the glycemic status on early postpartum In summary, to address the issue of GDM, FIGO recom-
testing, it should be assumed that women with GDM have mends the following:
the same or a higher level of future risk of diabetes and
cardiovascular disease as people with prediabetes and they ●● Public health focus: There should be greater international
should be advised to maintain a healthy lifestyle with an attention paid to GDM and to the links between maternal
appropriate diet, regular exercise, and normal body weight. health and noncommunicable diseases on the sustainable
Furthermore, to ensure optimal health before attempting developmental goals agenda. Public health measures to
their next pregnancy, they should seek consultation with increase awareness, access, affordability, and acceptance
healthcare providers knowledgeable about diabetes preven- of preconception counseling, and prenatal and post-
tion. Progression to diabetes is more common in women natal services for women of reproductive age must be
with a history of GDM compared with those without GDM prioritized.
history, despite equivalent degrees of IGT at baseline. Both ●● Universal testing: All pregnant women should be tested
intensive lifestyle and metformin have been shown to be for hyperglycemia during pregnancy using a one-step
highly effective in delaying or preventing diabetes in women procedure and FIGO encourages all countries and its
with IGT and a history of GDM and lowering the risk of pro- member associations to adapt and promote strategies to
gression from GDM to T2DM. ensure this.
Obstetricians, family physicians, internists, pediatricians, ●● Criteria for diagnosis: The WHO criteria for diagnosis
and other healthcare providers must link postpartum follow- of diabetes mellitus in pregnancy and the WHO and the
up of a GDM mother with the child’s vaccination and routine International Association of Diabetes in Pregnancy Study
pediatric care program to ensure continued follow-up and Groups (IADPSG) criteria for diagnosis of GDM should
engagement of the high-risk mother−child pair. be used when possible. Keeping in mind the resource
xvi Preface
constraints in many low-resource countries, alternate effective treatment options for GDM during the second
strategies described in the document should also be con- and third trimesters.
sidered equally acceptable. ●● Postpartum follow-up and linkage to care: Following a
●● Diagnosis of GDM: Diagnosis should ideally be based on GDM pregnancy, the postpartum period provides an
laboratory results of venous plasma samples that are prop- important platform to initiate beneficial health practices
erly collected, transported, and tested. Given the resource for both mother and child to reduce the future burden of
constraints in many low-resource countries, it is accept- several noncommunicable diseases. Obstetricians must
able to use a plasma-calibrated handheld glucometer for establish links with family physicians, internists, pedia-
diagnostic purposes. tricians, and other healthcare providers to support post-
●● Management of GDM: Management should be in accor- partum follow-up of GDM mothers and their children.
dance with available national resources and infrastructure A follow-up program linked to the child’s vaccination and
even if the specific diagnostic and treatment protocols are regular health check-up visits provides an opportunity for
not supported by high-quality evidence, as this is prefer- continued engagement with the high-risk mother−child
able to no care at all. pair.
●● Lifestyle management: Nutrition counseling and physical ●● Future research: There should be greater international
activity should be the primary tools in the management research collaboration to address the knowledge gaps
of GDM. Women with GDM must receive practical nutri- to better understand the links between maternal health
tional education and counseling that will empower them and noncommunicable diseases. Evidence-based find-
to choose the right quantity and quality of food and level ings are urgently needed to provide best practice stan-
of physical activity. They should be advised repeatedly dards for testing, management, and care of women
during pregnancy to continue the same healthy lifestyle with GDM. Cost-effectiveness models must be used in
after delivery to reduce the risk of future obesity, T2DM, countries with specific burden of disease and resources
and cardiovascular diseases. to make the best choices for testing and management
●● Pharmacological management: If lifestyle modifica- of GDM.
tion alone fails to achieve glucose control, metformin,
glyburide, or insulin should be considered as safe and Moshe Hod
Contributors
Salvatore Alberico Ahmet A. Baschat Eulàlia Brugués

Unit of Obstetrical Pathology Department of Obstetrics, Gynecology and Fundación DIABEM
Institute for Maternal and Child Health Reproductive Sciences Barcelona, Spain
IRCCS “Burlo Garofolo” School of Medicine
Trieste, Italy University of Maryland Luis Cabero
Baltimore, Maryland Hospital Vall de Hebron
Raul Artal
Universitat Autónoma de Barcelona
Department of Obstetrics, Gynecology and Itai Ben-David Barcelona, Spain
Women’s Health General Intensive Care Department and
School of Medicine Institute for Nutrition Research
Saint Louis University Leonie Kaye Callaway
Beilinson Hospital
St. Louis, Missouri Department of Obstetric Medicine
Rabin Medical Center University of Queensland
Sabaratnam Arulkumaran Petah Tikva, Israel Brisbane, Queensland, Australia
St George’s University of London and
London, United Kingdom and
Sackler School of Medicine
Eran Ashwal Royal Brisbane and Women’s Hospital
Tel Aviv University
Rabin Medical Center Herston, Queensland, Australia
Tel Aviv, Israel
Helen Schneider Hospital for Women
Petah Tikva, Israel Roberta Carboni
Rhonda Bentley-Lewis
Neonatal Intensive Care Unit
and Harvard Medical School
Puericulture Institute and Neonatal Section
Massachusetts General Hospital
Sackler Faculty of Medicine Azienda Ospedaliera Universitaria
Boston, Massachusetts
Tel Aviv University University of Cagliari
Tel Aviv, Israel Cagliari, Italy
Noa Bischitz
Laboratory of Teratology
Amir Aviram
Hadassah Medical School Ron Charach
Hebrew University Department of Obstetrics and Gynecology
Israeli Ministry of Health Soroka University Medical Center
Petah Tikva, Israel
Jerusalem, Israel Ben-Gurion University of the Negev
and Beer Sheva, Israel
Bartolomé Bonet
Sackler Faculty of Medicine Department of Pediatrics Lisa Chasan-Taber
Tel Aviv University Universitat Illes Balears Division of Biostatistics and Epidemiology
Tel Aviv, Israel Illes Balears, Spain School of Public Health and Health Sciences
Giulia Babucci University of Massachusetts
and
Department of Gynecology Amherst, Massachusetts
and Servicio de Pediatría
Centre of Perinatal and Reproductive Hospital Can Misses Graziano Clerici
Medicine Ibiza, Spain Department of Gynecology
University of Perugia and
Perugia, Italy María Bonet-Alavés Centre of Perinatal and Reproductive
Universitat Illes Balears Medicine
Komal Bajaj Illes Balears, Spain University of Perugia
Perugia, Italy
Albert Einstein College of Medicine and
Bronx, New York
Servicio de Pediatría
Rosa Corcoy
Jacob Bar Hospital Can Misses
Universidad Autònoma de Barcelona
Department of Obstetrics & Gynecology Ibiza, Spain
and
Edith Wolfson Medical Center
Matteo Andrea Bonomo Diabetes Unit
Holon, Israel
Diabetes Unit Department of Endocrinology, Diabetes and
and Niguarda Ca’ Granda Hospital Nutrition
Milano, Italy Hospital de la Santa Creu i Sant Pau
Sackler Faculty of Medicine Barcelona, Spain
Tel Aviv University
Laura D. Brown
Tel Aviv, Israel
Department of Pediatrics Donald R. Coustan
Helen Lorraine Graham Barrett School of Medicine Division of Maternal-Fetal Medicine
Royal Brisbane and Women’s Hospital University of Colorado Brown University
Herston, Queensland, Australia Aurora, Colorado Providence, Rhode Island
xvii
xviii Contributors
Maria Grazia Dalfrà Graziano Di Cianni Rinat Gabbay-Benziv

DPT Medicine Diabetes and Metabolic Diseases Unit Helen Schneider Hospital for Women
UOC Diabetology and Dietetic Livorno Hospital Rabin Medical Center
Padova University Livorno, Italy Petah Tikva, Israel
Padova, Italy
Gian Carlo Di Renzo
Department of Obstetrics and Gynecology Ronni Gamzu
Peter Damm
and Tel Aviv Sourasky Medical Center
Faculty of Health and Medical Sciences
Perinatal and Reproductive Medicine Center Tel Aviv University
Center for Pregnant Women with Diabetes,
and Midwifery School Tel Aviv, Israel
Rigshospitalet
The Institute of Clinical Medicine University Hospital
and Perugia, Italy
Eduard Gratacós
Department of Obstetrics, Rigshospitalet
and Hospital Clinic i Hospital Sant Joan de Deu
University of Copenhagen
Universitat de Barcelona
Copenhagen, Denmark Permanent International and European
Barcelona, Spain
School of Perinatal and Reproductive
Medicine (PREIS)
Harold W. de Valk
Florence, Italy
Department Internal Medicine Susan J. Gross
University Medical Center Albert Einstein College of Medicine
Utrecht, the Netherlands Celeste P. Durnwald Bronx, New York
Division of Maternal Fetal Medicine and
Department of Obstetrics and Gynecology Natera, Inc.
Marloes Dekker-Nitert University of Pennsylvania San Carlos, California
School of Medicine Philadelphia, Pennsylvania
Centre for Clinical Research
University of Queensland Eran Hadar
Brisbane, Queensland, Australia Ulf J. Eriksson
Department of Medical Cell Biology
Biomedical Center
Alberto de Leiva Petah Tikva, Israel
Uppsala University
Universitat Autònoma de Barcelona Uppsala, Sweden and
and
Department of Endocrinology, Diabetes and Sackler Faculty of Medicine
Nutrition Vassilios Fanos Tel Aviv University
Hospital de la Santa Creu i Sant Pau Neonatal Intensive Care Unit Tel Aviv, Israel
and Puericulture Institute and Neonatal Section
EDUAB-HSP, CIBER-BBN, ISCIII Azienda Ospedaliera Universitaria
and University of Cagliari David R. Hadden (deceased)
Fundación DIABEM Cagliari, Italy Regional Endocrinology and Diabetes
Barcelona, Spain Centre
Royal Victoria Hospital
Maria Farren Northern Ireland, United Kingdom
Alejandra de Leiva-Pérez UCD Centre for Human Reproduction
Fundación DIABEM Coombe Women and Infant’s University
and Hospital Ilana Jaye Halperin
Universitat Oberta de Catalunya Dublin, Ireland Division of Endocrinology and Metabolism
Barcelona, Spain Department of Medicine
Denice S. Feig Sunnybrook Health Sciences Centre
Department of Medicine, Obstetrics and University of Toronto
Fiona C. Denison
Gynecology Toronto, Ontario, Canada
Tommy’s Centre for Maternal and Fetal
Health and
MRC/University of Edinburgh Centre for Department of Health Policy, Management
and Evaluation Mark A. Hanson
Reproductive Health
University of Toronto Institute of Developmental Sciences
Queen’s Medical Research Institute
and Southampton General Hospital
Edinburgh, United Kingdom
Division of Endocrinology and Metabolism Southampton, United Kingdom
Mount Sinai Hospital
Gernot Desoye Toronto, Ontario, Canada
Department of Obstetrics and Gynaecology Avi Ben Haroush
Medical University of Graz Department of Obstetrics and Gynecology
Graz, Austria Francesc Figueras Helen Schneider Hospital for Women
Barcelona Center of Maternal-Fetal Medicine Rabin Medical Center
and Neonatology Petah Tikva, Israel
Angelica Dessì Hospital Clinic and Hospital Sant Joan de Deu
Neonatal Intensive Care Unit University of Barcelona
Puericulture Institute and Neonatal Section and William W. Hay, Jr.
Azienda Ospedaliera Universitaria Centre for Biomedical Research on Rare School of Medicine
University of Cagliari Diseases University of Colorado
Cagliari, Italy Barcelona, Spain Aurora, Colorado
Contributors xix
Emilio Herrera Michal Kovo Gianpaolo Maso

Faculties of Pharmacy and Medicine Department of Obstetrics and Gynecology Unit of Obstetrical Pathology
Universidad CEU San Pablo Edith Wolfson Medical Center Institute for Maternal and Child Health
Madrid, Spain Holon, Israel IRCCS “Burlo Garofolo”
Trieste, Italy
and
Liran Hiersch
Sackler Faculty of Medicine Elisabeth R. Mathiesen
Lis Hospital for Women
Tel Aviv University Faculty of Health and Medical Sciences
Tel Aviv Sourasky Medical Center
Tel Aviv, Israel Center for Pregnant Women with Diabetes,
Tel Aviv University
Petah Tikva, Israel Rigshospitalet
Aaron Kowalski The Institute of Clinical Medicine
Juvenile Diabetes Research Foundation and
Moshe Hod New York, New York Faculty of Health Sciences, Rigshospitalet
Department of Obstetrics and Gynecology Department of Endocrinology
Helen Schneider Hospital for Women Emilia Lacaria University of Copenhagen
Rabin Medical Center Diabetes and Metabolic Diseases Unit Copenhagen, Denmark
Petah Tikva, Israel Livorno Hospital
Livorno, Italy
and Dídac Mauricio
Sackler Faculty of Medicine Mark B. Landon Department of Endocrinology and Nutrition
Tel Aviv University Department of Obstetrics and Gynecology Hospital Germans Tries i Pujol
Tel Aviv, Israel College of Medicine Universitat Autònoma de Barcelona
The Ohio State University Badalona, Spain
Columbus, Ohio
Berthold Huppertz Harold David McIntyre
Institute of Cell Biology, Histology and Oded Langer University of Queensland
Embryology Department of Obstetrics and Gynecology St. Lucia, Queensland, Australia
Medical University of Graz St. Luke’s–Roosevelt Hospital Center
Graz, Austria and
and
University Hospital for Columbia University Mater Health Services
Lois Jovanovic New York, New York South Brisbane, Queensland. Australia
Keck School of Medicine
University of Southern California Annunziata Lapolla Federico Mecacci
Los Angeles, California DPT Medicine Obstetrical Pathology Department and High
UOC Diabetology and Dietetic Risk Pregnancy Unit
and Padova University University of Florence
University of California, Santa Barbara Padova, Italy Florence, Italy
and
Sansum Diabetes Research Institute Cristina Lencioni Hamutal Meiri
Santa Barbara, California Diabetes and Metabolic Diseases Unit ASPRE
Livorno Hospital Tel Aviv, Israel
Livorno, Italy
James G. Kahn
Philip R. Lee Institute for Health Policy Nir Melamed
Studies Nicky Lieberman Department of Obstetrics/Gynecology
Global Health Sciences Community Medicine Division Rabin Medical Center
and Clalit Health Services Petah Tikva, Israel
Global Health Economics Consortium Tel Aviv, Israel
University of California, San Francisco Giorgio Mello
San Francisco, California Nicolai Lohse Obstetrical Pathology Department and High
Department of Anesthesia Risk Pregnancy Unit
Copenhagen University Hospital, University of Florence
Ofra Kalter-Leibovici Rigshospitalet Florence, Italy
Gertner Institute for Epidemiology and Copenhagen, Denmark
Health Policy Research Paul Merlob
Ramat Gan, Israel Rahat Maitland
Department of Neonatology
Division of Women’s Health
Schneider Children Hospital
King’s College
Anil Kapur Petah Tikva, Israel
London, United Kingdom
World Diabetes Foundation
and
Gentofte, Denmark
Elliot Marseille Sackler School of Medicine
John L. Kitzmiller Health Strategies International Tel Aviv University
Good Samaritan Hospital Oakland, California Tel Aviv, Israel
San Jose, California
and
and Boyd E. Metzger
Global Health Economics Consortium Feinberg School of Medicine
Sansum Medical Research Institute University of California, San Francisco Northwestern University
Santa Barbara, California San Francisco, California Chicago, Illinois
xx Contributors
Delphine Mitanchez Michael J. Paidas Isabel Sánchez-Vera

Division of Neonatology Department of Obstetrics, Gynecology and School of Medicine
Department of Perinatology Reproductive Sciences University San Pablo
Armand Trousseau Hospital Yale University and
Sorbonne Universités New Haven, Connecticut School of Pharmacy
University Pierre et Marie Curie Urb Monteprincipe
Paris, France Lucilla Poston Boadilla del Monte
Department of Life Sciences and Medicine Madrid, Spain
Helen R. Murphy King’s College London
Wellcome Trust-MRC Institute of Metabolic London, United Kingdom Eyal Sheiner
Science Faculty of Health Sciences
University of Cambridge Metabolic Research Department of Obstetrics and Gynecology
Marlon Pragnell Soroka University Medical Center
Laboratories
Juvenile Diabetes Research Foundation Ben-Gurion University of the Negev
and
New York, New York Beer Sheva, Israel
NIHR Cambridge Biomedical Centre
Addenbrooke’s Hospital
Cambridge, United Kingdom Marianna Pina Rambaldi
Obstetrical Pathology Department and High Anat Shmuely
Risk Pregnancy Unit Rabin Medical Center
Angela Napoli Tel Aviv University
University of Florence
Department of Clinical and Molecular Petah Tikva, Israel
Florence, Italy
Medicine
Sapienza University of Rome
Roma, Italy Maria Rosaria Raspollini Umberto Simeoni
Division of Histology and Molecular Faculté de Pharmacie
Diagnostics Marseille-University
Kypros Nicolaides University of Florence
The Fetal Medicine Foundation Marseille, France
Florence, Italy
London, United Kingdom and
Rebecca M. Reynolds Division of Pediatrics
Jeremy Oats Endocrinology Unit and
Melbourne School of Population and UoE/BHF Centre for Cardiovascular DOHaD Research Unit
Global Health Science University of Lausanne
University of Melbourne Queen’s Medical Research Institute Lausanne, Switzerland
Burnley, Victoria, Australia Edinburgh, United Kingdom
Sara Ornaghi Lene Ringholm Pierre Singer

Department of Obstetrics and Gynecology Faculty of Health and Medical Sciences General Intensive Care Department and
University of Milano-Bicocca Center for Pregnant Women with Diabetes, Institute for Nutrition Research
Monza, Italy Rigshospitalet Beilinson Hospital
The Institute of Clinical Medicine Rabin Medical Center
and and Petah Tikva, Israel
Department of Obstetrics, Gynecology and Department of Endocrinology and
Reproductive Sciences University of Copenhagen Sackler School of Medicine
Yale Women and Children’s Center for Copenhagen, Denmark Tel Aviv University
Blood Disorders and Preeclampsia Tel Aviv, Israel
Advancement Drucilla Roberts
Yale University Department of Pathology
Harvard Medical School, Gabriella Solomon
New Haven, Connecticut
Massachusetts General Hospital Community Medical Division
Boston, Massachusetts Clalit Health Services
Asher Ornoy Tel Aviv, Israel
Laboratory of Teratology
Paul J. Rozance
Hadassah Medical School
Department of Pediatrics Zoe A. Stewart
Hebrew University
School of Medicine Welcome Trust-MRC Institute of Metabolic
Israeli Ministry of Health
University of Colorado Science
Jerusalem, Israel
Aurora, Colorado University of Cambridge Metabolic Research
Laboratories
Henar Ortega-Senovilla Laura Russo and
Faculties of Pharmacy and Medicine Diabetes and Metabolic Diseases Unit NIHR Cambridge Biomedical Research Centre
Universidad CEU San Pablo Livorno Hospital Addenbrooke’s Hospital
Madrid, Spain Livorno, Italy Cambridge, United Kingdom
Contributors xxi
Rina Su Tuangsit Wataganara Yariv Yogev

Department of Obstetrics and Gynecology Faculty of Medicine Siriraj Hospital Helen Schneider Hospital for Women
Peking University First Hospital Division of Maternal-Fetal Medicine Rabin Medical Center
Beijing, People’s Republic of China Department of Obstetrics and Gynecology Petah Tikva, Israel
Mahidol University
and
Kinneret Tenenbaum-Gavish Bangkok, Thailand
Rabin Medical Center Sackler Faculty of Medicine
Tel Aviv University Louise K. Weile Tel Aviv University
Petah Tikva, Israel Department of Gynaecology and Obstetrics Tel Aviv, Israel
Odense University Hospital
Odense, Denmark
Yoel Toledano
Division of Maternal Fetal Medicine Catherine Yzydorczyk
Helen Schneider Hospital for Women Parri Wentzel Faculté de Pharmacie
Rabin Medical Center Department of Medical Cell Biology Aix-Marseille University
Petah Tikva, Israel Biomedical Center Marseille, France
Uppsala University
Uppsala, Sweden
Michael Turner Weiwei Zhu
UCD Centre for Human Reproduction Stephanie R. Wesolowski National Institute of Hospital Administration
Coombe Women and Infant’s University Department of Pediatrics Beijing, People’s Republic of China
Hospital School of Medicine
Dublin, Ireland University of Colorado
Aurora, Colorado Moshe Zloczower
Gerard H.A. Visser Bruce Rappaport Faculty of Medicine
Department of Obstetrics Huixia Yang Endocrine Department and Diabetes in
Wilhelmina Children’s Hospital Department of Obstetrics and Gynecology Pregnancy Clinic
University Medical Center Peking University First Hospital Technion—Israel Institute of Technology
Utrecht, the Netherlands Beijing, People’s Republic of China Haifa, Israel
1 Introduction: Merging the
legacies and hypotheses—
Maternal medicine meets
fetal medicine
Moshe Hod, Kypros Nicolaides, Hamutal Meiri, and Nicky Lieberman
Such long-term effects on the health of mothers and their

Introduction children are mediated through epigenetic, physiological, endo-
This chapter was written in line with the need to revolution- crinological, and biochemical pathways and by imprinting of
ize maternal fetal medicine by returning the M (maternal) to responses to stress. They contribute to the increased postpreg-
the MFM subspecialty (maternal–fetal medicine) by introduc- nancy risk of developing noncommunicable diseases (NCDs)
ing a new paradigm of care composed of novel technologies that are passed on from one generation to the next through the
and comprehensive services in order to reduce maternal and critical period of pregnancy. A need for an all-encompassing
fetal morbidity and mortality. A three-floor service model is approach for improving maternal and fetal medicine is thus
introduced, composed of the prepregnancy clinic, the inverted required to interrupt the vicious cycle that starts at pregnancy
pyramid of antenatal care, and the postpregnancy clinic, each disorder in order to improve the maternal life in this genera-
with a combined methodology composed of existing and tion and the fetal life of the future generations. The holistic
novel testing procedures (such as preglycemic evaluation in approach is needed to merge the importance of maternal and
the prepregnancy clinic, free circulating DNA during the first fetal health in the MFM subspecialty. This chapter presents
trimester, or echocardiography of the newborn in the postna- a new three-floor holistic and multidisciplinary model for
tal service, among many others). All floors begin at the level of maternal and fetal medicine. The model’s first floor is pre-
community clinic/family obstetrician before the high-risk spe- pregnancy care and involves family planning and assessment
cialists are called to service and introduce a contingency man- of the prior risks for NCDs and their prepregnancy control
agement and prevention follow-up. The approach expresses and prevention. It continues through the introduction of the
the need to provide comprehensive service starting from a inverted pyramid of antenatal care that shifts the emphasis
traditional patient evaluation spanning medical and preg- from the third to the first trimester of pregnancy, offering a
nancy history and demography, biochemical and biophysical multidisciplinary screening and risk assessment followed by
markers, sonography, chemical blood tests, and introducing individually tailored prevention and management pathways.
“omics” to fetal medicine. This approach enables personal- The third floor is the postpregnancy health management to
ized medicine and a systematic method to focus the medical minimize long-term damage. This model of MFM care is pro-
attention on those who need it most, allowing the others to posed to improve maternal outcome and prevent short- and
have a less intensive medical involvement. It fits the new world long-term complications not only to the mothers but also to
of social media, computerized algorithms derived from mega their children and the coming generations.
databases, and the need to integrate all sources of information
and know-how to generate an evidence-based medical treat-
ment plan as required in today’s world of medicine. This chap- Bringing maternal–fetal medicine to
ter calls for introducing training and education to the new the new era of medicine
doctor generation and to systematically adjust the maternal–
fetal medicine (MFM) system of care in order to achieve the MFM was established a few decades ago as a multidisci-
required improvement in maternal and fetal health. plinary subspecialty dedicated to optimizing pregnancy
The first 9 months of life shapes the offspring’s adulthood and perinatal outcomes. The MFM subspecialty emerged
while simultaneously impacting maternal life after pregnancy. from the need to combine diagnosis and treatment of both
1
2 Introduction
the mother and her fetus in cases of high-risk pregnancies. and for pregnancy at an advanced maternal age.8 Finally,
The rapid advance in sonography, the introduction of MRI, there is a continuing trend across the developed world to
and the leap jump of measuring fetal DNA in maternal blood postpone family planning and attempt to conceive at an
have shifted the MFM emphasis to fetal medicine. This prog- advanced maternal age, with Italy leading the convoy with
ress occurs in parallel to increased maternal morbidity and 34.9% of women delivering at age 35 and older as reported
failures in the attempts to decrease maternal mortality rates by the EUROPRESTAT project evaluating pregnancy sta-
over the last few decades.1–3 Thus, there is a need to reem- tistics of 39 countries in Europe (http://www.europeristat.
phasize the maternal component (M) in the MFM specialty com/reports/european-perinatal-health-report-2010.htm).
by means of introducing new concepts and advanced diag- While this may in turn be a source of organ regeneration and
nostic procedures and policies coupled with changes in edu- extension of maternal longevity,9 it is also associated with
cation and training, in order to implement improvements increased risk for preeclampsia, intra uterine growth restric-
in healthcare services for pregnant women and enable the tion (IUGR), and preterm birth.10 Consequently, a higher
introduction of a personalized medical approach.1,2 percentage of pregnant women are at high risk for develop-
In this commentary, we aim to present a new paradigm ing serious maternal complications during pregnancy and
for healthcare service reorganization combining front-edge the postpartum period.1,5
technologies for early diagnostics, prevention, and treat- In the United States, the major etiologies of maternal
ments that can assist healthcare organizations in achieving mortality during pregnancy, labor, and delivery are CVDs,
reduced morbidity and mortality while optimizing cost ben- 14.6%; infection or sepsis, 13.6%; non-CVDs, 12.7%; cardio-
efit for the obstetrical care as a whole and pregnancy out- myopathy, 11.8%; hemorrhage, 11.4%; thrombotic pulmo-
comes in particular. The approach is patient centered and nary embolism, 9.6%; hypertensive disorders of pregnancy,
offers services to meet the individual patient needs. 9.4%; cerebrovascular accidents, 6.2%; amniotic fluid embo-
lism, 5.3%; and anesthesia complications, 0.7% as reported
by the National Center for Chronic Disease Prevention
The Neglected M in MFM and Health promotion, NCCDPHP (http://www.cdc.gov/
Although the maternal and fetal medicine subspecialty was reproductivehealth/MaternalInfantHealth/PMSS.html).
originally introduced to equally address fetal and/or mater- These leading causes of maternal morbidity and mortality
nal aspects of pregnancy management, the main focus of demonstrate that prepregnancy health (especially in the
MFM today is the diagnosis and treatment of fetal complica- context of CVDs) contributes to increased pregnancy com-
tions and improving neonatal outcome. The health of preg- plications. Further emphasis of the potential benefit of early
nant mothers is no longer getting sufficient attention. This diagnosis and prevention (especially of hypertensive disor-
situation may have been influenced by the rapid progress in ders during pregnancy) is attempted at reducing maternal
prenatal diagnosis of congenital and chromosomal anoma- mortality during pregnancy or immediately after delivery,
lies, the introduction of fetoscopic surgery for in utero treat- also demonstrating how the mode of delivery and postnatal
ments of fetal disorders and advanced imaging and Doppler care impacts pregnancy outcome.9,10
methodologies, and the profound impact of these develop- In the 2010 annual convention of the American Society
ments in reducing the incidence of major pregnancy disor- for Maternal and Fetal Medicine, M.E. D’Alton was the first
ders, stillbirth, and preterm birth.3 to address the question: “Where is the ‘M’ in Maternal Fetal
As the advanced treatment offered by neonatal intensive Medicine?”11 She urged to outline a specific plan for clinical,
care units progressed, premature babies with very low birth educational, and research initiatives in order to return the
weight are saved generating challenges for the postdelivery maternal “M” in MFM to the center of fetal and maternal
management of newborns.4 medical care. In 2012, a step forward was undertaken by the
At the same time, maternal morbidity rates have been ris- leading U.S. authorities in obstetrics and gynecology by pub-
ing.5 One major contributor, mainly in developed countries, lishing their recommendations to enhance education and
is the obesity epidemic6 causing increased rates of metabolic training in maternal care for MFM fellows, to improve medi-
disorders, diabetes, hypertensive disorders, and cardiovascu- cal care and management of pregnant women, and to address
lar diseases (CVDs), all of which are chronic in nature, in the critical research gaps in maternal medicine.4,11 However, to
general population. Women affected by these diseases have really return the focus to the maternal M in MFM, educa-
an elevated risk for pregnancy disorders, including hyper- tion and training alone are insufficient. A revolution in ser-
tension disorders in pregnancy, gestational diabetes (GDM), vice provision and introduction of the front-edge diagnostic
and prematurity as suggested by the American College and prevention technologies is required in order to shift the
of Obstetricians and Gynecologists (http://www.acog. emphasis to early diagnosis and prevention, to introduce an
org/Resources-And-Publications/Committee-Opinions/ individualized approach, to combine the front-edge genom-
Committee-on-Obstetric-Practice/Obesity-in-Pregnancy). ics into the standard of care, and to shift the paradigms of
The constant rise in the rates of delivery by cesarean section perinatal and antenatal care to emphasize maternal and
entails elevated risks of placenta accreta and hemorrhages.7 not just fetal aspects. And this will make the difference,
Furthermore, new assisted reproductive technologies save lives, and improve the use of financial resources to
opened the possibility for women to conceive among those the patients’ benefit. Such a paradigm shift is timely and is
with kidney, lung, heart, and other serious medical diseases required to improve maternal health and fetal outcome.
Bringing maternal–fetal medicine to the new era of medicine 3
This paradigm shift involves the introduction of multidis- these “programmed changes” are metabolic adaptations
ciplinary assessment procedures and novel test technologies to fetal undernutrition expressed in enhanced catabolism
with a focus on early diagnosis and prevention/management and self-consumption of substrates for energy supplies.15 A
of the major causes of maternal morbidities and mortality.12 In prolonged fetal adjustment period to undernutrition also
this commentary, we suggest the introduction of the inverted reduces endocrine concentration of fetal growth hormones,
pyramid of antenatal care for early (first trimester) screening via the reduced transfer of amino acids and glucose across
of a pregnant woman. The aim is to detect the development the placenta, due to decreased maternal insulin-like growth
of pregnancy disorders and outline the personalized path- factor. These changes are followed by reduced rates of fetal
way for the prevention of these disorders to improve preg- growth also creating a process of response to stress that is
nancy outcomes, as proposed by Nicolaides.13 It is proposed to repeated in adulthood life and thereby lead to metabolic
implement the inverted pyramid of prenatal care as the major disorders and CVDs.16
paradigm for antenatal care during pregnancy. Low birth weight was shown to be associated with
Additional components offered in the current com- increased rate of ischemic heart diseases in adulthood.
mentary is the introduction of new means and tools for Studies with three large cohorts (>16,000 individuals) in
implementation into the routine practice of pregnancy the United Kingdom have shown that mortality from isch-
management the methods for prepregnancy family planning emic heart disease later in life were twofold higher in those
and health evaluation. It will enable to assess the prior risks born <2.5 kg at birth compared to the ones born >4.3 kg.17
before conceiving and to plan ahead and provide appropri- Thin or stunted and small trunk babies who were born due
ate maternal care facilities. Using this framework will allow to in utero undernutrition, hypoxia, and other changes are
prior risk assessment and management before entering into predisposed to consequential diseases in the long term.18
the process of conceiving. Furthermore, increased mortality rates from coronary heart
Offering postpregnancy prevention and management can diseases are found among men born with a low birth weight,
then complete the whole framework of comprehensive ser- low placental weight, or narrow head circumference.19
vices for pregnant women. The prevalence of Diabetes Mellitous type 2 (type 2 DM)
Combining the progress in fetal medicine with innova- and impaired glucose tolerance later in life are threefold
tion in maternal health not only offers remerging the M and higher in people who were born with the smallest (<2.5 kg)
the F of the MFM specialty but will also bridge the discrep- birth weight compared to the people who weighed >4.3 kg
ancies and introduce new strategies for pregnancy manage- at birth.20,21 There is evidence that deficiency in insulin pro-
ment. This approach can assist healthcare organizations in duction and insulin resistance are both determined in utero
optimizing obstetrics care and pregnancy outcome.14 and that low-birth-weight babies develop in utero the “insu-
lin resistance syndrome” that prevailed in their adulthood,
causing an impaired glucose tolerance, hypertension, and
Bridging between maternal and fetal medicine: The high concentrations of triacylglycerol.21,22
maternal–fetal medicine hypotheses The extreme example of the long-term impact of nutrient
shortage in pregnancy was discovered with the Dutch study
The integration between the “M” and “F” components of
of individuals who were in utero during the Dutch famine
maternal and fetal medicine is required to enable compre-
of 1944–1945.23 This study provides evidence linking fetal
hensive care and management.12 During pregnancy, women’s
undernutrition to programmed insulin resistance and type 2
bodies and their fetuses are interconnected. Maternal heart
diabetes. Their glucose tolerance tests at age 50 years were all
adjustments occur in response to the increased physiological
higher than in those conceived before or after the famine.21,22
burden of pregnancy and the signaling of circulating factors
This study has also provided evidence for long-lasting epigen-
that are exchanged between the mother and the fetus having
etic effects transferred from the newborn to their progenies,
a positive/negative impact.14 These are just obvious elements
not through the mother but through the father, indicating
of the required integrative whole framework of the “M” and
the profound impact of undernutrition on the DNA meth-
“F” components of MFM.
ylation of germ cells associated with facilitated aging-related
Pregnancy shapes adulthood health: The Barker diseases for the generations to come.16,21 Another example is
hypothesis (fetal origins of adult diseases: programming the Chinese famine during 1954–1964, which was identified
and imprinting in utero) to be associated with a higher likelihood to develop metabolic
syndrome in adulthood.24 Based on all these changes, Time
Low birth weight: Increased risk for lifelong CVD and magazine published its series of articles on the way the first
diabetes Barker was the first to demonstrate how low 9 months shape the person’s health throughout life (http://
birth weight is associated with elevated risk for CVDs in
content.time.com/time/magazine/article/0,9171,2021065,00.
adulthood.12 He postulated that fetal shortage of nutrients
html#ixzz2s8IDDoru).
and oxygen due to placental insufficiency is associated
with fetal development of physiological pathways for stress
adjustment underlying recruitment of the same pathways Blood pressure and hypertension A multitude of studies
in adulthood, and leading to the development of higher have found a trend in which each 1 kg increase in birth
adulthood susceptibility to obesity, diabetes mellitus weight is associated with a fall of around 3.5 mmHg in blood
(DM), hypertension, and CVDs. According to Barker, pressure in adult life.25 There is a strong association between
4 Introduction
Adult life—
The barker hypothesis
Fetal/newborn
health
Late pregnancy—
The pedersen hypothesis
Pregnancy-induced Early and late pregnancy—

complications The freinkel hypothesis
Maternal
NCD
PET
PPH, Accreta
GDM
Preterm birth
VTE
Maternal
ng
health
gi M Type 2 diabetes
ir n the
B k FM
c
ba to M PET-related morbidity
M Cardiovascular morbidity
Figure 1.1 Maternal medicine meets fetal medicine; the vicious cycle—noncommunicable disease epidemic.
hypertension disorder in adulthood to low birth weight, diseases associated with pregnancy disorders in the next
thinness, stunting, and below-average head circumference.26 generation onward, is now recognized as the link between
All the aforementioned examples have demonstrated the origin of NCDs in neonatal life and adulthood diseases
how birth weight, in utero conditions, and epigenetic changes (Figure 1.1). It requires implementation of healthcare assess-
are associated with the increased adulthood morbidity from ment and preventive interventions before pregnancy to
NCDs leading to a vicious cycle for generations to come. reduce infant and maternal morbidity and mortality and
prevent developing NCDs later in life.29
Maternal aspects of placenta insufficiency
Increased maternal CVDs and decreased life expectancy Vicious cycle of the NCD epidemic (obesity, diabetes,
due to preeclampsia Another consequence of placental hypertension, metabolic syndrome): Fetal programming
insufficiency is preeclampsia, particularly the early form According to the World Health Organization (WHO), of the
of the disorder. McDonald et al., in their meta-analysis of 57 million people who died in 2008, 36 million died from
35,000 women, have shown that hypertension disorders in NCDs, stating that NCDs represent a “slow motion disas-
pregnancy are associated with increased maternal morbidity ter.”30 The four main chronic diseases responsible for most
from CVDs and DM 10 years later.26 Furthermore, Irgens NCD deaths are CVDs, including heart attacks and stroke
et al., using the Medical Birth Registry of Norway, have shown (17.3 million annually); cancer (7.6 million); respiratory dis-
in >600,000 women and their spouses a 10-year shortening of eases, such as chronic obstructive pulmonary disease and
maternal longevity following early preeclampsia and IUGR.27 asthma (4.2 million); and diabetes (1.3 million).31
Thus, placental insufficiency is programming high sus- Intermediate risk factors predisposing to NCDs include
ceptibility to CVDs and diabetes not only among babies hypertension, elevated blood glucose, hyperlipidemia, over-
born with lower birth weight but also among their mothers. weight, and obesity, which all can lead to the development
of CVDs.
NCDs and maternal morbidity The hypothesis about the developmental origins of health
Prepregnancy conditions of maternal health (obesity, dia- and disease put forward the concept that internal and exter-
betes, anemia, and undernutrition, kidney, blood, and heart nal environmental conditions during pregnancy cause
diseases) all impact maternal health during pregnancy.28 critical biochemical, endocrinological, and epigenetic modi-
Prepregnancy diabetes and GDM can cause macrosomia, fications in the DNA, cell differentiation, and formation of
obstructed labor, postpartum hemorrhage, and neonatal specific tissues in both the mother and her fetus/newborn.31
mortality due to prematurity, respiratory distress syndrome, While at birth these functional changes are currently not
hypoglycemia, etc. Maternal undernutrition can lead to fetal detected by conventional tests and are likely to be initially
metabolic and hormonal alterations causing lifelong sus- masked by systemic effects, the slow process of their devel-
ceptibility to certain diseases. At the same time, low birth opment into disorders may impact the health of the mothers
weight and accelerated growth during childhood have been and their children later in life.32
demonstrated as risk factors for CVD and type 2 DM. Epigenetic changes in DNA methylation and Cytosine-
This vicious cycle starting from prepregnancy health, Phosphate-Guanidin (CPG) Islands cause the silencing
influencing the outcome, which in turn causes adulthood or activation of certain genes that are essential for the
physiological function in early childhood and in adult life of developing pregnancy disorders in the early trimester of
and could lead to an accelerated DNA clocking and aging.33 pregnancy36 and the development of individualized pregnancy
Thus, epigenetic methods could shed light on in utero pro- management and disorder prevention and monitoring toward a
cesses that predispose individuals to diseases in adult life.31–33 timed delivery,38 including management of maternal and new-
The programmed in utero changes of the metabolism and born health for those who developed pregnancy disorders.37
physiology could lead to dysfunction and disease in adulthood. The model proposes a change in the organization of the
As such, the related pregnancy disorders such as preterm deliv- clinical services to pregnant women as a combination of
ery, IUGR, and preeclampsia can be considered as markers of hospital-based clinics and community clinics. This combined
increased risk of CVDs, obesity, and metabolic disorders, and paradigm of evaluation, prevention, treatment, and follow-up
GDM could be the source for obesity and DM or—overall— could provide high-quality perinatology services from pre-
the origin of NCDs. In fact, the American Heart Association pregnancy planning, through antenatal follow-up, to labor
(AHA) has identified women who develop hypertension and to delivery and through the postpartum management.
disorder during pregnancy and women who have GDM as
the two new high-risk groups for developing CVDs, which
require special management and monitoring as included in Clalit Health Services approach
the American Heart Association Stroke Council guidelines Clalit Healthcare Services is the largest health manage-
for the prevention of stroke in women (http://blog.heart.org/ ment organization (HMO) in Israel and one of the largest
preeclampsia-doubles-womens-stroke-risk-quadruples-later- in the world. It has more than 4 million insurees, operat-
high-blood-pressure-risk/). ing more than 100 community clinics, 40 regional women’s
In order to reduce the influence of epigenetic, biochemi- health centers, and 14 hospitals with some of the leading
cal, endocrinological, and physiological preconditioning of maternal–fetal medicine departments in Israel. The HMO
NCDs in the perinatal period, preventative measures should employs thousands of family physicians and gynecologists,
be introduced, including the provision of sufficient prenatal trains hundreds of interns and fellows, involves thousands of
care; prevention or optimal treatment of conditions such as nurses and many midwives, and manages advanced labora-
obesity, diabetes, and chronic hypertension; and also direct- tory services and testing infrastructure.
ing the attention at prepregnancy assessment of their prior As a leading HMO in Israel, the organization is in the
risks and family planning to assure women begin their preg- process of implementing the Clalit’s Maternal Medicine
nancy period with rich metabolic reservoirs and with a pre- Meet Fetal Medicine project that is built of three “floors”:
planned program for their pregnancy management based on
their prior risks. Pre-Pregnancy: Identifying NCDs and carriers of gene
In this way, it may be possible to inhibit negative epi- disorders before pregnancy
genetic, biochemical, physiological, and endocrinologi- The goal of the first floor is to improve maternal health and
cal programming. The importance of good maternal care prenatal outcome before the women get pregnant. While
beginning prior to conception and continuing during preg- part of this campaign involves promotion of health educa-
nancy and after delivery is therefore crucial to shape the tion to families, it also involves the assessment of the prior
health of mothers and their babies for life and to prevent risk of developing pregnancy disorders and offers appropri-
the impact of internal and external effects of long-lasting ate interventions with proven efficacy.38,39
changes, thus reducing the likelihood of NCD development In 2013, the AHA introduced the seven metrics to estab-
in adulthood. lish three health levels of well, intermediate, and sick status,
based on blood pressure, obesity, blood glucose and choles-
terol levels, smoking, physical activity, and healthy diets,
Reconnecting the M to the F in MFM adjusted to age and gender. The seven metrics are linked
The importance of improving maternal healthcare and ser- through a comprehensive algorithm to divide patients into
vices can leverage on recent achievements of the medical one of the three categories.40
research to introduce a strategic plan and policy. Preconception evaluation should follow this approach to
A healthy pregnancy begins before conception.34,365 It con- establish the high, intermediate, and low risk of developing
tinues during the gravid period with early recognition and pregnancy disorders, given that the same criteria that are
management of complications if they arise, with strategies considered as risks for hypertension disorders in pregnancy,
to prevent complications, planning for a timed delivery, and GDM, spontaneous preterm birth (SPB), and fetal growth
follow-up in the postdelivery period.37 Healthcare providers restriction are underlying risks for developing major CVDs
can help women prepare for pregnancy and for any potential and metabolic disorders.30,31,41
problems during pregnancy with postpregnancy manage- Thus, three categories of risk levels in pregnancy will be
ment of complications that were not previously prevented.37 defined, and stratification will be performed at prepregnancy
The Maternal Medicine Meets Fetal Medicine project of visit(s) to the family physician. Using patient interviews and
the Clalit Health Services proposes a new strategy and aims blood tests (including the hemoglobin A1C test) and obtain-
to establish a new paradigm of pregnancy care. It extended ing family, pregnancy, and medical history from the patients
from prepregnancy evaluation of prior risk of developing and from their electronic medical record, the family phy-
NCDs34,35 to the emphasis on a thorough evaluation of the risk sician will gather all the information available to establish
6 Introduction
The great obstetrical syndromes contingent management
Assessment Integrated clinic at 11–13 weeks Intervention
PE and SGA Obesity GDM PTL Aneuploidy Early PE and SGA Obesity GDM PTL
anatomic
scan
Aspirin Life style Diet Progesterone
History Prepregnancy Diagnosis History NT + fHCG
BMI +P
MAP Cx length APPA Insulin Cerdage
Weight gain
Uterine PI NIPT Oral hypo- Pessary?
GDM glycemic
FPG ≥ 5.1 mmol/L agents
PAPP-A, CVS
(92 mg/dL)
PLGF
Overt diabetes:
Proteinuria
• FPG ≥ 7.0 mmol/L (126 mg/dL)
• HbA1c ≥ 6.5%
• RPG ≥ 11.1 mmol/L (200 mg/dL)
Figure 1.2 The first trimester clinic and contingent management.
the prepregnancy risk for NCDs. Hadar et al.42 have demon- Inverted pyramid of antenatal care and personalized risk
strated that when glycemic control and obesity are optimized management
prior to gestation, smoking and drinking are avoided before The second floor of Clalit’s Maternal Medicine Meets Fetal
pregnancy, and moderate physical activity and a healthy diet Medicine project is proposed to follow the inverted pyramid
are adopted before conception—which could prevent GDM, of antenatal care suggested by Nicolaides.13 Unlike the cur-
congenital malformation, abortion, prenatal and neonatal rent antenatal care that focuses on a uniform service model
death, and adverse pregnancy outcomes.42 reacting to complications when they develop and focusing
The whole system will be accompanied by a team of on a uniform high-frequency visit plan during the third
genetic counselors in order to assess the prior risk of major trimester of pregnancy, the inverted pyramid of antenatal
genetic disorders. This approach was adopted by American care focuses on a thorough first trimester risk evaluation
College of Obstetricians and Gynecologists (ACOG) antena- and aims to implement a personalized approach for subse-
tal Care–Introducing the Inverted Pyramid for Pregnancy quent management according to the individual risk score
Management43 and is widely offered in Israel. Chips for blood (Figure 1.2).
test are available for all couples to identify carriers of major The approach applies to both those who are already man-
ethnic-associated DNA-based diseases. Couples identified as aged by the MFM specialized center due to prior risk and to
carriers will be directed for genetic counseling to consider the rest of the pregnant women population.
various assisted fertility technologies, including preimplan- The first prenatal visit aims to quantify the woman’s risk
tation Genetics diagnosis (PGD), whereas others are advised in this pregnancy in developing major pregnancy disorders.
to begin with spontaneous conception. The evaluation is based on the obstetrics and medical his-
For assessment of the major NCDs, the family physician/ tory, demographics, biochemical serum markers, and bio-
community gynecologist could prepare the prior risk evalu- physical parameters, including mean arterial blood pressure
ation and direct women to either dietitians and even to phys- and sonographic image of nuchal translucency and uterine
ical trainers to improve her prepregnancy conditions if she artery Doppler pulsatility index. All are entered into the risk
is at intermediate risk, or to an MFM specialist for glycemic algorithms to provide the patient risk for any of the chromo-
controls and other endocrinological and biochemical moni- somal aberrations, preeclampsia, and IUGR, SPB, and GDM.
toring, if her prior risk is high. Those at a low risk are reas- If performed at the community level, the inverted pyra-
sured and could start with their family planning. mid of antenatal care requires training and certification of
The success of this preconception level of care depends the community ob-gyn physicians to incorporate the wider
on the implementation of unified guidelines within the range of testing methodologies and equipment for provid-
entire complex of family physicians and obstetricians in the ing a multiple marker screening and the use of advanced
community, together with dietitians, physical trainers, and risk algorithms to enable the assessment of the patient risk
high-risk hospital clinics.42 They will have joint meetings to to a diversity of major pregnancy disorders. In the United
discuss cases and guidelines and will use unified medical Kingdom, this stage is carried out in hospital departments
records that will create a database for future quality evalua- that specialized in these procedures.
tion and research. As already reviewed by Hadar et al.,42 the Based on the specific pregnant woman’s medical and
effectiveness of prepregnancy care is highly dependent on obstetric history, her serum markers, blood pressure, and
such guidelines, education, and training, including patient sonographic examination, the risk score will be determined.
education.44
Accordingly, her pregnancy management will be adjusted Similar contingent approaches are being developed for
by a contingency management algorithm that will issue GDM and macrosomia.48 Metformin and insulin have been
an individualized pregnancy monitoring tailored for low-, shown to prevent GDM and macrosomia by 30%–40%.49
intermediate- and high-risk patients. In the Clalit version of the inverted pyramid, low-risk
While women at high risk are directed toward more fre- pregnancies will be managed in community clinics, by
quent visits and offered preventive care to prohibit/reduce general treating gynecologists. For high- and intermediate-
the risk of developing particular disorders while allowing risk pregnancies, follow-up will be offered via two potential
the majority of gravid women who are at low risk to enjoy pathways:
less medically intense pregnancy care, many women wish to
experience pregnancy as a healthy period. ●● Hospital maternal–fetal medicine clinics: High-risk preg-
The smaller group of women who are at a higher likeli- nancy with a multidisciplinary team including special-
hood of developing the threatening outcome will then be ized perinatologists, ultrasound experts and technicians,
directed to an intensive process of evaluation and, when pregnancy nurse educators, dietitians, genetic counselor,
appropriate, offered preventive treatment. The smaller size and perinatal laboratory personnel with immediate refer-
of this group enables to focus the efforts on those who need ral to invasive diagnostic tests.
it, while the majority will not require intensive medical ●● Women’s healthcare centers: A regional network of
attention. clinics that will have a dedicated service for high-risk
The model is shifting perinatal management from the women, with specialized perinatologists working in
treatment of complications after they develop to early risk tight cooperation with the regional obstetric depart-
identification during the first trimester of pregnancy to ment. It is important to note that in Israel today, some
enable prevention. For example, women who are at an ele- tests that are in the market and are important for a com-
vated risk for major pregnancy disorders, such as placental plete risk assessment as offered by high-risk algorithms
insufficiency, are offered the use of daily aspirin already in developed by the Fetal Medicine Foundation in London
the first trimester to prevent the development of preeclamp- are not yet included in the pregnancy tests covered by the
sia and IUGR, based on evidence that low-dose aspirin state insurance policy for pregnant women.* Thus Clalit
from the first trimester prevents preeclampsia, fetal growth offers a local additional alternative, through its fully
restriction, and stillbirth by at least 50%.44 It has been evalu- owned complementary insurance network of Mor for
ated that the cost of implementing such broad first trimester Women Health providing access to these new services.
evaluation and prevention is much cheaper than treating the The inverted pyramid of antenatal care proposes two
long-term impact of preeclampsia on mothers and their off- major additional clinic visits in the second trimester, at
spring, preventing further mortality and morbidity.45 This the 22nd and 32nd gestational weeks. The model sug-
approach was already taken in Australia on a small group gests a stepwise reduction in the number of patients at
of patients, which is further evaluated with the European the high-risk group during these additional visits, and
Commission (EC) project Aspirin for preeclampsia (ASPRE). the identification of the intermediate group, which can
The vast majority of women in the very-low-risk category be removed from close surveillance. Those that are iden-
can be reassured that preeclampsia and IUGR are unlikely. tified to develop these conditions in late pregnancy can
The power of the inverted pyramid and of the proposed be offered planning for a timed or early delivery.
contingent model of prevention could be demonstrated for
the case of Down syndrome. Early screening is based on Postpregnancy: Early prevention of NCDs by early
maternal age, biochemical markers, and nuchal translu- examination shortly after birth
cency. Based on the initial risk, a small proportion of women The third floor of Clalit’s Maternal Medicine Meets Fetal
at the highest risk are directed to interventional procedures Medicine project is offered for the postpregnancy period
(Chorionic Villus Sampling [CVS], or early amniocentesis) (Figure 1.3). The prevalence of various forms of NCDs—
and karyotyping. The intermediate-risk group is referred to the whole spectrum of metabolic syndrome and premature
cell-free fetal DNA testing to assess for chromosomal abnor- CVDs—is increased in women with a history of maternal
malities while the very-low-risk group is sent home. With placental syndromes (pregnancy-induced complications),
this approach, the fetal medicine foundation group was able including GDM, pregnancy-associated hypertensive disor-
to show prediction of more than 97% of all major trisomies ders (e.g., preeclampsia), fetal growth restriction, preterm
in the first trimester.46 labor, and placental abruption. 50 Following pregnancies
In the case of SPB, an algorithm at 11–13 weeks identi- with any of the aforementioned conditions, the informa-
fies a small, very-high-risk group with a history of SPB and a tion will be passed to the family physician, neonatolo-
short cervix that can benefit from the treatment of cerclage gist, and pediatrician, which will refer the patients to the
or followed up closely and offered prevention treatment with needed consultants—e.g., diabetologists, endocrinologists,
progesterone at the 22nd gestational week. Based on ran-
domized studies, evidence was found that the use of proges-
terone can prevent SPB by 45%.47 The vast majority of women * In this context, it is important to note that this community clinic–based
in the very-low-risk category can be reassured that SPB is solution should be adopted as the healthcare model of each country that
unlikely. should involve not just midwives but also obstetric gynecologists.
8 Introduction
Follow-up—Postpregnancy-induced complications
Type2DM
Normal Cardiovascular risk other
Community clinic Postnatal evaluation Community clinic

Family physician‐GP OGTT (6–12 weeks) Family physician‐GP
gynecologist other option gynecologist
Prevention Annual follow-up Treatment

FPG
Lifestyle medications HbA1C
other options OGTT
Other options
Planning new pregnancy

Prepregnancy evaluation
New pregnancy
Early evaluation (first visit)
FPG Treatment
HbA1c and/or
OGTT prevention
Other options
Figure 1.3 The postpregnancy clinic.
cardiologists, and nutritionists. Thus, the women will be evaluate the preconception risks to adjust patient’s glycemic
immediately managed aggressively in community clinics control and dietary, physical activity, and nonsmoking hab-
that will provide measures for the prediction and preven- its before conception. Combining the two floors will improve
tion of future pathological conditions, including lifestyle pre- and postpregnancy management and integrate the skills
management, medication, and family lifestyle advice and of the specialists in maternal and fetal medicine with inter-
support. nal medicine and family physicians to substantially improve
maternal and fetal outcome. The preconception service meets
the needs and habits of the young generation through com-
Summary munication and family planning. The inverted pyramid floor
forms the opportunity for the majority of pregnant women
NCDs such as CVDs, diabetes, obesity, and hypertension to experience pregnancy as a healthy period, offering them
have significant adverse impacts on maternal health and reassuring evidence to enjoy such an approach. The small
pregnancy outcomes, and through the imprinting mecha- proportion of women who are at a high risk and need close
nism of intrauterine programming, the burden of future surveillance and perhaps treatment will have the opportunity
NCDs is highly influenced by pregnancy disorders and to do so and will be offered a sensible service personalized
impacts the burden of NCDs in future generations.51,52 to their needs. Together the two floors are tailored better to
Research today has shown repeated cycles of vulnerabil- the public. The dedicated postpartum and child development
ity to NCDs through major pregnancy disorders such as low floor could offer better monitoring to women who experi-
birth weight, preeclampsia, GDM, and SPB that are propa- enced pregnancy complications in adjusting their lifestyle
gating the risk for NCDs to subsequent generations through and taking healthcare measures. It enables early intervention
epigenetic, physiological, endocrinological, and biochemical and diagnosis of newborn complications to reach sufficient
pathways. nutrient support, physical training, physiotherapy, and cog-
Clalit’s Maternal Medicine Meets Fetal Medicine project is nitive progress in a period of plasticity of the brain and of
based on three “floors.” At the core is the inverted pyramid of other tissues to achieve better outcomes.51,52
antenatal care that offers a systematic method to screen and
prevent pregnancy disorders according to the principles set up
by Nicolaides,13 involving first trimester risk score and con- Acknowledgment
tingency managements of low-, intermediate-, and high-risk
pregnancies according to individualized management plans. The funding for this review was partially obtained from
In the preconception period, Clalit’s Maternal Medicine the ASPRE project of the EC Seventh Framework project #
Meets Fetal Medicine project offers an additional floor to 601852 (KN, HM).
References 9
REFERENCES
1. Berg CJ, Callaghan WM, Syverson C, Henderson Z. Pregnancy- 25. Barker DJP, Godfrey KM, Osmond C, Bull A. The relation of fetal
related mortality in the United States, 1998 to 2005. Obstet length, ponderal index and head circumference to blood pres-
Gynecol 2010; 116: 1302–1309. sure and the risk of hypertension in adult life. Paediatr Perinat
2. Callaghan WM, Creanga AA, Kuklina EV. Severe maternal mor- Epidemiol 1992; 6: 35–44.
bidity among delivery and postpartum hospitalizations in the 26. McDonald SD, Malinowski A, Zhou Q et al. Cardiovascular
United States. Obstet Gynecol 2012; 120: 1029–1036. sequelae of preeclampsia/eclampsia: A systematic review and
3. King JC. Maternal mortality in the United States—Why is it meta-analyses. Am Heart J 2008; 156: 918–930.
important and what are we doing about it? Semin Perinatol 2012; 27. Irgens HU, Reisæter L, Irgens LM, Lie RT. Long term mortality
36: 14–18. of mothers and fathers after pre-eclampsia: Population based
4. Menacker F, Hamilton BE. Recent trends in cesarean delivery cohort study. Br Med J 2001; 323: 1213–1217.
in the United States. NCHS data brief, No. 35. Hyattsville, MD: 28. Keeling A. Halting the rise of non-communicable diseases: An
National Center for Health Statistics, 2010. urgent priority for women’s health. Diabetes Res Clin Pract 2011;
5. The American Board of Obstetrics and Gynecology. The Division 92: 143–144.
of Maternal-Fetal Medicine. AJOG’s guide to learning in mater- 29. World Health Organization. 2008–2013 Action plan for the global
nal fetal medicine. Dallas, TX, 2013. Available from: http://www. strategy for the prevention and control of non-communicable
abog.org/publications/2013GuidetoLearning-MFM.pdf, accessed diseases. Geneva, Switzerland: WHO, 2008. Available from:
on May 18, 2014. http://www.who.int/nmh/publications/9789241597418/en/index.
6. Artal R, Lockwood CJ, Brown HL. Weight gain recommendations html, accessed May 18, 2014.
in pregnancy and the obesity epidemic. Obstet Gynecol 2010; 30. Hannum G, Guinney J, Zhao L et al. Genome-wide methylation
115(1): 152–155. profiles reveal quantitative views of human aging rates. Mol Cell
7. American College of Obstetricians and Gynecologists. Obstetric 2013; 49: 359–367.
care consensus No. 1. Safe prevention of the primary cesarean 31. World Health Organization. Global status report on noncom-
delivery. Obstet Gynecol 2014; 123: 693–711. municable diseases 2010. Geneva, Switzerland: World Health
8. Schmidt L, Sobotka T, Bentzen JG, Nyboe Andersen A, on behalf Organization, 2011.
of the ESHRE Reproduction Society Task Force. Demographic 32. Barouki R, Gluckman PD, Grandjean P et al. Developmental ori-
and medical consequences of the postponement of parenthood. gins of non-communicable disease: Implications for research and
Hum Reprod Update 2012; 18: 29–43. public health. Environ Health 2012; 11: 42.
9. Gielchinsky Y, Laufer N, Weitman E et al. Pregnancy restores 33. Perera F, Herbstman J. Prenatal environmental exposures, epi-
the regenerative capacity of the aged liver via activation of an genetics and disease. Reprod Toxicol 2011; 31: 363–373.
mTORC1-controlled hyperplasia/hypertrophy switch. Genes Dev 34. The importance of preconception in the continuum of women’s
2010; 24: 543–548. health care. American Academy of Pediatrics and The American
10. Skjaerven R, Wilcox AJ, Klungsoyr K et al. Cardiovascular mor- College of Obstetricians and Gynecologists. Guidelines for
tality after pre-eclampsia in one child mothers: Prospective, Perinatal Care, 5th edn., No. 313. 2005, reaffirmed 2012.
population based cohort study. BMJ 2012; 345: e7677. pp. 1–2.
11. D’Alton ME, Bonanno CA, Berkowitz RL et al. Putting the “M” 35. CDC. Recommendations to improving preconception health and
back in maternal-fetal medicine. Am J Obstet Gynecol 2012; health care—United States. MMWR 2006; 55(No. RR-06): 1–23.
pii: S0002-9378(12)02187-4. 36. Vogel JP, Souza JP, Mori R et al. Maternal complications and
12. Hod M, Liberman N. Maternal fetal medicine—How we can perinatal mortality: Findings of the World Health Organization
practically connect the “M” to the “F”? Best Pract Res Clin Obstet Multicountry Survey on Maternal and Newborn Health. WHO
Gynaecol 2015; 29(2): 270–283. Multicountry Survey on Maternal and Newborn Health Research
13. Nicolaides KH. A model for a new pyramid of prenatal care Network. BJOG 2014; 121(Suppl. 1): 76–88.
based on the 11 to 13 weeks’ assessment. Prenat Diagn 2011; 37. Blenning CE, Paladine H. An approach to the postpartum office
31: 3–6. visit. Am Fam Physician December 15, 2005; 72: 2491–2496,
14. VanHook AM. Regeneration switch. Sci Signal 2010; 3(114): Review.
ec88. 38. Moos MK, Cefalo RC. Preconceptional health promotion: A
15. Barker DJP. Mothers, Babies and Health in Later Life, Churchill focus for obstetric care. Am J Perinatol 1987; 4: 63–67.
Livingstone: Edinburgh, Scotland, 1998. 39. Adams EM, Bruce C, Shulman MS et al. The PRAMS working
16. Osmond C, Barker DJP, Winter PD et al. Early growth and death group: Pregnancy planning and preconceptional counseling.
from cardiovascular disease in women. Br Med J 1993; 307: Obstet Gynecol 1993; 82: 955.
1519–1524. 40. Go AS, Mozaffarian D, Roger VL et al., for the American Heart
17. Martyn CN, Barker DJP, Osmond C. Mothers’ pelvic size, fetal Association Statistics Committee and Stroke Statistics Sub
growth, and death from stroke and coronary heart disease in men Committee. Heart disease and stroke statistics 2014 update: A
in the UK. Lancet 1996; 348: 1264–1268. report from the American Heart association. Circulation 2014;
18. Kathryn G, Dewey KG, Begu K. Long-term consequences of 129: e28–e292.
stunting in early life. Matern Child Nutr 2011; 7(5s): 5–18. 41. Ness RB, Markovic N, Bass D et al. Family history of hyper-
19. Barker DJ, Osmond C, Forsen TJ et al. Trajectories of growth tension, heart disease and stroke among women who develop
among children who have coronary events as adults. N Engl J hypertension in pregnancy. American College of Obstetricians
Med 2005; 353: 1802–1809. and Gynecologists. Obstet Gynecol 2003; 102: 1366–1371.
20. Phipps K, Barker DJP, Hales CN et al. Fetal growth and impaired 42. Hadar E, Ashwal E, Hod M. The preconception period as an
glucose tolerance in men and women. Diabetologia 1993; 36: opportunity for prediction and prevention on non-communicable
225–228. diseases. Best Pract Res Clin Obstet Gynaecol 2015; 29: 54–62.
21. Barker DJP, Hales CN, Fall CHD et al. Type 2 (non-insulin depen- 43. American College of Obstetricians and Gynecologists. Prenatal
dent) diabetes mellitus, hypertension and hyperlipidaemia (syn- and preconceptional carrier screening for genetic diseases in indi-
drome X): Relation to reduced fetal growth. Diabetologia 1993; viduals of Eastern European Jewish descent. ACOG Committee
36: 62–67. Opinion No. 298. Obstet Gynecol 2004; 104: 425–428.
22. Wehkalampi K, Muurinen M, Wirta SB et al. Altered methyla- 44. Roberge S, Villa P, Nicolaides K et al. Early administration of low-
tion of IGF2 Locus 20 years after preterm birth at very low birth dose aspirin for the prevention of preterm and term preeclamp-
weight. PLOS ONE 2013; 8(6): e67379. sia: A systematic review and meta-analysis. Fetal Diagn Ther
23. Ravelli ACJ, van der Meulen JHP, Michels RPJ et al. Glucose tol- 2012; 31: 141–146.
erance in adults after prenatal exposure to the Dutch famine. 45. Shmueli A, Meiri H, Gonen R. Economic assessment of screening
Lancet 1998; 351: 173–177. for pre-eclampsia. Prenat Diagn 2012; 32: 29–38.
24. Li Y, Jaddoe VW, Qi L et al. Exposure to the Chinese famine 46. Gil MM, Akolekar R, Quezada MS et al. Analysis of cell-free DNA
in early life and the risk of metabolic syndrome in adulthood. in maternal blood in screening for aneuploidies: Meta-analysis.
Diabetes Care 2011; 34: 1014–1018. Fetal Diagn Ther 2014; 35: 156–173.
10 Introduction
47. Romero R, Nicolaides K, Conde-Agudelo A et al. Vaginal proges- 50. Siddiqui N, Hladunewich M. Understanding the link between the
terone in women with an asymptomatic sonographic short cer- placenta and future cardiovascular disease. Trends Cardiovasc
vix in the midtrimester decreases preterm delivery and neonatal Med 2011; 21: 188–193.
morbidity: A systematic review and metaanalysis of individual 51. Zwicker JG, Harris SR. Quality of life of formerly preterm and
patient data. Am J Obstet Gynecol 2012; 206: 124.e1–124.e19. very low birth weight infants from preschool age to adulthood.
48. Metzger BE, Lowe LP, Dyer AR et al., HAPO Study Cooperative Pediatrics 2008; 121: e366–e376.
Research Group. Hyperglycemia and adverse pregnancy out- 52. Vederhus BJ, Markestad T, Eide GE et al. Health related quality
comes. N Engl J Med 2008; 358: 1991–2002. of life after extremely preterm birth: A matched controlled cohort
49. Cotarelo A, Zaman HT, Jovanovič L, Hod M. Technology and study. Health Qual Life Outcomes 2010; 8: 53.
pregnancy. Diabetes Technol Ther 2014; 16(Suppl. 1): S68–S77.
2 History of diabetic pregnancy
David R. Hadden
sugar,” and in early Chinese and Japanese writings “the urine

Introduction of diabetics was very large in amount and so sweet that it
One hundred years ago, the medical literature on diabetic attracted dogs.”1,2
pregnancy was very limited. Pregnancy itself was no less fre- After the European Renaissance, the first physician to
quent, but the outcome was affected by so many other major rediscover and record the sweetness of the urine in diabetes
problems that the influence of a medical disorder of a chronic was Thomas Willis in London (1679): “The diabetes or piss-
nature was both unrecognized and disregarded. Diabetes ing evil … in our age given to good fellowship and guzzling
mellitus was also less prevalent, due both to demographic down of unalloyed wine.” And Mathew Dobson 100 years
differences in the age of the population and to epidemiologi- later in Liverpool first demonstrated chemically the presence
cal factors—mainly the absence of any effective treatment so of sugar in the urine of diabetic patients. The demonstration
that young people with diabetes had a life expectancy of only by Oskar Minkowski (1889) that removal of the pancreas in
a few years. The diagnosis of diabetes depended on the dem- a dog unexpectedly resulted in uncontrolled polyuria—the
onstration of sugar in the urine and the well-known symp- urine sugar attracted flies in the laboratory to the puddles
toms of thirst, polyuria, and weight loss, but there was no on the floor—was the significant observation that eventu-
accurate measurement to assess severity, and the distinction ally led to the extraction of insulin from the pancreatic islets
between what are now known as type 1 and type 2 diabetes in Toronto in 1922.3 The story of the discovery of insulin is
was only anecdotal. There was no documentation of the spe- a remarkable record of disappointment: it was almost dis-
cific long-term complications of hyperglycemia in the eyes, covered in 1906 by Zuelzer in Berlin, and then in 1912 by
nerves, heart, kidneys, or blood vessels. Scott in Chicago, but was actually extracted by Paulesco in
Romania in 1920. However, the world recognizes the story
of the Toronto group—including Banting, Best, Collip,
Early history of diabetes and Macleod—as the definitive discovery, and in 1923, the
Nobel Prize in Physiology or Medicine was awarded to two
Diabetes was well recognized as a medical disorder of them, Frederick Banting and JJR Macleod.4
>2000 years ago, and some well-known references are worth Up until then, the only effective treatment for diabetes had
quoting. The ancient Egyptian Ebers papyrus, dating to been dietary, and it was well known that restriction of food
1500 BC, records abnormal polyuria; the Greek father of would ameliorate the symptoms. John Rollo had demon-
medicine Hippocrates (466–377 BC) mentioned “making strated this with his patient Captain Meredith in the army in
water too often” and Aristotle also referred to “wasting of the Ireland in 1797, who obeyed his doctor’s advice, documented
body.” Aretaeus of Cappadocia (AD 30–90) in Asia Minor the reduction in urine volume and subsequent weight loss,
(now Turkey) is credited as the first to use the name “dia- and even extracted sugar from the urine by evaporation. The
betes,” which is Greek for a siphon, meaning water passing dietary approach was carried to its logical extreme by the
through the body: “diabetes is a wasting of the flesh and overenthusiastic approach of FM Allen in New York (1919),
limbs into urine – the nature of the disease is chronic, but whose starvation therapy often temporarily returned the
the patient is short lived … thirst unquenchable, the mouth blood glucose to normal, but only succeeded in extending
parched and the body dry ….” The famous Arabian physi- life for a year or so in the severe juvenile cases, all of whom
cian Avicenna (AD 980–1027) recorded further important became skeletally thin. Dr. Elliott Joslin is remembered as
observations that maintained and extended the previous the Boston physician who bridged the period immediately
Greek knowledge through what became known in Europe before insulin’s discovery and the exciting clinical dem-
as the Dark Ages: he described the irregular appetite, men- onstration of its effectiveness in the following decade.5 In
tal exhaustion, loss of sexual function, carbuncles, and London, Dr. Robin Lawrence, diabetic himself, on dietary
other complications. There are also references to diabetes therapy only in his early twenties, recorded how his life was
in ancient Hindu texts (AD 500) as a “disease of the rich, saved in 1923 by a telegram from his doctor in Kings College
brought about by gluttony or over-indulgence in flour and Hospital: “I’ve got insulin, and it works – come back quick.”
11
He survived for many years and became the leading diabetes successfully treated by diet only (which was probably type 2,
specialist in England.6 rather than type 1) are of considerable theoretical interest,
These two doctors, Joslin in Boston and Lawrence in but it is perhaps important that these cases were not often
London, became the leaders of the revolution that would reported in the literature and may well have been missed due
take place in both the opportunity for and the outcome of to not even testing the urine for sugar.
pregnancy in diabetic women. In the preinsulin days, and for some time after, death of
the mother during or soon after pregnancy from uncon-
trolled diabetes was the major risk. But maternal mortality
Pregnancy and diabetes before the was high for many reasons unrelated to diabetes, and retro-
discovery of insulin spective analysis of data from England and Wales between
1850 and 1937 shows that poor interventional obstetric care
A full historical review of fertility and of the outcome of with increased risk of puerperal sepsis was more important
pregnancy in different parts of the world is beyond the scope than social or economic deprivation.10 The maternal mortal-
of this chapter, but there are a number of aspects that are of ity rates for Scandinavian countries were much lower, and
particular relevance to the story of diabetes. Medical history it is now clear that this was due to better overall obstetric
in particular is constrained by publication bias, and there management in the prevention of sepsis; in the United States,
is much more available data regarding Europe and North maternal mortality between 1921 and 1924 was 6.8 per 1000
America than in other parts of the world. The geographi- births, in England and Wales 3.9 per 1000 births, and in the
cal and ethnic differences in the distribution, development, Netherlands only 2.5 per 1000 births.8 These differences at
and management of diabetes in different places at different national level have been widely discussed, but must be borne
times would be of great interest to review, but as the data in mind when considering the isolated effect of maternal dia-
are patchy and both diabetic and obstetric treatments often betes over those years.
poorly defined, it may be that “History followed different Overall perinatal mortality (death of the fetus after
courses for different peoples, because of differences among 28 weeks or within 7 days of delivery) has shown a more
peoples’ environments, not because of biological differences consistent fall over the same period of time in all Western
among peoples themselves.”7 There are certainly both envi- countries. Most of the decline was in postneonatal mortality
ronmental and genetic reasons for the differing prevalence related to the rising standards of living and nutrition but also
and incidence of diabetes in different countries, as much to improved public health measures—broadly speaking, the
as for the different outcomes of pregnancy, but the interna- predominant form of infant mortality in Western countries
tional historical study of these factors is still in its infancy. was postneonatal in the nineteenth century and neonatal
The collection of vital statistics first became avail- in the twentieth. There was not a close link between neo-
able at varying times in the developed Western countries. natal and maternal mortality, but there were very consider-
The Scandinavian countries were first Sweden (1749) and able differences in each of these measures between countries
Denmark (1801), England and Wales followed (1838), and at the time of discovery of insulin (Table 2.1). The overall
then Russia (1867); although the process was initiated in infant mortality rates in Scandinavian countries were per-
the United States in 1880, it did not become complete until sistently lower than in England and Wales, or Belgium,
1933.8 Fertility rates have varied as much as death rates and between 1920 and 1965, although all countries show a steady
migration in different countries, so that population dynam- exponential decline.8 As perinatal mortality is now used as
ics will have a considerable effect on reported statistics for a
single condition such as diabetes in pregnancy. The classi-
cal Malthusian checks on death rate—disease, famine, and Table 2.1 Overall maternal mortality and infant and
war—and the effects of celibacy and restraint on birth rate neonatal mortality for selected countries at the time
will have more effect on the overall outcome statistics of of discovery of insulin
pregnancy in diabetic mothers than the diabetes itself. The
general fertility rate for England and Wales was about 130 Maternal Infant Neonatal
live births per 1000 women between the ages of 15 and 44 in deaths, deaths, deaths,
1840 but is now only half that rate. At present, the total fertil- 1921–1924, 1924, per 1924,
per 1000 1000 per 1000
ity rate (average number of children born per woman) varies
Country births births births
from 2.1 in Western Europe to 6.7 in West Africa.9 However,
there is no doubt that untreated diabetes must have been vir- The Netherlands 2.5 67.3 18.6
tually incompatible with successful pregnancy before about Japan 3.3 166.4 67.5
1850. In 1856, Blott in Paris wrote that “True diabetes was England/Wales 3.9 75.1 33.1
inconsistent with conception,” and certainly the then short Australia 4.5 57.1 29.8
life expectancy of a young woman with what we now call United States 6.8 70.8 38.6
type 1 diabetes before the discovery of insulin would sup-
Source: Loudon, I., Death in Childbirth: An International Study
port that statement. Recent speculation on the possible of Maternal Care and Maternal Mortality 1800–1950,
nutritional causes of the present-day epidemic of type 2 Clarendon Press, Oxford, U.K., 1992, pp. 1–622.
diabetes in older patients means that any data on diabetes
Important early publications 13
a main comparator for the outcome of diabetic pregnancy, only a few Caucasian.24 Subsequent studies in many parts
it is important to bear these long-standing historical trends of the world have extended the recognition of what has now
in mind. become, in some places, an epidemic of hyperglycemia in
Congenital malformations are also an important compar- pregnancy. Jorgen Pedersen also used the term gestational
ator for obstetrical results, but the recognition of a possible diabetes in his monograph in 1967, but preferred to so clas-
link with maternal diabetes is much more recent: anecdotal sify a mother only after delivery, when he had demonstrated
accounts in small series in the 1940s were not supported until that her abnormal glucose tolerance in pregnancy had actu-
the report by the U.K. Medical Research Council in 195511 ally returned to normal postpartum; this rigorous defini-
and the larger series from Copenhagen in 1964.12 Historical tion has proved too difficult to achieve in practice.25,26 The
records on the frequency of congenital malformations true definition of hyperglycemia in pregnancy judged by the
are very incomplete, and it was not until the International internationally acceptable 75 g oral glucose tolerance test
Clearinghouse for Birth Defects began to operate after 1974 awaits the results of the large Hyperglycemia and Adverse
that any baseline data on the prevalence of congenital mal- Pregnancy Outcome study.27 The enthusiasm of the team at
formations became possible.13 It is still difficult to compare Northwestern University, Chicago, led by Norbert Freinkel
results for specifically identified diabetic pregnancies with and subsequently by Boyd Metzger has ensured that the con-
overall national malformation rates where the collection of cept of gestational diabetes is now firmly imprinted on the
cases is much less detailed.14 Other obstetrical complications obstetric mind, as well as having established a major place
such as preeclampsia appear today to be more common in as an epidemiological tool to study not only the immediate
diabetic pregnancy, but it is difficult to trace this possible outcome of pregnancy but also the long-term effects on both
interrelationship back to the days before organized antenatal mother and baby of the relatively short phase of hyperglyce-
care. Some of the cases where maternal death occurred in a mia during the latter part of the pregnancy.
diabetic pregnancy may have been due to eclampsia rather
than diabetic coma.
Important early publications
Gestational diabetes The historical development of understanding in obstetric,
metabolic, and pediatric disciplines over the past 100 years
The concept of gestational diabetes, actually meaning hyper- is perhaps best illustrated by several more extensive quota-
glycemia due to the pregnancy itself but in practice defined tions and commentaries on seminal papers from the early
as “carbohydrate intolerance of varying severity with onset literature: Bennewitz HG, Diabetes mellitus—A symptom of
or first recognition during pregnancy,” is also recent.15 In the pregnancy [translated from Latin].28
very first recorded case, Bennewitz, in 1823, considered that This is the first reference to diabetes in pregnancy.
the diabetes was actually a symptom of the pregnancy, and Although the patient was young, the clearly described onset
as the symptoms and glycosuria disappeared after at least of her symptoms during the pregnancy would now classify
two successive pregnancies, he had some evidence to support this as gestational diabetes. Is it possible that she only sur-
his views.16 That lesser degrees of maternal hyperglycemia vived because she has a milder case who responded to diet,
were also a risk to pregnancy outcome dates back to studies while all the more severe type 1 diabetic patients died?
in the 1940s in the United States17,18 and Scotland,19 which Henry Gottleib Bennewitz publicly defended his the-
showed increased perinatal mortality some years before the sis for the degree of doctor of medicine at the University
recognition of clinical diabetes mellitus. This led to the term of Berlin on June 24, 1824. It is a simple case report and
prediabetes in pregnancy and to poorly defined concepts of review of the literature on the causes and treatments of
temporary and latent diabetes. The first prospective study diabetes known at that time. His Greek derivation of the
of carbohydrate metabolism in pregnancy was established word diabetes and his one-line definition of the symp-
in Boston in 1954, using a 50 g 1-hour screening test, which toms are unchanged today: “Urine differing in quality and
has subsequently been widely adopted in the United States.20 quantity from the normal … accompanied by unquench-
O’Sullivan21 first used the name gestational diabetes in 1961, able thirst and eventual wasting.” Before giving the case
following the term metagestational diabetes used by Dr. JP history, he summarized his belief that the diabetic con-
Hoet in 1954 after his early studies in Louvain, Belgium.22 dition was in some way a symptom of the pregnancy, or
At that time, the U.S. emphasis was on establishing crite- due to the pregnancy. He noted that “Other disorders …
ria for the 100 g oral glucose tolerance test in pregnancy as began to break out as the pregnancy matured … the little
an index of the subsequent risk of the mother developing fires which had hidden beneath the smouldering deceiving
established diabetes, and the well-known O’Sullivan cri- ashes broke forth and devoured again the woman’s condi-
teria were derived on this basis.23 At about the same time, tion in the most wretched manner.” He was convinced that
Mestman, in southern California, began to identify the very “The disease appeared along with pregnancy, and at the
considerably increased perinatal mortality associated with very same time …; when pregnancy appeared, it appeared;
abnormal oral glucose tolerance in the obstetric popula- while pregnancy lasted, it lasted; it terminated soon after
tion of Los Angeles County Hospital, which then comprised the pregnancy.” He showed a degree of humility when he
>60% Latino mothers with the rest African-American and remarked that his patient must be something of a rare bird.
The case history commences on November 13, 1823, were usually of a large size: at least 10 survived and only 3
when Frederica Pape, aged 22, was admitted at 7 months in miscarriages are recorded; another 20 pregnancies seem to
her fifth pregnancy to the Berlin Infirmary. The first three have occurred before the recorded cases, so some of these
pregnancies appear to have been unremarkable, but in the mothers must represent late-onset type 2 or gestational dia-
fourth in 1822 she had an onset of thirst and polyuria that betes, and these seemed to have a better prognosis for both
had resolved spontaneously after delivery. These symptoms mother and child.
returned at an unspecified time in her fifth pregnancy: she
had “a really unquenchable thirst—she consumed more So far as is known, all, with one exception, were mul-
than six Berlin measures of beer or spring water, although tipara, the pregnancy of highest number being the tenth.
the quantity of urine greatly exceeded the amount of liquid They cannot be read without giving a strong impression
consumed, and the urine itself smelt like stale beer. Her voice of the great gravity of the complication, but they are not
was weak, skin dry, face cold and she complained of a drag- sufficiently numerous to justify any statistical argument
ging pain in her back.” based on the number of occurrences.
Treatment was more a matter of belief than of under-
standing, but apart from having withdrawn 360 mL of The histories further show that
venous blood all at once (the equivalent of 36 10 mL routine
blood tests today) and taking a high-protein diet, probably • Diabetes may come on during the pregnancy.
deficient in vitamins, she must have benefited from the rest • Diabetes may occur only during pregnancy, being
and care. The measurement of 2 oz of sugar in 16 lb (224 oz) absent at other times.
of urine, which is equivalent to about 1% glycosuria, was • Diabetes may cease with the termination of the preg-
Bennewitz’s only biochemical evidence of diabetes mellitus. nancy, recurring some time afterwards.
From about 32 to 36 weeks, the patient had a recurrent sore • Pregnancy may occur during diabetes.
throat and increased abdominal distension such that twins • Pregnancy and parturition may be apparently unaf-
were suspected. When examined on December 28, 1823, fected in its healthy progress by diabetes.
the cervix was dilating and the fetal head already partially • Pregnancy is very liable to be interrupted in its course,
descended. On December 29, she had an obstructed labor, and probably always by the death of the foetus.
and the child died intrapartum, probably due to delay in the
second stage. Bennewitz remarked that the baby was of “such Whitfield Williams was professor of obstetrics at Johns
robust and healthy character whom you would have thought Hopkins University and wrote the first major American
Hercules had begotten.” The infant weighed 12 lb, a fact wit- textbook on obstetrics, which still survives today in the eigh-
nessed carefully. Postpartum, in spite of continued dieting, teenth edition. He was concerned that the demonstration of
sweating and purging, and the application of eight leeches, sugar in the urine in pregnancy would be overinterpreted.
the patient’s strength improved daily, and sugar disappeared “I know of no complication of pregnancy the significance
from her urine. “With nature to preserve and treat her, we of which is more variously interpreted than the presence of
dismissed our patient cured.” sugar in the urine of pregnant women.” Williams blamed
Unfortunately, there is no record of the woman’s sub- Matthews Duncan29 for concluding that the detection of
sequent health, perhaps because Dr. Bennewitz presented sugar in the urine constituted one of the most serious compli-
his thesis within 6 months and, having been successful in cations of pregnancy, as Duncan’s views were accepted with-
obtaining his doctorate, dropped out of academic medicine. out question, although they were based on a small series of
This pregnancy would certainly qualify as “carbohydrate 22 pregnancies in 16 women collected from the then medical
intolerance of varying severity with onset or first recogni- literature over 60 years, and his own small experience in
tion during pregnancy”—which was the definition agreed Aberdeen. Williams30 presented six case reports to illustrate
for gestational diabetes at the first workshop–conference in the various conditions in which sugar may be observed in the
Chicago in 1980. urine of pregnant women: simple lactosuria, transient gly-
Matthews Duncan graduated in Aberdeen and became cosuria (two cases), alimentary glycosuria, recurrent glycos-
one of the leading obstetricians of his day. This compila- uria, and mild diabetes. All resulted in a normal pregnancy
tion of cases from the literature, from anecdotal reports, outcome (although all the recorded birth weights were >8 lb).
and from his own experience first identified the serious He then analyzed the urinary records of 3000 consecutive
problem of diabetes to the obstetrical world. He recorded patients in the obstetrical department of Johns Hopkins
at least 22 pregnancies in 15 mothers between the ages of Hospital, in 167 of whom sugar had been demonstrated by
21 and 38 (the data are confused in places): the mother Fehling’s solution. He concluded that 137 of these repre-
survived the pregnancy for long enough to become preg- sented definite postpartum lactosuria, being recognized
nant again in 9 instances, in 5 died at the delivery, and in only during lactation, and that almost all the others who had
6 within a few months. The cause of maternal death was been recognized in late pregnancy were similar. He was able
usually diabetic coma, although it is not possible to exclude to accurately distinguish glucose from lactose in a few cases
eclampsia, and some must also have developed puerperal and found only 2 of the 167 cases had definite glycosuria and
sepsis, and one died from exacerbation of tuberculosis. could thus be considered to have mild diabetes complicat-
Out of the 22 babies, 12 died, usually in utero, and they ing pregnancy. This may be the first evidence of screening
References 15
for gestational diabetes, suggesting a rather low prevalence treated with insulin (probably the more severe and often
in hospital practice in Baltimore, MD, nearly 100 years ago. referred cases), the perinatal mortality was still 57%.31 The
The major difficulty in the bedside measurement of 13 perinatal deaths included 6 stillbirths, 2 intrapartum
reducing sugars by Fehling’s test is no longer apparent, as deaths, and 5 early neonatal deaths; of the 10 living children,
all test strips now use a glucose oxidase system and recog- 3 were asphyxiated at birth, 1 weighed only 1500 g, and 1
nize only glucosuria (lactosuria will still occur but no longer was 5250 g. Histological examination of the pancreas in two
causes medical concern). Whitfield Williams then tabulated full-weight fetuses showed a pronounced increase in the size
all reported cases (81) of diabetes complicating pregnancy and number of the islets of Langerhans. Dr. Brandstrup,
from 1826 to 1907: he considered 15 cases to be doubtful, as who was in charge of these mothers’ care during that time,
glycosuria disappeared after delivery (including the famous set the scene for the future advances made by his successor
patient first reported by Bennewitz in 1826, although he had Dr. Jorgen Pedersen after the war.
not read the full case report in the original Latin). He calcu- Brandstrup noted that most of his patients had been
lated an overall immediate maternal mortality of 27%, with considered to be well adjusted with insulin treatment, but
an additional 23% of mothers dying within the following that they still had high levels of blood sugar for the greater
2 years. He concluded: “Pregnancy may occur in diabetic part of the day. He had previously undertaken physiologi-
women, or diabetes may become manifest during pregnancy; cal studies in pregnant rabbits on the passage of carbohy-
either is a serious complication, although the prognosis is drates across the placenta after intravenous injection and
not so alarming as is frequently stated.” had shown that while glucose and pentoses passed across
Joslin was the first internist to specialize in diabetes by a process of slow diffusion, the placental membrane was
and wrote the first textbook on the subject. In 1915, 6 years almost impermeable to disaccharides, including saccharose
before the discovery of insulin, he was able to describe and lactose.32 He described one case treated in 1927, illus-
seven personal cases of moderate or severe diabetes asso- trated by a 24-hour curve for blood sugar, who had been
ciated with pregnancy. He wished to take a more hopeful treated with two doses of insulin daily, felt well, and was
view, but admitted that little progress had been made. Of his looked upon as treated adequately, but he was unhappy with
seven cases, four were dead—one by suicide, one with ure- the level of control achieved:
mic manifestations (eclampsia), one of diabetic coma while
under the care of a clairvoyant, and the fourth, having sur- The blood sugar is seen to keep at very high levels through
vived one pregnancy with a healthy child, died of pulmo- a great part of the day. This feature is typical of the severe
nary tuberculosis 2 months after losing her second child. cases of diabetes under treatment with insulin, and it
But he was pleased that of the three remaining cases, one explains why the children are subject to intrauterine
was in exceptionally good health, free from sugar, and had a obesity through excessive supply of sugar also now in
normal child, another in a tolerable condition having been the epoch of insulin therapy. But these children are not
pregnant three times but with only one child now living, and only fat: they are large too. They present a condition of
the remaining case alive although severely ill with diabetes universal macrosomia … it seems probable that it is the
6 years after confinement. He closed his paper with an opti- maternal hyperglycaemia alone that brings about the
mistic comment: “It is certainly true that with the improve- pathologic–anatomical changes in the child.
ments in the treatment of diabetic patients [he meant strict
diet], diabetic women will be less likely to avoid pregnancy.”
The immediate postinsulin period was marked by some Conclusion
euphoria by both patients and their doctors, but it took
a long time for the very considerable fear of pregnancy to There is no doubt that had insulin not been discovered in
diminish and to some extent that fear remains to the pres- 1922, then the present-day outlook for successful pregnancy
ent day. A careful retrospective assessment of those early in a diabetic mother would still remain very poor because of
years of insulin at the Rigshospitalet in Copenhagen from continued maternal hyperglycemia, in spite of the enormous
1926 to 1938 showed that although there had been no mater- improvements in social, medical, and obstetrical care that
nal deaths in 22 pregnancies in 19 diabetic women mostly have occurred in the intervening years.
REFERENCES
1. Peel J. A historical review of diabetes and pregnancy. Obstet 6. Laurence RD, Oakley WG. Diabetic pregnancy. Q J Med 1942;
Gynaecol Br Comm 1972; 79: 385–395. 11: 45–54.
2. Reece EA. The history of diabetes mellitus. In: Reece EA, Coustan 7. Diamond J. Guns, Germs and Steel: The Fates of Human Societies,
DR, eds., Diabetes Mellitus in Pregnancy, 2nd ed., Churchill Norton & Co.: New York, 1997, p. 25.
Livingstone: New York, 1995, pp. 1–10. 8. Loudon I. Death in Childbirth: An International Study of Maternal
3. Banting FG, Best CH. The internal secretion of the pancreas. Care and Maternal Mortality 1800–1950, Clarendon Press:
J Lab Clin Med 1922; 7: 256–271. Oxford, U.K., 1992, pp. 1–622.
4. Bliss M. The Discovery of Insulin, Paul Harris Publishing: 9. Chamberlain G. Birth rates. In: Turnbull A, Chamberlain G, eds.,
Edinburgh, Scotland, 1983, pp. 20–58. Obstetrics, Churchill Livingstone: Edinburgh, Scotland, 1989,
5. Joslin EP. Pregnancy and diabetes mellitus. Boston Med Surg J pp. 1105–1110.
1915; 173: 841–849.
10. Turnbull A. Maternal mortality. In: Turnbull A, Chamberlain G, 20. Wilkerson HLC, Remein QR. Studies of abnormal carbohydrate
eds., Obstetrics, Churchill Livingstone: Edinburgh, Scotland, metabolism in pregnancy. Diabetes 1957; 6: 324–329.
1989, pp. 1121–1132. 21. O’Sullivan JB. Gestational diabetes. Unsuspected, asymptom-
11. Medical Research Council Conference on Diabetes and atic diabetes in pregnancy. N Engl J Med 1961; 264: 1082–1085.
Pregnancy. The use of hormones in the management of preg- 22. Hoet JP. Carbohydrate metabolism during pregnancy. Diabetes
nancy in diabetes. Lancet 1955; ii: 833–836. 1954; 3: 1–12.
12. Molsted-Pedersen L, Tygstrup I, Pederson J. Congenital malfor- 23. O’Sullivan JB, Mahan C. Criteria for the oral glucose tolerance
mations in newborn infants of diabetic women. Lancet 1964; i: test in pregnancy. Diabetes 1964; 13: 278–285.
1124–1126. 24. Mestman JH, Anderson GU, Barton P. Carbohydrate metabo-
13. International Clearinghouse for Birth Defects Monitoring Systems. lism in pregnancy. Am J Obstet Gynecol 1971; 109: 41–45.
Congenital Malformations Worldwide, Elsevier: Amsterdam, the 25. Pederson J. Diabetes og gravid: En introduktion. Ugeskr Laeger
Netherlands, 1991, pp. 1–8. 1951; 113: 1771–1777.
14. Kalter H. Of Diabetic Mothers and Their Babies: An Examination 26. Pedersen J. The Pregnant Diabetic and Her Newborn: Problems and
of Maternal Diabetes on Offspring, Perinatal Development Management, Munksgaard: Copenhagen, Denmark, 1967, p. 46.
and Survival, Harwood Academic Publishers: Amsterdam, the 27. HAPO Study Cooperative Research Group. The Hyperglycemia
Netherlands, 2000, pp. 95–111. and Adverse Pregnancy Outcome (HAPO) study. Int J Gynecol
15. Freinkel N. Of pregnancy and progeny. The Banting Lecture Obstet 2002; 78: 69–77.
1980. Diabetes 1980; 29: 1023–1035. 28. Bennewitz HG. De diabete mellito, gravidatatis symptomate, MD
16. Hadden DR. The development of diabetes and its relation to thesis, University of Berlin, Berlin, Germany, 1824 [translated into
pregnancy: The long-term and short-term historical viewpoint. In: English, deposited at the Wellcome Museum of the History of
Sutherland HW, Stowers JM, Pearson DWM, eds., Carbohydrate Medicine, Euston Road, London, U.K., 1987].
Metabolism in Pregnancy and the Newborn II, Springer-Verlag: 29. Duncan JM. On puerperal diabetes. Trans Obstet Soc Lond 1882;
London, U.K., 1989, pp. 1–8. 24: 256–285.
17. Miller HC. The effect of the prediabetic state on the survival of 30. Williams JW. The clinical significance of glycosuria in pregnant
the fetus and the birthweight of the newborn infant. N Engl J Med women. Am J Med Sci 1909; 137: 1–26.
1945; 233: 376–378. 31. Brandstrup E, Okkels H. Pregnancy complicated with diabetes.
18. Hurwitz D, Jensen D. Carbohydrate metabolism in normal preg- Acta Obstet Gynecol Scand 1938; 18: 136–163.
nancy. N Engl J Med 1946; 234: 327–329. 32. Brandstrup E. On the passage of some substances from mother
19. Gilbert JAL, Dunlop DM. Diabetic fertility, maternal mortality to fetus in the last part of pregnancy. Acta Obstet Gynecol Scand
and foetal loss rate. Br Med J 1949; i: 48–51. 1930; 10: 251–287.
3 Metabolism in normal pregnancy
Emilio Herrera and Henar Ortega-Senovilla
facilitated diffusion according to concentration-dependent

Introduction kinetics, thanks to the presence of a high number of glucose
During pregnancy, the mother adapts her metabolism to transporters, particularly GLUT1.11,12 The fetus does not syn-
ensure the continuous supply of nutrients to the fetus in thesize glucose, but uses it as its main oxidative substrate.
order to sustain its exponential growth. Of the nutrients that This causes fetal glycemia normally to be lower than that of
cross the placenta, glucose is quantitatively the most impor- its mother, allowing a positive maternal–fetal glucose gradi-
tant, followed by amino acids. Although lipids cross the pla- ent, which facilitates its placental transfer.
centa in much lower quantities, maternal lipid metabolism An increased contribution of carbohydrate to oxidative
is consistently and intensely affected during pregnancy in metabolism occurs in late pregnancy, which is commensurate
order to satisfy both maternal and fetal needs. Fetal growth with the increased rate of glucose production. The 24-hour
and development also depend on other essential nutrients, respiratory quotient (RQ) has been shown to be higher in
like vitamins. The metabolism of certain vitamins is there- late pregnancy than postpartum.13 Since glucose utilization
fore affected during pregnancy to ensure their proper avail- by maternal tissues during late pregnancy is reduced,14 the
ability to the fetus. The purpose of this chapter is to review higher RQ seems to reflect the obligatory use of glucose by
the main changes in carbohydrate, amino acid, lipid, and the fetus,2 which uses an estimated 17–26 g glucose/day in
vitamin metabolisms that take place throughout pregnancy late pregnancy,15 such values being well within the increment
under normal conditions. in carbohydrate oxidation found in pregnancy.
Carbohydrate metabolism Protein and amino acid metabolism

Glucose is the primary energy source of fetoplacental tissues. The accretion of protein is essential for fetal growth and
During early pregnancy, both glucose tolerance and insu- must be sustained by the active transfer of amino acids from
lin sensitivity are normal or greater than normal1 and the maternal circulation. There is no evidence that pregnant
insulin responses to oral glucose are also greater than nor- women store protein during early pregnancy, when fetal
mal,2 whereas hepatic basal glucose production is normal.3 needs are scarce. Therefore, the increased requirements
However, by the third trimester of pregnancy, a progressive of late pregnancy must be met by metabolic adjustments
insulin resistance develops,4,5 and an increase in basal and that increase both dietary protein utilization and nitrogen
postprandial insulin concentrations is seen.6 During late retention in order to satisfy fetal demands. Protein metabo-
pregnancy, the mother develops hypoglycemia, which is espe- lism changes gradually throughout gestation so that nitro-
cially conspicuous under fasting conditions, when the rate gen conservation for fetal growth achieves its full potential
of gluconeogenesis from different substrates is increased.7,8 during the last quarter of pregnancy.16,17 Nitrogen balance
The use of different substrates for the increased gluconeo- studies showed that the rate of maternal nitrogen retention
genesis is variable: the conversion of glycerol rather than between 20 and 40 weeks of gestation was greater than the
other more classical gluconeogenic substrates, like pyruvate predicted need,18 leading to the proposal that the mother
or alanine, to glucose is especially intense.9 The development gains additional protein in her own tissues. The increased
of maternal hypoglycemia, despite the increased gluconeo- nitrogen retention in late pregnancy is due to a reduction in
genesis and the reduced consumption of glucose by mater- urinary nitrogen excretion as a consequence of decreased
nal tissues caused by her insulin-resistant condition, is the urea synthesis,16 which is a reflection of amino acid oxida-
result of the high rate of placental transfer of glucose, which tion. It is therefore suggested that, during late pregnancy,
is greater than that of any other substrate.10,11 The predomi- there is a shift in the partitioning of amino acids toward
nance of glucose transfer by the placenta is brought about by net tissue deposition and away from oxidation.19 Thus, in
17
late pregnancy, nitrogen balance is modified to allow a more uteroplacental blood flow, and maternal concentrations of
efficient use of dietary proteins.20 amino acids, 24 all of which change as gestation advances
Although these alterations in protein metabolism dur- and are dependent on maternal health.26
ing late pregnancy favor nitrogen conservation, pregnancy
is associated with hypoaminoacidemia, which is especially
evident during fasting, is present at early gestation, and per- Lipid metabolism
sists throughout pregnancy.21,22 Since insulin infusion in
nonpregnant adults decreases both plasma amino acid lev- Accumulation of fat depots in maternal tissues and mater-
els and protein breakdown, it is proposed that the decrease nal hyperlipidemia are characteristic features during normal
in plasma amino acids found during normal pregnancy pregnancy. Although lipids cross the placenta with difficulty,
is not associated with the insulin resistance condition of essential fatty acids (EFA) and long-chain polyunsaturated
pregnancy. Thus, maternal hypoaminoacidemia reflects fatty acids (LCPUFA) are needed for fetal growth and devel-
increased placental amino acid uptake. Additionally, mater- opment and must arrive from maternal circulation. Thus,
nal oxidation of branched-chain amino acids decreases in throughout pregnancy, there are major changes in lipid
late pregnancy, increasing their availability for transfer to metabolism.
the fetus.23
In contrast to glucose, the concentrations of most amino
acids in fetal plasma are higher than those found in the Adipose tissue metabolism
mother, because placental transfer of amino acids is carried Fat accumulation takes place during the first two-thirds of
out by an active process, using selective transporters and gestation27,28 and represents an important contribution to
metabolic energy.24,25 This capacity to concentrate amino the increase in maternal body weight that take place dur-
acids on the fetal side of the placenta against the concentra- ing pregnancy.29 It is the result of both hyperphagia and
tion gradient is clearly seen in the fed and 24-hour fasted increased lipid synthesis and is driven by the greater respon-
rat. As shown in Figure 3.1, under fed conditions, maternal siveness to insulin by adipose tissue that occurs during early
total plasma amino acid levels are similar in 20-day pregnant pregnancy.30 During this early stage of pregnancy, there is
rats and sex- and age-matched virgin animals, whereas the also an increase in adipose tissue lipoprotein lipase (LPL)
levels in fetal plasma are already higher than in the mother. activity as reported in women31 and in rats,32 which cata-
However, after fasting, the decline of plasma amino acids lyzes the hydrolysis of circulating triacylglycerols (TAGs)
in the late pregnant rat is greater than that seen in virgin that are carried in TAG-rich lipoproteins (i.e., chylomicrons
animals, whereas fetal total plasma amino acid concentra- and very-low-density lipoproteins [VLDLs]). The hydrolytic
tion remains the same as when fed. Thus, under fasting, the products, nonesterified fatty acids (NEFAs) and glycerol, are
fetal/maternal total amino acid ratio becomes even higher taken up by the subjacent tissue,33 and overall these changes
than when fed, showing the efficiency of the placenta in facilitate the accumulation of lipids in maternal depots.
transferring amino acids against the gradient. A multiplic- Further increments in maternal fat depots cease dur-
ity of factors affects the overall placental amino acid deliv- ing the third trimester of gestation as a consequence of two
ery rates, including the activity and location of the amino changes: (1) a decrease in the LPL activity,34 which mainly
acid transporter systems, changes in placental surface area, corresponds to that present in adipose tissue35 and causes a
decline in the hydrolysis and uptake of TAG circulating in
the TAG-rich lipoproteins, and (2) an increased activity of
10 enzymes that cause the lipolysis of adipose tissue TAG stores,
b b which is especially manifest under fasting conditions.36,37
8 The placental transfer of the products of adipose tissue
Concentration of plasma
lipolysis released into the circulation, NEFA and glycerol, is

amino acids (mM)
6 a quantitatively low,38 and therefore their main destination is

c a
maternal liver. In liver, NEFAs are converted into acyl-CoA,
4 d and glycerol into glycerol-3-phosphate, which are partially
reesterified for the synthesis of TAG. These are released back
2 into the circulation in the form of VLDL, as maternal liver
production is increased. In addition, glycerol is also used
0 as a preferential substrate for gluconeogenesis, and NEFAs
Virgin Pregnant Fetus are used for β-oxidation, leading to energy production and
ketone body synthesis. These pathways are markedly increased
Fed 24 hour fasted
under fasting conditions in late pregnancy.8,39 Ketone bodies
easily cross the placenta.40 Although not synthesized by the
Figure 3.1 Plasma concentration of total amino acids in fed
and 24-hour fasted virgin and 20-day pregnant rats and their
fetus, they reach the same concentration in fetal circulation as
fetuses. Letters above each bar correspond to the statistical in the mother’s.41 In contrast to what occurs in adults, ketone
comparison between the groups: different letters indicate bodies can be used by the fetus not only as energetic fuels but
statistical differences (p < 0.05). also as substrates for brain lipids.42,43
Lipid metabolism 19
Liver
NEFA Ketone bodies Muscle

TAG
Glycerol Glucose
Adipose
tissue
CO2 + ATP
Proteins
Amino
Maternal acids
glucose-dependent
tissues
Fetus
Figure 3.2 Schematic representation of maternal response to starvation during late pregnancy. Enhanced adipose tissue lipolysis
increases the availability of glycerol in the liver, where it is used as the preferential substrate for gluconeogenesis, and of non-
esterified fatty acids (NEFA) for ketogenesis. This mechanism enables the mother to preserve other gluconeogenic substrates,
such as amino acids (mainly, alanine), and ensures their availability to the fetus while producing glucose for the fetus and her
own needs. Note: TAG, triacylglycerols.
Thus, as shown in Figure 3.2, both the mother and the fasting conditions, where maternal hypoglycemia seems to
fetus benefit from the increased adipose tissue lipolytic be responsible for an increase in catecholamine production44
activity during late pregnancy, especially during periods of and consequently for increasing the breakdown of fat depot
fasting. The preferential conversion of glycerol to glucose and promoting the use of NEFA and glycerol by maternal tis-
allows the preservation of other gluconeogenic substrates sues while preserving glucose and amino acids for the fetus.
like alanine and other amino acids for transfer to the fetus.
The active production of ketone bodies from fatty acids by
fasting maternal liver allows both their transfer to the fetus Hyperlipidemia
and their use by maternal tissues such as skeletal muscle as Hyperlipidemia normally develops during the last third of
alternative fuels. Ketone bodies also save glucose for its use gestation and mainly corresponds to increases in TAGs, with
by maternal tissues, like the nervous system that depends smaller rises in phospholipids and cholesterol.29,34 As well as
on glucose, and for its transfer by the placenta. the increase in VLDL levels described earlier, the increase
Pregnancy hormones may contribute to some of the in plasma TAGs also results from their proportional enrich-
changes taking place in adipose tissue metabolism. Figure 3.3 ment in both low-density lipoprotein (LDL) and high-density
summarizes the main pathways of adipose tissue metabo- lipoprotein (HDL),34 lipoproteins that are normally poor in
lism, showing how representative hormones exert active con- TAGs. Such changes in the maternal lipoprotein profile and
trol on them. Insulin augments LPL activity and therefore composition are the result of the combined action of several
increases the uptake of circulating TAG-rich lipoproteins, factors, which are schematically summarized in Figure 3.4:
whereas it decreases the adipocytes’ lipolytic activity. This (1) the arrival of larger amounts of the adipose tissue lipolytic
later effect is achieved by increasing the conversion of cAMP products, NEFA and glycerol, at the liver, which facilitates the
into AMP and consequently decreasing the activator of the hepatic synthesis of TAGs, and their subsequent release into
protein kinase A, which is responsible for the breakdown of the circulation as VLDL; (2) a decreased removal of VLDL
intracellular TAG through the phosphorylation and activa- from the circulation as a consequence of the reduced adi-
tion of several key proteins like perilipin, adipocyte TAG pose tissue LPL activity; (3) an increase in cholesteryl ester
lipase, and hormone-sensitive lipase (HSL). transfer protein activity that takes place at midgestation,45
On the other hand, lipolytic hormones increase the activ- facilitating the transfer of cholesterol from LDL and HDL
ity of the lipolytic cascade by increasing the activity of ade- to VLDL in exchange for TAGs; and (4) the intense decrease
nylate cyclase and consequent production of cAMP, as is the in hepatic lipase (HL),34 which results in less conversion of
case for glucagon and catecholamines, by interacting with the relatively buoyant HDL2b TAG-rich particles into smaller
their respective receptors (Figure 3.3). TAG-poor and cholesterol-rich particles (HDL3), allowing
In early pregnancy, increased estrogen, progesterone, and the former to accumulate in the circulation.34
insulin favor lipid deposition and inhibit lipolysis. However, Plasma estrogen increases hepatic VLDL production46
in late pregnancy, placental hormones like human chorionic and decreases HL expression and activity.47 Consequently,
somatotropin add to the effect of the insulin resistance by the increase in its concentration during gestation will add
promoting lipolysis and fat mobilization. This shift from to the effects of the insulin resistance48 and contribute to the
an anabolic to a catabolic state is further accelerated under maternal hypertriacylglycerolemia.
Catecholamines
Insulin
Glucagon +
AC
+
α γ
β
ATP β-AR
(IRS/PKI3-Akt)
+ PDE
TAG cAMP
+
LPL 5’AMP P
VLDL
PKA
PKA
P
NEFA P
ATGL HSL
ACS P P
ACIL-CoA Perilipin
P P P
ATGL HSL MAGL

MAG MGAT TAG MAG P
DAG Glycerol
DAG TAG NEFA

P
DGAT
P
P P
P
NEFA-AFABP P Glycerol
NEFA-albumin
Figure 3.3 Overview of adipose tissue lipolysis and the uptake of circulating triacylglycerol (TAG)-rich lipoproteins (mainly very-
low-density lipoprotein [VLDL]) and their control by insulin, glucagon, and catecholamines. The binding of glucagon or catechol-
amines to their respective receptors (in case of catecholamines to the β-adrenergic receptors, β-AR) activates adenylate cyclase (AC)
and increases the production of cyclic AMP (cAMP). This leads to the activation of protein kinase A (PKA), which phosphorylates
and activates several key proteins including perilipin, adipocyte triacylglycerol lipase (ATGL), and hormone-sensitive lipase (HSL).
Phosphorylation of perilipin modifies its barrier function, resulting in a restructuring of the surface of the lipid droplet, and allows
access by the lipases to the lipid droplets. The translocation of ATGL and HSL from the cytosol to the lipid droplet results in the
sequential hydrolysis of TAGs to diacylglycerols (DAG) and then to monoacylglycerols (MAG), which are hydrolyzed by monoac-
ylglycerol lipase (MAGL) producing nonesterified fatty acids (NEFAs) and glycerol. Glycerol is directly released into the circulation;
the NEFAs are bound to adipocyte fatty acid binding protein, which facilitates their intracellular trafficking from the droplet surface
to the plasma membrane as part of their efflux from the cell. In plasma, the NEFAs circulate bound to other binding proteins (fatty
acid binding proteins), of which albumin is the most abundant. Insulin also has many effects, most of which oppose the effects of
glucagon and catecholamines. It stimulates lipoprotein lipase (LPL) activity, increasing the hydrolysis of circulating TAGs of TAG-
rich lipoproteins (mainly VLDL), to facilitate the uptake of released NEFA, which are converted into TAG inside the cell, and has an
antilipolytic action by its activation on the phosphodiesterase (PDE) activity, which converts cAMP into 5′-AMP. This mechanism
reduces intracellular concentrations of cAMP, resulting in an inhibition of lipolysis. Conditions causing an insulin resistance, as is
the case in late pregnancy, prevent these two effects of insulin, i.e., they decrease LPL activity and increase lipolytic activity. These
actions are exaggerated at late pregnancy by the actions of catecholamines and lipolytic hormones produced by the placenta. ACS,
acyl-CoA synthetase; MGAT, acyl-CoA monoacylglycerol acyltransferase; DGAT, acyl-CoA diacylglycerol acyltransferase.
Benefits of maternal hypertriacylglycerolemia to the fasting conditions, the maternal liver switches from an
fetus and newborn exporter organ to an importer of plasma TAGs, which
Although maternal TAGs do not directly cross the placenta,49 may be used as substrates for ketogenesis. This allows the
we think that there are several ways by which the fetus and exaggerated increase in plasma ketone bodies, which, as
newborn may benefit from maternal hypertriacylglycerolemia: discussed earlier, save glucose in maternal tissues and
cross the placental barrier providing a substrate for the
1. Use of TAGs as metabolic fuels. Although adult liver fetus.
does not normally express LPL activity, studies in the 2. Placental transfer of polyunsaturated fatty acids (PUFA).
rat have shown that under fasting conditions during late EFA from the maternal diet and LCPUFA, derived either
pregnancy, there is a marked increase in liver LPL activ- from the diet or by endogenous synthesis from EFA, are
ity.50 This liver LPL seems to be the result of the washout mainly transported in their esterified form in mater-
of LPL from extrahepatic tissues carried out by the rem- nal plasma lipoproteins rather than as NEFA.51 The
nants of the TAG-rich lipoproteins. In this way, under placenta expresses receptors for all the major plasma
Lipid metabolism 21
Liver
NEFA
TG TG
Glycerol
Adipose
tissue
LPL
VLDL
(TG)
TG TG
CE CE
CETP
LDL CE HL
HDL2b HDL3a
(TG)
TG
EFA, LCPUFA
Placenta
Fetus
Figure 3.4 Schematic representation of the relationship of adipose tissue lipolytic activity with lipoprotein metabolism during
late pregnancy, and its role as a source of essential (EFA) and long-chain polyunsaturated fatty acids (LCPUFA) for the fetus. CE,
cholesterol esters; CETP, cholesterol ester transfer protein; HDLs, high-density lipoproteins; LDLs, low-density lipoproteins; VLDLs,
very-low-density lipoproteins; LPL, lipoprotein lipase; HL, hepatic lipase; TG, triacylglycerols.
lipoproteins. It has different lipolytic activities, includ- Adipocytokines and fetal growth
ing LPL, phospholipase A 2, and an intracellular lipase, Adipose tissue secretes several specific proteins called adi-
and it also expresses fatty acid binding proteins (for pocytokines that play critical roles in energy homeostasis
a review, see Herrera et al.52). Thus, esterified PUFA in in adults, and some of them also modulate insulin action in
maternal plasma lipoproteins are taken up either intact different tissues.54,55 Changes in the plasma levels of some of
through the placenta’s receptors or after hydrolysis as these adipocytokines during pregnancy are involved in the
their constituent fatty acids. Within the placental tissue, maternal changes of insulin sensitivity and are known to
the fatty acids are reesterified to be subsequently hydro- contribute to fetal growth. Here we will describe four adipo-
lyzed along with those taken up intact and finally dif- cytokines that are known to possess this property: adipocyte
fused into the fetal plasma by a mechanism that is not fatty acid binding protein (AFABP), leptin, adiponectin, and
yet completely understood. This process, together with retinol binding protein 4 (RBP4).
the direct transfer of NEFA and the intrinsic placental AFABP is responsible for the intracellular trafficking of
fatty acid metabolism, determines the actual rate of the fatty acids and interacts with HSL increasing its lipolytic
selective placental fatty acid transfer, which is essential activity.56 This novel adipocytokine is present in human
for fetal development. serum55 and its concentrations are associated with insulin
3. Contribution to milk synthesis in preparation for lactation. resistance and type 2 diabetes,57 being higher in pregnant
Around parturition, there is a rapid rise in mammary gland women with gestational diabetes mellitus (GDM) than
LPL activity53 that, together with the low LPL activity in in control subjects.58 The concentration of AFABP in cord
adipose tissue,35 drives circulating TAGs to the mammary serum of control subjects is higher than in the correspond-
gland. This caused a rapid disappearance of maternal hyper- ing maternal serum, and in GDM fetuses, values correlate
triacylglycerolemia,34 and EFA and LCPUFA from maternal with neonatal fat mass.58
circulation are taken up by the mammary gland for milk Leptin is produced by adipose tissue as well as by other
synthesis to become available to the suckled newborn, con- tissues, including the placenta,59,60 its concentration in
tributing to its normal development. maternal plasma being increased during late pregnancy.60
Besides increasing adipose tissue lipolysis, leptin modulates DNA and RNA synthesis. Thus, it is obvious that these
glucose metabolism and insulin sensitivity, and maternal folate-dependent reactions are essential for fetal growth and
hyperleptinemia seems to contribute to the availability of development and for maternal well-being.
nutrients to the fetus. Although the relationship between Pregnancy is associated with an increased folate demand
fetal leptin and birth weight has not been clearly established, and, in some cases, leads to overt folate deficiency. The increase
leptin concentration in serum from the umbilical vein has in folate requirement during pregnancy is due to the growth
a strong positive correlation with neonatal fat mass61 and is of the fetus and uteroplacental organs. Circulating folate con-
considered a marker of adiposity in the fetus. centrations decline in pregnant women who are not supple-
RBP4 is another adipocytokine that has been shown to mented with folic acid.68 Possible causes for the declines in
regulate glucose metabolism by reducing insulin sensitiv- blood folate include an increased folate demand for the fetus,
ity.62 During pregnancy, RBP4 is also synthesized by the increased folate catabolism, increased folate clearance and
placenta, and its concentration in maternal plasma increases excretion, decreased folate absorption, hormonal influence on
progressively with gestational age as the placenta expands.63 folate metabolism as a physiological response to pregnancy,
The concentration of RBP4 in maternal and cord blood and low folate intake.69,70 There is a strong positive associa-
serum is higher in pregnant women with GDM than in con- tion between folate concentrations in maternal plasma, cord
trols,58 and it has been proposed that this contributes to the plasma, and placenta, suggesting that transplacental folate
higher insulin resistance found in women with GDM. delivery depends upon maternal plasma folate concentrations.
Adiponectin is another adipocytokine, but unlike the The transfer of 5-methyltetrahydrofolate (the main form of
foregoing examples, it increases insulin sensitivity. During folate in plasma) from the maternal to the fetal perfusate is
normal pregnancy, no significant alterations in adiponectin not saturable in a range well above typical physiological con-
levels have been found,64 but its levels negatively correlate centrations71; also it is favored by the folate receptor, which is
with insulin resistance65 and they are lower in women with abundant in the epithelial cells of placenta.72
GDM than in control women.66 Cord blood adiponectin Studies conducted in recent years led to the recognition
concentrations increase with gestational age and correlate that supplementing the maternal diet with folic acid reduced
positively with birth weights,67 suggesting a possible role in the prevalence of folate deficiency in pregnancy and pre-
fetal adiposity and development. However, cord blood adipo- vented pregnancy-related disorders. Data from these stud-
nectin concentrations in women with GDM are lower than ies suggest that 200–300 μg folic acid per day is needed in
in controls, and no relationship is found between this vari- addition to dietary folate to maintain normal folate status
able and neonatal fat mass or birth weight.58 and to prevent folate deficiency during pregnancy.73,74
Vitamin C
Vitamin metabolism in pregnancy
In addition to the prevention of scurvy, vitamin C (or ascor-
Adequate maternal micronutrient and vitamin status is bic acid) has numerous other functions and is a cofactor for
especially critical during pregnancy and lactation. Several several enzyme systems. For humans, ascorbic acid is an
micronutrient deficiencies (like iron, iodine, zinc) are well essential vitamin, with an important antioxidant function.
established as contributors to abnormal prenatal develop- As antioxidant defense systems are important to protect tis-
ment or pregnancy outcome. Less well recognized for their sues and cells from damage caused by oxidative stress, an
importance are deficiencies of vitamins. Evidence is accu- imbalance between increased oxidative stress and decreased
mulating that maternal antioxidant status is important to antioxidant defenses impairs fetal growth.75 Thus, pregnant
prevent abnormal pregnancy outcomes. In lactation, the women utilize a defense mechanism, composed of antioxi-
maternal status of several of these vitamins affects their dant enzymes and nutrients including vitamin C, against
concentration in breast milk. The main cause of multiple oxidative stress and free radical damage. It is believed that
vitamin deficiencies is a poor-quality diet, even though gene ascorbic acid, after conversion to dehydroascorbic acid,
polymorphism can also impair vitamin absorption or alter crosses the placenta to enter fetal circulation. Once in the
their metabolism and cause vitamin deficiency. fetal circulation, dehydroascorbic acid is reduced back to
We summarize here the changes in the metabolism of the ascorbic acid and is maintained in high concentrations on
vitamins during pregnancy that have the highest implica- the fetal side in the placenta.76 Maternal serum vitamin C
tions in fetal growth and development. levels during the second trimester of gestation are correlated
with birth weight and length in full-term babies.77
Hydrophilic vitamins
Folic acid
Lipophilic vitamins
Folates act in a number of one-carbon transfer reactions, Because lipophilic vitamins are fat soluble, they share
including purine and thymidylate biosynthesis, amino acid several common mechanisms with other lipophilic sub-
metabolism, and formate oxidation. Purine and thymidy stances in terms of their metabolism and transfer to the
late biosyntheses are fundamental requirements underlying offspring. Although lipophilic vitamins are essential during
Lipophilic vitamins 23
intrauterine and early postnatal life, little is known about Vitamin A

their placental transfer during pregnancy or their uptake by Vitamin A exists in several forms in animal tissues: retinol,
mammary gland during lactation. retinal, retinoic acid, and retinyl esters, mainly as retinyl
palmitate. All forms of vitamin A are hydrophobic com-
pounds, which are highly unstable in the presence of O2.
Vitamin D A diet deficient in either retinol or the provitamin A carot-
Vitamin D metabolites have numerous potential physiological enoids that can be metabolized to retinol results in impaired
and pharmacological actions, but their principal physiologi- growth, night blindness, and, ultimately, xerophthalmia
cal function is maintaining serum calcium and phosphorus and blindness. We now know that there are two metabolites
concentrations in a range that supports cellular processes, of vitamin A, retinoic acid and retinal, that are responsible
neuromuscular function, and bone mineralization. for growth and development by regulating gene expression,
Significant changes in maternal vitamin D and calcium whereas retinal and its isomers are responsible for the visual
metabolism occur during pregnancy to provide the calcium function of vitamin A.
needed for fetal bone mineral accretion. Fetal concentra- During pregnancy, concentrations of retinol in maternal
tions of 1,25-dihydroxy-vitamin D (1,25(OH)2D3, the active plasma fall as gestation advances80 (Figure 3.5), which prob-
metabolite of vitamin D3) are low, whereas maternal levels ably reflects the increasing demands of the rapidly growing
are strikingly elevated during pregnancy.78 This increase maternal and fetal tissues. It has been shown that retinol
in maternal 1,25(OH)2D3 levels appears to be caused by plays an essential role in the development of organs such as
increased production rather than decreased clearance, but the lungs, heart, and skeleton, while retinoic acid enables the
the precise source of the increased 1,25(OH)2D3 synthesis setting up of the vascular and nervous system and is involved
has yet to be fully defined. 1,25(OH)2D3 can pass through as a morphogenic agent during embryonic development.81,82
the placental barrier in either direction to sustain the active Multiple fetal anomalies occur in vitamin A deficiency, as
transport of calcium across the placenta during late gesta- well as in knockout mice deficient in the receptor for retinoic
tion. Approximately, 25–30 g of calcium is transferred to the acid. An excess of vitamin A also induces the same type of
fetal skeleton by the end of pregnancy, most during the last abnormality: the importance of the abnormality depends on
trimester. It has been estimated that the fetus accumulates the period of gestation and the duration of the excessive or
up to 250 mg/dL calcium during the third trimester. The deficient supply.83
three possible calcium sources that could supply the mother The mechanism of transfer of vitamin A from mother to
with the necessary calcium to support fetal growth include fetus remains unknown, but it allows vitamin A levels in the
increased intestinal absorption from the diet, increased fetus to remain unaffected by alterations in maternal vitamin
renal conservation, and increased bone mobilization.79 A status, except in conditions of deficiency or excess.84 The
To date, there is no evidence to indicate a beneficial placenta’s vitamin A content increases in the last trimester of
effect of vitamin D intake during pregnancy above amounts pregnancy, thanks to the supply of vitamin A from maternal
routinely required to prevent vitamin D deficiency among stores (i.e., liver).85 Studies in rats showed that in early gestation,
nonpregnant women. maternal RBP (retinol binding protein) is transported across
40 2
c
a a
35
a 1.5
a
b
30 b
Vitamin E (µmol/L)
Retinol (µmol/L)
bd
1
a
25 ad
0.5
20
0 0
First Second Third Postpartum Postlactation
trimester trimester trimester
Figure 3.5 Plasma levels of vitamin E (α- and γ-tocopherol) and vitamin A (retinol) in the different trimesters of pregnancy,
6–8 days postpartum and postlactation, in healthy women. Data are expressed as means ± SEM. Statistical comparisons between
the different times are shown by the letters above the points. Different letters for the corresponding vitamin between the groups
indicate statistical significance (p < 0.05).
the placenta and delivers retinol, whereas in late gestation, a which circulates as its nonesterified alcohol form in serum
different mechanism appears to be operating because fetal lipoproteins. Changes in plasma α-tocopherol levels during
liver is capable of synthesizing RBP.86 Studies in vivo show that pregnancy parallel maternal hyperlipidemia (see earlier in
maternal RBP does not cross the placental membrane barrier this chapter). In contrast, γ-tocopherol reaches a maximum
in the last trimester of gestation and cannot enter fetal circula- concentration in maternal plasma at midgestation. The rea-
tion,87 and homozygous RBP-null mutant mice are viable and son for the different concentration patterns (between α- and
fertile.88 In humans, serum apo-RBP (retinol-free) concentra- γ-tocopherol) during pregnancy is unknown, but could be
tion appears to be elevated during pregnancy, suggesting that related to differences in their tissue uptake and intracellular
pregnancy may alter the affinity of RBP for retinol or induce metabolism.
the binding of the vitamin to other uncharacterized proteins.89 α-Tocopherol concentration in the plasma of human
Moreover, other forms of vitamin A, such as retinyl esters and fetuses is lower than in their mothers, but rises toward the
retinoic acid, can also be taken up at the placental barrier. end of pregnancy. Since α-tocopherol is carried in plasma
Nevertheless, there are no significant correlations between mainly by VLDL and LDL, its uptake and handling by the
maternal and cord plasma concentrations of retinol or carot- placenta are similar to that of the other lipoprotein lipophilic
enoids under normal conditions, although some authors components (see earlier in this chapter). The placenta also
report a weak but statistically significant correlation when expresses α-tocopherol transfer protein (α-TTP), and in a
the concentrations of retinol in cord and maternal plasma manner similar to its role in liver, it may actively contribute
are low.90 Published studies in humans show that maternal to the specific transfer of α-tocopherol to the fetus.98,99
subclinical vitamin A deficiency is related to neonatal sub- Despite the existence of all these processes, efforts to inves-
clinical vitamin A deficiency and to low birth weight90,91; tigate the kinetics of the transfer of vitamin E by isolated
furthermore, adequate vitamin A at delivery is positively human placental systems have found that it is specific for
associated with birth outcomes and children’s neurodevel- the natural stereoisomer, RRR-α-tocopherol, rather than any
opment in later childhood.92 other form of vitamin E, but that its rate is very low. This justi-
The situation with vitamin A in early lactation is peculiar. fies the consistent finding of much lower α-tocopherol levels in
Because of the limited transplacental transfer, the infant’s fetal plasma and red blood cells than in maternal ones.
liver stores of vitamin A at birth are small even in well- During lactation, vitamin E intake through milk is the
nourished populations, so newborns are highly dependent on way of supplying the newborn with an essential defense
dietary intake of the vitamin to establish proper tissue stores, against oxygen toxicity and of stimulating the development
maintain rapid growth, and develop their immune system. of its immune system. A good supply of vitamin E to the
Colostrum contains a higher vitamin A concentration than offspring is therefore particularly critical in this period. The
milk and has an important role to play in providing the ini- increase in vitamin E content in body tissues of the offspring
tial protection from vitamin A deficiency in the newborn.93 following birth is attributed to the ingestion of colostrum and
The timing of colostrum ingestion seems to play a role in the milk, emphasizing the limited placental vitamin E transfer
efficiency of intestinal vitamin absorption, thus colostrum and the importance of milk consumption. Colostrum con-
feeding on the day of birth is important for the establish- tains a higher concentration of vitamin E than does milk,100
ment of absorptive mechanism allowing intestinal transport which may imply an active uptake by the mammary gland in
of fat-soluble vitamins. Moreover, breast milk is also a good compensation for the limited placental transport. A decline
source of vitamin A and clinical vitamin A deficiency is rare in maternal circulating vitamin E concentration is noticed at
in breast-fed infants during their first year of life. the end of gestation or in early lactation; this decrease coin-
Debate surrounds the use of retinol supplements during cides with the high concentration of α-tocopherol present
pregnancy. The use of retinol supplements in well-nourished in colostrum, and the two events are probably related. The
mothers does not affect fetus concentrations. High doses mechanism of transfer from blood into milk is not com-
of retinol are teratogenic, and in some countries, pregnant pletely understood. Perhaps the transfer of vitamin E into
women are advised to avoid retinol-containing supple- milk occurs through a protein-mediated transport: the pres-
ments.94 However, this advice may lead to vitamin A defi- ence of an α-TTP-like mechanism in the mammary gland
ciency.95 Serum retinol is a relatively insensitive indicator of cannot be excluded nor can the presence of a lipoprotein
body vitamin A status: only 1% of the body’s reserves cir- scavenger receptor class B1 (SR-BI receptor) in the mammary
culate in the plasma, and homeostatic mechanisms control gland, which could be involved in the uptake of α-tocopherol
concentrations via retinol binding protein concentrations. from HDL. In addition, the high concentration of vitamin E
found in colostrum compared to mature milk might be due
to an increase in the activity of mammary LDL receptors and,
Vitamin E thus, to an important uptake of LDL by the mammary gland
Dietary vitamin E is present as tocopherol, mainly α- and around parturition. LPL also seems to modulate the mam-
γ-tocopherol, and tocopheryl esters. Plasma levels of vitamin mary gland uptake of α-tocopherol from VLDL and chylo-
E in the mother increase significantly from the first trimes- microns.101 In contrast to placental transfer of tocopherols,
ter of gestation to reach a maximum in the third trimester which remains low even when maternal serum levels are high,
of gestation (Figure 3.5).80,96,97 Unlike vitamin A, there is no the transfer through colostrum and milk can be increased by
specific carrier protein in the serum to transport vitamin E, increasing the ingestion of vitamin E by the mother.
References 25
It is important to note that tocopherol is able to affect the main destination of released NEFA and glycerol is the
the metabolism of vitamin A in several tissues and may play liver, where they are used for the synthesis of TAGs. The glyc-
a role in tissue retinol homeostasis. It has been shown to erol is also used as a gluconeogenic substrate and the NEFAs
modulate the levels of retinol and total vitamin A in tissues are also oxidized for ketogenesis.
such as the liver, kidney, and intestine. In vitro, tocopherol Adipocytokines secreted by adipose tissue play critical
exerts an inhibiting or stimulating action (depending on the roles in energy homeostasis in adults, and some of them
tissue) on the hydrolysis of retinyl palmitate. modulate insulin action. Circulating levels of some of these
adipocytokines are modified during pregnancy, are involved
in the maternal changes of insulin sensitivity, and are
Summary appeared to contribute to fetal growth. This is the case for
AFABP, leptin, adiponectin, and RBP4.
Maternal metabolic adaptations during pregnancy are Major changes also occur in vitamin metabolism.
mainly directed toward maintaining a continuous availabil- Vitamins A and E are the most affected. Maternal plasma
ity of substrates to enable fetal growth. Glucose, used as a retinol falls as gestation progresses, whereas vitamin E
primary energy source of fetoplacental tissues, is quantita- levels increase parallel to the increase in plasma lipids.
tively the most important substrate crossing the placenta. Transplacental transfer of these vitamins is limited, but both
During late pregnancy, the mother develops hypoglycemia the fetus and the newborn need them. They are both taken
as a result of the high rate of placental transfer, despite there up by the mammary gland, and their high concentration in
being more active gluconeogenesis and a reduced consump- colostrum seems to play an important role in the extrauter-
tion of glucose. ine adaptations of the suckled newborn.
Amino acids cross the placenta against the concentra-
tion gradient, thanks to an active process. Fetal growth is
sustained by the transfer of amino acids from the maternal Acknowledgments
circulation. Protein metabolism changes gradually through-
out gestation, and although during late pregnancy there is The results reported in this chapter were obtained
increased nitrogen retention, the mother develops hypoami- with grants from the Spanish Ministry of Science and
noacidemia, which is especially evident during fasting. Innovation (Grant SAF2012-39273) and Fundación
Fat depot accumulation and maternal hyperlipidemia are Ramón Areces of Spain (CIVP16A1835). The authors
characteristic features of pregnancy. During late pregnancy, thank pp-science-editing.com for editing and linguistic
maternal adipose tissue lipolytic activity is increased, and revision of the chapter.
REFERENCES
1. Catalano PM, Tyzbir ED, Roman NM, Amini SB, Sims EA. 10. Herrera E, Amusquivar E. Lipid metabolism in the fetus and the
Longitudinal changes in insulin release and insulin resistance newborn. Diabetes Metab Res Rev 2000; 16: 202–210.
in nonobese pregnant women. Am J Obstet Gynecol 1991; 165: 11. Hay WW, Jr. Placental-fetal glucose exchange and fetal glucose
1667–1672. metabolism. Trans Am Clin Climatol Assoc 2006; 117: 321–339.
2. Butte NF. Carbohydrate and lipid metabolism in pregnancy: 12. Jansson T, Wennergren M, Illsley NP. Glucose transporter pro-
Normal compared with gestational diabetes mellitus. Am J Clin tein expression in human placenta throughout gestation and in
Nutr 2000; 71: 1256S–1261S. intrauterine growth retardation. J Clin Endocrinol Metab 1993;
3. Catalano PM, Tyzbir ED, Wolfe RR et al. Longitudinal changes in 77: 1554–1562.
basal hepatic glucose production and suppression during insulin 13. Butte NF, Hopkinson JM, Mehta N, Moon JK, Smith EO.
infusion in normal pregnant women. Am J Obstet Gynecol 1992; Adjustments in energy expenditure and substrate utilization
167: 913–919. during late pregnancy and lactation. Am J Clin Nutr 1999; 69:
4. Buchanan TA, Metzger BE, Freinkel N, Bergman RN. Insulin 299–307.
sensitivity and B-cell responsiveness to glucose during late preg- 14. Williams CM, Pipe NG, Coltart TM. A longitudinal study of
nancy in lean and moderately obese women with normal glu- adipose tissue glucose utilization during pregnancy and the
cose tolerance or mild gestational diabetes. Am J Obstet Gynecol puerperium in normal subjects. Hum Nutr Clin Nutr 1986; 40:
1990; 162: 1008–1014. 15–23.
5. Ryan EA, O’Sullivan MJ, Skyler JS. Insulin action during preg- 15. Hay WW, Jr. Placental supply of energy and protein substrates to
nancy. Studies with the euglycemic clamp technique. Diabetes the fetus. Acta Paediatr Suppl 1994; 405: 13–19.
1985; 34: 380–389. 16. King JC. Physiology of pregnancy and nutrient metabolism. Am J
6. Lesser KB, Carpenter MW. Metabolic changes associated with Clin Nutr 2000; 71: 1218S–1225S.
normal pregnancy and pregnancy complicated by diabetes mel- 17. Kalhan SC. Protein metabolism in pregnancy. Am J Clin Nutr
litus. Semin Perinatol 1994; 18: 399–406. 2000; 71: 1249S–1255S.
7. Kalhan S, Rossi K, Gruca L, Burkett E, O’Brien A. Glucose turn- 18. Calloway DH. Nitrogen balance during pregnancy. In: Winick
over and gluconeogenesis in human pregnancy. J Clin Invest M, ed., Nutrition and Fetal Development, John Wiley & Sons:
1997; 100: 1775–1781. Philadelphia, PA, 1974, pp. 79–94.
8. Herrera E, Knopp RH, Freinkel N. Carbohydrate metabolism in 19. Duggleby SL, Jackson AA. Protein, amino acid and nitrogen
pregnancy. VI. Plasma fuels, insulin, liver composition, gluconeo- metabolism during pregnancy: How might the mother meet the
genesis, and nitrogen metabolism during late gestation in the fed needs of her fetus? Curr Opin Clin Nutr Metab Care 2002; 5:
and fasted rat. J Clin Invest 1969; 48: 2260–2272. 503–509.
9. Herrera E. Metabolic adaptations in pregnancy and their implica- 20. Mojtahedi M, de Groot LC, Boekholt HA, van Raaij JM. Nitrogen
tions for the availability of substrates to the fetus. Eur J Clin Nutr balance of healthy Dutch women before and during pregnancy.
2000; 54: S47–S51. Am J Clin Nutr 2002; 75: 1078–1083.
21. Metzger BE, Unger RH, Freinkel N. Carbohydrate metabolism 44. Herrera E, Knopp RH, Freinkel N. Urinary excretion of epineph-
in pregnancy. XIV. Relationships between circulating glucagon, rine and norepinephrine during fasting in late pregnancy in the
insulin, glucose and amino acids in response to a “mixed meal” rat. Endocrinology 1969; 84: 447–450.
in late pregnancy. Metabolism 1977; 26: 151–156. 45. Iglesias A, Montelongo A, Herrera E, Lasuncion MA. Changes in
22. Schoengold DM, de Fiore RH, Parlett RC. Free amino acids in cholesteryl ester transfer protein activity during normal gestation
plasma throughout pregnancy. Am J Obstet Gynecol 1978; 131: and postpartum. Clin Biochem 1994; 27: 63–68.
490–499. 46. Knopp RH, Bonet B, Lasuncion MA, Montelongo A, Herrera E.
23. Fitch WL, King JC. Plasma amino acid, glucose, and insulin Lipoprotein metabolism in pregnancy. In: Herrera E, Knopp RH,
responses to moderate-protein and high-protein test meals in eds., Perinatal Biochemistry, CRC Press: Boca Raton, FL, 2006,
pregnant, nonpregnant, and gestational diabetic women. Am J pp. 19–51.
Clin Nutr 1987; 46: 243–249. 47. Brinton EA. Oral estrogen replacement therapy in postmeno-
24. Regnault TR, de Vrijer B, Battaglia FC. Transport and metabo- pausal women selectively raises levels and production rates of
lism of amino acids in placenta. Endocrine 2002; 19: 23–41. lipoprotein A-I and lowers hepatic lipase activity without lower-
25. Avagliano L, Garo C, Marconi AM. Placental amino acids trans- ing the fractional catabolic rate. Arterioscler Thromb Vasc Biol
port in intrauterine growth restriction. J Pregnancy 2012; 2012: 1996; 16: 431–440.
972562–972567. 48. Ramos P, Herrera E. Reversion of insulin resistance in the rat dur-
26. Herrera E. Metabolic changes in diabetic pregnancy. In: Djelmis ing late pregnancy by 72-h glucose infusion. Am J Physiol 1995;
J, Desoye G, Ivanisevic M, eds., Diabetology of Pregnancy, 269: E858–E863.
Karger: Basel, Switzerland, 2005, pp. 34–45. 49. Herrera E, Lasuncion MA. Maternal-fetal transfer of lipid metabo-
27. Villar J, Cogswell M, Kestler E et al. Effect of fat and fat-free lites. In: Polin RA, Fox WW, Abman SH, eds., Fetal and Neonatal
mass deposition during pregnancy on birth weight. Am J Obstet Physiology, 3rd ed., Saunders: Philadelphia, PA, 2004, pp.
Gynecol 1992; 167: 1344–1352. 375–388.
28. Lopez-Luna P, Maier I, Herrera E. Carcass and tissue fat content 50. Lopez-Luna P, Olea J, Herrera E. Effect of starvation on lipopro-
in the pregnant rat. Biol Neonate 1991; 60: 29–38. tein lipase activity in different tissues during gestation in the rat.
29. Herrera E, Lasuncion MA, Gomez-Coronado D et al. Role of Biochim Biophys Acta 1994; 1215: 275–279.
lipoprotein lipase activity on lipoprotein metabolism and the 51. Herrera E. Lipid metabolism in pregnancy and its consequences
fate of circulating triglycerides in pregnancy. Am J Obstet in the fetus and newborn. Endocrine 2002; 19: 43–55.
Gynecol 1988; 158: 1575–1583. 52. Herrera E, Amusquivar E, López-Soldado I, Ortega H. Maternal
30. Ramos MP, Crespo-Solans MD, del Campo S, Cacho J, Herrera lipid metabolism and placental lipid transfer. Horm Res 2006; 65:
E. Fat accumulation in the rat during early pregnancy is 59–64.
modulated by enhanced insulin responsiveness. Am J Physiol 53. Ramirez I, Llobera M, Herrera E. Circulating triacylglycerols,
Endocrinol Metab 2003; 285: E318–E328. lipoproteins, and tissue lipoprotein lipase activities in rat mothers
31. Rebuffe-Scrive M, Enk L, Crona N et al. Fat cell metabolism and offspring during the perinatal period: Effect of postmaturity.
in different regions in women. Effect of menstrual cycle, preg- Metabolism 1983; 32: 333–341.
nancy, and lactation. J Clin Invest 1985; 75: 1973–1976. 54. Sethi JK, Vidal-Puig AJ. Thematic review series: Adipocyte biol-
32. Knopp RH, Boroush MA, O’Sullivan JB. Lipid metabolism in ogy. Adipose tissue function and plasticity orchestrate nutritional
pregnancy. II. Postheparin lipolytic activity and hypertriglyc- adaptation. J Lipid Res 2007; 48: 1253–1262.
eridemia in the pregnant rat. Metabolism 1975; 24: 481–493. 55. Poulos SP, Hausman DB, Hausman GJ. The development and endo-
33. Lasuncion MA, Herrera E. Changes with starvation in the rat of crine functions of adipose tissue. Mol Cell Endocrinol 2010; 323:
the lipoprotein lipase activity and hydrolysis of triacylglycerols 20–34.
from triacylglycerol-rich lipoproteins in adipose tissue prepara- 56. Smith AJ, Thompson BR, Sanders MA, Bernlohr DA. Interaction
tions. Biochem J 1983; 210: 639–643. of the adipocyte fatty acid-binding protein with the hormone-
34. Alvarez JJ, Montelongo A, Iglesias A, Lasuncion MA, Herrera sensitive lipase: Regulation by fatty acids and phosphorylation.
E. Longitudinal study on lipoprotein profile, high density lipo- J Biol Chem 2007; 282: 32424–32432.
protein subclass, and postheparin lipases during gestation in 57. Tso AW, Xu A, Sham PC et al. Serum adipocyte fatty acid bind-
women. J Lipid Res 1996; 37: 299–308. ing protein as a new biomarker predicting the development of
35. Martin-Hidalgo A, Holm C, Belfrage P, Schotz MC, Herrera E. type 2 diabetes: A 10-year prospective study in a Chinese cohort.
Lipoprotein lipase and hormone-sensitive lipase activity and Diabetes Care 2007; 30: 2667–2672.
mRNA in rat adipose tissue during pregnancy. Am J Physiol 58. Ortega-Senovilla H, Schaefer-Graf U, Meitzner K et al. Gestational
1994; 266: E930–E935. diabetes mellitus causes changes in the concentrations of adipo-
36. Williams C, Coltart TM. Adipose tissue metabolism in preg- cyte fatty acid-binding protein and other adipocytokines in cord
nancy: The lipolytic effect of human placental lactogen. Br J blood. Diabetes Care 2011; 34: 2061–2066.
Obstet Gynaecol 1978; 85: 43–46. 59. Masuzaki H, Ogawa Y, Sagawa N et al. Nonadipose tissue pro-
37. Knopp RH, Herrera E, Freinkel N. Carbohydrate metabolism duction of leptin: Leptin as a novel placenta-derived hormone in
in pregnancy. 8. Metabolism of adipose tissue isolated from humans. Nat Med 1997; 3: 1029–1033.
fed and fasted pregnant rats during late gestation. J Clin Invest 60. Hartmann BW, Wagenbichler P, Soregi G. Maternal and umbil-
1970; 49: 1438–1446. ical-cord serum leptin concentrations in normal, full-term preg-
38. Herrera E, Lasunción MA. Maternal-fetal transfer of lipid nancies. N Engl J Med 1997; 337: 863.
metabolites. In: Polin RA, Fox WF, Steven HA, eds., Fetal and 61. Okereke NC, Uvena-Celebrezze J, Hutson-Presley L, Amini SB,
Neonatal Physiology, 4th ed., Elsevier Saunders: Philadelphia, Catalano PM. The effect of gender and gestational diabetes mel-
PA, 2011, pp. 441–454. litus on cord leptin concentration. Am J Obstet Gynecol 2002;
39. Zorzano A, Lasuncion MA, Herrera E. Role of the availability of 187: 798–803.
substrates on hepatic and renal gluconeogenesis in the fasted 62. Graham TE, Yang Q, Bluher M et al. Retinol-binding protein
late pregnant rat. Metabolism 1986; 35: 297–303. 4 and insulin resistance in lean, obese, and diabetic subjects.
40. Alonso de la Torre SR, Serrano MA, Medina JM. Carrier- N Engl J Med 2006; 354: 2552–2563.
mediated beta- d -hydroxybutyrate transport in brush-border 63. Johansson S, Gustafson AL, Donovan M, Eriksson U, Dencker L.
membrane vesicles from rat placenta. Pediatr Res 1992; 32: Retinoid binding proteins-expression patterns in the human pla-
317–323. centa. Placenta 1999; 20: 459–465.
41. Herrera E, Gomez-Coronado D, Lasuncion MA. Lipid metabo- 64. Mazaki-Tovi S, Kanety H, Pariente C et al. Maternal serum adi-
lism in pregnancy. Biol Neonate 1987; 51: 70–77. ponectin levels during human pregnancy. J Perinatol 2007; 27:
42. Page MA, Williamson DH. Enzymes of ketone-body utilisation 77–81.
in human brain. Lancet 1971; 2: 66–68. 65. López-Bermejo A, Casano-Sancho P, Fernández-Real JM et al.
43. Shambaugh GE, Koehler RR, Yokoo H. Fetal fuels. III: Ketone Both intrauterine growth restriction and postnatal growth influ-
utilization by fetal hepatocyte. Am J Physiol 1978; 235: ence childhood serum concentrations of adiponectin. Clin
E330–E337. Endocrinol 2004; 61: 339–346.
References 27
66. Ategbo JM, Grissa O, Yessoufou A et al. Modulation of adipo- 84. Underwood BA. Maternal vitamin A status and its importance in
kines and cytokines in gestational diabetes and macrosomia. infancy and early childhood. Am J Clin Nutr 1994; 59: S517–S524.
J Clin Endocrinol Metab 2006; 91: 4137–4143. 85. Ismadi SD, Olson JA. Dynamics of the fetal distribution and trans-
67. Kadowaki K, Waguri M, Nakanishi I et al. Adiponectin concen- fer of vitamin A between rat fetuses and their mother. Int J Vitam
tration in umbilical cord serum is positively associated with the Nutr Res 1982; 52: 112–119.
weight ratio of fetus to placenta. J Clin Endocrinol Metab 2006; 86. Takahashi YI, Smith JE, Goodman DS. Vitamin A and retinol-
91: 5090–5094. binding protein metabolism during fetal development in the rat.
68. Cikot RJ, Steegers-Theunissen RP, Thomas CM et al. Longitudinal Am J Physiol 1977; 233: E263–E272.
vitamin and homocysteine levels in normal pregnancy. Br J Nutr 87. Quadro L, Hamberger L, Gottesman ME et al. Transplacental
2001; 85: 49–58. delivery of retinoid: The role of retinol-binding protein and lipo-
69. Caudill MA, Gregory JF, Hutson AD, Bailey LB. Folate catabo- protein retinyl ester. Am J Physiol Endocrinol Metab 2004; 286:
lism in pregnant and nonpregnant women with controlled folate E844–E851.
intakes. J Nutr 1998; 128: 204–208. 88. Dancis J, Levitz M, Katz J et al. Transfer and metabolism of retinol
70. Higgins JR, Quinlivan EP, McPartlin J et al. The relationship by the perfused human placenta. Pediatr Res 1992; 32: 195–199.
between increased folate catabolism and the increased require- 89. Sapin V, Alexandre MC, Chaib S et al. Effect of vitamin A status
ment for folate in pregnancy. BJOG 2000; 107: 1149–1154. at the end of term pregnancy on the saturation of retinol binding
71. Bisseling TM, Steegers EA, van den Heuvel JJ et al. Placental folate protein with retinol. Am J Clin Nutr 2000; 71: 537–543.
transport and binding are not impaired in pregnancies compli- 90. Gazala E, Sarov B, Hershkovitz E et al. Retinol concentration in
cated by fetal growth restriction. Placenta 2004; 25: 588–593. maternal and cord serum: Its relation to birth weight in healthy
72. Piedrahita JA, Oetama B, Bennett GD et al. Mice lacking the folic mother-infant pairs. Early Hum Dev 2003; 71: 19–28.
acid-binding protein Folbp1 are defective in early embryonic 91. Ghebremeskel K, Burns L, Burden TJ et al. Vitamin A and related
development. Nat Genet 1999; 23: 228–232. essential nutrients in cord blood: Relationships with anthropo-
73. Rayburn WF, Stanley JR, Garrett ME. Periconceptional folate metric measurements at birth. Early Hum Dev 1994; 39: 177–188.
intake and neural tube defects. J Am Coll Nutr 1996; 15: 121–125. 92. Zhang X, Chen K, Wei XP et al. Perinatal vitamin A status in rela-
74. Swain RA, St Clair L. The role of folic acid in deficiency states and tion to neurodevelopmental outcome at two years of age. Int J
prevention of disease. J Fam Pract 1997; 44: 138–144. Vitam Nutr Res 2009; 79: 238–249.
75. Myatt L, Eis AL, Brockman DE et al. Differential localization 93. Schweigert FJ, Bathe K, Chen F, Buscher U, Dudenhausen JW.
of superoxide dismutase isoforms in placental villous tissue of Effect of the state of lactation in humans on carotenoid levels in
normotensive, pre-eclamptic, and intrauterine growth-restricted milk, blood plasma and plasma lipoprotein fractions. Eur J Nutr
pregnancies. J Histochem Cytochem 1997; 45: 1433–1438. 2004; 43: 39–44.
76. Guajardo L, Beharry KD, Modanlou HD, Aranda JV. Ascorbic 94. Keusch GT. Vitamin A supplements: Too good not to be true.
acid concentrations in umbilical cord vein and arteries of preterm N Engl J Med 1990; 323: 985–987.
and term newborns. Biol Neonate 1995; 68: 1–9. 95. Sanders TAB. Vitamin A and pregnancy. Lancet 1990; 336: 1375.
77. Lee BE, Hong YC, Lee KH et al. Influence of maternal serum 96. Gonzalez-Corbella MJ, Lopez-Sabater MC, Castellote-Bargallo
levels of vitamins C and E during the second trimester on birth AI, Campoy-Folgoso C, Rivero-Urgell M. Influence of caesarean
weight and length. Eur J Clin Nutr 2004; 58: 1365–1371. delivery and maternal factors on fat-soluble vitamins in blood
78. Kovacs CS, Kronenberg HM. Maternal-fetal calcium and bone from cord and neonates. Early Hum Dev 1998; 53: S121–S134.
metabolism during pregnancy, puerperium, and lactation. 97. De Vriese SR, Dhont M, Christophe AB. Oxidative stability of
Endocrinol Rev 1997; 18: 832–872. low density lipoproteins and vitamin E levels increase in maternal
79. Ritchie LD, Fung EB, Halloran BP et al. A longitudinal study of blood during normal pregnancy. Lipids 2001; 36: 361–366.
calcium homeostasis during human pregnancy and lactation and 98. Jishage K, Arita M, Igarashi K et al. Alpha-tocopherol transfer pro-
after resumption of menses. Am J Clin Nutr 1998; 67: 693–701. tein is important for the normal development of placental labyrin-
80. Herrera E, Ortega H, Alvino G et al. Relationship between thine trophoblasts in mice. J Biol Chem 2001; 276: 1669–1672.
plasma fatty acid profile and antioxidant vitamins during normal 99. Kaempf-Rotzoll DE, Horiguchi M, Hashiguchi K et al. Human
pregnancy. Eur J Clin Nutr 2004; 58: 1231–1238. placental trophoblast cells express alpha-tocopherol transfer pro-
81. Zachman RD. Role of vitamin A in lung development. J Nutr tein. Placenta 2003; 24: 439–444.
1995; 125: 1634S–1638S. 100. Barbas C, Herrera E. Lipid composition and vitamin E content in
82. Rohwedel J, Guan K, Wobus AM. Induction of cellular differen- human colostrum and mature milk. J Physiol Biochem 1998; 54:
tiation by retinoic acid in vitro. Cells Tissues Organs 1999; 165: 167–173.
190–202. 101. Martinez S, Barbas C, Herrera E. Uptake of alpha-tocopherol by
83. Ackermans MM, Zhou H, Carels CE, Wagener FA, Von den Hoff the mammary gland but not by white adipose tissue is dependent
JW. Vitamin A and clefting: Putative biological mechanisms. Nutr on lipoprotein lipase activity around parturition and during lacta-
Rev 2011; 69: 613–624. tion in the rat. Metabolism 2002; 51: 1444–1451.
4 Intermediary metabolism in
pregnancies complicated by
gestational diabetes
Bartolomé Bonet, María Bonet-Alavés, and Isabel Sánchez-Vera
triglycerides and cholesterol reach the highest levels.1,2,10,11

Changes in intermediary metabolism The high plasma levels of triglycerides seem to be secondary
during pregnancy to both an increased hepatic synthesis of very-low-density
lipoproteins–triglycerides (VLDL-TG) and a decreased lipo-
As has been shown in the previous chapters, major changes protein lipase (LPL) activity in the adipose tissue in late ges-
take place in the intermediary metabolism throughout preg- tation.10,11 These changes are due to some of the hormonal
nancy: changes that will facilitate the fetal needs of energy changes observed during pregnancy. The increased produc-
and precursors for fetal and placental growth, as well as tion of VLDL-TG is secondary to the elevation of estradiol
for placental hormone synthesis. From a metabolic point that takes place during gestation, as increased levels of estra-
of view, during gestation there are two different periods. In diol stimulate the hepatic synthesis of triglycerides.12,13 Both
the first half of pregnancy, during the embryo development the human placental lactogen, which reaches its maximum
period, there are maternal changes that lead to storage of concentration by the end of pregnancy, and insulin resis-
energy and nutrients.1–3 During this time of pregnancy, there tance increase adipose tissue lipolysis, leading to an abun-
is both increased appetite and normal or increased insulin dant supply of fatty acids to the liver.14,15 The lower levels of
sensitivity. These changes will facilitate glucose and lipid adiponectin observed as pregnancy advances16,17 may also
uptake by adipose tissue, increasing the lipid stores.1–3 In play a role in the increased hepatic synthesis of triglycerides,
fact, during the first half of pregnancy, most women show an as under this condition, in the liver, there is a decrease in
increase in adipose tissue mass.4 Nevertheless, as pregnancy beta-oxidation18 and, therefore, the fatty acids are derived
advances and the fetal–placental unit is rapidly growing, a toward reesterification and synthesis of triglycerides. The
marked shift in the metabolic pathways is observed. This high plasma levels of triglycerides are used as a source of
second period is characterized by a state of maternal insulin energy for the maternal tissues, sparing glucose for the fetus
resistance, decreasing the uptake of glucose in the insulin- and the maternal tissues where glucose is the only source
sensitive tissues, mainly the white adipose tissue and muscle of energy. In addition, they are used by the placenta, where
(Figure 4.1, panel b). Such a condition facilitates the supply LPL is present19; therefore, the placenta is able to hydrolyze
of glucose toward the fetus, where the daily glucose require- the VLDL-TG, releasing fatty acids, which are taken up by
ments are very high (30–50 g of glucose/day).5 The insulin placental cells.19 In fact, several studies have shown a posi-
resistance is also responsible of the blunted curve that takes tive relationship between maternal plasma triglycerides and
place after an oral glucose tolerance test or a regular meal.6–8 birth weight.20,21
Such a curve causes higher and more prolonged plasma lev- The elevation of the plasma levels of cholesterol, mainly
els of glucose after a meal facilitating a higher glucose supply from low-density lipoproteins (LDLs),1,2,10,11 will facilitate
toward the fetus, as the transfer of glucose through the pla- the substrate for the elevated synthesis of steroid hormones
centa is via passive glucose diffusion and, therefore, concen- that takes place during pregnancy. By the end of pregnancy,
tration dependent.9 plasma levels of estradiol are almost a thousand times higher
than in nonpregnant women and progesterone 10 times
Plasma lipid changes during pregnancy higher.22 Cholesterol is also a source of precursors for the
high synthesis of cell membranes that takes place in the fast-
During gestation, there are also relevant changes in lipid growing fetus.
metabolism (Figure 4.1, panel b), changes most marked As occurs in carbohydrates and lipid metabolism, there
in the second half of pregnancy, when both plasma are changes in protein metabolism throughout pregnancy.
28
Mechanisms of pregnancy-induced insulin resistance 29
Adipocyte Adipocyte
Adipocyte
Triglycerides Triglycerides
Triglycerides
hPL
+ + hPL
Cortisol Cortisol +
Insulin Insulin Cortisol
Catecholamines – Catecholamines – Insulin
Catecholamines –
VLDL-TG VLDL-TG
VLDL-TG
FFA Glycerol FA FFA Glycerol FA

FFA Glycerol FA
Peripheral Peripheral
Peripheral
tissues tissues
LPL LPL tissues
LPL
Insulin + Insulin +
Insulin +
FFA Glycerol-3P FFA Glycerol-3P
FFA Glycerol-3P
Adiponectin Acyl-CoA Glucose Glucose
Adiponectin Acyl-CoA Adiponectin Acyl-CoA Glucose
+ Triglycerides VLDL-TG + Triglycerides VLDL-TG
– – + Triglycerides VLDL-TG
–
Insulin Acetyl-CoA Insulin Acetyl-CoA Insulin
LPL Acetyl-CoA LPL
Placenta
ATP Placenta
ATP KB KB ATP KB
(a) (b) (c)
Figure 4.1 Intermediary metabolism in nonpregnant women (panel a), normal pregnancies (panel b), and pregnancies with GDM
(panel c). The thickness of the lines relates to the degree of stimulation (solid lines) or inhibition (dash lines) for both the metabolic
pathways and the hormonal effects.
These changes develop gradually during gestation. By the Plasma changes

end of pregnancy, when fetal growth is maximal, nitrogen Estrogens
retention is four times higher than in early pregnancy,23 sug- Placenta Progesterone
hPL
gesting that amino acids are conserved for tissue synthesis. Insulin removal
Nevertheless, the sum of total plasma amino acids declines Second half
Cortisol Insulin resistance
by 15%–25%, reflecting the enhanced uptake of amino acids of gestation
by the placenta,24 where there is an energy-dependent active
transport,24,25 providing a higher concentration of amino TNF alpha
Increased adipose
acids in fetal plasma. Such a mechanism facilitates the sup- tissue mass Adiponectin
ply of amino acids for the rapid protein accretion that takes
place in the fetus by the end of pregnancy. Figure 4.2 Different factors from the placenta and the adi-
Overall the metabolic changes observed during preg- pose tissue involved in pregnancy-induced insulin resistance.
nancy are directed to supply nutrients to the placental–fetal
unit, allowing in a short period of time the fast growing of placenta and the increased adipose tissue promote impor-
both elements. tant changes in the maternal milieu: hormonal, inflamma-
tory, and metabolic. As pregnancy advances, there is an
increase in the plasma levels of placental lactogen and estro-
Mechanisms of pregnancy-induced gens. The first condition leads to increased lipolysis and to
insulin resistance higher plasma levels of free fatty acids15 that promote periph-
eral insulin resistance.27 The elevation of estradiol and the
The mechanisms involved in pregnancy-induced insulin decrease in adiponectin promote an increase in the hepatic
resistance, although not completely understood, seem to synthesis and release of triglycerides, such as VLDL-TG; such
be related to the placenta and the increased adipose tissue a condition causes peripheral insulin resistance by increas-
mass that develops during pregnancy (Figure 4.2). In preg- ing the cell metabolism of fatty acids, raising the cellular lev-
nant rats, there is increased degradation of insulin by the els of NADH and ATP. This intracellular condition lowers
placenta26; therefore, insulin removal is accelerated, caus- the glucokinase activity and the cell ability to phosphorylate
ing a higher activity in the pancreatic beta cells. A similar glucose, decreasing the cell uptake of this substance. The
phenomenon may occur in human pregnancies. Both the placenta and the higher adipose mass increase the plasma
30 Intermediary metabolism in pregnancies complicated by gestational diabetes
levels of TNF-alpha, a condition that modifies the intracel- insulin resistance,40,41 exacerbating the pregnancy-induced
lular signaling of insulin, causing insulin resistance.28,29 The insulin resistance and the risk of developing GDM.
insulin-sensitizing effects of adiponectin30 are plumbed by
the decreased plasma levels of adiponectin observed in the
second half of pregnancy. The higher concentration of corti- Glucose alterations in GDM
sol found in pregnancy increases the hepatic glucose produc-
tion and decreases the peripheral tissue glucose utilization, Independent of the mechanisms involved, in GDM, there is
further increasing insulin resistance. The high plasma levels a relative lack of insulin during a period of time with high
of progesterone found during the second half of pregnancy insulin needs to compensate the insulin resistance that
also has been related with the pregnancy-induced insulin develops in the third trimester of pregnancy. When GDM
resistance.31 develops, in the maternal tissues, the insulin-dependent
glucose uptake is further decreased and hyperglycemia
develops. Because the placental transfer of glucose is con-
Mechanisms leading to the centration dependent9 under conditions of maternal hyper-
glycemia and placental normal function, there is increased
development of gestational diabetes placental transfer of glucose (Figure 4.1, panel c), and fetal
The mechanisms leading to the development of gestational hyperglycemia and, secondary to this alteration, hyper-
diabetes mellitus (GDM) have not been fully defined but insulinism develop. As insulin is one of the main growth
are probably related to both an exacerbation of the beta-cell factors during fetal life,42 hyperinsulinemia promotes fetal
dysfunction in subjects genetically predisposed to beta-cell growth, macrosomia, and complications secondary to the
alterations and to the insulin resistance that takes place as delivery of a large baby, mainly both maternal perineal
gestation advances. In that sense, GDM will act like a type 2 damage and birth trauma, including shoulder dystocia and
diabetes mellitus (DM). Regarding the beta-cell dysfunction, Erb’s palsy. Once the umbilical supply of glucose is sud-
several mechanisms could be involved in this process. High denly arrested after delivery, in the newborn, the remain-
progesterone levels may play a relevant role.32,33 Recently, in ing hyperinsulinism increases the risk of hypoglycemia.
models of knockout mice,33 it has been shown that the lack The macrosomic newborn will need monitoring of blood
of progesterone receptors is associated with a higher insulin glucose within the first hour after birth and early feeds
secretion by beta cells. Therefore, the high levels of progester- and occasionally may require intravenous administration
one that develop during pregnancy may damage these cells. of glucose.43 Hypoglycemia in the newborn, if not cor-
The hyperlipidemia observed during pregnancy may also rected, may lead to brain damage.44 Fetal macrosomia also
decrease the capability of beta cells to secrete insulin.34,35 increases the risk of obesity, type 2 DM, and cardiovascular
Although fatty acids may induce insulin secretion,36 under diseases later on in life.45,46
certain circumstances, prolonged elevated levels of fatty
acids may damage the beta cell, decreasing insulin secretion.
In fact, in an experimental animal model during pregnancy, Lipid alterations in GDM
a decrease in plasma free fatty acids (FFA) and triglycerides
increases insulin secretion after an oral glucose tolerance In GDM, as occurs in other conditions of insulin resistance
test.37 Recently, higher plasma levels of cholesterol have been and beta-cell dysfunction, there is an increase in plasma
linked to islet dysfunction38; therefore, pregnancy-induced levels of triglycerides and cholesterol. This effect should be
hypercholesterolemia may play a role in the beta-cell failure added to the physiological hyperlipidemia induced by preg-
associated with GDM. nancy10,11 (Figure 4.1, panel c). Therefore, women with GDM
The higher food intake that develops during early preg- have higher plasma levels of triglycerides and cholesterol
nancy may cause beta-cell hyperplasia. In certain geneti- than found in normal pregnancies.47–49 Hypertriglyceridemia
cally predisposed subjects, this higher supply of glucose and found in GDM may also play a role in the fetal macrosomia
fatty acids to the beta cell may increase the cell metabolism, observed in these pregnancies, as several authors have shown
leading to increased beta-cell apoptosis and cell death. 34,39 a positive correlation between the plasma levels of triglycer-
Such a phenomenon would compromise the capability ides and birth weight.20,21
of the beta cells to provide enough insulin in a period of Nevertheless, these increased plasma levels of both
increased insulin requirements causing development of triglycerides and cholesterol, as occur in other condi-
GDM. Further research is needed for a better understand- tions through the life, may have deleterious effects on the
ing of the beta-cell dysfunction observed in GDM, in order normal development of pregnancy and on fetal growth.
to develop methods to improve this function and decrease Increased levels of triglycerides have been associated to
the incidence of GDM. structural alterations in LDL.50,51 High triglycerides make
Because the insulin resistance that takes place in the sec- LDL particles smaller and denser. Such particles are more
ond half of pregnancy plays a key role in the development of susceptible to oxidation. Through pregnancy, the maternal
GDM, any condition susceptible to exacerbating this resis- milieu changes, developing conditions that both increase
tance may play a role in the development of GDM. Obesity and decrease LDL susceptibility to oxidation. Higher lev-
and the associated metabolic alterations also promote els of triglycerides, cholesterol, glucose, and progesterone
Amino acid alterations in pregnancies complicated by gestational diabetes 31
are prooxidant,52,53 but this effect may be blunted by higher Fetal weight
levels of vitamin E and estradiol,22 two powerful antioxi- Macrosomia Normal weight Intrauterine growth
dants whose levels are increased in pregnancy.22 In fact, our retardation
group and others have shown that in normal pregnancies, Glucose
as pregnancy advances, despite increased plasma levels of
triglycerides and cholesterol, LDL susceptibility to oxidation Triglycerides
decreases,22,54 a phenomenon that is partially explained by
LDL oxidation
the high levels of estradiol observed at the end of pregnancy.
Nevertheless, under conditions of exacerbated dyslipidemia, Placental damage
the prooxidant effects may lead to increased LDL oxidation
and to the consequences associated with this process. An Figure 4.4 Potential factors affecting fetal growth in preg-
increased LDL oxidation may have relevant consequences nancies complicated by gestational diabetes mellitus. Factors
for the placenta as well as for the fetus. In a cell culture model increasing fetal growth are shown in yellow and conditions
of human placental trophoblasts and macrophages, LDL oxi- decreasing fetal growth in red.
dation is cytotoxic.55 If this process occurred during preg-
nancy, it could damage the placenta. If such damage were
extensive, the placental capability to transfer nutrients and streaks in the fetal arteries.60,61 As a consequence, children
oxygen to the fetus might be decreased, compromising fetal from GDM pregnancies may have a higher incidence of
growth. In fact, our group has shown a correlation between these alterations than newborns from normal pregnancies.
LDL susceptibility to oxidation and birth weight,22 suggest- Further studies are needed to demonstrate if stillborns from
ing that in conditions where LDL oxidation is increased, the GDM pregnancies also have increased incidence of fatty
fetal growth is compromised. Furthermore, when oxidized streaks in their arteries.
LDL is taken up by human trophoblasts, through scavenger To sum up, in GDM, there is a combination of factors that
receptors and not by the LDL receptor, despite increasing may affect the nutrient supply to the fetus (Figure 4.4). Under
the intracellular concentration of cholesterol, progesterone certain conditions, increased supply of glucose and triglyc-
secretion is decreased56 (Figure 4.3). These data suggest that erides toward the fetus may lead to increased fetal growth
the metabolism of cholesterol from oxidized LDL does not and macrosomia. Nevertheless, under certain conditions,
follow the physiological pathways required for hormonal dyslipidemia and hyperglycemia found in GDM may lead
synthesis. Therefore, under circumstances of increased LDL to increased LDL oxidation, placental damage, and vascular
oxidation, there may be a lack of cholesterol for the placenta, dysfunction; such a condition would decrease the transfer of
decreasing the placental synthesis of steroidal hormones as nutrients toward the fetus and lead to intrauterine growth
well as the transfer of cholesterol to the fetus. As occurs in retardation (Figure 4.4). At present, most efforts are directed
other conditions, increased LDL oxidation may also lead to toward blood glucose normalization, and little attention
vascular dysfunction,57–59 decreasing the vascular blood flow has been paid to the dyslipidemia and to the LDL oxidation
and the nutrient transfer to the fetus. Therefore, increased associated with this process; in fact in our institution, in
LDL oxidation could affect the transfer of nutrients and oxy- GDM pregnancies with optimal metabolic control, we have
gen to the fetus either damaging the placenta or decreasing found a higher incidence of intrauterine growth retardation
the placental blood flow. (13.5%) than macrosomic infants (9.1%).49 Further studies
Finally, conditions of increased plasma levels of choles- are needed to obtain a better understanding of the role of
terol and triglycerides have been linked to increased fatty dyslipidemia in the maternal and fetal complications associ-
ated with GDM. There is a need of studies to determine why
some pregnancies complicated by GDM causes intrauterine
growth retardation instead of large infants.
350
300
250
Amino acid alterations in pregnancies
200 complicated by gestational diabetes
150
In GDM, there is an increase in the number of essential and
100 nonessential amino acids in umbilical venous and arterial
50 concentration,62 compared to the values found in normal
0 pregnancies. The higher plasma levels of fetal amino acids
0 20 40 80 do not seem to be related to a higher concentration in mater-
Native LDL OX LDL
nal plasma, as only ornithine has been shown to increase in
plasma from pregnant women with GDM.62 More recently,
Figure 4.3 Progesterone secretion in trophoblast incubated studies analyzing maternal protein and amino acid metab-
with increasing concentrations of normal or oxidized low- olism by stable isotope methodologies did not find signifi-
density lipoprotein (for details, see reference 52). cant differences in either treated63 or untreated GDM64,65
32 Intermediary metabolism in pregnancies complicated by gestational diabetes
compared to normal pregnancies. However, the elevation taken up across the microvillous plasma membranes, is rap-
observed in the plasma amino acid concentration in the idly transferred to the fetus, contributing to the accelerated
umbilical cord, but not in maternal circulation, suggests fetal growth in these patients.66,67 More studies are needed
that placental amino acid exchange and/or fetal–placental for a further clarification of the role of amino acid metabolic
metabolism is altered in GDM.62 alterations associated with GDM.
Studies in vitro show that among the different amino acid In conclusion, GDM is a condition of increased mater-
transporters, the expression of system A, which mediates the nal metabolic fuels; such a condition in the majority of
transfer of neutral amino acids such as alanine, serine, and pregnancies leads to a higher fetal growth, although under
glutamine, is increased in diabetic pregnancies.66,67 This, in certain conditions, the increased fuels, mainly lipids,
turn, could increase the uptake and delivery of neutral amino through oxidative process and increased free radical gen-
acids into the fetus. However, it does not seem to be the pri- eration may lead to placental and vascular damage, com-
mary cause of accelerated fetal growth. Other transporters promising the normal development of pregnancy and fetal
such as the specific system for leucine (system L) have also growth.
been shown to be increased in microvillous plasma mem-
branes isolated from GDM pregnancies with large babies
for their gestational age.66,67 Nevertheless, other authors Acknowledgments
did not find an increased activity of these transporters.68 It
is remarkable that leucine has been proven to be an effec- I dedicate this chapter to my former mentor, Bob Knopp, who
tive stimulus for fetal insulin secretion in human pancreas introduced me to the world of pregnancy and lipids, but most
studied in vitro.69 In vivo studies applying stable isotope importantly taught me with his example how to take care of
techniques have provided evidence to suggest that leucine, patients in a loving and compassionate way (B. Bonet).
REFERENCES
1. Knopp RH, Montes A, Childs M, Li JR, Mabuchi H. Metabolic 15. Grumbach MM, Kaplan SL, Abramo CL. Plasma free fatty acid
adjustments in normal and diabetic pregnancy. Clin Obstet response to the administration of chorionic “growth hormone
Gynecol 1981; 24: 21–49. prolactin.” J Clin Endocrinol Metab 1966; 26: 478.
2. Knopp RH, Bonet B, Zhu XD. Lipid metabolism in pregnancy. 16. Retnakaran R, Hanley AJG, Raif N, Connelly PW, Sermer M,
In: Cowett RM, ed., Principles of Perinatal-Neonatal Metabolism, Zinman B. Reduced adiponectin concentration in women with
Springer: Berlin, Germany, 1998, pp. 221–258. gestational diabetes. Diabetes Care 2004; 27: 799–800.
3. Catalano PM, Tyzbir ED, Wolfe RR, Roman NM, Amino SB, Sims 17. Worda C, Leipold H, Gruber C, Kautzky-Willer A, Knofler M,
EAH. Longitudinal changes in basal hepatic glucose produc- Bancher-Todesca D. Decreased plasma adiponectin concentra-
tion and suppression during insulin infusion in normal pregnant tions in women with gestational diabetes mellitus. Am J Obstet
woman. Am J Obstet Gynecol 1992; 167: 913–919. Gynecol 2004; 191: 2120–2124.
4. Hytten RE, Thomson AM, Taggart N. Total body water in normal 18. Kershaw EE, Flier JS. Adipose tissue as an endocrine organ. J Clin
pregnancy. BJOG 1966; 73: 553–561. Endocrinol Metab 2004; 89: 2548–2556.
5. Hay WW. Glucose metabolism in the fetal-placental unit. In: 19. Bonet B, Brunzell JD, Gown AM, Knopp RH. Metabolism of very-
Cowett RM, ed., Principles of Perinatal-Neonatal Metabolism, low density lipoprotein triglyceride by human placental cells: The
Springer: Berlin, Germany, 1998, pp. 337–367. role of the lipoprotein lipase. Metabolism 1992; 41: 596–603.
6. Kalhan S, Rossi K, Gruca L, Burkett E, O’Brien A. Glucose turn- 20. Knopp RH, Magee MS, Walden CE, Bonet B, Benedetti TJ.
over and gluconeogenesis in human pregnancy. J Clin Invest Prediction of infant birthweight by gestational diabetes screen-
1997; 100: 1775–1781. ing test: Importance of plasma triglyceride. Diabetes Care 1992;
7. Spelleacy WN, Goetz FC. Plasma insulin in normal late preg- 15(11): 1605–1613.
nancy. N Engl J Med 1963; 268: 988–991. 21. DiCianni G, Miccoli R, Volpe L et al. Maternal triglyceride levels
8. Bleicher SJ, O’Sullivan JB, Freinkel N. Carbohydrate metabolism and newborn weight in pregnant women with normal glucose
in pregnancy. V. The interrelations of glucose, insulin, and free tolerance. Diabet Med 2005; 22: 21–25.
fatty acids in late pregnancy and postpartum. N Engl J Med 1964; 22. Sánchez-Vera I, Bonet B, Viana M, Quintanar A, López-Salva A.
271: 866–872. Increased low-density lipoprotein susceptibility to oxidation in
9. Hay WW. Placental nutrient metabolism and transport. In: pregnancies and fetal growth restriction. Obstet Gynecol 2005;
Herrera E, Knopp RH, eds., Perinatal Biochemistry, CRC Press: 106: 345–351.
Boca Raton, FL, 1992, pp. 93–130. 23. King JC. Physiology of pregnancy and nutrient metabolism. Am J
10. Knopp RH, Bonet B, Lasunción MA, Montelongo A, Herrera E. Clin Nutr 2000; 71: 1218S–1225S.
Lipoprotein metabolism in pregnancy. Lipoprotein metabolism in 24. Phelps RL, Metzger BE, Freinkel N. Carbohydrate metabolism
pregnancy. In: Herrera E, Knopp RH, eds., Perinatal Biochemistry, in pregnancy. XVII. Diurnal profiles of plasma glucose, insulin,
CRC Press: Boca Raton, FL, 1992, pp. 19–51. free fatty acids, triglycerides, cholesterol, and individual amino
11. Knopp RH, Montes A, Warth MR. Carbohydrate and lipid metab- acids in late normal pregnancy. Am J Obstet Gynecol 1981; 140:
olism in normal pregnancy. In: Food and Nutrition Boards, eds., 730–736.
Laboratory Indices of Nutritional Status in Pregnancy, National 25. Cariappa R, Heath-Monning E, Smith CH. Isoforms of amino
Academy of Sciences: Washington, DC, 1978, p. 35. acid transporters in placental syncytiotrophoblast: Plasma mem-
12. Knopp RH, Zhu XD, Bonet B. Effects of estrogens on lipo- brane localization and potential role in maternal/fetal transport.
protein metabolism and cardiovascular disease in women. Placenta 2003; 24: 713–726.
Atherosclerosis 1994; 110: S83–S91. 26. Goodner CJ, Freinkel N. Carbohydrate metabolism in preg-
13. Sacks FM, Gerhard M, Walsh BW. Sex hormones, lipoproteins, nancy: The degradation of insulin by extracts of maternal and
and vascular reactivity. Curr Opin Lipidol 1995; 6: 161–166. fetal structures in the pregnant rat. Endocrinology 1959; 65:
14. Desoye G, Schweditsch O, Pfeiffer KP, Zechner R, Kostner GM. 957–967.
Correlation of hormones with lipid and lipoprotein levels dur- 27. Sivan E, Homko CJ, Whittaker PG, Reece EA, Chen X, Boden G.
ing normal pregnancy and postpartum. J Clin Endocrinol Metab Free fatty acids and insulin resistance during pregnancy. J Clin
1987: 64: 704–712. Endocrinol Metab 1998; 83: 2338–2342.
References 33
28. Kirwan JP, Hauguel-De Mouzon S, Lepercq J et al. TNF-alpha is 51. Sattar N, Greer IA, Louden J, Lindsay G, McConnell M, Shepherd
a predictor of insulin resistance in human pregnancy. Diabetes J, Packard CJ. Lipoprotein subfraction changes in normal preg-
2002; 51: 2207–2213. nancy: Threshold effect of plasma triglyceride on appearance
29. Hotamisligil GS, Murray DL, Choy LN, Spiegelman BM. Tumor of small, dense low density lipoprotein. J Clin Endocrinol Metab
necrosis factor alpha inhibits signaling from the insulin receptor. 1997; 82(8): 2483–2491.
Proc Natl Acad Sci USA 1994; 91: 4854–4858. 52. Cominacini L, Pastorino AM, Garbin U, Campagnola M, de Santis
30. Kadowaki T, Yamauchi T, Kubota N, Hara K, Ueki K, Tobe K. A, Davoli A, Faccini G, Bertozzo L, Pasini F, Pasini AF. The suscep-
Adiponectin and adiponectin receptors in insulin resistance, tibility of low-density lipoprotein to in vitro oxidative is increased
diabetes, and the metabolic syndrome. J Clin Invest 2006; 116: in hypercholesterolemic patients. Nutrition 1994; 10(6): 564–566.
1784–1792. 53. Liu ML, Ylitalo K, Vakkilainen J, Nuotio I, Valkonen M, Lahdenper
31. Kalchoff RK. Metabolic effects of progesterone. Am J Obstet A, Viikari J, Taskinen MR. Susceptibility of LDL to oxidation
Gynecol 1982; 142: 735–738. in vitro and antioxidant capacity in familial combined hyperlip-
32. Branisteanu DD, Mathieu C. Progesterone in gestational diabetes idemia: Comparison of patients with different lipid phenotypes.
mellitus: Guilty or not guilty? Trends Endrocrinol Metab 2003; 14: Ann Med 2002; 34(1): 48–54.
54–55. 54. De Vriese SR, Shont M, Christophe AB. Oxidative stability of low
33. Picard F, Wanatabe M, Schoonjans K, Lydon J, O’Malley BW, density lipoproteins and vitamin E levels increase in maternal
Auwerx J. Progesterone receptor knockout mice have an blood during normal pregnancy. Lipids 2001; 36: 361–366.
improved glucose homeostasis secondary to beta-cell prolifera- 55. Bonet B, Hauge-Gillenwater H, Zhu XD, Knopp RH. LDL oxida-
tion. Proc Natl Acad Sci USA 2002; 99: 15644–15648. tion and human placental trophoblast and macrophage cytotox-
34. Kasuga M. Insulin resistance and pancreatic beta-cell failure. icity. Proc Soc Exp Biol Med 1998; 217: 203–211.
J Clin Invest 2006; 116: 1756–1760. 56. Bonet B, Chait A, Gown AM, Knopp RH. Metabolism of modi-
35. McGarry DJ, Dobbins RL. Fatty acids, lipotoxicity and insulin fied LDL by cultured human placental cells. Atherosclerosis 1995;
secretion. Diabetologia 1999; 42: 128–138. 112: 125–136.
36. Rojo-Martinez G, Esteva I, Ruiz de Aldana MS et al. Dietary fatty 57. Tanner FC, Noll G, Boulanger CM, Lüsscher TF. Oxidized low
acids and insulin secretion: A population-based study. Eur J Clin density lipoproteins inhibit relaxations of porcine coronary arter-
Nutr 2006; 60(10): 1195–1200. ies: Role of scavenger receptor and endothelium-derived nitric
37. Sanchez-Vera I, Bonet B, Viana M, Herrera E, Indart A. Effect of oxide. Circulation 1991; 83: 2012–2020.
acipimox on plasma lipids and glucose/insulin in pregnant rats. 58. Vergnani L. Effect of native and oxidized low-density lipopro-
Int J Exp Diabetes Res 2002; 3: 233–239. tein on endothelial nitric oxide and superoxide production.
38. Brunham LR, Kruit JK, Verchere BC, Hayden MR. Cholesterol in Circulation 2000; 101: 1261–1266.
islet dysfunction and type 2 diabetes. J Clin Invest 2008; 118: 59. Boulanger CM, Tanner FC, Dea MI, Hahn AW, Werner A, Luscher
403–408. TF. Oxidized LDL induce mRNA expression and release of endo-
39. Prentki M, Nolan CJ. Islet beta-cell failure in type 2 diabetes. thelin from human and porcine endothelium. Circ Res 1992; 70:
J Clin Invest 2006; 116: 1802–1812. 1191–1197.
40. Kanh BB, Flier JS. Obesity and insulin resistance. J Clin Invest 60. Napoli C, D’Armiento FP, Mancini FP, Witztum JL, Palumbo G,
2000; 106: 473–481. Palinski W. Fatty streak formation occurs in human fetal aor-
41. King JC. Maternal obesity, metabolism and pregnancy outcomes. tas and is greatly enhanced by maternal hypercholesterolemia:
Annu Rev Nutr 2006; 26: 271–291. Intimal accumulation of LDL and its oxidation precede monocyte
42. Osler M, Pederson J. The body composition of newborn infants recruitment into early atherosclerotic lesions. J Clin Invest 1997;
of diabetic mothers. Pediatrics 1960; 26: 985–992. 100: 2680–2690.
43. Bonet B. Macrosomia and neonatology. In: Ovesen PG, Jensen 61. Napoli C, Glass CK, Witztum JL, Deutsch R, D’Armiento FP,
DM, eds., Maternal Obesity and Pregnancy, Springer-Verlag: Palinski W. Influence of maternal hypercholesterolaemia dur-
Berlin, Germany, 2012, pp. 247–254. ing pregnancy on progression of early atheroesclerotic lesions in
44. Chase HP, Marlow RA, Dabiere CS, Welch NN. Hypoglycemia childhood: Fate of early lesions in children (FELIC) study. Lancet
and brain development. Pediatrics 1973; 52: 513–520. 1999; 354: 1234–1241.
45. Boney CM, Verma A, Tucker R, Vohr BR. Metabolic syndrome 62. Cetin I, Nobile de Santis MS, Tarico E et al. Maternal and fetal
in childhood: Association with birth weight, maternal obe- amino acid concentrations in normal pregnancies and pregnan-
sity and gestational diabetes mellitus. Pediatrics 2005; 115: cies with gestational diabetes mellitus. Am J Obstet Gynecol
290–296. 2005; 192: 610–617.
46. McCance D, Pettit D, Hanson R, Jacobson K, Knowler W, Bennett 63. Zimmer DM, Golichowski AM, Karn CA, Brechtel G, Baron A,
P. Birth weight and non-insulin dependent diabetes: Thrifty geno- Denne SC. Glucose and amino acid turnover in untreated gesta-
type, thrifty phenotype or surviving baby genotype? Br Med J tional diabetes. Diabetes Care 1996; 19(6): 591–596.
1994; 308: 942–945. 64. Butte NF, Hsu HW, Thotathuchery M, Wong WW, Khoury J,
47. Knopp RH, Chapman M, Bergelin R, Wahl PW, Warth MR, Reeds P. Protein metabolism in insulin-treated gestational diabe-
Irvine S. Relationships of lipoprotein lipids to mild fasting hyper- tes. Diabetes Care 1999; 22: 806–811.
glycemia and diabetes in pregnancy. Diabetes Care 1980; 3: 65. Hod M, Lapidot A. Dynamic parameters of maternal amino acid
416–420. metabolism and fetal growth. Isr J Med Sci 1996; 32: 530–536.
48. Hollingsworth DR, Grundy SM. Pregnancy-associated hypertri- 66. Jansson T, Ekstrand Y, Björn C, Wennergren M, Powell TL. Alterations
glyceridemia in normal and diabetic women. Differences in insu- in the activity of placental amino acid transporters in pregnancies
lin-dependent, non-insulin-dependent, and gestational diabetes. complicated by diabetes. Diabetes 2002; 51: 2214–2219.
Diabetes 1982; 31: 1092–1097. 67. Jansson T, Powell TL. Human placental transport in altered fetal
49. Sanchez-Vera I, Bonet B, Viana M, Quintanar A, Martin MD, growth: Does the placenta function as a nutrient sensor?—A
Blanco P, Donnay S, Albi M. Changes in plasma lipids and review. Placenta 2006; 27(suppl A, Trophoblast Research): S91–S97.
increased low-density lipoprotein susceptibility to oxidation in 68. Kuruvilla AG, D’Souza SW, Glazier JD, Mahendran D, Maresh
pregnancies complicated by gestational diabetes: Consequences MJ. Altered activity of the system A amino acid transporter in
of obesity. Metabolism 2007; 56: 1527–1533. microvillous membrane vesicles from placentas of macrosomic
50. Silliman K, Shore V, Forte TM. Hypertriglyceridemia during late babies born to diabetic women. J Clin Invest 1994; 94: 689–695.
pregnancy is associated with the formation of small dense low- 69. Milner RDG, Ashworth MA, Barson AJ. Insulin release from
density lipoproteins and the presence of large buoyant high- human foetal pancreas in response to glucose, leucine and argi-
density lipoproteins. Metabolism 1994; 43: 1035–1041. nine. J Endocrinol 1972; 52: 497–505.
5 Nutrient delivery and
metabolism in the fetus
William W. Hay, Jr., Paul J. Rozance, Stephanie R. Wesolowski, and Laura D. Brown
preterm infants, indicating that the lack of marked fatty

Introduction acid oxidation in the fetus is primarily due to the ready
Fetuses of diabetic mothers have markedly different supply and oxidation of glucose, lactate, and amino acids,
growth rates and develop considerably different body as well as the limited amounts of enzymes and transport-
compositions despite all having a mother with diabetes ers in fetal tissues necessary to deliver fatty acids into
mellitus during her pregnancy. Fetuses of poorly con- the mitochondria. Hormonal regulation of metabolic
trolled diabetics who have wide swings in meal-a ssociated substrate utilization and growth in the fetus, including
plasma concentrations of glucose and fatty acids tend to the effects of insulin and the insulin-like growth factors
be macrosomic, with large amounts of subcutaneous adi- (IGFs), is important but secondary to the supply of nutri-
pose tissue. Such conditions are most commonly asso- ent substrates.1–4
ciated with gestational diabetics. In contrast, severely
diabetic pregnant women, particularly those with vas-
cular disorders and hypertension, frequently produce
smaller placentas that transfer fewer nutrients to the
Role of the placenta in nutrient
fetus; their fetuses tend to be growth restricted and rela- transfer to the fetus
tively devoid of skeletal muscle and body fat. To appreci-
ate how such disparate patterns of growth can occur, it The placenta plays a key role in nutrient transfer to the
is important to understand the basic aspects of nutrient fetus. The placenta contains membrane transporter
transport to the fetus. In the following discussion, data proteins for glucose, lactate, and fatty acids that facili-
from a variety of animal models, principally sheep, are tate their transport to the fetus by concentration gra-
used to augment and support the more limited informa- dients. The placenta also actively concentrates amino
tion from humans. acids in the trophoblast cells of the placenta and then
transfers the amino acids to the fetal plasma, processes
aided by the unique positioning of specific amino acid
transporter proteins and systems on the maternal-f acing
Nutrients for the fetus and fetal-facing trophoblast membranes. The placenta
also consumes nutrient substrates at a very high meta-
The principal metabolic nutrients in the fetus are glucose bolic rate, producing part of the transplacental nutri-
and amino acids. Glucose (including its metabolic prod- ent substrate gradient for glucose and fatty acids, as
uct lactate, which is produced primarily in the placenta) well as specific metabolic products of glucose, lipid, and
serves as the principal substrate in the fetus for mainte- amino acid metabolism that then provide a unique fetal
nance energy production and expenditure, energy storage plasma nutrient milieu. This creates a unique placental–
in glycogen and adipose tissue, and the energy require- fetal metabolic interaction, in which certain substrates
ments of protein synthesis and growth. Amino acids, transported directly to the fetus by the placenta are then
while primarily providing the structural basis for protein metabolized into products for both fetal and, in turn,
synthesis and growth, also serve as oxidative substrates placental metabolism. Collectively, these nutrient met-
for energy production, particularly when glucose supply is abolic and transport processes provide a unique fetal
restricted. Fatty acids also are taken up by the fetus, where nutrient metabolic milieu that supports the essential tis-
they are primarily used for structural components of sue and organ-specific metabolic pathways necessary for
membranes and for growth of adipose tissue. In humans, the high growth rate and rapidly changing development
fatty acid oxidation occurs readily after birth, even in of the fetus.1–4
34
Nutrient supply and fetal metabolic rate 35
Nutrient supply and fetal metabolic rate fatty acid oxidation in the fetus. At markedly reduced rates
of glucose supply to the fetus, fetal glucose utilization rates
Estimates of carbon supply to the fetus are compared (GUR) decrease proportionally, until about 50% of normal
with requirements for energy production and storage in GUR, when GUR is maintained by the development of
Table 5.1. The fetal glucose/oxygen metabolic quotient, an glucose production, initially glycogenolysis, but then also
index of that fraction of fetal oxygen consumption that gluconeogenesis.8,9 Fetal oxygen consumption remains
could be accounted for by complete oxidation of net fetal near normal under such conditions of reduced nutrient
glucose uptake from the placenta, is <1.0.4 Furthermore, supply from the placenta, indicating an increase in the
the fraction of fetal glucose utilization that actually pro- reciprocal oxidation of other substrates, such as glucose
duces CO2 is only c. 0.5–0.6 (Table 5.2). 5–7 Thus, carbon released from glycogen and gluconeogenesis from amino
substrates other than glucose are metabolized to meet the acids. Simultaneously, there is an associated decrease in
oxidative requirements imposed by the rate of fetal oxygen fetal protein accretion rates, which limits the requirement
consumption; as noted earlier, these include lactate and for energy substrates to maintain protein synthesis and net
amino acids principally, given the limited evidence for protein balance.10
Table 5.1 Estimated human fetal nutrient substrate balance in late gestation
Carbon (g/kg/day) Calories (kcal/kg/day)

Requirement
Accretion in carcass: nonfat (human) 3.2 32
Accretion in carcass: fat (human) 3.5 33
Excretion as CO2 4.4 0
Excretion as urea 0.2 2
Excretion as glutamate 0.3 2
Heat (measured as O2 consumption) 0.0 50
Total 11.6 119
Uptake
Amino acids 3.9 45
Glucose 3.7 26
Lactate 1.7 21
Fatty acids 1.1–2.2 17–34
Total 10.4–11.5 109–126
Sources: Reprinted from An Introduction to Fetal Physiology, Battaglia, F.C. and Meschia, G., Copyright
1986, with permission from Elsevier; Reprinted from Fetal and Neonatal Physiology, 4th ed.,
Hay, W.W., Jr., Brown, L.D., and Regnault, T.R.H., Fetal requirements and placental transfer of
nitrogenous compounds, pp. 585–602, Copyright 2011, with permission from Elsevier;
Sparks, J.W. et al., J. Clin. Invest., 70, 179, 1982.
Table 5.2 Fetal carbon substrate oxidation in relation to fetal oxygen

consumption (VO2)a
Carbon for oxidation

Substrate Oxidation fraction (mmol/min/kg) Fraction of fetal VO2
Glucose 0.55 0.09 0.29
Lactate 0.72 0.14 0.50
Amino acids 0.30 0.03 0.09
Total 0.88
Sources: Reprinted from Fetal and Neonatal Physiology, 4th ed., Hay, W.W., Jr., Brown, L.D., and
Regnault, T.R.H., Fetal requirements and placental transfer of nitrogenous compounds,
pp. 585–602, Copyright 2011, with permission from Elsevier; Hay, W.W., Jr. et al., Am. J.
Physiol., 256, E704, 1989; DiGiacomo, J.E. and Hay, W.W., Jr., Metabolism, 39, 193, 1990;
Sparks, J.W. et al., J. Clin. Invest., 70, 179, 1982.
a Estimates derived from data in fetal sheep in late gestation.
36 Nutrient delivery and metabolism in the fetus
Similar to nutrient deprivation, excess delivery of nutri- storage in adipose tissue also are directly related to fetal
ents to the fetus, such as with experimental glucose infusion glucose supply and uptake. Thus, it is not surprising that
into the fetus or mother, decreases amino acid oxidation, but fetuses of hyperglycemic, diabetic mothers tend to con-
has little effect on fetal metabolic rate. A maximal increase tain maximal amounts of hepatic and muscle glycogen and
of c. 15% has been observed in fetal sheep infused directly excess body fat than do fetuses of more normally glycemic
with glucose and insulin.7 The balance of excess glucose mothers, whether they are diabetic or not.17–20
consumption under these conditions maximizes glycogen The concentration of glucose in the fetal plasma declines
stores, and in those fetuses that can produce abundant fat, relative to that in the maternal plasma over the second half
such as the human, augments the growth of adipose tissue. of gestation. This increases the transplacental glucose con-
Excess fetal amino acids are used primarily for oxidation, centration gradient in later gestation, providing a greater
substituting their carbon for that of glucose and lipids, and driving force to supply glucose for the increasing glucose
not excess growth of lean body mass.11–13 Thus, fetuses of dia- requirements of the growing fetus.21 The decrease in fetal
betic mothers tend primarily to be macrosomic (i.e., obese) glucose concentration over the second half of gestation rep-
due to increased storage of glucose and lipids as fat in adi- resents an absolute increase in glucose clearance by the fetus.
pose tissue without increased oxidation of these maternally At least three principal mechanisms are responsible for this
derived fuels/substrates. Similarly, excess insulin in the pres- increase in glucose clearance: The size, cellularity, and glu-
ence of excess amino acids has been shown in fetal sheep to cose metabolic rate of the brain increase relative to other
produce the same linear growth rate with only minimally fetal tissues and organs; there is progressive development of
increased production of lean mass as occurs with normal fetal insulin secretion by the expanding mass of pancreatic
insulin concentrations and amino acid supplies.14,15 islets and beta cells; and there is increased growth of insu-
lin-sensitive tissues, primarily skeletal muscle, and also the
heart and adipose tissue.
Fetal carbohydrate supply and
metabolism Placental glucose transport
The rate of glucose transfer from maternal to fetal plasma Glucose transporters
and the net rate of fetal glucose uptake are directly related Placental glucose uptake and transfer are mediated by Na+-
to the maternal glucose concentration (Figure 5.1a).16 Fetal dependent transport systems on both the maternal-facing
growth rate, glycogen deposition, and fat production and microvillus and fetal-facing basal plasma membranes of the
100 7 10
Fetal glucose uptake (mg/min/kg)
Uterine glucose uptake (mg/min)
Uteroplacental glucose uptake
80 Vmax 6 8
(mg/min/kg of fetus)
GM = 70
GF = 12
60 5 6
40 4 4 GM = 50
20 3 GF = 30 2
km ks
0 2 0
0 40 80 120 160 0 20 40 60 80 0 10 20 30 40
Maternal arterial plasma Maternal arterial plasma Fetal arterial plasma
(a) glucose (mg/dL) (b) glucose (mg/dL) (c) glucose (mg/dL)
Figure 5.1 (a) Schematic representation of the effect of maternal glucose concentration on uterine glucose uptake, based on
experiments in which glucose was infused into pregnant sheep after an overnight fast to produce a large variety of maternal arterial
blood glucose concentrations. Fick principle measurements were then made of net uterine glucose uptake rates versus the maternal
arterial blood glucose concentration, which shows saturation kinetics with an approximate km value in the physiological range of
maternal glucose concentration (about 50–60 mg/dL). (b) Fetal glucose uptake (net transfer of glucose from placenta to fetal circu-
lation) plotted versus maternal arterial glucose concentration, showing a saturable dependence of fetal glucose uptake on maternal
glucose concentration. In addition, this relationship is left-shifted as fetal glucose concentration is decreased, showing that as fetal
glucose concentration is decreased relative to that of the mother, which increases the maternal–fetal glucose concentration gradi-
ent, placental-to-fetal glucose transfer increases. (c) Net rate of uteroplacental glucose consumption in sheep, expressed per kilo-
gram of fetus, plotted versus fetal arterial plasma glucose. Solid line: values measured while maternal arterial plasma glucose was
clamped at about 70 mg/dL. Dotted line: values measured while maternal arterial plasma glucose was clamped at about 50 mg/
dL. These data show that although maternal glucose concentration determines glucose entry into the uteroplacenta and fetus,
actual uteroplacental glucose consumption is regulated largely by the fetal glucose concentration. (Reproduced from Hay, W.W. Jr.,
Placental function, in: Gluckman, P.D. and Heymann, M.A., eds., Pediatrics and Perinatology: The Scientific Basis, 2nd ed., Edward
Arnold, London, U.K., 1996, pp. 213–227. With permission; [a] Adapted from data in Hay, W.W., Jr. and Meznarich, H.K., Proc.
Soc. Exp. Biol. Med., 190, 63, 1988; [b and c] Adapted from data in Hay, W.W., Jr. et al., Am. J. Physiol., 258, R569, 1990.)
Placental glucose transport 37
syncytiotrophoblast.22 The predominant molecular isoforms IGF-1 can upregulate GLUT1, GLUT3, GLUT8, and GLUT9
of glucose transporters (GLUT) in the placenta are GLUT1 sufficiently to correct some aspects of placental insufficiency
and GLUT3,23 although GLUT8 also has been found in the and reduced placental glucose uptake and transport.41
placenta and in sheep is reduced in amount with intrauterine Increasing maternal feed intake to meet calculated energy
growth restriction (IUGR),24 as has GLUT9, both isoforms requirements in previously nutrient-restricted ewes dur-
of which appear to be increased in diabetic pregnancies.25 ing the second half of gestation increases placental mass
Participation of GLUT3 in glucose uptake and transport has and fetal weight and the abundance of GLUT1, an adapta-
not been confirmed in vivo in the human placenta, although tion not observed if maternal food intake is increased above
in sheep placenta, cytochalasin binding assays indicate that requirements,42 the opposite of what occurs with nutrient
GLUT3 might account for as much as 40% of glucose uptake restriction that results in IUGR—inhibition of placental
by the end of gestation.26 Other studies in sheep indicate that mammalian target of rapamycin (mTOR) and insulin/IGF-1
GLUT3 is localized close to mitochondria where it might signaling resulting in downregulation of placental nutrient
serve to support placental glucose uptake and energy pro- transporters may link maternal undernutrition to restricted
duction at low glucose concentrations, thus preserving pla- fetal growth.43
cental function.27 In other studies in human placental tissue,
GLUT3 appears to peak earlier in gestation and declines Kinetics of glucose uptake and transport by the placenta
toward term.28 GLUT3 also regulates intraplacental glucose Although the effect of the maternal glucose concentration on
transport that, in mice, at least, appears reciprocally related net placental-to-fetal glucose transfer demonstrates satura-
to leucine transport.29 Studies in human placentas indicate tion kinetics,16 this relationship does not necessarily define
that GLUT3 is upregulated with intrauterine growth restric- the quantitative characteristics of placental-to-fetal glucose
tion,30 as also found with the original classical studies of transport capacity, because as maternal glucose concentra-
placental hypoxia in vitro.31 Further studies are indicated in tion and placental glucose transport are increased, both
diabetic pregnancies, as there is evidence that fetal macro- fetal glucose concentration and fetal GUR increase. Other
somia is associated with increased expression of GLUT3.32 studies in which glucose was infused directly into the fetus
In vivo studies of human placentas from high-altitude have shown degrees of increase (slope) and saturation of
hypoxia with IUGR indicate reduced GLUT1 expression.33 fetal GUR occurring at about the same fetal glucose con-
With GLUT3 upregulated under similar conditions, this centrations as determined by maternal glucose infusions.44
would set the stage for enhanced placental glucose con- Thus, the maternal glucose infusion approach reflects fetal
sumption but reduced glucose supply to the fetus, perhaps glucose consumption kinetics as well as those of placental-
contributing to fetal growth restriction. Other studies have to-fetal glucose transfer. To address this experimental prob-
noted that corticotropin-releasing hormone (CRH) also lem, different studies have used glucose clamp procedures to
affects GLUT1 (increases) and GLUT3 (decreases) via spe- regulate the maternal-to-fetal glucose concentration gradi-
cific receptors in trophoblast cells, though physiological ent at different maternal and fetal glucose concentrations.45
significance of these changes has not been determined.34 As shown in Figure 5.1b, placental-to-fetal glucose transfer
Related to the effects of CRH, both GLUT1 and GLUT3 are is sensitive to a change in fetal glucose concentration and
downregulated by glucocorticoids, perhaps playing a role in thus the maternal-to-fetal glucose concentration gradient,
how maternal stress can lead to placental insufficiency and regardless of the maternal glucose level.46
intrauterine growth restriction.35 Still other studies have At almost any maternal glucose concentration, therefore
amplified the impact of placental GLUT1 and GLUT3 regu- uteroplacental glucose consumption rate (the metabolism
lation by systemic hormones. For example, one study noted of glucose by the uteroplacenta) is directly related to the
that maternal growth hormone treatment increases fetal fetal glucose concentration (Figure 5.1c). In fact, when the
growth, in part, by enhancing placental nutrient transporter transplacental glucose concentration is abolished, c. 75%
protein expression and hence fetal nutrient supply, as well as of the glucose consumed by the uteroplacenta is supplied
trophoblast proliferation and differentiation, and may have by the fetal circulation.47 These observations imply that the
the potential to ameliorate intrauterine growth restriction.36 fetal side of the uteroplacenta is markedly more perme-
Despite considerable study, the regulation of placen- able to glucose than the maternal side. They also indicate
tal GLUT expression and activity remain poorly defined. that changes in fetal glucose concentration have a strong
Placental GLUT1 is acutely upregulated by hypoxia and influence on placental glucose flux and metabolism. The
hyperglycemia and downregulated by hypoglycemia,37,38 importance of this regulation of placental-to-fetal glucose
while hypoglycemia leads to downregulation, although one transfer and net uteroplacental glucose consumption by
study found upregulation of GLUT1 in the growth-restricted fetal glucose concentration is highlighted by observations
rat placenta.39 Chronic changes in glycemia generally are in chronically hypoglycemic pregnant sheep in which fetal
associated with diminished expression and glucose trans- glucogenesis developed.8 Fetal glucose production contrib-
port.40 In vitro studies indicate that changes in GLUT con- utes glucose molecules to the fetal glucose pool and appears
centrations are related to transport capacity, but this has not to be sufficient in amount to sustain fetal glucose utilization
been demonstrated in vivo, except in experiments in which at near-normal rates. As a result, the placental-to-fetal glu-
the transporters were competitively blocked by pharmaco- cose concentration gradient and the placental-to-fetal
logic inhibitors. A recent study in mice also indicates that glucose transfer rate are relatively reduced; under these
Table 5.3 Metabolic rates in the fetus that account for glucose utilizationa
Glucose utilization rate (mg/min/kg fetus) Percent of total

Whole fetus (sheep, measured) 5.0 100
Whole fetus (human, estimated) 6.0–8.0 100
Brain (sheep, measured) 0.8 16
Brain (human, estimated) 4.0 50–67
Heart (sheep, measured) 0.65 13
Lungs (sheep, estimated) 0.1 2
Liver (sheep, measured) 0.1 2
Red blood cells (human, estimated) 0.1 2
Gut (sheep, estimated) unknown unknown
Carcass/skeletal muscle (estimated, sheep) 3.25 65
Total of organs accounted for
Sheep 5.0 100
Human 8.2 103–137
Sources: Reprinted from An Introduction to Fetal Physiology, Battaglia, F.C. and Meschia, G., Copyright 1986, with permission
from Elsevier; Reprinted from Fetal and Neonatal Physiology, 4th ed., Hay, W.W., Jr., Brown, L.D., and Regnault, T.R.H.,
Fetal requirements and placental transfer of nitrogenous compounds, pp. 585–602, Copyright 2011, with permission
from Elsevier; Sparks, J.W. et al., J. Clin. Invest., 70, 179, 1982.
a Based on data in fetal sheep and estimates for human fetuses for the brain.
circumstances, uteroplacental glucose consumption is is directly related to glucose deprivation.50 Table 5.3 presents
maintained at near-normal rates for the level of maternal estimated utilization rates of glucose in several fetal organs
glycemia. Fetal glucose production can compensate for a and the remaining carcass of fetal sheep in late gestation. All
reduced maternal glucose supply and sustain placental as organs are dependent on the plasma glucose concentration
well as fetal glucose utilization requirements. for their specific rate of glucose uptake, while the skeletal
muscle, heart, and liver develop insulin sensitivity in later
gestation. The acute effect of increased fetal plasma insulin
Gestational changes in placental glucose transfer concentrations results in increased fetal glucose utilization
Placental glucose transport increases markedly over gesta- and decreased fetal plasma glucose concentrations.7 It still
tion. In sheep, the increase in transport capacity accounts is not known to what extent basal insulin concentration
for c. 60% of the increase in placental glucose transport, with affects glucose uptake by specific organs and tissues in the
an increase in the transplacental glucose concentration gra- fetus. However, an acute decrease in the fetal plasma insulin
dient accounting for the remaining 40%.47 This increased concentration, such as by somatostatin infusion (studies in
transport capacity most likely reflects the growth of the fetal sheep), does not affect measurements of the fetal GUR.
surface area of the trophoblast and increased numbers of These procedures do, however, lead to an increase in the fetal
GLUT.26,37 It has not been determined if an increase in tro- glucose concentration,51 which may be a consequence of
phoblast membrane GLUT concentration occurs as well. The decreased glucose disposal or increased glucose production.
regulation of this developmental increase in GLUT abun- Thus, the basal plasma insulin concentration in the fetus
dance, other than by remodeling of the trophoblast mem- appears to regulate glucose production but not utilization;
brane surface area, remains unknown. the latter is more under the control of the plasma glucose
concentration.
A chronic decrease of fetal plasma insulin concentration,
Fetal glucose uptake and utilization however, either by pancreatectomy or streptozotocin injec-
GUR in near-term fetal sheep is c. 5–7 mg/min/kg.48 This tion (a drug that leads to destruction of the pancreatic beta
value is similar to those measured in term human new- cells) into the fetus,52–54 results in an increased fetal plasma
born infants using stable isotope tracer methodology49 and glucose concentration and decreased glucose utilization.55
is about half the value that occurs at midgestation in fetal As discussed previously, fetal hyperglycemia decreases pla-
sheep16 when fetal growth, protein turnover rate, and frac- cental-to-fetal glucose transfer. Chronic hyperglycemia in
tional protein synthetic rates also are about twice those fetal sheep also is associated with decreased peripheral tis-
closer to term. The high correlation between fetal glucose sue insulin sensitivity and glucose utilization capacity,56 as
utilization and growth rates indicates that glucose probably well as the potential release of insulin’s normal inhibition of
serves a major role as the energy supply for the protein syn- hepatic glucose production. Examples of metabolic effects of
thesis required for growth. Indeed, fetal growth restriction increased glucose supply to the fetus are shown in Table 5.4.1
Placental glucose transport 39
is upregulated by hypoglycemia and hypoinsulinemia

Table 5.4 Fetal responses to increased glucose supply
in skeletal muscle and adipose tissue, while there is no
change in the brain. In contrast, hyperglycemia appears to
Acute: mild–moderate downregulate GLUT1 protein concentrations in most tis-
Increased glucose uptake sues.37,59 Insulin-responsive GLUT4 protein is upregulated
Increased insulin production and secretion by hypoglycemia, 37 but in response to hyperglycemia, it is
Hyperinsulinemia initially upregulated and then downregulated toward nor-
Increased glucose utilization mal, or to less than normal, levels in skeletal muscle and
Increased fetal oxygen consumption adipose tissue.59,60 Acute hyperinsulinemia decreases fetal
Mild arterial hypoxemia
glucose concentrations and increases whole fetal GUR, 8 but
it has been difficult to demonstrate in which organs this
Respiratory acidosis
increased GUR takes place. Hyperinsulinemia also appears
Increased placental lactate production
to have acute effects on increasing protein concentrations
Increased fetal lactate uptake and utilization
for both GLUT1 and GLUT4.59,60 In humans, there is con-
Acute: severe siderable evidence for insulin lowering the plasma glucose
Arterial hypoxemia concentration in the preterm and term newborn. The prin-
Hypoinsulinemia cipal action of insulin in the human fetus is uncertain, but
Increased erythropoietin
a unique effect, due to fatty acid supplies, is to promote
lipid formation and deposition in adipose tissue. In this sit-
Increased fetal oxygen consumption
uation, substrate supply (glucose, fatty acids, triglycerides,
Metabolic acidosis
and glycerol) probably is as or more important than insulin
Decreased placental perfusion
itself. Different studies among species, tissues studied, ges-
Fetal demise
tational ages, and conditions of glycemia and insulinemia
Chronic show considerable variability and complexity of changes in
Decreased insulin secretion and/or synthesis GLUT concentrations during fetal life.
Decreased peripheral insulin sensitivity
Increased ratio of placental glucose consumption to
placental glucose transfer
Kinetics of the glucose utilization rate in the fetus
The capacity for glucose utilization in the human fetus
Source: Reproduced from Hay, W.W., Jr. et al., Nutrition and devel- can only be estimated from measurements in infants born
opment of the fetus: Carbohydrate and lipid metabolism,
preterm or in animal models such as the sheep. In preterm
in: Duggan, C.P., Watkins, J.B., Koletzko, B., and Walker,
W.A., eds., Nutrition in Pediatrics: Basic Science and humans, doubling or even tripling of GUR from basal is
Clinical Applications, 5th ed., PMPH-USA, Ltd., Shelton, possible.61 This GUR capacity is variable, however, as
CT, in press, 2015, Chapter 28. increased entry of glucose into the fetal plasma from the
placenta increases fetal glucose concentration and insulin
secretion, which, in turn, augment fetal glucose utilization,
thus limiting further increases in fetal glucose concentra-
As a result of chronic fetal glucose deprivation, from
tion. Glucose and insulin clamp experiments in fetal sheep,
whatever cause, fetal growth rate diminishes and insu-
in which glucose or insulin or both are infused until GUR
lin concentrations are reduced. However, fetuses from
reaches maximal rates, have shown that plasma glucose and
placental insufficiency have increased basal and insulin-
insulin concentrations act independently (i.e., additively) to
stimulated GUR. 57,58 In these fetuses, acute insulin has
increase glucose utilization and oxidation.8 Furthermore,
little effect on protein metabolism.12 These results indicate
while both glucose and insulin enhance fetal GUR accord-
that the predominant growth-regulating effect of insulin
ing to saturation kinetics, it is not known how this effect
in the fetus is its capacity to enhance glucose utilization
is partitioned among different fetal organs. Despite wide
rather than protein metabolism. Furthermore, tissues in
changes in glucose utilization, the relative proportion of
the growth-restricted fetus may have differences in insu-
glucose oxidized during short-term 3- to 4-hour studies—c.
lin sensitivity in late gestation, with peripheral tissues
55%—does not change significantly over the entire range of
remaining insulin sensitive for glucose disposal and the
glucose utilized. Furthermore, because oxygen consump-
liver becoming insulin resistant to suppression of glucose
tion, and thus the fetal metabolic rate, does not vary sig-
production. 57,58
nificantly, if at all, under these circumstances, oxidation of
other carbon substrates, such as amino acids and lactate,
Fetal glucose transporters must increase in compensation. Indeed, acute hypoglyce-
GLUT1 is found throughout the fetal tissues and on all mia in fetal sheep leads to a near doubling of the rate of leu-
endothelial cells and probably accounts for the major- cine oxidation relative to the rate of leucine disposal from
ity of basal tissue glucose uptake from the fetal plasma. the plasma.62 In the human fetus, glucose and/or insulin
GLUT4 is found in the heart, adipose tissue, and skeletal may promote lipogenesis more than oxidation. This may
muscle. In the fetal sheep, GLUT1 protein concentration occur because the human fetus naturally produces adipose
tissue in late gestation, similar to adult humans in whom Insulin, IGF, and other growth factors
higher rates of glucose utilization are partitioned more into
glucose storage in adipose tissue fat than into oxidation. Glucose acts at the transcriptional level to regulate the syn-
Glucose carbon also contributes significantly to the forma- thesis and thus plasma concentrations of both IGF-1 and
tion of glycogen, which is stored primarily in skeletal mus- IGF-2.78,79 Conversely, infusion of IGF-1 into late-gestation
cle, but also in the heart, liver, and lung, and to the carbon fetal sheep promotes both glucose and amino acid utiliza-
contained in amino acids and synthesized proteins. tion.79 In turn, increased fetal plasma IGF-1 concentrations
have been shown to either inhibit fetal protein breakdown80
or promote fetal protein synthesis and accretion,79 depending
Fetal insulin secretion on the availability of circulating amino acids. These observa-
tions indicate that IGF-1 indirectly enhances the capacity for
Glucose-stimulated fetal insulin secretion (measured as an glucose to promote the fetal nitrogen balance and growth.81
acute increase in fetal plasma insulin concentration) increases Likewise, acutely increased insulin concentrations in fetal
more than fivefold during the second half of gestation in fetal sheep promote glucose utilization, amino acid utilization,
sheep and has been documented in third-trimester human and net nitrogen balance,82,83 as well as activate proteins in
fetuses.63,64 Fetal insulin secretion can be modified by the the mitogen-activated protein (MAP) kinase cascade.84 Such
degree, duration, and pattern of changes in the fetal plasma effects are probably short-lived, as chronically raising fetal
glucose concentration. Sustained, marked, relatively constant insulin concentrations above normal does not increase fetal
hyperglycemia actually decreases fetal insulin secretion, while growth beyond normal rates.15
pulsatile hyperglycemia increases fetal insulin secretion and
fetal β-cells.56,65–68 This is consistent with findings in gesta-
tional diabetic women in whom there is a strong tendency Fetal glucose carbon contribution to
to develop increasingly exaggerated, meal-associated hyper-
glycemia in late gestation and associated increases in new- glycogen formation
born insulin secretion.69 Thus, a principal cause of enhanced Many fetal tissues, including the placenta, as well as the
fetal insulin secretion is variability in the magnitude and the brain, liver, lung, heart, and skeletal muscle, produce gly-
intermittent nature of fetal glucose concentration. Fatty acids cogen over the second half of gestation.85 Liver glycogen
also stimulate fetal insulin secretion; their concentrations are content increases with gestational age (Figure 5.2) and is
increased in pregnant diabetics and in their fetuses in late the most important store of glycogen for systemic glucose
gestation, perhaps contributing to augmented insulin secre- needs, because only the liver contains sufficient glucose-
tion.66 Chronic hypoglycemia diminishes fetal insulin secre- 6-phosphatase for the release of glucose into the circula-
tion and β-cell mass.70 Increased amino acid supply augments, tion.86 Skeletal muscle glycogen content increases during late
and decreased amino acid supply attenuates, fetal insulin gestation, whereas lung glycogen content decreases with loss
secretion.71,72 of glycogen-containing alveolar epithelium, development of
type II pneumocytes, and onset of surfactant production.87
Cardiac glycogen concentrations decrease with gestation as
Effect of other hormones on fetal cellular hypertrophy develops. Despite this decrease, the car-
glucose metabolism diac glycogen content is essential for postnatal cardiac energy
supply and cellular function; in fact, deficits of cardiac glyco-
Fetal thyroid hormone indirectly enhances fetal glucose uti- gen are associated with shortened survival time during peri-
lization by increasing the fetal metabolic rate (oxygen con- ods of anoxia.87 Fetal glycogen synthetic rates vary from low,
sumption).73 Changes in fetal plasma cortisol concentrations steady rates of accumulation in species with relatively long
during late gestation have little effect on fetal glucose con- gestations, such as the human and sheep, to exceptionally
centrations or on the rates of glucose utilization.74 However, high rates in species such as the rat that have relatively short
fetal plasma cortisol concentrations do increase in very late gestations. In larger, more slow-growing fetuses (e.g., sheep,
gestation, at which time cortisol-dependent increases in fetal monkey, human), glycogen synthesis by the liver accounts
hepatic glycogenolytic and gluconeogenic enzyme activities for only a small (<10%) portion of fetal glucose utilization.88
develop. These may enhance the glucogenic capacity of the
fetus, thereby contributing to the endogenous glucose pro-
duction observed in normal fetuses just before term and at Fetal glucogenesis
the time of delivery.75 Glucagon and circulating catechol-
amines (adrenal epinephrine and spillover norepineph- Tracer studies in humans89 and sheep16 have shown the same
rine from peripheral nerve endings) are normally present specific activity or enrichment ratio of tracer glucose to
in modest concentrations in the fetal plasma, but they do nonlabeled glucose in fetal as well as in maternal plasma when
stimulate fetal glucogenesis when infused into the fetus. the tracer glucose is infused into the mother and is trans-
Catecholamines promote glucose production at physiologi- ported across the placenta to the fetus. This demonstrates
cal levels,76 but glucagon must reach relatively high concen- that the only source of glucose in the fetus is from the mater-
trations in the fetal plasma to do this.77 nal plasma. Otherwise, new glucose production into the fetal
Fetal lipid metabolism 41
120
100
Liver glycogen or total carbohydrate

(mg glucose/g wet weight)
Monkey
80
Rabbit
60
Sheep
Rat
40
Man
Guinea
pig
20 Dog
Pig
0
0 20 40 60 80 100
0 1 9 17 25
Stage of gestation (%)
Age (days after birth)
The vertical line indicates both term and time of birth.

Man
Rhesus monkey
Sheep
Pig
Dog
Rat
Rabbit
Guinea pig
Figure 5.2 Liver glycogen in various species before and after birth. Hepatic glycogen content in several species is shown to
increase with gestational age, decrease precipitously during the immediate postnatal period, and increase again with a normal
neonatal diet. (From Shelly, H.J., Br. Med. Bull., 17, 137, 1961. With permission.)
plasma either from the fetus itself or from the placenta would Fetal lipid metabolism
dilute the tracer glucose coming from the mother along with
unlabeled glucose, thus lowering the fetal labeled glucose- Placental lipid metabolism and fetal lipid supply
specific activity or enrichment ratio. Furthermore, studies in The amount and type of fatty acid or complex lipid trans-
fetal sheep have shown that the net uptake of glucose by the ported by the placenta varies among species. Placental lipid
fetus from the placenta invariably is equal to the fetal GUR, transport capacity is greatest in the hemochorial placenta of
independently measured with glucose tracers.48,57,90,91 Thus, the human, guinea pig, and rabbit and least in the epithelio-
there is no evidence for fetal glucose production under nor- chorial placenta of the ruminant and the endotheliochorial
mal conditions. Also, there is little if any fetal glucogenesis placenta of the carnivores.95 There are many lipid substances
under the conditions of short-term (1–4 hours) changes in in the plasma that are transported across the placenta that
maternal and fetal glucose concentrations, placental-to-fetal are essential for placental and fetal development, even if they
glucose transfer, and fetal GUR. Measurable rates of fetal glu- do not contribute to nutritional or energy metabolism. Also,
cose production only develop significantly after prolonged brown fat is common to all fetuses; it is essential for postna-
periods (several days) of decreased fetal glucose supply and tal thermogenesis, even if the neonate is not “fat” with white
sustained fetal hypoglycemia and hypoinsulinemia.51–53,92–94 adipose tissue. Furthermore, many lipid substances enter-
The capacity of the fetus to make new glucose molecules from ing the fetus are qualitatively different from those taken up
nonglucose substrates (e.g., lactate, amino acids, and glycerol) by the uterus and uteroplacenta, implying active placental
varies considerably among species. In nearly all cases, this metabolism of individual lipid substances. More complex
appears to be a late gestational development, augmented by pathways include lipoprotein dissociation by placental
hormonal signals including decreased insulin and increased lipoprotein lipase and phospholipase activity.96 Maternal
cortisol and catecholamines that activate phosphoenolpyru- plasma triglycerides are hydrolyzed by these enzymes and
vate carboxykinase, the rate-limiting step for gluconeogen- the released fatty acids are then taken up by the placenta. In
esis, and glucose-6-phosphatase, the enzyme necessary for the trophoblast cells, the fatty acids are then reesterified and
the release of glucose from the liver into the circulation.54,56 further hydrolyzed, facilitating their diffusion into the fetal
Maternal Placenta Fetal Fetal body

circulation circulation
Lipoproteins 20
Percentage body
Lipoprotein 16 Adipose
lipase tissue
content
Mono-and Lipoproteins 12
diglycerides 8
Triglycerides
4
Esterification Hydrolysis
Free fatty acids (FFAs) FFAs 32 36 40
Essential fatty acids (EFAs) Gestational age (weeks)
Phospholipid EFAs
metabolism PUFAs Membrane growth
Polyunsaturated fatty acids (PUFAs)
Keto acids Keto acids Oxidation (brain, liver, muscle)
Glycerol Glycerol Triglyceride synthesis, glucogenesis
Figure 5.3 Schematic of placental–fetal interrelationships in humans for various aspects of placental lipid metabolism, fetal lipid
uptake and metabolism, and fetal lipogenesis into adipose tissue. (Adapted from Hay, W.W., Jr. et al., Nutrition and development
of the fetus: Carbohydrate and lipid metabolism, in: Duggan, C.P., Watkins, J.B., Koletzko, B., and Walker, W.A., eds., Nutrition in
Pediatrics: Basic Science and Clinical Applications, 5th ed., PMPH-USA, Ltd., Shelton, CT, 2015 (in press), Chapter 28; Hay, W.W.,
Jr., Placenta, 16, 19, 1995.)
circulation. Triglycerides also are taken up and metabolized from increased placental lipid transfer or increased de
by complex processes in the placenta, including metabolic novo lipogenesis fueled by increased insulin and glucose
pathways of oxidation, and chain-lengthening, synthe- concentrations.104
sis, and interconversion pathways, followed by release into In humans, umbilical venous–arterial fatty acid concen-
the fetal plasma as free fatty acids (FFAs) or lipoproteins.97 tration differences in cord blood samples show that the net
A schema of placental lipid uptake, metabolism, trans- flux of nonesterified fatty acids into the fetus from the mater-
port, and metabolic interaction with the fetus is shown in nal circulation can account for the fetal requirement of fatty
Figure 5.3.3,97 Recent reviews summarize information about acids during the end of pregnancy.105 Similar estimates in
placental lipid transport and fetal lipid metabolism and nonhuman primates (e.g., monkey)106 indicate that in placen-
accumulation.98–100 tal fatty acid, transfer across the placenta could account for
The most important factor that regulates the flux of lipid all fetal fat deposition in late gestation. Overall, therefore, it
into the placenta and into the various pathways of transport appears that there is a relatively direct relationship between
and metabolism is the concentration of maternal plasma the permeability of the placenta to lipids, especially fatty
lipids, including FFAs and triglycerides. For example, pla- acids, and the adiposity of the fetus at term. Human fetuses
cental triglyceride content increases in women who are develop the most body fat, c. 15%–18% of body weight by the
fasting, who deliver preterm infants, or who have diabe- end of gestation. Guinea pig and rabbit fetuses whose moth-
tes mellitus—all conditions in which maternal plasma FFA ers were fed enriched lab chow are second at c. 12% and c.
concentrations are increased. Fatty acid transport into the 7%, respectively. The sheep fetus develops only about 3%
fetal circulation is largely determined by the transplacental body fat by term, because there appears to practically no fatty
gradient of FFAs relative to available circulating binding acid transfer except for essential fatty acids across the ovine pla-
protein concentrations and the activity and availability of centa. Other small mammals produce even less fetal body fat
their binding sites. In humans, fetal patterns of essential content, such as the mouse at only about 1% (Figure 5.4).
fatty acids and structural lipids correlate directly with the
fatty acid/lipid composition of the maternal plasma (and,
indirectly, the maternal diet).98–100 Fetal lipid metabolism
The fetal impact of maternal plasma FFA and lipid con- Physiological changes that develop in the fetus in late
centrations is further reflected in the fetal lipid content gestation and increase nutrient utilization, such as the
and adipose tissue development.101 Fatter (more macroso- increase in plasma insulin concentration, act to enhance net
mic) human fetuses develop in pregnant women who have maternal-to-fetal fatty acid and lipid transport by increas-
higher BMIs and plasma glucose, triglyceride, and insulin ing fatty acid utilization in the fetus (largely to develop
concentrations, particularly among women with diabetes adipose tissue).107 Increased utilization of fatty acids by
during pregnancy.102 Reducing maternal plasma glucose fetal tissues lowers fetal plasma fatty acid concentrations
and lipid concentrations with a complex carbohydrate diet relative to those in the maternal plasma and increases the
also reduces fetal adiposity.103 Recent data also indicate maternal-to-fetal fatty acid concentration gradients.108 For
that infants from mothers with gestational diabetes melli- example, human maternal venous blood concentrations of
tus have increased neonatal hepatic fat, which could result fatty acids are directly related to the umbilical artery FFA
Placental amino acid uptake and transport to the fetus 43
20
15
Percentage body fat

10
0
Human Guinea Rabbit Sheep Calf Cat Monkey Pig Rat
pig
Figure 5.4 Fetal fat content at term as a percent of fetal body weight among species. (Adapted from Hay, W.W., Jr. et al., Nutrition
and development of the fetus: Carbohydrate and lipid metabolism, in: Duggan, C.P., Watkins, J.B., Koletzko, B., and Walker,
W.A., eds., Nutrition in Pediatrics: Basic Science and Clinical Applications, 5th ed., PMPH-USA, Ltd., Shelton, CT, 2015 (in press),
Chapter 28; Hay, W.W. Jr., Placenta, 16, 19, 1995; Reprinted from An Introduction to Fetal Physiology, Battaglia, F.C. and Meschia, G.,
Copyright 1986, with permission from Elsevier.)
concentrations and the umbilical vein–artery concentra- exchange, but amino acid transport systems also increase
tion differences of FFA.109 their transport activity over gestation.114,115 For example,
Recent evidence supports the unique roles for leptin studies that have evaluated intralitter variation in mice
in fetal metabolism, including enhancement of growth.110 have shown that the lightest placenta of a litter adapts
Umbilical cord leptin concentrations correlate directly with to meet fetal growth demands by increasing surface area
birth weight, and also length and head circumference. They for nutrient exchange as well as amino acid transporter
are higher in macrosomic infants of diabetic mothers and activity.115
lower in infants with IUGR, indicating that the fetal meta- Several factors regulate the expression and activity of
bolic and hormonal milieu independent of fetal fat content placental amino acid transporters. Insulin, IGF-1, inter-
may regulate leptin synthesis and circulating concentra- leukin-6, and tumor necrosis factor-α all have been shown
tions. Interestingly, cord blood leptin levels also predicted to stimulate placental system A transporter activity,117–119
weight gain independent of insulin concentrations over which is the amino acid transporter system responsible
the first 2 years of life. The effect of leptin levels on weight for the uptake of small nonessential neutral amino acids.
gain was independent of birth weight and was still evident Less is known about the regulation of placental system L
at 24 months. Thus, the wide variation in growth during transporter activity, which exchanges neutral amino acids
infancy may be partly regulated by leptin levels present in for essential amino acids against their concentration gra-
utero in response to maternal diet and metabolism and by dient, although a recent study of fetal IUGR produced in
fetal fat accumulation.111 mice by uterine artery ligation could be reversed in part by
IGF-1 gene transfection of the placenta with increased sys-
tem L activity.120 Insulin has been shown to increase system
L activity,117 while a 4-day adiponectin infusion into preg-
Placental amino acid uptake and nant mice downregulated system L activity and decreased
transport to the fetus fetal growth.121 Mechanisms by which these factors regu-
late amino acid transporter activity are currently being
Growth and regulation of placental amino acid elucidated, though evidence supports the involvement of
transport capacity mTOR and signal transducer and activator of transcription
In human and in many other species, fetal body weight in 3 nutrient sensing pathways.117,118
late gestation correlates positively with placental weight.112
As pregnancy advances, the increasing protein synthetic
and nitrogen balance demands of the growing fetus are Fetal amino acid uptake
met by an appropriate increase in placental size and Amino acids are actively concentrated in the trophoblast
amino acid transport capacity. This enhanced transport intracellular matrix by Na+/K+-adenosine triphosphatase-
is facilitated by increases in uteroplacental blood flow, F- and H+-dependent transporter proteins at the maternal-
trophoblast membrane exchange area, and transporter facing microvillous membrane of the trophoblast and then
concentrations in the trophoblast membranes.113 Fetal transported into the fetal plasma.122 Thus, fetal amino acid
growth not only is supported by increased surface area for concentrations are generally higher than maternal levels.123
Maternal Placenta Fetal Fetal

circulation circulation tissues
Direct transport
Ala Ala
Metabolism NH3 Liver
Lactate Lactate Glucose

NH3 NH3
Glycogen
Nitrogen retention
Urea Urea excretion Urea
Metabolic cycles CO2
Gln Gln
CO2 Glu
Oxidation
Ser Gly Gly
Protein Muscle
MeTHF Ser synthesis
CO2 CO2
KIC
Protein
synthesis
Leu Leu
KIC KIC
Unique keto acids
Insulin Pancreas
Signals
Arg Arg
e
Nitric oxid
Vasculature
Figure 5.5 Schematic representation of a variety of placental–fetal metabolic interactions with respect to amino acid uptake
by the placenta, metabolism in the trophoblast cells, direct transfer to the fetus, signaling of fetal vascular and metabolic pro-
cesses, and utilization in fetal tissues. Ala, alanine; Gln, glutamine; Glu, glutamate; Ser, serine; Gly, glycine; Leu, leucine; KIC,
α-ketoisocaproic acid; Arg, arginine; MeTHF, methyl-tetrahydrofolate; NH3, ammonia. (Adapted from Hay, W.W., Jr., Placenta, 16,
19, 1995; Reprinted from Fetal and Neonatal Physiology, 4th edn., Hay, W.W., Jr., Brown, L.D., and Regnault, T.R.H., Fetal require-
ments and placental transfer of nitrogenous compounds, pp. 585–602, Copyright 2011, with permission from Elsevier.)
This active transport process is decreased by hypoxia and can account for c. 0.46 g/day/kg of nitrogen excretion,
inhibitors of protein synthesis,124,125 while hypoamino- which is c. 25% of fetal nitrogen uptake in amino acids.
acidemia increases transport, indicating that synthesis of Such fetal urea production rates are large, exceeding neo-
the transporters is in part responsible for their functional natal and adult weight-specific rates, indicating relatively
state.126 Net total fetal amino acid uptake can account for up rapid protein turnover and oxidation in the fetus. Direct
to 30%–40% of the combined carbon requirements for oxida- measurement of fetal amino acid oxidation using carbon-
tive metabolism and deposition in fetal protein, glycogen, labeled isotopic tracers has determined fetal oxidation rates
and fat, providing 100% of the fetal nitrogen require- for leucine (c. 25% of utilization), lysine (c. 10% of utiliza-
ments.127,128 The placenta and fetus also interact in a variety of tion), and glycine (c. 13% of utilization).130 These studies also
ways to ensure amino acid supply to a large and complex set demonstrated that the fetal oxidation/disposal rate ratio
of vital developmental, metabolic, and signaling processes that is directly related to the excess umbilical uptake of these
are unique to fetal growth and development (Figure 5.5).5 amino acids above that required for protein accretion and
to their plasma concentrations.127,131 Similarly, when amino
acids are exogenously infused into fetal lambs during late
Fetal amino acid metabolism gestation, the rates of leucine oxidation increase and the
glucose/oxygen metabolic quotient decreases, indicating
Fetal amino acid oxidation that the fetus will increase amino acid oxidation rates in
Evidence for a relatively high rate of fetal oxidation of lieu of glucose depending on substrate availability.13
amino acids comes from three observations: Amino acids
are taken up by the fetus in excess of their rate of deposi-
tion in fetal protein5; fetal urea production rates are quite Fetal protein synthesis and turnover
high129; and fetal infusions of carbon-labeled amino acids The net umbilical uptake rates of several nonessential amino
have produced fetal production and excretion of labeled acids are less than their total rate of utilization, emphasiz-
carbon dioxide.62 The urea production rate in fetal sheep ing the need for a relatively high rate of fetal amino acid
References 45
0.3 to the higher metabolic rate and GUR at that stage of ges-
tation.131,132 Thus, protein synthesis per millimole of oxygen
consumed is quite constant from midgestation until term
(Figure 5.6).
Skeletal muscle in the fetal sheep is a major consumer
0.2 of both essential and nonessential amino acids from the
circulation,133 reflecting the relatively high rate of protein
synthesis and nitrogen accretion of the fetus. Insulin and
ks
(per day)
IGF-1 both have been shown to increase blood flow to the
fetal hindlimb (used as an estimate of skeletal muscle-
specific metabolic rates),133,134 promoting uptake of amino
0.1
acids by the hindlimb. Euglycemic hyperinsulinemia
also has been shown to promote the net uptake of several
kG
amino acids as well as increase glucose utilization by the
(per day)
fetal hindlimb.134,135 Therefore, it is reasonable to hypoth-
esize that insulin-induced glucose utilization by muscle
0 increases net protein balance, in part by substituting car-
80 100 120 140 bon from glucose for that of amino acids in the tricarbox-
Fetal age (days) ylic acid cycle.136
Figure 5.6 Fractional rate of protein synthesis (kS) over
gestation in fetal sheep studied with leucine (●) and lysine (O)
radioactive tracers compared with the fractional rate of growth
(kG) in the lower portion of the figure. (Reprinted from An
Acknowledgments
Introduction to Fetal Physiology, Battaglia, F.C. and Meschia, G., The preparation of this chapter was supported in part by
Copyright 1986, with permission from Elsevier; Meier, P.R.
et al., Am. J. Physiol., 240, E320, 1981; Kennaugh, J.M. et al., the Bill & Melinda Gates Foundation Grand Challenges
Pediatr. Res., 22, 688, 1987.) Exploration Grant OPP1061082 (WWH, PI), NIH Training
Grant T32HD007186-32 (WWH, PI and PD), NIH
production.5,130 Protein synthetic rates also are high com- K12HD068372 (WWH PD), NIH UL1TR001082 (WWH,
pared to the adult. Fractional protein synthetic rate (kS) and Co-PD), NIH R01DK088139 and K08HD060688 (PJR, PI;
fractional growth rate (kG) are higher in the midgestation WWH, Co-I), NIH/UCDenver K12HD057022 (LDB, PI),
fetus compared to the near-term fetus, both proportional NIH KO1DK090199, and NIH R03DK102972 (SRW, PI).
REFERENCES
1. Hay WW, Jr., Brown LD, Thorn S, Rozance PJ. Nutrition and 10. Limesand SW, Rozance PJ, Brown LD, Hay WW, Jr. Effects of
development of the fetus: Carbohydrate and lipid metabo- chronic hypoglycemia and euglycemic correction on lysine
lism. In: Duggan CP, Watkins JB, Koletzko B, Walker WA, metabolism in fetal sheep. Am J Physiol Endocrinol Metab April
eds., Nutrition in Pediatrics: Basic Science and Clinical 2009; 296(4): E879–E887.
Applications, 5th ed., PMPH-USA, Ltd.: Shelton, CT, 2015, in 11. Rozance PJ, Crispo MM, Barry JS et al. Prolonged maternal amino
press, Chapter 28. acid infusion in late-gestation pregnant sheep increases fetal
2. Hay WW, Jr. Energy and substrate requirements of the placenta amino acid oxidation. Am J Physiol Endocrinol Metab September
and fetus. Proc Nutr Soc 1991; 50: 321–336. 2009; 297(3): E638–E646.
3. Hay WW, Jr. Metabolic interrelationships of placenta and fetus. 12. Brown LD, Rozance PJ, Thorn SR et al. Acute supplementation of
Placenta 1995; 16: 19–30. amino acids increases net protein accretion in IUGR fetal sheep.
4. Battaglia FC, Meschia G. An Introduction to Fetal Physiology, Am J Physiol Endocrinol Metab August 1, 2012; 303(3): E352–E364.
Academic Press: Orlando, FL, 1986. 13. Maliszewski AM, Gadhia MM, O’Meara MC et al. Prolonged
5. Hay WW, Jr., Brown LD, Regnault TRH. Fetal requirements and infusion of amino acids increases leucine oxidation in fetal
placental transfer of nitrogenous compounds. In: Polin RA, Fox sheep. Am J Physiol Endocrinol Metab June 15, 2012; 302(12):
WW, Abman SH, eds., Fetal and Neonatal Physiology, 4th ed., E1483–E1492.
Elsevier: Philadelphia, PA, 2011, pp. 585–602. 14. Susa JB, McCormick KL, Widness JA et al. Chronic hyperinsu-
6. Sparks JW, Hay WW, Jr., Bonds D et al. Simultaneous measure- linemia in the fetal rhesus monkey: Effects on fetal growth and
ments of lactate turnover rate and umbilical lactate update in the composition. Diabetes 1979; 28: 1058–1063.
fetal lamb. J Clin Invest 1982; 70: 179–192. 15. Fowden AL, Hughes P, Comline RS. The effects of insulin on
7. Hay WW, Jr., DiGiacomo JE, Meznarich HK et al. Effects of glu- the growth rate of the sheep fetus during late gestation. Q J Exp
cose and insulin on fetal glucose oxidation and oxygen consump- Physiol 1989; 74: 703–714.
tion. Am J Physiol 1989; 256: E704–E713. 16. Hay WW, Jr., Meznarich HK. Effect of maternal glucose concen-
8. DiGiacomo JE, Hay WW, Jr. Fetal glucose metabolism and oxy- tration on uteroplacental glucose consumption and transfer in
gen consumption during sustained maternal and fetal hypoglyce- pregnant sheep. Proc Soc Exp Biol Med 1988; 190: 63–69.
mia. Metabolism 1990; 39: 193–202. 17. Osler M, Pedersen J. The body composition of newborn infants
9. Narkewicz MR, Carver TD, Hay WW, Jr. Induction of cytosolic of diabetic mothers. Pediatrics 1960; 26: 985–992.
phosphoenolpyruvate carboxykinase in the ovine fetal liver by 18. Mughal MZ, Eelloo J, Roberts SA et al. Body composition and
chronic fetal hypoglycemia and hypoinsulinemia. Pediatr Res bone status of children born to mothers with type 1 diabetes mel-
May 1993; 33(5): 493–496. litus. Arch Dis Child 2010; 95: 281–285.
19. Pribylová H, Rázová M. Glycogen reserves in the subcutaneous 43. Kavitha JV, Rosario FJ, Nijland MJ et al. Down-regulation of pla-
adipose tissue and energy metabolism in newborns of diabetic cental mTOR, insulin/IGF-I signaling, and nutrient transporters in
mothers. Biol Neonate 1981; 39: 22–27. response to maternal nutrient restriction in the baboon. FASEB J
20. Catalano PM, Thomas A, Huston-Presley L et al. Increased fetal 2014; 28: 1294–1305.
adiposity: A very sensitive marker of abnormal in utero develop- 44. Hay WW, Jr., Meznarich HK, DiGiacomo JE et al. Effects of insu-
ment. Am J Obstet Gynecol 2003; 189: 1698–1704. lin and glucose concentrations on glucose utilization in fetal
21. Molina RD, Meschia G, Battaglia FC et al. Maturation of placental sheep. Pediatr Res 1988; 23: 281–287.
glucose transfer capacity in the ovine pregnancy. Am J Physiol 45. Hay WW, Jr., Molina RD, DiGiacomo JE et al. Model of placental
1991; 261: R697–R704. glucose consumption and glucose transfer. Am J Physiol 1990;
22. Illsley NP. Glucose transporters in the human placenta. Placenta 258: R569–R577.
January 2000; 21(1): 14–22. 46. Hay WW, Jr. Placental function. In: Gluckman PD, Heymann
23. Hahn T, Desoye G. Ontogeny of glucose transport systems in MA, eds., Pediatrics and Perinatology: The Scientific Basis, 2nd
the placenta and its progenitor tissues. Early Pregnancy 1996; 2: ed., Edward Arnold: London, U.K., 1996, pp. 213–227.
168–182. 47. Molina RD, Meschia G, Battaglia FC et al. Gestational matura-
24. Limesand SW, Regnault TR, Hay WW, Jr. Characterization of glu- tion of placental glucose transfer capacity in sheep. Am J Physiol
cose transporter 8 (GLUT8) in the ovine placenta of normal and 1991; 261: R697–R704.
growth restricted fetuses. Placenta 2004; 25: 70–77. 48. Hay WW, Jr., Sparks JW, Wilkening RB et al. Fetal glucose uptake
25. Bibee KP, Illsley NP, Moley KH. Asymmetric syncytial expression and utilization as functions of maternal glucose concentration.
of GLUT9 splice variants in human term placenta and alterations Am J Physiol 1984; 246: E237–E242.
in diabetic pregnancies. Reprod Sci 2011; 18: 20–27. 49. Kalhan SC, Savin SM, Adam PAJ. Measurement of glucose turn-
26. Ehrhardt RA, Bell AW. Developmental increases in glucose trans- over in the human newborn with glucose-1–13C. J Clin Endocrinol
porter concentration in the sheep placenta. Am J Physiol 1997; Metab 1976; 43: 704–707.
273: R1132–R1141. 50. Carver TD, Quick AN, Jr., Teng CC et al. Leucine metabolism
27. Wooding FB, Fowden AL, Bell AW et al. Localisation of glucose in chronically hypoglycemic, hypoinsulinemic growth restricted
transport in the ruminant placenta: Implications for sequential fetal sheep. Am J Physiol 1997; 272: E107–E117.
use of transporter isoforms. Placenta 2005; 26: 626–640. 51. DiGiacomo JE, Hay WW, Jr. Effect of hypoinsulinemia and hyper-
28. Brown K, Heller DS, Zamudio S et al. Glucose transporter 3 glycemia on fetal glucose utilization and oxidation. Am J Physiol
(GLUT3) protein expression in human placenta across gestation. 1990; 259: E506–E512.
Placenta 2011; 32: 1041–1049. 52. Hay WW, Jr., Meznarich HK, Fowden AL. Effect of streptozotocin
29. Ganguly A, Collis L, Devaskar SU. Placental glucose and amino of rates of ovine fetal glucose utilization, oxidation and produc-
acid transport in calorie-restricted wild-type and Glut3 null het- tion in the sheep fetus. Metabolism 1988; 38: 30–37.
erozygous mice. Endocrinology 2012; 153: 3995–4007. 53. Fowden AL, Hay WW, Jr. The effects of pancreatectomy on the
30. Janzen C, Lei MY, Cho J et al. Placental glucose transporter 3 rates of glucose utilization, oxidation and production in the sheep
(GLUT3) is up-regulated in human pregnancies complicated by fetus. Q J Exp Physiol 1988; 73: 973–984.
late-onset intrauterine growth restriction. Placenta 2013; 34: 54. Fowden AL. The endocrine regulation of fetal metabolism and
1072–1078. growth. In: Gluckman PD, Johnston BM, Nathanielsz PW, eds.,
31. Esterman A, Greco MA, Mitani Y et al. The effect of hypoxia Advances in Fetal Physiology: Reviews in Honor of GC Liggins,
on human trophoblast in culture: Morphology, glucose transport Perinatology Press: Ithaca, NY, 1989, pp. 229–243.
and metabolism. Placenta 1997; 18: 129–136. 55. Fowden AL, Forhead AJ. Insulin deficiency alters the metabolic
32. Kainulainen H, Jarvinen T, Heinonen PK. Placental glucose trans- and endocrine responses to undernutrition in fetal sheep near
porters in fetal intrauterine growth retardation and macrosomia. term. Endocrinology August 2012; 153(8): 4008–4018.
Gynecol Obstet Invest 1997; 44: 89–92. 56. Carver TD, Anderson SM, Aldoretta PW et al. Glucose suppres-
33. Zamudio S, Baumann MU, Illsley NP. Effects of chronic hypoxia sion of insulin secretion in chronically hyperglycemic fetal sheep.
in vivo on the expression of human placental glucose transport- Pediatr Res 1995; 38: 754–762.
ers. Placenta 2006; 27: 49–55. 57. Thorn SR, Brown LD, Rozance PJ et al. Increased hepatic glu-
34. Gao L, Lv C, Xu C et al. Differential regulation of glucose cose production in fetal sheep with intrauterine growth restric-
transporters mediated by CRH receptor type 1 and type 2 tion is not suppressed by insulin. Diabetes January 2013; 62(1):
in human placental trophoblasts. Endocrinology 2012; 153: 65–73.
1464–1471. 58. Thorn SR, Rozance PJ, Brown LD et al. The intrauterine growth
35. Hahn T, Barth S, Graf R et al. Placental glucose transporter restriction phenotype: Fetal adaptations and potential implica-
expression is regulated by glucocorticoids. J Clin Endocrinol tions for later life insulin resistance and diabetes. Semin Reprod
Metab 1999; 84: 1445–1452. Med May 2011; 29(3): 225–236.
36. Tung E, Roberts CT, Heinemann GK et al. Increased placental 59. Anderson MS, He J, Flowers-Ziegler J et al. Effects of selec-
nutrient transporter expression at midgestation after maternal tive hyperglycemia and hyperinsulinemia on glucose trans-
growth hormone treatment in pigs: A placental mechanism for porters in fetal ovine skeletal muscle. Am J Physiol 2001; 50:
increased fetal growth. Biol Reprod 2012; 87: 1–8. R1256–R1263.
37. Das UG, Schroeder RE, Hay WW, Jr. et al. Time-dependent and 60. Anderson MS, Ziegler JA, Das UG et al. Glucose transporter pro-
tissue-specific effects of circulating glucose on fetal ovine glu- tein responses to selective hyperglycemia or hyperinsulinemia in
cose transporters. Am J Physiol 1999; 276: R809–R817. fetal sheep. Am J Physiol 2001; 281: R1545–R1552.
38. Baumann MU, Zamudio S, Illsley NP. Hypoxic upregulation of 61. Zarlengo KM, Battaglia FC, Fennessey P et al. Relationship
glucose transporters in BeWo choriocarcinoma cells is mediated between glucose utilization rate and glucose concentration in
by hypoxia-inducible factor-1. Am J Physiol Cell Physiol 2007; preterm infants. Biol Neonate 1986; 49: 181–189.
293: C477–C485. 62. van Veen LCP, Teng C, Hay WW, Jr. et al. Leucine disposal and
39. Reid GJ, Flozak AS, Simmons RA. Increased expression of glu- oxidation rates in the fetal lamb. Metabolism 1987; 36: 48–53.
cose transporter protein-1 (GLUT-1) in the growth retarded pla- 63. Aldoretta PW, Gresores A, Carver TD et al. Maturation of
centa. J Soc Gynecol Invest 1995; 2: 193. glucose-stimulated insulin secretion. Biol Neonate 1998; 73:
40. Osmond DT, Nolan CJ, King RG et al. Effects of gestational diabe- 375–386.
tes on human placental glucose uptake, transfer, and utilisation. 64. Nicolini U, Hubinont C, Santolaya J et al. Effects of fetal intrave-
Diabetologia 2000; 43: 576–582. nous glucose challenge in normal and growth retarded fetuses.
41. Jones HN, Crombleholme T, Habli M. Adenoviral-mediated pla- Horm Metab Res August 1990; 22(8): 426–430.
cental gene transfer of IGF-1 corrects placental insufficiency via 65. Kervran A, Randon J. Development of insulin release by fetal rat
enhanced placental glucose transport mechanisms. PLOS ONE pancreas in vitro. Diabetes 1980; 29: 673–678.
2013; 8(9): e74632. 66. Kervran A, Guillaume M, Jost A. The endocrine pancreas of the
42. Dandrea J, Wilson V, Gopalakrishnan G et al. Maternal nutri- fetus of diabetic pregnant rat. Diabetologia 1978; 15: 387–393.
tional manipulation of placental growth and glucose transporter 67. Carver TD, Anderson SM, Aldoretta PW et al. Effect of low-level
1 (GLUT-1) abundance in sheep. Reproduction 2001; 122: plus marked ‘pulsatile’ hyperglycemia on insulin secretion in fetal
793–800. sheep. Am J Physiol 1996; 271: E865–E871.
References 47
68. Frost MS, Zehri AH, Limesand SW et al. Differential effects 93. Thorn SR, Sekar SM, Lavezzi JR et al. A physiological increase in
of chronic pulsatile versus chronic constant maternal hyperinsulin suppresses gluconeogenic gene activation in fetal sheep
glycemia on fetal pancreatic β-cells. J Pregnancy 2012; 2012: with sustained hypoglycemia. Am J Physiol Regul Integr Comp
812094. Physiol October 15, 2012; 303(8): R861–R869.
69. Freinkel N, Phelps NL, Metzger BE. Intermediary metabolism 94. Rozance PJ, Limesand SW, Barry JS et al. Chronic late-gestation
during normal pregnancy. In: Sutherland HW, Stowers JM, eds., hypoglycemia upregulates hepatic PEPCK associated with
Carbohydrate Metabolism in Pregnancy and the Newborn, increased PGC1alpha mRNA and phosphorylated CREB in fetal
Springer-Verlag: New York, 1979, pp. 1–31. sheep. Am J Physiol Endocrinol Metab February 2008; 294(2):
70. Lavezzi JR, Thorn SR, O’Meara MC et al. Increased fetal insu- E365–E370.
lin concentrations for one week fail to improve insulin secretion 95. DiGiacomo JE, Hay WW, Jr. Regulation of placental glucose
or β-cell mass in fetal sheep with chronically reduced glucose transfer and consumption by fetal glucose production. Pediatr
supply. Am J Physiol Regul Integr Comp Physiol January 1, 2013; Res 1989; 25: 429–434.
304(1): R50–R58. 96. Widdowson EM. Growth and composition of the human fetus
71. Gadhia MM, Maliszewski AM, O’Meara MC et al. Increased and newborn. In: Assali NS, ed., Biology of Gestation, Vol. 2,
amino acid supply potentiates glucose-stimulated insulin secre- Academic Press: New York, 1968, pp. 1–48.
tion but does not increase β-cell mass in fetal sheep. Am J Physiol 97. Rice GE, Wong MH, Farrugia W, Scott KF. Contribution of type II
Endocrinol Metab February 15, 2013; 304(4): E352–E362. phospholipase A2 to in vitro phospholipase A2 enzymatic activ-
72. Cherif H, Reusens B, Ahn MT et al. Effects of taurine on the insu- ity in human term placenta. J Endocrinol 1998; 157: 25–31.
lin secretion of rat fetal islets from dams fed a low-protein diet. 98. Coleman RA. Placental metabolism and transport of lipid. Fed
J Endocrinol November 1998; 159(2): 341–348. Proc 1986; 45: 2519–2523.
73. Fowden AL, Silver MA. The effects of thyroid hormones on oxy- 99. Auestad N. Diverse roles of lipids in neonatal physiology and
gen and glucose metabolism in the sheep fetus during late gesta- development. In: Neu J, ed., Gastroenterology and Nutrition:
tion. J Physiol 1995; 482: 203–213. Neonatology Questions and Controversies (Polin RA, Consulting
74. Barnes RJ, Comline RS, Silver M. Effect of cortisol on liver gly- eds.), Saunders, Elsevier: Philadelphia, PA, 2008, pp. 253–277.
cogen concentrations in hypophysectomized, adrenalectomized 100. Haggarty P. Effect of placental function on fatty acid require-
and normal foetal lambs during late or prolonged gestation. ments during pregnancy. Eur J Clin Nutr 2004; 58: 1559–1570.
J Physiol 1978; 275: 567–579. 101. Haggarty P. Fatty acid supply to the human fetus. Annu Rev Nutr
75. Fowden AL, Comline RS, Silver M. The effects of cortisol on the 2010; 30: 237–255.
concentration of glycogen in different tissues in the chronically 102. Magnusson AL, Waterman IJ, Wennergren M et al. Triglyceride
catheterized fetal pig. Q J Exp Physiol 1985; 70: 23–32. hydrolase activities and expression of fatty acid binding proteins
76. Padbury JF, Ludlow JK, Ervin MG et al. Thresholds for physiologi- in the human placenta in pregnancies complicated by intrauter-
cal effects of plasma catecholamines in fetal sheep. Am J Physiol ine growth restriction and diabetes. J Clin Endocrinol Metab
1992; 252: E530–E537. 2004; 89: 4607–4614.
77. Devaskar SU, Ganguli S, Styer D et al. Glucagon and glucose 103. Harmon KA, Gerard L, Jensen DR et al. Continuous glucose
dynamics in sheep: Evidence for glucagon resistance in fetus. Am profiles in obese and normal-weight pregnant women on a con-
J Physiol 1984; 246: E256–E265. trolled diet: Metabolic determinants of fetal growth. Diabetes
78. Oliver MH, Harding JE, Breier BH et al. Glucose but not mixed Care 2011; 34: 2198–2204.
amino acid infusion regulates plasma insulin-like growth factor-I 104. Hernandez TL, Van Pelt RE, Anderson MA et al. A higher-complex
concentrations in fetal sheep. Pediatr Res 1993; 34: 62–65. carbohydrate diet in gestational diabetes mellitus achieves glu-
79. Shen W, Wisniowski P, Ahmed L et al. Protein anabolic effects cose targets and lowers postprandial lipids: A randomized cross-
of insulin and IGF-I in the ovine fetus. Am J Physiol Endocrinol over study. Diabetes Care 2014; 37: 1254–1262.
Metab April 2003; 284(4): E748–E756. 105. Brumbaugh DE, Tearse P, Cree-Green M et al. Intrahepatic fat is
80. Liechty EA, Boyle DW, Moorehead H et al. Effects of circulating increased in the neonatal offspring of obese women with gesta-
IGF-I on glucose and amino acid kinetics in the ovine fetus. Am J tional diabetes. J Pediatr May 2013; 162(5): 930–936.
Physiol 1996; 271: E177–E185. 106. Elphick MC, Hull D, Sanders RR. Concentrations of free fatty
81. Liechty EA, Boyle DW, Moorehead H et al. Glucose and amino acid acids in maternal and umbilical cord blood during elective cesar-
kinetic response to graded infusion of rhIGF-I in the late gestation ean section. Br J Obstet Gynaecol 1976; 83: 539–544.
ovine fetus. Am J Physiol September 1999; 277(3 Pt. 1): E537–E543. 107. Portman OW, Behrman RE, Soltys P. Transfer of free fatty acids
82. Liechty EA, Boyle DA, Moorehead H et al. Effect of hyperinsu- across the primate placenta. Am J Physiol 1969; 216: 143–147.
linemia on ovine fetal leucine kinetics during prolonged maternal 108. Sparks JW, Girard J, Battaglia FC. An estimate of the caloric require-
fasting. Am J Physiol 1992; 263: E696–E702. ments of the human fetus. Biol Neonate 1980; 38: 113–119.
83. Thureen PJ, Scheer B, Anderson SM et al. Effect of hyperinsu- 109. Haggarty P, Page K, Abramovich DR et al. Long-chain polyunsat-
linemia on amino acid utilization in the ovine fetus. Am J Physiol urated fatty acid transport across the perfused human placenta.
2000; 279: E1294–E1304. Placenta 1997; 18: 635–642.
84. Stephens E, Thureen PJ, Goalstone ML et al. Fetal hyperinsu- 110. Hendrickse W, Stammers JP, Hull D. The transfer of free fatty
linemia increases farnesylation of p21 Ras in fetal tissues. Am J acids across the human placenta. Br J Obstet Gynaecol 1985; 92:
Physiol 2001; 281: E217–E223. 945–953.
85. Shelley HJ. Glycogen reserves and their changes at birth and in 111. Hoggard N, Hunter L, Lea RG et al. Ontogeny of the expression
anoxia. Br Med Bull 1961; 17: 137–143. of leptin and its receptor in the murine fetus and placenta. Br J
86. Dawkins MJR. Biochemical aspects of developing function in Nutr 2000; 83: 317–326.
newborn mammalian liver. Br Med Bull 1961; 22: 28–33. 112. Ong KK, Ahmed ML, Sherrif A et al. Cord blood leptin is associ-
87. Shellhase E, Kuroki Y, Emrie PA et al. Expression of pulmonary ated with size at birth and predicts infancy weight gain in humans.
surfactant apoproteins in the developing rat lung. Clin Res 1989; ALSPAC Study Team. Avon Longitudinal Study of Pregnancy and
37: 208A. Childhood. J Clin Endocrinol Metab 1999; 84: 1145–1148.
88. Mott JC. The ability of young mammals to withstand total oxygen 113. Molteni RA, Stys SJ, Battaglia FC. Relationship of fetal and pla-
lack. Br Med Bull 1961; 17: 144–148. cental weight in human beings: Fetal/placental weight ratios at
89. Sparks JW. Augmentation of glucose supply. Semin Perinatol various gestational ages and birth weight distributions. J Reprod
1979; 3: 141–155. Med November 1978; 21(5): 327–334.
90. Marconi A, Cetin E, Davoli A et al. An evaluation of fetal gluco- 114. Regnault TR, Friedman JE, Wilkening RB et al. Fetoplacental
genesis in intrauterine growth retarded pregnancies: Steady state transport and utilization of amino acids in IUGR—A review.
fetal and maternal enrichments of plasma glucose at cordocente- Placenta April 2005; 26(Suppl. A): S52–S62.
sis. Metabolism 1993; 42: 860–864. 115. Mahendran D, Byrne S, Donnai P et al. Na+ Transport, H+ con-
91. Hay WW, Jr., Sparks JW, Quissel BJ et al. Simultaneous measure- centration gradient dissipation, and system A amino acid
ments of umbilical glucose uptake, fetal utilization rate, and fetal transporter activity in purified microvillous plasma membrane
turnover rate of glucose. Am J Physiol 1981; 240: E662–E668. isolated from first-trimester human placenta: Comparison with
92. Gleason CA, Rudolph AM. Gluconeogenesis by the fetal sheep the term microvillous membrane. Am J Obstet Gynecol 1994;
liver in vivo. J Dev Physiol 1985; 7: 185–194. 171: 1534–1540.
116. Novak DA, Beveridge MJ, Malandro M et al. Ontogeny of amino 125. Milley JR. Protein synthesis during hypoxia in fetal lambs. Am J
acid transport system A in rat placenta. Placenta 1996; 17: Physiol 1987; 252: E519–E524.
643–651. 126. Milley JR. Uptake of exogenous substrates during hypoxia in fetal
117. Coan PM, Angiolini E, Sandovici I et al. Adaptations in placental lambs. Am J Physiol 1988; 254: E572–E578.
nutrient transfer capacity to meet fetal growth demands depend 127. Smith CH. Incubation techniques and investigation of placental
on placental size in mice. J Physiol September 15, 2008; 586(Pt. transport mechanisms in vitro. Placenta 1981; 2: 163–168.
18): 4567–4576. 128. Lemons JA, Adcock EW, 3rd, Jones MD, Jr. et al. Umbilical uptake
118. Roos S, Lagerlöf O, Wennergren M et al. Regulation of amino of amino acids in the unstressed fetal lamb. J Clin Invest 1976; 58:
acid transporters by glucose and growth factors in cultured 1428–1434.
primary human trophoblast cells is mediated by mTOR sig- 129. Marconi AM, Battaglia FC, Meschia G et al. A comparison of
naling. Am J Physiol Cell Physiol September 2009; 297(3): amino acid arteriovenous differences across the liver, hindlimb
C723–C731. and placenta in the fetal lamb. Am J Physiol 1989; 257: E909–E915.
119. Jones HN, Jansson T, Powell TL. IL-6 stimulates system A amino 130. Hay WW Jr., Brown LD, Regnault TRH. Fetal requirements and
acid transporter activity in trophoblast cells through STAT3 placental transfer of nitrogenous compounds. In: Polin RA, Fox
and increased expression of SNAT2. Am J Physiol Cell Physiol WW, Abman SH, eds., Fetal and Neonatal Physiology, 4th ed.,
November 2009; 297(5): C1228–C1235. Elsevier: Philadelphia, PA, 2001, pp. 585–602.
120. Sferruzzi-Perri AN, Owens JA, Standen P et al. Early treatment 131. Gresham EL, James EJ, Raye JR et al. Production and excretion of
of the pregnant guinea pig with IGFs promotes placental trans- urea by the fetal lamb. Pediatrics 1972; 50: 372–379.
port and nutrient partitioning near term. Am J Physiol Endocrinol 132. Battaglia FC, Meschia G. Fetal nutrition. Annu Rev Nutr 1988; 8:
Metab March 2007; 292(3): E668–E676. 43–61.
121. Jones H, Crombleholme T, Habli M. Regulation of amino acid 133. Meier PR, Peterson RG, Bonds DR et al. Rates of protein synthesis
transporters by adenoviral-mediated human insulin-like growth and turnover in fetal life. Am J Physiol 1981; 240: E320–E324.
factor-1 in a mouse model of placental insufficiency in vivo and 134. Kennaugh JM, Bell AW, Teng C et al. Ontogenetic changes in the
the human trophoblast line BeWo in vitro. Placenta February rates of protein synthesis and leucine oxidation during fetal life.
2014; 35(2): 132–138. Pediatr Res 1987; 22: 688–692.
122. Rosario FJ, Schumacher MA, Jiang J et al. Chronic maternal infu- 135. Wilkening RB, Boyle DW, Teng C et al. Amino acid uptake by
sion of full-length adiponectin in pregnant mice down-regulates fetal ovine hindlimb under normal and euglycemic hyperinsulin-
placental amino acid transporter activity and expression and emic states. Am J Physiol 1994; 266: E72–E78.
decreases fetal growth. J Physiol March 15, 2012; 590(Pt. 6): 136. Boyle DW, Denne SC, Moorehead H et al. Effect of rhIGF-I infu-
1495–1509. sion on whole fetal and fetal skeletal muscle protein metabolism
123. Smith CH, Moe AJ, Ganapathy V et al. Nutrient transport path- in sheep. Am J Physiol December 1998; 275(6 Pt. 1): E1082–E1091.
ways across the epithelium of the placenta. Annu Rev Nutr 1992; 137. Wilkening RB, Molina RD, Battaglia FC et al. Effect of insulin on
12: 183–206. glucose/oxygen and lactate/oxygen quotients across the hindlimb
124. Cetin I, Marconi AM, Corbetta C et al. Fetal amino acids in nor- of fetal lambs. Biol Neonate 1987; 51(1): 18–23.
mal pregnancies and in pregnancies complicated by intrauter- 138. Liechty EA, Boyle DW, Moorehead H et al. Increased fetal
ine growth retardation. Early Hum Dev June–July 1992; 29(1–3): glucose concentration decreases ovine fetal leucine oxidation
183–186. independent of insulin. Am J Physiol 1993; 265: E617–E623.
6 Pathogenesis of gestational
diabetes mellitus
Yariv Yogev
screening challenge tests. A detailed discussion of variations

Introduction in the diagnostic criteria is beyond the scope of this chapter,
Gestational diabetes mellitus (GDM) is characterized by but the main issue is that the diagnosis of GDM is based on
carbohydrate intolerance of variable severity, with onset or the screening of a large number of apparently healthy young
first recognition during pregnancy. This definition applies women.
whether or not there is a need for insulin and whether or not As in type 1 diabetes mellitus, GDM is associated with
it disappears after the pregnancy. It does not apply to gravid both insulin resistance and impaired insulin secretion.2–4
patients with previously diagnosed diabetes.1 A detailed dis- The two disorders also share the same risk factors, have a cor-
cussion of glucose regulation in pregnancy is beyond the responding prevalence within a given population, and have
scope of this chapter. However, two points are important the same genetic susceptibility; therefore, they are assumed
for the discussion that follows. First, pregnancy is normally to be etiologically indistinct, with one preceding the other.
attended by progressive insulin resistance that begins near In this chapter, the development of insulin resistance
midpregnancy and progresses through the third trimester to during pregnancy, hormones, and newly discovered factors
levels that approximate the insulin resistance seen in indi- associated with insulin resistance and secretion, the insulin-
viduals with type 2 diabetes. The insulin resistance appears signaling system during normal and diabetic pregnancy,
to result from a combination of increased maternal adiposity and metabolic predictors of diabetes will be discussed.
and the insulin-desensitizing effects of hormonal products
of the placenta. The fact that insulin resistance rapidly abates
following delivery suggests that the major contributors to Insulin sensitivity and resistance in
this state of resistance are placental hormones. Second, pregnancy
pancreatic β cells normally increase their insulin secretion
to compensate for the insulin resistance of pregnancy. As a The majority of women with GDM appear to have β-cell
result, changes in circulating glucose levels over the course of dysfunction that occurs on a background of chronic insulin
pregnancy are quite small compared with the large changes resistance. As noted earlier, pregnancy normally induces quite
in insulin sensitivity. Robust plasticity of β-cell function in marked insulin resistance. This physiological insulin resis-
the face of progressive insulin resistance is the hallmark of tance also occurs in women with GDM. However, it occurs on
normal glucose regulation during pregnancy. a background of chronic insulin resistance to which the insu-
Although pregnancy is a carbohydrate-intolerant state, lin resistance of pregnancy is partially additive. As a result,
only a small proportion of pregnant women (3%–5%) develop pregnant women with GDM tend to have even greater insulin
GDM. As pregnancy advances, the increasing tissue resistance resistance than normal pregnant women.
to insulin creates a demand for more insulin. In the great The cellular mechanisms underlying insulin resistance in
majority of women, insulin requirements are readily met, so normal and diabetic pregnancy are still unknown. The mea-
the balance between insulin resistance and insulin supply is surement of fasting insulin concentrations and the calcula-
maintained. However, if resistance becomes dominant due to tion of fasting insulin/glucose ratios can provide a qualitative
impaired insulin secretion, hyperglycemia develops. In the but not a quantitative estimation of insulin sensitivity. In
majority of such cases, it develops in the last half of pregnancy, nonpregnant patients, hyperinsulinemic–euglycemic clamps5
with insulin resistance increasing progressively until delivery, and minimal model analysis of intravenous glucose tolerance
when, in most cases, it rapidly disappears. tests (IVGTTs)6,7 have been used to obtain quantitative data
Controversy still exists about the screening and diagno- about insulin action. The IVGTT model provides data on the
sis of GDM. In the majority of cases, carbohydrate intoler- glucose infusion that is required to maintain euglycemia dur-
ance is asymptomatic and can be detected only by routine ing constant insulin infusion. However, its use in pregnancy
49
50 Pathogenesis of gestational diabetes mellitus
is limited owing to the change in the relationship between Estrogen and progesterone
common measures of body size, such as total body weight and In early pregnancy, both progesterone and estrogen rise
body surface area. Catalano et al.8,9 were the first to conduct a but their effects on insulin activity are counterbalanced.
prospective longitudinal study using the hyperinsulinemic– Progesterone causes insulin resistance, whereas estrogen
euglycemic clamp model in obese and nonobese gravid is protective.20 An IVGTT given to estrogen-treated rats
women with normal glucose tolerance tests. They found a 47% showed a significant decrease in glucose concentrations and
decrease in insulin sensitivity in obese gravid women and a a twofold increase in insulin concentration21; the addition of
56% decrease in lean gravid women. Differences in whole- progesterone was associated with a 70% increase in the insu-
body insulin sensitivity tend to be small in the third trimes- lin response to a glucose challenge test, but there were no
ter, owing to the marked effects of pregnancy itself on insulin alterations in glucose tolerance.22 In cultured rat adipocyte
resistance. Nonetheless, precise and direct measures of insu- tissue treated with estrogen, there was no effect on glucose
lin sensitivity applied during the third trimester have identi- transport, but maximum insulin binding was increased.
fied, in women with GDM, exaggerated resistance to insulin’s However, progesterone decreased both maximum glucose
ability to stimulate glucose utilization.9 transport and insulin binding.20,21
The development of resistance to the glucose-lowering Gonzalez et al.23 evaluated the role played by proges-
effects of insulin is a normal phenomenon of pregnancy. terone and/or 17beta-estradiol on sensitivity to insulin
In a pioneer study, Burt et al.10 demonstrated that pregnant action that took place during pregnancy. Ovariectomized
women experience fewer hypoglycemic events in response rats were treated with different doses of progesterone and/
to insulin infusion than nongravid women. Accordingly, or 17beta-estradiol in order to simulate the plasma levels
later research found women with normal pregnancies had in normal pregnancy rats. A hyperinsulinemic–euglycemic
progressively exaggerated insulin responses to ingested glu- clamp was used to measure insulin sensitivity. The results
cose, together with a slightly decreased glucose tolerance.11,12 suggested that the absence of female steroid hormones leads
Using the IVGTT model, Buchanan et al.13 and Cousins to decreased insulin sensitivity. Thus, the rise in insulin sen-
et al.14 demonstrated a significant (70%) reduction in insulin sitivity during early pregnancy, when plasma concentrations
sensitivity during the second trimester of normal pregnancy, of 17beta-estradiol and progesterone are low, could be due to
with a return to normal values shortly after delivery. Ryan 17beta-estradiol. However, during late pregnancy, when both
et al.2 were the first to report quantitative differences in insu- plasma concentrations of 17beta-estradiol and progesterone
lin sensitivity between normal and diabetic pregnancies. are high, the role of 17beta-estradiol may serve to antagonize
Other researchers noted that insulin sensitivity was lower in the effect of progesterone, diminishing insulin sensitivity.23
patients with GDM than in patients with normal pregnan-
cies at 12–14 weeks of gestation, before the point of maximal
physiological insulin resistance; however, the difference was Cortisol
not statistically significant. By the third trimester, insulin Cortisol levels increase as pregnancy advances and by the
resistance was similar in the two groups.8,14 end of pregnancy concentrations are threefold higher than
Much effort has been invested to identify the tissues that in the nonpregnant state.24 Rizza et al.,25 in a clamp study,
contribute to the insulin resistance of pregnancy. Findings, demonstrated that under infusion of high amounts of corti-
in animal models, indicate a 40% reduction in insulin- sol, hepatic glucose production increased and insulin sensi-
mediated glucose utilization by skeletal muscle and a similar tivity decreased. Findings in a skeletal muscle model showed
effect in cardiac muscle and fat cells.15,16 that an excess of glucocorticoid is characterized by decreased
It remains unclear whether hepatic insulin sensitivity is total tyrosine phosphorylation of the insulin receptor (IR);
altered during gestation. Kalhan et al.17 and Cowett et al.18 therefore, it seems logical that glucocorticoid-induced insu-
noted no significant differences in basal glucose produc- lin resistance is related to a postreceptor mechanism. In a
tion in pregnant women at term compared to nonpregnant study by Ahmed and Shalayel,26 30 pregnant women with
control subjects when the data were expressed per kilo- GDM and 30 pregnant women with impaired glucose toler-
gram of body weight; however, expression of the data in ance (IGT) were compared with 30 pregnant women with
relation to pregravid weight yielded an increase in hepatic normal glucose tolerance. The GDM and IGT groups were
glucose production in late pregnancy.19 Furthermore, in found to have significantly higher levels of serum cortisol
hyperinsulinemic–euglycemic clamp studies, hepatic glu- than the control group.
cose production was significantly less suppressed in lean and
obese patients with GDM than in the control group.8,9
Prolactin
During pregnancy, maternal prolactin levels increase 7–10-
Hormonal effect in normal and fold. Gustafson et al.27 reported that the basal insulin con-
diabetic pregnancy centration and postchallenge glucose and insulin responses
were greater in women with hyperprolactinemia than in
Reproductive hormones tend to increase during pregnancy, healthy controls. These findings were supported by studies
most of them contribute to insulin resistance and altered showing that the culture of pancreatic islet cells with prolac-
beta-cell function. tin induces an increase in insulin secretion.28 Skouby et al.29
Other factors affecting gestational diabetes mellitus 51
investigated the relationship between the deterioration in shown that leptin directly affects whole body insulin sensitiv-
glucose tolerance and plasma prolactin levels in patients with ity by regulating the efficiency of insulin-mediated glucose
normal and diabetic pregnancies. Oral glucose tolerance metabolism by skeletal muscle37 and by hepatic regulation of
tests (OGTTs) were performed in late pregnancy and post- gluconeogenesis.38 Leptin may also wield an acute inhibitory
partum. In late pregnancy, the GDM group had significantly effect on insulin secretion.39 Yamashita et al.40 suggested that
elevated fasting glucose levels compared to the controls, and, an alteration in leptin action might play a role in GDM and
after glucose challenge, their insulin responses were sig- fetal overgrowth weight gain. They found that pregnant mice
nificantly diminished and the suppression of glucagon less treated with leptin had markedly lower glucose levels than
pronounced. These differences in glucose metabolism were controls during glucose and insulin challenge tests. However,
markedly reduced in the early postpartum period. There was despite the reduced energy intake and improved glucose tol-
no difference in basal prolactin concentrations between the erance, fetal overgrowth was not reduced. Results provide
two groups at either time point. The prolactin levels were evidence that leptin administration during late gestation can
also not altered during the OGTTs, and there was no cor- reduce adiposity and improve glucose tolerance in the model
relation between the deterioration in glucose tolerance and of spontaneous GDM. These data suggest that alterations
the prolactin concentrations in either group. Thus, abnormal in placental leptin levels may contribute to the regulation
prolactin levels are not of pathophysiologic importance in of fetal growth independently of maternal glucose levels.
the development of GDM. Kautzky-Willer et al.41 measured plasma concentrations
of leptin and beta-cell hormones during fasting and after an
oral glucose load (OGTT of 75 g) in pregnant women with
Human placental lactogen GDM and normal glucose tolerance at 28 weeks gestation
Human placental lactogen (hPL) levels rise at the beginning and in women who were not pregnant. Plasma leptin con-
of the second trimester, causing a decrease in phosphory- centration was higher in the women with GDM than in the
lation of insulin receptor substrate (IRS)-1 and profound women with normal glucose tolerance and higher in both
insulin resistance.20 Beck and Daughday30 demonstrated these groups than in the nonpregnant controls. No change
that overnight infusion of hPL results in abnormal glucose in plasma leptin concentrations was induced by OGTT in
tolerance and increased insulin and glucose concentration any group. Basal insulin release was higher in women with
in response to an oral glucose challenge. Accordingly, Brelje GDM than in women with normal glucose tolerance. The
et al.31 found that in islet cell culture, hPL directly stimulates authors concluded that women with GDM and no change in
insulin secretion. This may indicate that hPL directly reg- plasma leptin on oral glucose loading have increased plasma
ulates islet cell function and is probably the principal hor- leptin concentrations during and after pregnancy. Vitoratos
mone responsible for the increase in islet function observed et al.42 investigated the changes in leptin levels and the rela-
during normal pregnancy.31 tionship between leptin substance and insulin and glucose
in pregnant women with GDM. Plasma leptin levels were
measured in peripheral vein blood samples from healthy
Leptin and diabetic women at 29 and 33 weeks gestation. Results
Leptin is a 16 kDa protein encoded by the ob/ob (obesity) gene showed a correlation of plasma leptin levels with fasting
secreted by adipocyte tissue. It can modulate energy expen- plasma insulin levels and plasma glucose levels measured
diture by direct action on the hypothalamus. Fasting insu- 1 hour after oral administration of 50 g of glucose. Serum
lin and leptin concentrations correlate closely with body fat, leptin levels were significantly higher in the women with
making leptin a good marker of obesity and insulin resistance. GDM than in the women with uncomplicated pregnancies.
As receptors to leptin are found in skeletal muscle, the liver, The GDM group also showed a significant, positive corre-
the pancreas, adipocyte tissue, the uterus, and the placenta, lation of serum leptin levels with glycosylated hemoglobin
it may be responsible for both peripheral and central insulin levels, fasting serum insulin levels, and plasma glucose lev-
resistance. Reductions in leptin concentrations are caused by els measured 1 hour after administration of 50 g of glucose.
weight loss, fasting, and starvation; leptin concentrations are Thus, levels of leptin are elevated in women with GDM, and
increased with weight gain and hyperinsulinemia. leptin metabolism depends on insulin levels and the sever-
In animal models, using hyperinsulinemic–euglycemic ity of the diabetes. Wiznitzer et al.43 reported that umbilical
clamp studies, infusion of leptin was found to increase the cord leptin concentration was an independent risk factor for
glucose infusion rate.32 Leptin levels are significantly higher fetal macrosomia in nondiabetic pregnant women.
in pregnancy than in the nonpregnant state, especially dur-
ing the second and third trimesters,33–35 and this change in
circulating leptin concentrations is generally consistent Other factors affecting gestational
with changes in maternal fat stores and glucose metabolism. diabetes mellitus
Results of studies by Laivuori et al.36 suggest that pregnancy-
associated increases in maternal plasma leptin may result Tumor necrosis factor-alpha
from an upregulation of adipocyte leptin synthesis in the Tumor necrosis factor-alpha (TNF-α) has been implicated
presence of increasing insulin resistance and hyperinsu- in the pathogenesis of insulin resistance in type 2 diabetes mel-
linemia in the latter half of pregnancy. Investigators have also litus, but only limited data are available with regard to GDM.
Coughlan et al.44 investigated the effect of exogenous glucose level was extremely high in comparison to serum levels in
on the release of TNF-α from placental and adipose tissue children and adults and was positively correlated to fetal
obtained from normal and diabetic pregnant women. They birth weights. No correlation was found between cord adi-
found significantly greater TNF-α release under conditions of ponectin levels and maternal body mass index, cord leptin,
high glucose concentrations in the GDM group. As TNF-α has or insulin levels. Cord adiponectin levels were significantly
been implicated in the regulation of glucose and lipid metabo- higher compared with maternal levels at birth, and no cor-
lism, and in insulin resistance, these data are consistent with relation was found between cord and maternal adiponectin
the hypothesis that TNF-α is involved in the pathogenesis and/ levels. There were no significant differences between adipo-
or progression of GDM. nectin levels at birth and 4 days postpartum. These find-
Catalano et al.45 reported that changes in insulin sensi- ings indicate that adiponectin in cord blood is derived from
tivity from early to late pregnancy correlated with a gradual fetal and not from placental or maternal tissues. The high
increase in TNF-α levels, which in turn correlated with the adiponectin levels in newborns compared with adults may
percentage change in body weight. be the result of deficient negative feedback on adiponectin
production stemming from the lack of adipocyte hyper-
trophy, low percentage of body fat, or a different distribu-
Adrenomedullin tion of fat storage in newborns. Adiponectin may emerge
Adrenomedullin is a newly discovered hypotensive peptide as a significant factor in carbohydrate–fat metabolism and
involved in the insulin regulatory system, and it may play in the development of insulin resistance during pregnancy.
a role in modifying diabetes in pregnancy. Di Iorio et al.46 Data suggest that there are decreased adiponectin levels in
studied its correlation to GDM. Adrenomedullin concen- women with GDM compared with healthy control subjects.
trations were measured in maternal and fetal plasma and This finding supports the concept of a common pathogenesis
in amniotic fluid in diabetic and nondiabetic pregnancies. between type 2 diabetes and GDM. Although adiponectin
Overall amniotic fluid concentration was higher in the preg- level appears to rise throughout pregnancy, its contribution
nant diabetic women (type 1 or GDM), but there was no to gestation remains unclear.
difference between the group in maternal and fetal plasma
levels. These findings suggest that placental adrenomedullin
production is upregulated in diabetic pregnancy and that it Pancreatic beta-cell function in
may be important to prevent excessive vasoconstriction of
placental vessels. normal pregnancy and gestational
diabetes mellitus
Adiponectin Insulin is the main hormone controlling blood glucose con-
Adiponectin is an adipose tissue hormone, which is a spe- centration. Most commonly, assessment of beta-cell func-
cific plasma protein that is secreted by adipocytes. It may tion is made by measuring the fasting insulin concentration
facilitate the regulation of the glucose and lipid metabolism. or the response to glucose infusion. Fasting plasma insulin
Adiponectin decreases the hepatic glucose production and increases gradually during pregnancy—by the third trimes-
insulin resistance by upregulating fatty acid oxidation.47 ter, levels are twofold higher than before pregnancy. Patients
Adiponectin also suppresses the secretion of TNF-α by adi- with GDM have fasting insulin levels equal to or higher than
pose tissue, a factor that is known to contribute to insulin those of women with nondiabetic pregnancies, with the
resistance.48 Studies have shown that adiponectin serum lev- highest levels occurring in obese women with GDM.
els were decreased in obese subjects49 and patients with type During normal pregnancy, oral and intravenous glucose
2 diabetes.50 In studies with rhesus monkeys, adiponectin tolerance deteriorates only slightly, despite the reduction in
plasma levels were significantly decreased with the progres- insulin sensitivity.13 The main mechanism responsible for that
sion of obesity and insulin resistance.51 In all probability, phenomenon is a gradual increase in insulin secretion by the
adiponectin increases insulin sensitivity by enhancing the beta cells. Kuhl12 reported a hyperbolic relationship between
β-oxidation of free fatty acids and by decreasing the intra- insulin sensitivity and beta-cell responsiveness to glucose in
cellular concentrations of triglycerides.52,53 In patients with both pregnant and nonpregnant women, pointing to a role for
type 2 diabetes, who have the same risk factors for GDM, the beta cells in pathological states such as GDM and demon-
i.e., obesity, maternal age, ethnic origin, and family history, strating the magnitude of the change in insulin secretion that
lower serum levels of adiponectin were detected. In mice, is necessary to maintain glucose tolerance. The mechanism
the intravenous administration of adiponectin was associ- responsible for increase insulin secretion during pregnancy is
ated with loss of weight and reduced plasma concentrations not well understood. A major contributing factor is the increase
of fatty acids54; the proportion of total body fat mass was cor- in the beta-cell mass, a combination of hyperplasia and hyper-
related negatively with adiponectin serum levels.55 The data trophy.59 The increased beta-cell mass can contribute to the
suggests that low plasma adiponectin concentration during increased fasting insulin concentration despite normal or
even early pregnancy may be associated with subsequent lowered fasting glucose concentrations in late pregnancy, and
development of GDM.56–58 Levels of adiponectin have been the enhanced insulin response to glucose during pregnancy
assessed in fetal cord at delivery.58 A cord blood adiponectin (twofold to threefold above nonpregnant levels).
Pancreatic beta-cell function in normal pregnancy and gestational diabetes mellitus 53
In GDM, the early insulin response to OGTT (15–30 GDM may be similar to the genes conferring risk of type 2 dia-
minutes after glucose ingestion) is reduced compared to betes, whereas in others, novel genes may contribute to GDM.
nondiabetic pregnant control women, suggesting a defect in
the beta-cell response.60 First-phase beta-cell responses to
glucose infusion in GDM patients has also been reported to Insulin signaling system in normal pregnancy and in
be reduced. GDM tends to be milder in women with a nor- gestational diabetes mellitus
mal beta-cell response, and they are at relatively low risk for The action of insulin is triggered when it binds to the IR. The
developing diabetes.61 IR belongs to the IGF receptor family, which possesses an
intrinsic tyrosine kinase (TK) activity. The receptor is com-
Genetics, immunology, and gestational diabetes posed of two alpha-subunits, each linked to a beta-subunit
mellitus and to each other by disulfide bonds; only the beta-subunit
Some patients with GDM manifest evidence for autoimmu- has enzymatic TK activity. When insulin binds to the recep-
nity toward beta cells (insulin autoantibodies and anti–islet tor, the conformational change activates the beta-subunit
cell antibodies); however, the prevalence of such autoim- and autophosphorylation begins. Thus, the activation of TK
munity has been reported to be extremely low (<10%).62,63 enzyme leads to increased tyrosine phosphorylation of cellu-
Mutations in the glucokinase gene occur in no more than lar substrates. IRS-1, a cytosolic protein, binds to the phos-
5% of patients with GDM.64 The inheritance of GDM was phorylated intracellular substrates, thereby transmitting the
studied in a group of 100 women with previous GDM.65 The insulin signal downstream. The distribution of the IRS pro-
women were reinvestigated 11 years postpartum and 60% teins tends to be tissue specific: IRS-2 is more copious in the
were found to have either IGT or type 2 diabetes. An inves- liver and pancreas, whereas both IRS-1 and IRS-2 are widely
tigation of their parents showed that a substantial propor- expressed in skeletal muscle. Insulin stimulates the activation
tion had neither parent affected with IGT or type 2 diabetes, and binding of the lipid kinase enzyme, phosphatidylinositol
which suggests a polygenic inheritance or environmental (PI)-3-kinase, and its binding to IRS-1. The formation of PI is
influence rather than autosomal dominance inheritance mandatory for insulin action on glucose transport. Knockout
with high penetration rates. In addition, animal studies of the IRS-1 gene causes only a moderate increase in insulin
have shown that prenatal exposure to a diabetic intrauterine resistance due to increased insulin secretion, but not overt
milieu increases the risk of GDM. diabetes. In women with GDM, the skeletal muscle contains
Harder et al.66 reported that the prevalence of type 2 dia- lower levels of IRS-1 protein and significantly less insulin-
betes was significantly greater in mothers than in fathers of stimulated IRS-1 tyrosine phosphorylation, while levels of the
women with GDM, and there was also significant aggrega- IRS-2 protein are increased. These findings suggest that the
tion of type 2 diabetes in the maternal–grandmaternal line insulin resistance of GDM may be exerted through a decrease
compared to the paternal–grandpaternal line. Therefore, that in the insulin resistance cascade at the level of the IRS pro-
may suggest that a history of type 2 diabetes on the mother’s teins. The increased IRS-2 level may be a compensation for the
side might be considered as a particular risk factor for GDM. reduced IRS-1 level.69 Glucose uptake by cells is mediated by
The possible genetic background of GDM remains a family of membrane proteins, GLUT1–GLUT4, which have
unclear. In particular, its association with human leukocyte a significant sequence similarity. GLUT4 is the main insulin-
antigen (HLA) class II polymorphism has been poorly stud- sensitive glucose transport, expressed uniquely in skeletal and
ied, and the results are conflicting. In an attempt to clarify cardiac muscles and adipose tissue. Garvey et al.70 reported
these discrepancies, Vambergue et al.67 reported that the that in rectus abdominis taken from lean and obese women
distribution of HLA class II polymorphism was not signifi- with GDM, GLUT4 content was similar. In GDM, GLUT4
cantly different between GDM and IGT samples, and there gene expression is normal in skeletal muscles. To the extent
was no significant variation in DRB1*03 and DRB1*04 allele that these muscles are representative of the total muscle mass,
frequencies. These data provide further evidence that type 1 insulin resistance in skeletal muscle may involve impaired
or insulin-dependent diabetes mellitus HLA class II suscep- GLUT4 function or translocation, but not its depletion, as
tibility alleles cannot serve as genetic markers for suscepti- observed in adipose tissue. Garvey et al.71 demonstrated that
bility to glucose intolerance during pregnancy. the insulin-stimulated glucose transport in adipocyte tissue
Ober et al.68 studied the restriction fragment length poly- was reduced by 60% at term in women with GDM compared to
morphisms near “candidate diabetogenic genes” in order to nondiabetic pregnant women. Moreover, the GLUT4 content
identify molecular markers for GDM genes. Genotypes for in adipocytes was profoundly depleted in 50% of the GDM
the insulin hypervariable region, insulin-like growth factor 2 group. The whole group exhibited a novel abnormality in
(IGF2), IR, and glucose transporter (GLUT)1 were studied in GLUT4 subcellular distribution, an accumulation of GLUT4
GDM and control subjects. The results supported the hypoth- in membranes cofractionating with plasma membranes
eses that GDM has heterogeneous phenotypic and genotypic and high-density microsomes in basal cells, and absence of
features and that the risk for GDM in black and Caucasian translocation in response to insulin. These data suggest that
subjects is not related to obesity per se but to interactions abnormalities in cellular traffic or targeting relegate GLUT4
between obesity and IR alleles. In Caucasian women, IR and to a membrane compartment from which insulin cannot
IGF2 alleles interact to confer an additional risk for GDM. recruit transporters to the cell surface. This has important
Thus, in some women, genes responsible for susceptibility to implications for skeletal muscle insulin resistance in GDM.
The membrane protein plasma cell membrane glycoprotein-1 and it corresponds to the prevalence of type 2 diabetes and
(PC-1) has been identified as an inhibitor of insulin receptor IGT within a given population. The predominant patho-
TK (IRTK) activity. Shao et al.69 investigated IR function and genic factor in GDM could be inadequate insulin secre-
PC-1 levels in the muscle from three groups of obese subjects: tion. It has been convincingly demonstrated that GDM
women with GDM, pregnant women with normal glucose tol- occurs as a result of a combination of insulin resistance
erance, and nonpregnant control subjects. No significant dif- and decreased insulin secretion. The similar frequencies of
ferences were found in basal IR tyrosine phosphorylation or HLA-DR2, HLA-DR3, and HLA-DR4 antigens in healthy
IRTK activity among the three groups. After maximal insulin pregnant women and women with GDM, and the low
stimulation, IRTK activity increased in all subjects, but was prevalence of markers for autoimmune destruction of the
lower in women with GDM by 25% and 39% compared with beta cells in GDM, rule out the possibility that GDM has
pregnant and nonpregnant control subjects, respectively. an autoimmune origin. Pregnancy is associated with pro-
Similarly, IR tyrosine phosphorylation was significantly found hormonal changes that have a direct effect on carbo-
decreased in the subjects with GDM compared to the other hydrate tolerance. In early pregnancy, both progesterone
two groups. Treatment of the IR with alkaline phosphatase and estrogen levels rise, but their actions on insulin are
to dephosphorylate serine/threonine residues significantly counterbalanced, as progesterone causes insulin resistance
increased insulin-stimulated IRTK activity in the pregnant and estrogen is protective. In the second trimester, hPL,
control and GDM subjects, but the rates were still lower than cortisol, and prolactin levels all rise, causing decreased
in the nonpregnant controls. PC-1 content in the muscle from phosphorylation of IRS-1 and profound insulin resistance.
GDM subjects was increased by 63% compared with preg- In most subjects, pancreatic insulin secretion rise to meet
nant control subjects and by 206% compared with nonpreg- this need, but in those with underlying beta-cell defects,
nant control subjects. PC-1 content was negatively correlated hyperglycemia ensues. In women with GDM, the insulin
with IR phosphorylation and IRTK. Increased PC-1 content resistance of pregnancy is exaggerated, especially if fast-
in the pregnant control and GDM groups suggests an exces- ing hyperglycemia is present, and is related to additional
sive phosphorylation of serine/threonine residues in muscle defective tyrosine phosphorylation of the IR beta-subunit.
IR, both of which may contribute to the pregnancy-associated Recent research suggests that the postreceptor mecha-
decrease in IRTK activity. In GDM, changes worsened, even nisms that contribute to insulin resistance of pregnancy
when controlling for obesity. These postreceptor defects in are multifactorial but are exerted at the beta-subunit of the
insulin signaling may contribute to the pathogenesis of GDM IR and at the level of IRS-1. The resistance to insulin-medi-
and the increased risk for type 2 diabetes later in life. ated glucose transport appears to be greater in the skeletal
Receptor autophosphorylation has also been reported to muscle from GDM subjects than from pregnant subjects.
be impaired in GDM subjects, a finding consistent with their There is also a modest but significant decrease in the maxi-
increased insulin resistance.70 In addition, overexpression mal IR tyrosine phosphorylation in the muscle from obese
of membrane plasma cell differentiation factor-1 (i.e., PC-1) GDM subjects. Results also suggest that increased IR ser-
may play a role in developing insulin resistance by inhibit- ine/threonine phosphorylation and PC-1 could underlie
ing the TK activity of the IR.71 In GDM patients, PC-1 lev- the insulin resistance of pregnancy and contribute to the
els were significantly higher in skeletal muscle compared to pathogenesis of GDM.
nondiabetic pregnant women.72 Whether additional defects are exerted further
downstream from IRS-1 remains to be investigated. GDM
is a predictor of diabetes (mainly type 2) later in life. The
Summary cumulative incidence of type 2 diabetes is 50% at 5 years.
GDM is also a predictor, or even an early manifestation,
GDM is carbohydrate intolerance resulting in hypergly- of the metabolic (insulin resistance) syndrome. GDM is a
cemia of variable severity with onset or first recognition cardiovascular risk factor, and affected patients should be
during pregnancy. The incidence of GDM is 0.15%–15%, screened to prevent late complications.
REFERENCES
1. Metzger BE, Coustan DR. Summary and recommendations of 5. DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique:
the Fourth International Workshop–Conference on Gestational A method for quantifying insulin secretion and resistance. Am J
Diabetes Mellitus. Diabetes Care 1998; 21: B161–B167. Physiol 1979; 237: E241–E243.
2. Ryan EA, O’Sullivan MJ, Skyler JS. Insulin action during preg- 6. Bergman RN, Ider YZ, Bowden CR, Cobelli C. Quantitative
nancy: Studies with the euglycemic clamp technique. Diabetes estimation of insulin sensitivity. Am J Physiol 1979; 236:
1985; 34: 380–389. E667–E677.
3. Catalano PM, Tyzbir ED, Wolfe RR et al. Carbohydrate metabo- 7. Bergman RN. The Lilly Lecture 1989. Toward a physiological
lism during pregnancy in control subjects and women with ges- understanding of glucose tolerance: Minimal model approach.
tational diabetes. Am J Physiol 1993; 264: E60–E67. Diabetes 1989; 38: 1512–1528.
4. Kuhl C. Insulin secretion and insulin resistance in pregnancy and 8. Catalano PM, Tyzbir ED, Roman NM et al. Longitudinal changes
GDM: Implications for diagnosis and management. Diabetes in insulin release and insulin resistance in nonobese pregnant
1991; 40: 18–24. women. Am J Obstet Gynecol 1991; 165: 1667–1672.
References 55
9. Catalano PM, Huston L, Amini SB, Kalham SC. Longitudinal 33. Henson MC, Swan KF, O’Neil JS. Expression of placental leptin
change in glucose metabolism during pregnancy in obese women and leptin receptor transcripts in early pregnancy and at term.
with normal glucose tolerance and gestational diabetes mellitus. Obstet Gynecol 1998; 92: 1020–1028.
Am J Obstet Gynecol 1999; 180: 903–916. 34. Masuzaki H, Ogawa Y, Sagawa N. Nonadipose tissue production
10. Burt RL. Peripheral utilization of glucose in pregnancy. Insulin of leptin: Leptin as a novel placenta-derived hormone in humans.
tolerance. Obstet Gynecol 1956; 2: 558–664. Nat Med 1997; 3: 1029–1033.
11. Spellacy WN, Goetz FC. Plasma insulin in normal late preg- 35. Highman TJ, Friedman JE, Huston LP et al. Longitudinal changes
nancy. N Engl J Med 1963; 268: 988–991. in maternal serum leptin concentrations body composition and
12. Kuhl C. Glucose metabolism during and after pregnancy in nor- resting metabolic rate in pregnancy. Am J Obstet Gynecol 1999;
mal and gestational diabetic women. Acta Endocrinol 1975; 79: 178: 1010–1015.
709–719. 36. Laivuori H, Kaaja R, Koistinen H et al. Leptin during and after
13. Buchanan TA, Metzger BE, Freinkel N, Bergman RN. Insulin preeclamptic or normal pregnancy: Its relation to serum insulin
sensitivity and B-cell responsiveness to glucose during late preg- and insulin sensitivity. Metabolism 2000; 49: 259–263.
nancy in lean and moderately obese women with normal glu- 37. Cohen B, Novick D, Rubinstein M. Modulation of insulin activi-
cose tolerance or mild gestational diabetes. Am J Obstet Gynecol ties by leptin. Science 1996; 274: 1185–1188.
1990; 162: 1008–1014. 38. Rossetti L, Massillon D, Barzilai N et al. Short term effects of
14. Cousins L, Rea C, Crawford M. Longitudinal characterization of leptin on hepatic gluconeogenesis and in vivo insulin action.
insulin sensitivity and body fat in normal and gestational diabetic J Biol Chem 1997; 272: 27758–27763.
pregnancies. Diabetes 1988; 37: 251A (abstract). 39. Ceddia RB, Koistinen HA, Zierath JR et al. Analysis of paradoxi-
15. Hauguel S, Leturque A, Gilbert M, Girard J. Effects of pregnancy cal observations on the association between leptin and insulin
and fasting on muscle glucose utilization in the rabbit. Am J resistance. FASEB J 2002; 16: 1163–1176.
Obstet Gynecol 1988; 158: 1215–1218. 40. Yamashita H, Shao J, Ishizuka T et al. Leptin administration
16. Leturque A, Ferre P, Burnol AF et al. Glucose utilization rates and prevents spontaneous gestational diabetes in heterozygous
insulin sensitivity in vivo in tissues of virgin and pregnant rats. Lepr (db/+) mice: Effects on placental leptin and fetal growth.
Diabetes 1986; 35: 172–177. Endocrinology 2001; 142: 2888–2897.
17. Kalhan SC, D’Angelo LJ, Savin SM, Adan SM. Glucose produc- 41. Kautzky-Willer A, Pacini G, Tura A et al. Increased plasma leptin
tion in pregnant women at term gestation: Sources of glucose for in gestational diabetes. Diabetologia 2001; 44: 164–172.
the human fetus. J Clin Invest 1979; 63: 388–394. 42. Vitoratos N, Salamalekis E, Kassanos D et al. Maternal plasma
18. Cowett RA, Susa JB, Kahn CB et al. Glucose kinetics in non- leptin levels and their relationship to insulin and glucose in gesta-
diabetic and diabetic women during third trimester of pregnancy. tional-onset diabetes. Gynecol Obstet Invest 2001; 51: 17–21.
Am J Obstet Gynecol 1983; 146: 773–780. 43. Wiznitzer A, Furman B, Zuili I et al. Cord leptin level and fetal
19. Catalano PM, Ishizika T, Friedman JE. Glucose metabolism in macrosomia. Obstet Gynecol 2000; 96: 707–713.
pregnancy. In: Cowett RM, ed., Principles of Perinatal Neonatal 44. Coughlan MT, Oliva K, Georgiou HM et al. Glucose-induced
Metabolism, 2nd edn., Springer-Verlag: New York, 1998, release of tumor necrosis factor-alpha from human placental
pp. 183–206. and adipose tissues in gestational diabetes mellitus. Diabet Med
20. Ryan EA, Ennes L. Role of gestational hormones in the induction 2001; 18: 921–927.
of insulin resistance. J Clin Endocrinol Metab 1988; 67: 341–347. 45. Catalano P, Highman T, Huston L, Friedman J. Relationship
21. Costrini NV, Kalkhoff RK. Relative effect of pregnancy estradiol between reproductive hormones/TNF-a and longitudinal changes
and progesterone on plasma insulin and pancreatic islet insulin in insulin sensitivity during gestation. Diabetes 1996; 45: 175a.
secretion. J Clin Invest 1971; 50: 992–999. 46. Di Iorio R, Marinoni E, Urban G et al. Fetomaternal adreno-
22. Kalkhoff RK, Jacobson M, Lemper D. Progesterone, pregnancy medullin levels in diabetic pregnancy. Horm Metab Res 2001;
and the augmented plasma insulin response. J Clin Endocrinol 33: 486–490.
1970; 31: 24–28. 47. Chandran M, Phillips SA, Ciaraldi T et al. Adiponectin: More
23. Gonzalez C, Alonso A, Alvarez N et al. Role of 17 beta-estradiol than just another fat cell hormone? Diabetes Care 2003; 26:
and/or progesterone on insulin sensitivity in the rat: Implications 2442–2450.
during pregnancy. J Endocrinol 2000; 166: 283–289. 48. Hotamisligil GS. The role of TNF- and TNG- receptors in obesity
24. Gibson M, Tulchinski D. The maternal adrenal. In: Tulchinski and insulin resistance. J Intern Med 1999; 245: 621–625.
D, Ryan KJ, eds., Maternal-Fetal Endocrinology, WB Saunders: 49. Arita Y, Kihara S, Ouchi N et al. Paradoxical decrease of an adipose-
Philadelphia, PA, 1980, pp. 129–143. specific protein, adiponectin, in obesity. Biochem Biophys Res
25. Rizza RA, Mandarino LJ, Gerich JE. Cortisol induced insulin resis- Commun 1999; 257: 79–83.
tance in man: Impaired suppression of glucose production and 50. Weyer C, Funahashi T, Tanaka S et al. Hypoadiponectinemia in
stimulation of glucose utilization due to a postreceptor defect of obesity and type 2 diabetes: Close association with insulin resis-
insulin action. Clin Endocrinol Metab 1982; 54: 131–138. tance and hyperinsulinemia. J Clin Endocrinol Metab 2001; 86:
26. Ahmed SA, Shalayel MH. Role of cortisol in the deterioration of 1930–1935.
glucose tolerance in Sudanese pregnant women. East Afr Med J 51. Hotta K, Funahashi T, Bodkin NL et al. Circulating concentra-
1999; 76: 465–467. tions of the adipocyte protein adiponectin are decreased in
27. Gustafson AB, Banasiak MF, Kalkhoff RK. Correlation of hyperp- parallel with reduced insulin sensitivity during the progres-
rolactinemia with altered plasma insulin and glucose: Similarity sion to type 2 diabetes in rhesus monkeys. Diabetes 2001; 50:
to effects of late human pregnancy. J Clin Endocrinol Metab 1126–1133.
1980; 51: 242–246. 52. Hu E, Liang P, Spiegelman BM. AdipoQ is a novel adipose-
28. Sorenson RL, Brelje TC, Roth C. Effect of steroid and lactogenic specific gene dysregulated in obesity. J Biol Chem 1996; 271:
hormones on islet of Langerhans: A new hypothesis for the role 10697–10703.
of pregnancy steroids in the adaptation of islets to pregnancy. 53. Yamauchi T, Kamon J, Waki H et al. The fat derived hormone adi-
Endocrinology 1993; 133: 2227–2233. ponectin reverses insulin resistance associated with both lipoat-
29. Skouby SO, Kuhl C, Hornnes PJ, Andersen AN. Prolactin and rophy and obesity. Nat Med 2001; 7: 941–946.
glucose tolerance in normal and gestational diabetic pregnancy. 54. Kubota N, Terauchi Y, Yamauchi T et al. Disruption of adipo-
Obstet Gynecol 1986; 67: 17–20. nectin causes insulin resistance and neointimal formation. J Biol
30. Beck P, Daughday WH. Human placental lactogen: Studies of Chem 2002; 277: 25863–25866.
its acute metabolic effects and disposition in normal man. J Clin 55. Yang WS, Lee WJ, Funahashi T. Weight reduction increases
Invest 1967; 46: 103–110. plasma levels of an adipose-derived anti-inflammatory protein,
31. Brelje TC, Scharp DW, Lacy PE et al. Effect of homologous pla- adiponectins. J Clin Endocrinol Metab 2001; 86: 3815–3819.
cental lactogens, prolactins, and growth hormones on islet B-cell 56. Ranheim T, Haugen F, Staff AC et al. Adiponectin is reduced
division and insulin secretion in rat, mouse, and human islets: in gestational diabetes mellitus in normal weight women. Acta
Implication for placental lactogen regulation of islet function dur- Obstet Gynecol Scand 2004; 83(4): 341–347.
ing pregnancy. Endocrinology 1993; 132: 879–887. 57. Worda C, Leipold H, Gruber C et al. Decreased plasma adipo-
32. Sivitz WI, Walsh SA, Morgan DA et al. Effect of leptin on insulin nectins concentrations in women with gestational diabetes mel-
sensitivity in normal rats. J Clin Invest 1997; 138: 3395–3401. litus. Am J Obstet Gynecol 2004; 191(6): 2120–2124.
58. Williams MA, Qiu C, Muy-Rivera M et al. Plasma adiponectins 66. Harder T, Franke K, Kohlhoff R, Plagemann A. Maternal and
concentrations in early pregnancy and subsequent risk of ges- paternal family history of diabetes in women with gestational
tational diabetes mellitus. J Clin Endocrinol Metab 2004; 89(5): diabetes or insulin-dependent diabetes mellitus type I. Gynecol
2306–2311. Obstet Invest 2001; 51: 160–164.
59. Van Assche FA, Aerts L, De Prins F. A morphological study of the 67. Vambergue A, Fajardy I, Bianchi F et al. Gestational diabetes mel-
endocrine pancreas in human pregnancy. Br J Obstet Gynecol litus and HLA class II (-DQ, -DR) association: The Digest Study.
1978; 85: 818–820. Eur J Immunogenet 1997; 24: 385–394.
60. Swinn RA, Warham NJ, Gregory R et al. Excessive secretion of 68. Ober C, Xiang KS, Thisted RA et al. Increased risk for gestational
insulin precursors characterizes and predicts gestational diabe- diabetes mellitus associated with insulin receptor and insulin-
tes. Diabetes 1995; 44: 911–915. like growth factor II restriction fragment length polymorphisms.
61. Kjos SL, Peters RK, Xiang A et al. Predicting future diabetes in Genet Epidemiol 1989; 5: 559–569.
Latino women with gestational diabetes: Utility of early postpar- 69. Shao J, Catalano PM, Yamashita H et al. Decreased insulin recep-
tum glucose tolerance testing. Diabetes 1995; 44: 586–591. tor tyrosine kinase activity and plasma cell membrane glycopro-
62. Damm P, Kuhl C, Buschard K et al. Prevalence and predictive tein-1 overexpression in skeletal muscle from obese women with
value of women with islet cell antibodies and insulin autoanti- gestational diabetes mellitus (GDM): Evidence for increased ser-
bodies in women with gestational diabetes. Diabet Med 1994; ine/threonine phosphorylation in pregnancy and GDM. Diabetes
11: 558–563. 2000; 49: 603–610.
63. Catalano PM, Tyzbir ED, Simms EAH. Incidence and significance 70. Garvey WT, Maianu L, Hancock JA et al. Gene expression of
of islet cell antibodies in women with previous gestational diabe- GLUT4 in skeletal muscle from insulin-resistant patients with
tes. Diabetes Care 1990; 113: 478–483. obesity, IGT, GDM, and NIDDM. Diabetes 1992; 41: 465–475.
64. Stoffel M, Bell KL, Blacburn CL et al. Identification of glucokinase 71. Garvey WT, Maianu L, Zhu J-H et al. Multiple defects in the adi-
mutations in subjects with gestational diabetes mellitus. Diabetes pocyte glucose transport system cause cellular insulin resistance
1993; 42: 937–940. in gestational diabetes. Diabetes 1993; 42: 1773–1785.
65. McLeallan JAS, Barrow BA, Levy JC et al. Prevalence of diabe- 72. Goldfine ID, Maddux BA, Youngrem JF et al. Membrane glyco-
tes mellitus and impaired glucose tolerance in parents of women protein PC-1 and insulin resistance. J Cell Biochem 1998; 182:
with gestational diabetes. Diabetologia 1995; 38: 693–698. 177–184.
7 Autoimmunity in gestational
diabetes mellitus
Alberto de Leiva, Dídac Mauricio, and Rosa Corcoy
first-degree relatives of patients with DM-1 (FDRs-DM-1).17

Immunobiology of GDM In type 1A diabetes, a selective destruction of the insulin-
Pregnancy represents a distinct immunologic state; the fetus producing cells occurred, mediated by T cells.
acts as an allograft to the mother, needing protection against
potential rejection.1,2 Humoral immunoreactivity does not
change much during pregnancy, with the exception of low-
Diabetes-related antibodies and
ered immunoglobulin G concentration at late phase, probably pregnancy
explained by placental transport.3 Regarding cellular immu-
nity, reduction,4,5 elevation,6 and no variation7 in the number Prevalence and titers
of different lymphocytic populations have been reported. The Circulating islet cell antibodies, originally described by indi-
final effect of pregnancy on previously active autoimmune rect immunofluorescence in 1974,18 have been demonstrated
processes is controversial,8,9 and multiple autoimmune dis- in the great majority of individuals with autoimmune diabe-
turbances may be manifested during pregnancy.10 tes, both at the preclinical state and at the onset of clinically
In diabetic pregnancy, immunological abnormalities overt diabetes, and they persist in the circulation for a long
occurring in diabetes are superimposed on immunological time. Most investigated islet cell AAs include AAs to islet cell
changes of pregnancy, eventually influencing maternal and cytoplasm (islet cell AAs [ICAs]), to native insulin (insulin
fetal outcomes. The interplay between sex hormones and AAs [IAAs]), to GAD (GADA),19–21 and to tyrosine phospha-
the immune system is complex and multifactorial, affect- tases (insulinoma-associated antigens IA-2A and IA-2β)22,23:
ing many organs and systems. The mechanisms of estrogen
and progesterone modulation of individual components of ICAs. The prevalence of cytoplasmic islet cell antibodies
the immune system have been extensively studied in vitro in GDM has been reported to range between 0.98%
but not in humans. Older concepts of pregnancy as a state and 14.7% in Caucasians,1–3,14–32 higher than in the
of systemic immunosuppression are oversimplified. A more control group,14,25,27,33–35 except in studies with low
useful model may be the view of pregnancy as a modulated statistical power.36–38 Characteristically, ICA titers are
immunologic condition. As pregnancy advances, T-cell low when compared with subjects with new-onset type
activity, natural killer activity, and possibly B-cell activ- 1A diabetes and their FDRs.28,33–39
ity are reduced, whereas α-defensin levels and monocytic, IAAs. The presence of IAAs in sera of subjects with DM-1
dendritic cell and polymorphonuclear cell activities are before initiating insulin treatment was first reported in
increased. In general, levels of inflammatory cytokines are 1982.40 Since then, IAAs have been shown in 20%–50%
reduced, whereas levels of cytokines that induce phagocytic- of patients with DM-1 of recent diagnosis. IAA posi-
cell recruitment or activity are increased; these alterations tivity displays a strong negative association with age:
do not necessarily follow a clear Th1 or Th2 phenotype. IAAs are positive in 90% of children who develop
Gestational diabetes mellitus (GDM) is defined as an DM-1 before the age of 5 years, but in less than 40%
impairment of glucose tolerance first recognized at the index after the age of 12 years.41 This result is in accordance
pregnancy.11 For this category of women, an increased risk with the low prevalence of IAAs in women with GDM
of progression to type 2 diabetes mellitus (DM-2) has been (0%–5.9%), even when addressed just by a low num-
repeatedly reported.12–15 Nevertheless, a subset of women ber of studies,25,27,36–39 with only one of them reporting
with GDM depicts one or several autoantibodies (AA) against a higher prevalence than in the control population.27
various pancreatic islet cell autoantigens, typically detected When positive, IAA titers are also lower than those of
in type 1 diabetes mellitus (DM-1)16 as well as in high-risk patients with DM-1 and their FDRs.38 We have inves-
subjects for the development of the disease, in particular, tigated the frequency of IAAs in women with GDM
57
58 Autoimmunity in gestational diabetes mellitus
using a radiobinding assay. We observed that only FDRs-DM-1.35,60 Table 7.1 illustrates the prevalence of
0.98% (2 of 203) of consecutive women at diagnosis of DRAs in GDM, reported in multiple publications.
GDM displayed IAAs in their sera before initiation of A prospective German multicentric study has shown
treatment (registered frequency in control subjects, 0 that GADAs in women with GDM bound fewer epi-
of 106 individuals; frequency in FDRs-DM-1, 4.7%). topes than GADAs in FDRs-DM-1.60 Whereas the
Interestingly, in the subgroup of ICA (+) women with prevalence and titers of GADAs to the major GAD65
GDM, the prevalence of IAAs was higher than in ICA COOH-terminal middle epitope were not significantly
(−) ones (11% vs. 0.7%), so that in women with GDM and different (91% vs. 100%, 88.6% vs. 100%), AAs to epit-
ICA positivity, the prevalence of IAAs was not differ- opes GAD65-midb (middle b) and GAD67 were less
ent than in FDRs-DM-1.25 Quite a different issue is the frequently detected in patients with GDM (57% vs. 90%,
appearance of insulin antibodies (IAs) in women with 38% vs. 79%). AAs reactive to all epitopes of GAD65
GDM treated with exogenous insulin. Our group has tested in this study were also less frequent in subjects
reported that 44% of women receiving human insulin with GDM, in comparison to FDRs-DM-1 (65% vs. 23%,
therapy for GDM develop IAs, which may persist up to p < 0.01). The observed reduction of binding to GAD epi-
24 months after delivery.42 Similar findings have been topes in women with GDM, in comparison with young
described after treatment with lispro-insulin.43 FDRs-DM-1, is also seen in latent autoimmune diabetes
GADAs and IA-2As. GAD is the biosynthesizing enzyme of adulthood (LADA) patients.61 Similar to LADA, in
of γ-aminobutyric acid. Pancreatic β-cells express this GDM, the combined presence of various AAs is infre-
enzyme, among other cell types. The prevalence of these quent,25,27,34,48,49 which is consistent with the concept of a
AAs in subjects with new-onset DM-1 is 60%–70% for slow progression form of autoimmune diabetes.60
GADAs, 40% for IA-2As, and 20% for IA-2βAs.44,45 All Transplacental passage of DRAs and related effects. DRAs
DRAAs could be useful for the screening of autoim- are transferred to the fetus.42,62–67 They are mainly
mune diabetes. Because ICA measurement is time con- IgG and are actively transported by the placenta.68,69
suming and semiquantitative, currently it is usually Maternal and fetal levels are highly correlated for
used for confirmation purposes. The combination of IAs,42,62,64–66 ICAs,46 GADAs,62,70 and IA-2As.62 This
two antibodies offers a good yield with GADAs/IA-2As transplacental passage of antibodies lies behind the
and GADAs/IAAs being used in adults and children, higher prevalence of DRA in cord blood from infants
respectively.46 Our group has reported that nonisotopic of diabetic mothers compared with infants of diabetic
alternatives for GADA65 and IA-2As are suitable for fathers.67 Postnatal elimination of transplacentally
the precise use of estimation of risk prediction and acquired DRAs has been prospectively investigated in
diagnosis of autoimmune diabetes.47 infants born to type 1 diabetic mothers, with the mean
Some reports have shown that the prevalence of elimination time being 3.1 months for ICAs and IAs,
IA-2As in GDM ranges from 0% to 6.2%.27,34,38,48,49 4.5 months for GADAs, and 4.3 months for IA-2As.72
Several articles have compared the prevalence of IA-2As
in women with GDM with that of a control population, Menon et al.64 reported that the concentration of ani-
the figures being higher in two articles25,34 and similar mal insulin in cord serum correlated with birth weight,
in the other two.27,38 These antibodies are not frequent in but this observation has not been confirmed by other
this age range50 and are associated with rapid progression authors.42,65,66,71,72 The underlying mechanism would be
to severe insulinopenia.51 IA-2A titers in women with an IA-facilitated maternal insulin transfer and, as a con-
GDM are lower than in children with type 1A diabetes.48 sequence, an enhancement of fetal hyperinsulinism and
The prevalence of GADAs in GDM has been reported to related morbidity.73 Maternal insulin–IA complexes have
range from 0% to 10.8%,25,27,34,35,38,48,49,52,53,59,60–63,99 with been associated with toxemia,27,74,75 HELLP syndrome,76
an article from India describing a prevalence of 41% for hypoglycemia,27,77 respiratory distress,27 and high hemato-
of GADAs/IA-2As.59 Several reasons may account for crit in the newborn.76 Nevertheless, other investigators have
these heterogeneous findings. First, each population has not observed a relationship between maternal levels of IAs
a different genetic and environmental background that and cord blood levels of insulin, 32–33 split proinsulin,72
confer different risks. Second, in each population, dif- or C-peptide.65 In our experience, fetal outcomes were not
ferent ethnic groups may have different susceptibility influenced by the presence or titers of IAs, IAAs, or ICAs.78
for GDM and to β-cell autoimmunity. Third, method- In the nonobese diabetic mouse, the transplacental
ological issues, such as study design or laboratory pro- passage of maternal AAs increases the risk of experimen-
cedures, may justify these wide differences. As in the tal autoimmune diabetes in the offspring,79 whereas the
case of ICAs, their frequency has been reported to be transmission of AAs against activated T cells is protective.80
higher than in the control population in some25,27,34,35,62 Interestingly, in infants of diabetic mothers, the presence of
but not all articles,38,54,55,58,63 probably because of the GADAs and IA-2As, but not of IAs, in cord blood samples, is
low statistical power of the latter. As for other diabetes- associated with a lower risk of developing DRAs and DM-1.81
related antibodies (DRAs), titers of GADAs in GDM However, when present in infants of diabetic fathers, DRAs
have also been reported to be lower when compared are not protective, which seems to imply an active fetus self-
with type 1A diabetes, autoimmune prediabetes, and immune response in this case.81
Heterogeneity of autoimmune diabetes 59
Table 7.1 Prevalence of diabetes-related antibodies in women with

gestational diabetes mellitus (%)
Author n ICA IAA GADA I-A2A

Steel et al.26 50 10
Ginsberg-Fellner et al.29 88 35
Fallucca et al.27 39 5
Freinkel et al.14 160 7.5
Catalano et al.30 187 1.6
Bell et al.96 181 2.8
Stangenberg et al.31 55 1.8
Mauricio et al.33 307 12.4
Ziegler et al.39 55 11
Damm et al.36 139 2.9 0
Tuomilehto et al.52 112 5.0
Beischer et al.53 734 1.8
Mauricio et al.25 203 1
Petersen et al.54 139 2.2
Dozio et al.38 145 10 3.0 0 0
Fuchtenbusch et al.28 437 8.5 9.5 6.2
Fallucca et al.55 83 3.6
Whittingham et al.48 98 3 4 1
Panczel et al.109 68 14.7
Kinalski et al.34 156 5.1 7.0 3.2
Mitchell et al.56 100 6
Bartha et al.99 102 0.98 10.8
Kousta et al.57 321 4.0
Weng et al.58 66 4.5
Balaji et al.59 86 41 GADA/IA-2
Albareda et al.49 535 14 1.5 0.2
Bo et al.35 123 6.5 4.1
Source: de Leiva, A. et al., Diabetes Care, 30(Suppl. 2), 127, 2007. With permission.
Heterogeneity of autoimmune prevalence of alleles DR3, DQB1*0201 and DR4, and

DQB1*0302, with the proportion of heterozygotes declin-
diabetes ing with age at diagnosis.84 Children with the allele HLA
DR2, DQB1*0602, almost never develop diabetes, whereas
Autoimmune diabetes is caused by the destruction of β-cells
this allele confers a much lower protection for adult-onset
of pancreatic islets by an immune-mediated process, pro-
autoimmune diabetes.
moted by the interaction of genetic and environmental
A minority of patients with DM-1 have no known etiol-
factors.82 AAs against pancreatic β-cell antigens precede
ogy and no evidence of autoimmunity (type 1B diabetes,
the clinical onset of DM-1.82 Patients with DM-1 have an
idiopathic DM-1); most of these patients are of African or
increased risk of other autoimmune disorders, including
Asian origin. It is well known that autoimmunity against
Graves’ disease, thyroiditis, Addison’s disease, celiac disease,
pancreatic islets may evolve in some instances as a highly
and pernicious anemia.
aggressive process responsible of extreme insulinopenia,
Age not only modifies the risk of autoimmune diabe-
whereas in other occasions, it leads to a slow and nonag-
tes but also the presence of AAs, the intensity of beta-cell
gressive process, practically asymptomatic, recognized by
destruction, the rate of progression to overt diabetes, and
humoral autoimmunity markers.
the degree of residual insulin secretion. Approximately
During recent years, islet AAs have been demonstrated
30% of subjects with classic autoimmune diabetes (type 1A
in the sera of a significant fraction (5%–20%) of individu-
diabetes) present after age 35 years.83 In Caucasians, child-
als with phenotypical characteristics of DM-2. As a result,
hood autoimmune diabetes is associated with an increased
the term LADA has been incorporated to define this new α-chains that have an arginine residue 52 and β-chains
clinical variant of diabetes.85–87 that lack an aspartic acid at residue 57. The highest genetic
These patients with adult-onset autoimmune diabetes can risk corresponds to those persons in whom all four
be initially misclassified as having DM-2. Characteristically, HLA-DQ combinations meet this criterion (heterozygous
they display a lower rate of metabolic syndrome than for HLA DRB1*04-DQA1*0301-DQB1*0302 and DRB1*03-
patients with DM-2.88 The distinction between adult-onset DQA1*0501-DQB1*0201), with an absolute lifetime risk
DM-1 and LADA is sometimes difficult, but, characteris- for DM 1A in the general population of 1:12. On the con-
tically, patients with LADA evolve slowly toward insulin trary, persons who are protected are those with DRB1*15-
requirement (within 6 years), and older patients with LADA DQA1*0201-DQB1*0602 (Asp57+) haplotypes.95 People
show even a slower progression.86,89 The presence of GADA carrying out the B1*0401 and 0405 subtypes of DRB1*04
indicates a strong possibility of requiring insulin treatment are susceptible, whereas the *0403 and *0406 subtypes are
earlier than expected in classical-type diabetes. protective.
Some investigators believe that patterns of HLA suscep- So far, the genes of HLA complex have been the
tibility in LADA patients are similar to those reported for most investigated genetic factors in autoimmune GDM
classic DM-1, supporting the hypothesis that the genetics (AIGDM). The information obtained from various reports
of autoimmune diabetes in children and adults are dif- showed discordant results.14,36,37,96–98 In these protocols,
ferentiated by only relatively few age-dependent genetic the number of investigated subjects was small, and the
effects and the slower progression of the insulin deficiency analyzed populations quite heterogeneous. Rubinstein
in adults may be related to subtle variation in the HLA class et al. depicted a strong association between HLA DR3/DR4
II gene associations.90,91 On the contrary, other researchers and islet autoimmunity of women with GDM. 31 A similar
have proposed that LADA patients differ genetically from observation was provided by Freinkel et al., showing a
classical DM-1.92,93 Two distinct monocyte gene expression twofold increase in the frequency of DR3 and DR4 alleles
clusters have been identified in autoimmune diabetes. One in women with GDM.14 Ferber et al.98 investigated 184
cluster, comprising 12 pro-inflammatory cytokine/com- German women with GDM; when compared with another
pound genes, was detected in 60% of LADA patients and group of 254 nondiabetic unrelated subjects, no elevation
28% of adult-onset type 1 diabetic subjects but in only 10% in the frequency of any HLA class allele was observed.
of juvenile-onset type 1 diabetic patients. A second cluster, Nevertheless, DR3 allele frequency was increased in GDM
comprising 10 chemotaxis, adhesion, motility, and metabo- women with positivity to islet cell antibodies, particu-
lism genes, was identified in 43% of juvenile autoimmune larly GADA (p = 0.002), as well as DR4 and DQB1*0302
diabetes, 33% of LADA patients, and only 9% of adult-onset (p = 0.009). Those who had a DR3/DR4-containing gen-
type 1 diabetic individuals.92 otype are 60% of women who are islet antibody positive
and 74% of women who developed DM-1 postpartum.
Combining the determination of susceptible HLA alleles
Genetic markers in (DR3, DR4) with islet AA measurement increased the sen-
autoimmune GDM sitivity of identifying women with GDM developing post-
partum DM-1 to 92%. Damm et al. showed a trend toward
Although genetic markers hold a most relevant promise for an increased frequency of DR3/DR4 and a decrease fre-
the future, they are of only limited clinical value in the eval- quency of DR2 in women with GDM evolving to DM-1. 36
uation and management of diabetic patients. To screen for On the opposite, several reports could not found associa-
the genetic susceptibility of type 1A diabetes, HLA typing is tion between increased prevalence of class II alleles and
most useful. The HLA complex on chromosome 6p21.3 is a the presence of humoral islet cell autoimmune markers in
major susceptibility locus, IDDM1. The HLA complex con- women with GDM.73,45,99
tains class I and II genes that code for several polypeptide C. Murgia et al., from the University of Cagliari, have
chains. The class I genes are HLA-A, HLA-B, and HLA-C. reported a very high prevalence and very unusual dis-
The loci of class II genes are designated by three letters: tribution of antibodies in GDM patients of Sardinia (the
the first, -D-, indicates the class, the second (M, O, P, Q, R) population with the highest frequency of HLA DR3). They
the family, and the third (A or B) the chain. Both classes of determined anti-GAD65, IA-2A, and IAA in a group of 62
molecules are heterodimers: class I exhibits an α-chain and GDM patients and in 56 controls. Frequency of AAs was
β2-microglobulin and class II exhibits α- and β-chains. The 38.8% (7.1% for the control population), being IA-2A the
function of the HLA molecules is to present short peptides most frequent antibody (29.0% in GDM patients, 7.1% in
to T cells to initiate the immune response. Multiple genetic controls). The prevalence of GAD-65 AA was 3.2% (absent
reports have demonstrated an association between various in controls). IAAs were positive in 14.5% of women with
HLA alleles and autoimmune disorders. GDM (absent in controls). Eight percent of all GDM
In Caucasian patients with DM-1, HLA-D genes con- cases were positive for two antibodies (0% in controls).98
tribute as much as 50% of the genetic susceptibility.94 Pregestational weight and BMI of DRA-positive GDM
HLA-DQ genes appear critical to the HLA-associated risk patients were lower than of DRA-negative GDM patients,
of DM-1A. In any individual, four possible DQ dimers are in coincidence with other reports. One plausible expla-
encoded; positive risks for the disease are associated with nation of these unique frequency and distribution of
Autoimmune GDM and LADA 61
pancreatic AAs among women presented with GDM in displayed fewer features of insulin resistance, required
Sardinia is their distinct genetic background. more frequent insulin therapy than negative women, and
Anti-IA2 correlates with DQ8 and/or DR4/DQB1*0201 presumably had presymptomatic DM-1. Nevertheless,
and is negatively associated with DR3/DQB1*02012; IAA other authors have reported that women with GDM with or
and ICA are found with higher frequency in individuals without GADA positivity at follow-up display similar clini-
positive for DR4 and DQ8, while anti-GAD65 is more com- cal characteristics with the exception of BMI.57
monly found in type 1 diabetic patients with DR3 and/or It may be hypothesized that women with AIGDM have
DQB1*0201 haplotypes.100 LADA, which it has been demonstrated not to be the case.
LADA is neither the same than classic DM-1A with onset
in adulthood; although ICA and GADA are common
Autoimmune GDM and LADA in LADA, both anti-IA2 and IAA are much less common in
LADA than in DM-1.101
S. Bo et al. investigated the differences in clinical charac- Our group has recently reported the characteristics of
teristics of women with GDM, whether associated with LADA in Spanish population in terms of β-cell autoimmu-
islet cell autoimmunity or not.35 A total of 207 women nity, insulin secretion, and clinical aspects.102 The study was
composed the total group of investigated subjects; 12.5% initially part of the Action LADA Project, which was locally
were carrying AAs in their sera, either ICAs or GADAs, extended for an additional period of 3 years. Cases of LADA
usually at low titers. Women with GDM with autoimmune were defined as patients age 30–70 years at the diagnosis of
markers apparently presented a lower risk for the develop- diabetes, who did not require insulin for at least 6 months
ment of conventional GDM (were younger, had lower pre- after diagnosis, with GADA or IA-2A positivity. A total of
pregnancy BMI, lower prevalence of diabetes in their FDRs, 82 consecutive patients with LADA, 78 patients with DM-1,
lower waist circumference, lower fasting plasma insulin, and 480 patients with type 2 diabetes were included in the
and lower weight increase during pregnancy). The rate study. Table 7.2 shows the main clinical parameters of the
of insulin treatment during pregnancy was significantly three study groups. The investigation of metabolic charac-
higher in the group positive for AAs. These observations teristics depicted for patients with LADA an intermediate
led the authors to the conclusion that women with AIGDM phenotype.
Table 7.2 Clinical characteristics, including fasting C-peptide concentrations of patients with type 1 diabetes,
latent autoimmune diabetes in adults, or type 2 diabetes
P value for group comparisons

T1DM LADA T2DM
(n = 78) (n = 82) (n = 480) T1DM—LADA T2DM—LADA
Sex (n, women) 33 (42.3%) 30 (36.6%) 174 (36.3%) 0.518 1.000
Age (years) 45.5 ± 11.9 53.4 ± 12.3 56.07 ± 11.0 <0.001 0.076
Age at diagnosis 42.6 ± 11.4 49.2 ± 11.7 53.0 ± 10.5 <0.001 0.011
Disease duration (months) 10 (6–39) 38 (6–59) 24 (6–47) <0.001 0.006
BMI (kg/m2) 25.6 ±4.92 26.9 ±4.4 29.9 ± 5.4 0.031 <0.001
Waist (cm) 90.4 ± 12.5 94.8 ± 16.6 101.6 ± 14.4 0.003 <0.001
SBP (mmHg) 123.2 ± 24.0 131.4 ± 19.1 138.6 ± 21.5 0.036 0.002
DBP (mmHg) 74.0 ± 15.1 79.5 ± 9.9 82.3 ± 11.7 0.001 0.035
Antihypertensive treatment 11 (14.1%) 30 (36.6%) 248 (51.7%) 0.001 0.012
Triglycerides (mg/dL) 90.7 ± 48.4 121.5 ± 84.3 163.3 ± 120.8 0.022 <0.001
Total cholesterol (mg/dL) 192.4 ± 37.4 202.4 ± 61.6 218.8 ± 37.9 0.420 0.284
HDL cholesterol (mg/dL) 67.7 ± 23.2 59.8 ± 17.5 50 ± 15.8 0.037 <0.001
LDL cholesterol (mg/dL) 113.2 ± 33.5 118.7 ±36.0 122.3 ± 35.5 0.321 0.192
Metabolic syndrome (n, %) 11 (15.5) 28 (37.3) 303 (67.2) 0.005 0.000
Fasting glycemia (mg/dL) 178.3 ± 80.7 164.8 ± 59.8 148.8 ± 56.3 0.529 0.010
HbAlc (%) 8.9 ± 2.7 8.1 ± 2.1 7.6 ± 1.9 0.098 0.007
C-peptide (pmol/L) 261.0 ± 234.0 587.0 ± 515.8 978.0 ± 480.7 <0.001 <0.001
Insulin treatment (alone or combined 78 (100.0) 37 (45.7) 91 (19.00) <0.001 <0.001
with other agents) (n, %)
Source: Mollo, A. et al., Diabetes Metab. Res. Rev., 29, 446, 2013. With permission.
The prevalence of metabolic syndrome was higher GDM-related autoantibodies

among patients with LADA than in patients with T1DM,
but lower in patients with T2DM. Fasting plasma glucose and the risk of maternal glucose
and HbA1c were higher in patients with LADA than in intolerance/diabetes mellitus
patients with T2DM. Hazard ratio for insulin treatment
in LADA compared with T2DM was quite elevated (mean The majority of reports have agreed that positivity for DRA
value of 8.34; p < 0.001). A faster progression to insulin during pregnancy increases the maternal risk of glucose
treatment from the diagnosis of diabetes in patients with intolerance/diabetes (Table 7.3).
LADA when compared with patients with T2DM was In the first investigation on the prevalence of ICAs in
clearly observed (Figure 7.1). GDM, three of five ICA (+) gestational diabetic women
If immunotherapy becomes a reality for autoimmune developed classic type 1A diabetes shortly after pregnancy.26
diabetes, the investigation of both patients with LADA and Additional studies have confirmed an increased risk of
women with AIGDM will be of great importance to iden- diabetes33 or glucose intolerance30 in women positive for
tify subjects at high risk for T1DM who might benefit from ICAs. The association has also been extended to other DRAs:
immune intervention. New initiatives to be developed in positivity for either ICA/IA-2A or GADAs increases the risk
these two distinct types of autoimmune diabetes—LADA of DM-1 2 years after delivery, with the risk increasing with
and AIGDM—may include, among others, (1) the assess- the number of positive antibodies.25
ment of AAs to IA-2β, which have been associated with a There are also articles reporting a lack of association of
very high risk of diabetes in ICA/IA-2A-positive relatives ICAs32 or GADAs58 with abnormal glucose tolerance at short
of DM-1 subjects103; (2) the use of the novel multi-antigen term after delivery that could be attributable to a low sta-
single radioimmunoassay, able to evaluate GADA, IA-2A, tistical power of the studies. Nevertheless, it is important to
IAA, and islet beta-cell zinc cation efflux transporter highlight that only two of these articles performed statistical
(ZnT8) AAs104; and (3) investigation of the humoral immu- adjustment for other predictors.25,32
noreactivity against IA-2(256-760), the marker with the Our team have prospectively followed a group of 470
highest sensitivity for detection of humoral IA-2 immu- women with GDM from the time of diagnosis until the time
noreactivity (IA-2 immunoreactivity has been underes- of the first postpartum oral glucose tolerance test (OGTT),
timated, so far, in patients with LADA and AIGDM).105 and up to 7 years afterward; another group of 200 women
Furthermore, diagnosing LADA and AIGDM at the initial depicting normal OGTT at pregnancy, AAs negative, and
stage will be quite useful to initiate insulin earlier, facili- similar age and BMI acted as control. Interestingly, the
tating improved glycemic control sooner as well as pre- prevalence of all investigated AAs (ICA, GADA, IA-2A)
serving the residual beta cell function in these two distinct markedly increased after delivery. In the systematic follow-
subclasses of autoimmune diabetes.106 up postpartum, the analysis of the curve of survival free of
DM interval revealed an abrupt decline for the AA + group
at the very early phase of the curve, remained afterward in
a parallel fashion to the graphic displaying the behavior of
Spanish LADA study the AA-negative group. In fact, about 80% of AA-positive
1.0 women evolved to clinically overt diabetes within the first
24 months postpartum (Figure 7.2a and b). Investigators
from the Münich Diabetes Research Institution also found
0.8 increased risk of progression to DM in the group of AA+
Proportion not requiring insulin
women with GDM presenting DR3 and DR4 haplotypes. In

this context, the association to GADA appears to have the
0.6
strongest influence.25
In addition to ICAs being predictive of diabetes at the
first assessment after delivery, 33 our team reported an
0.4
impairment of the acute insulin response to intravenous
glucose in women with GDM with positivity for ICAs and
0.2 normal glucose tolerance after delivery.107 The response was
superimposable to that of ICA(+) FDRs-DM-1 (Figure 7.3).
LADA
Interestingly, a Finnish study on FDRs of patients with
T2 DM
0.0 LADA demonstrated similar metabolic features that were
described by us in women with GDM and positivity for
0.00 20.00 40.00 60.00
ICAs. These individuals (family members of patients with
Time to insulin (months)
LADA) exhibited decreased insulin secretion, associated
Figure 7.1 Spanish autoimmune diabetes of adulthood study. with an increased prevalence of risk genotypes.108
(From Mollo, A. et al., Diabetes Metab. Res. Rev., 29, 446, At longer follow-up, an increased risk of DM-1 has been
2013. With permission.) demonstrated for ICA, 27,29,36,47,109 GADA, 27,54 and positivity
GDM-related autoantibodies and the risk of maternal glucose intolerance/diabetes mellitus 63
Table 7.3 Abnormal glucose tolerance at follow-up in women with gestational diabetes mellitus
and diabetes-related antibodies
Author Follow-up ICAs GADAs IA-2As Several DRAs

Steel et al.26 1 year Predictive of type 1
diabetes
Ginsberg- Years Predictive of type 1
Fellner diabetes
et al.29
Stowers et al. Up to Not predictive of
(1985) 22 years the final state of
glucose
tolerance
Catalano Up to 4 years Predictive of IGT
et al.30
Stangenberg 2–4 months Not predictive of
et al.31 abnormal OGTTa
Mauricio Months Predictive of
et al.33 diabetes
Damm et al.36 Up to Predictive of type 1
11 years diabetes
Beischer Years GADAs at
et al.53 follow-up
associated with
type 1 and type 2
diabetes
Petersen Up to Predictive of type 1 Predictive of type 1 DFJA positivity predictive
et al.54 11 years diabetes diabetes of type 1 diabetes.
Fuchtenbusch Up to 5 years Predictive of type 1 Predictive of type 1 Not predictive Risk of type 1 diabetes
et al.28 diabetesa diabetes of type 1 increases with the
diabetesa number of DRAs.a
Panczel Up to Predictive of type 1
et al.109 14 years diabetes
Kousta et al.57 Up to GADAs at
45 months follow-up not
associated with
different FBG or
HOMA
estimations of
insulin secretion
and sensitivity
Weng et al.58 1 year No association with
diabetes/IGT
Lapolla et al.37 5 years DRA positivity, borderline
in association to type 1
diabetes.
Albareda Up to DRA positivity not
et al.49 11 years predictive of diabetes,
type 1 diabetes, or type
2 diabetes.a
Järvelä et al. Up to 7 years Predictive of type 1 Predictive of type 1 Not predictive Number of DRAs
(2006) diabetesa diabetesa of type 1 predictive of type 1
diabetesa diabetes.a
Löbnner Up to GAD and/or IA-2 positivity
et al.110 11 years predictive of diabetes.a
Source: de Leiva, A. et al., Diabetes Care, 30(Suppl. 2), 127, 2007. With permission.
Note: For groups with several articles on the subject, only articles providing new information on the predictive ability of diabetes-related anti-
bodies are included.
Abbreviations: FBG, fasting blood glucose; HOMA, homeostasis model assessment; IGT, impaired glucose tolerance; OGTT, oral glucose toler-
ance test.
a Adjusted for other predictors.
100
Cumulative survival free from GDM (%)

0 Ab
1 Ab 80
1.2
2 Ab AA+
1 3 Ab AA–
60
0.8
0.6 40
0.4
20
0.2
0 0
0 2 4 6 8 10 12 0 2 4 6 8 10 12
(a) Follow-up (years) (b) Follow-up (years)
Figure 7.2 (a) GDM: progression to DM at postpartum follow-up. (b) Gestational diabetes mellitus (GDM): cumulative survival
free from GDM. ([a] Adapted from Mauricio, D. et al., Diabetes Metab. Res. Rev., 17(6), 422, 2001; [b] modified from Albareda, M.
et al., Diabetologia, 41, 106, 1998.)
800 *
** ICA+ GDM ICA+ relatives Control

**
600
Insulin (pmol/L)
n 9 9 12
** 1 + 3 insulin (pmol/L) 634 (241)* 635 (387) 1021 (391)
Area under the insulin curve 2422 (874)* 2690 (1402) 4155 (1381)
400 0– 10 minutes (pmol/L/min)
K value 1.76 (0.69)* 2.11 (1.07) 2.7 (0.91)
*p < 0.02 versus control group, and NS versus relatives. Numbers in
200
parenthesis indicate SD.
0
0 1 3 5 10
Minutes
Figure 7.3 Gestational diabetes mellitus: islet cell autoantibody, IVGTT, insulin release. (From Mauricio, D. et al., Diabet. Med.,
12, 1009, 1995. With permission.)
for one or more islet cell antibodies, 54,110–112 with the risk of autoimmune markers. This has already been observed
increasing with the number of antibodies, 27 but not for in FDRs of type 1 diabetic subjects. Another possibility
IA-2. 25,27 Even for ICAs and GADAs, not all articles find a points to the appearance of DM-1 shortly after pregnancy.
positive association between DRA positivity and diabetes Alternatively, progression to an established state of glucose
at follow-up, which in some cases, 28,27 but not others,49,57 intolerance could occur, which may eventually progress
can be attributed to a low statistical power. For example, to further deterioration of the insulin-secretory capacity
in our population, despite the aforementioned association as a result of the progression of autoimmune destruction
of ICA positivity with postpartum abnormal glucose toler- of beta cells. If the autoimmune process evolves, the time
ance, DRA positivity (ICA/GADA/IA-2, alone or in any course and degree of destruction may vary. This can result
combination) was not predictive of diabetes at midterm in a full and rapid progression to late-onset DM-1, with the
follow-up.49 As in the case of short-term follow-up, only typical onset of the disease. Alternatively, the process may
some studies have adjusted for other predictors27,49,110 progress slowly, resulting in a long subclinical period with
In our opinion, the beta-cell secretory capacity of an further progression to a non-insulin-dependent state man-
otherwise normal glucose-tolerant woman may be chal- ifested as LADA. Moreover, the course of the autoimmune
lenged by the higher functional demand associated with destruction of the beta-cell mass may be accelerated at any
the state of pregnancy. After delivery, the autoimmune time point in patients with LADA resulting in a rapid-onset
process may follow different pathways. One would be that insulin-dependent form of DM-1 (Figure 7.4).25
of the restoration of normal glucose tolerance when preg- These considerations were accepted by the Consensus
nancy is over, with eventual persistence or disappearance Panel at the Fifth International Workshop-Conference on
References 65
NGT patients with slowly evolving DM-1. We consider that this

condition may display the various types of expression of
Autoimmune GDM the immune reactivity against the β-cell. AIGDM appears
to be the result of the variable expression of autoimmu-
nity against the beta cell, challenged by the higher func-
NGT tional demand associated with the insulin-resistant state of
pregnancy.115
There are different time-course patterns in the progres-
IGT sion of AIGDM: from the restoration of normal glucose
tolerance when pregnancy is over (even with eventual dis-
appearance of autoimmune markers) to the appearance of
DM-1 shortly postpartum, to an established state of glucose
LADA/slowly progressive type 1 DM Type 1 DM
intolerance that may eventually progress slowly, and to a non-
insulin-dependent state, manifested as LADA. Furthermore,
Figure 7.4 Autoimmunity in gestational diabetes mellitus: the course of the autoimmune destruction of the residual
postpartum follow-up. (From Mauricio, D. et al., Diabetes beta cell mass may be accelerated at any time point resulting
Metab. Rev., 12, 275, 1996. With permission.) in a rapid-onset form of DM-1.
With these considerations, we have proposed that
GDM (Chicago, 2007). In the final report of the confer- AIGDM be regarded as a distinct clinical entity.115 This pro-
ence, it was included the proposal that women with GDM posal does not only make reference to a special subtype of
depicting AA should be subjected during the follow-up to GDM for academic or classification purposes, but it also
a closer surveillance due to their high risk to develop rapid addresses a peculiar and complex prediabetic status, suscep-
progression to insulinopenic diabetes.113 Theoretically, this tible of future new strategies for diabetes prevention. Due
subgroup of women can be clearly considered as predia- to its potential high risk for progression to clinically overt
betic type 1. Although initiatives to prevent DM-1 have been insulinopenia, women with AIGDM may be considered can-
unsuccessful so far, AA-associated GDM means a distinct didates for immune interventions.
category of candidates for such purpose. One illustration
of the need that women with GDM depicting AA should be
closely followed during pregnancy has been shown by the
group of DM and pregnancy of the University of Padova, Acknowledgments
reported in 2013. Three women attended either at pregnancy We thank all participants of the Spanish Pre-Diabetes Study
or at the postpartum displayed high GAD antibodies titers. Group; Professor Lluis Cabero and Dr. Joan Adelantado,
They developed persistent autoimmune diabetes requiring from the Department of Obstetrics and Gynecology,
insulin treatment (two of them during pregnancy and the Universitat Autònoma de Barcelona; Drs. Mercedes
third after partum).114 Albareda, Montserrat Balsells, Jaume Binimelis, Mercedes
Codina, Gemma Ginovart, Eugenia Mato, Josefa Morales,
Xavier Palomer, Sandra Piquer, and Manel Puig-Domingo,
Autoimmune GDM: A distinct for their collaboration in different phases of the research
prediabetic stage team activities and related publications; Eulalia Brugués
MSc (Pharmacy), for the efficient help in the preparation
Approximately, 10% of women with GDM displays humoral of the chapter; all registered nurses of the Department of
autoimmune markers against pancreatic cells in association Endocrinology and Diabetes, Obstetrics and Gynecology,
with glucose intolerance at pregnancy. and Pediatrics of the Hospital de la Santa Creu i Sant Pau
Freinkel et al.24 foresaw the evolution of what may be for the excellent services offered to all women and their
defined as AIGDM, when they wrote in 1987, that GDM newborns attended at the institution, and related to this
entails genotypic and phenotypic diversity and may include publication.
REFERENCES
1. Gleicher N. Pregnancy and autoimmunity. Acta Haematol 1986; 5. Bulmer R, Hancock W. Depletion of circulating T lymphocytes in
76: 68–77. pregnancy. Clin Exp Immunol 1977; 29: 302–307.
2. Lewis JE, Coulam CB, Moore S. Immunologic mechanisms in the 6. Clements PJ, Yu DTY, Levy J, Pearson CM. Human lymphocytes
maternal-fetal relationship. Mayo Clin Proc 1986; 61: 655–665. subpopulation: Effects on pregnancy. Proc Soc Exp Biol Med 1976;
3. Mauroluis GB, Buckley RH, Younger GB. Serum immunoglobulin 152: 664–672.
levels during normal pregnancy. Am J Obstet Gynecol 1971; 109: 7. Dodson MG, Kerman RH, Lange CF, Stefani SS, O’Leary JA. T and
971–976. B cells in pregnancy. Obstet Gynecol 1976; 49: 229–303.
4. Galluzzo A, Giordano C, Bompiani GF. Cell-mediated immunity 8. Shonfeld Y, Schwarz RS. Immunologic and genetic factors in auto-
in diabetic pregnancy. In: Andreani D, Bompiani G, Di Mario U, immune diseases. N Engl J Med 1984; 311: 1019–1029.
Page Faulk W, Galluzzo A, eds., Immunobiology of Normal and 9. Giordano C. Immunobiology of normal and diabetic pregnancy.
Diabetic Pregnancy, Wiley: Chichester, U.K., 1990, pp. 273–281. Immunol Today 1990; 11: 301–303.
10. Torry DS, McIntyre JA. The role of the antibody in pregnancy. In: 30. Catalano PM, Tyzbir ED, Sims EA. Incidence and significance of
Andreani D, Bompiani G, Di Mario U, Page Faulk W, Galluzo A, islet cell antibodies in women with previous gestational diabetes.
eds., Immunobiology of Normal and Diabetic Pregnancy, Wiley: Diabetes Care 1990; 13: 478–482.
Chichester, U.K., 1990, pp. 39–57. 31. Stangenberg M, Agarwal N, Rahman F, Sheth K, Al Sedeiry S,
11. Metzger BE, Coustan DM, and the Organizing Committee. De Vol E. Frequency of HLA genes and islet cell antibodies (ICA)
Summary and recommendations of the Fourth International and result of postpartum oral glucose tolerance tests (OGTT) in
Workshop–Conference on Gestational Diabetes Mellitus. Diabetes Saudi Arabian women with abnormal OGTT during pregnancy.
Care 1998; 21(Suppl. 2): B161–B167. Diabetes Res 1990; 14: 9–13.
12. Kjos SL, Buchanan TA. Gestational diabetes mellitus. N Engl J Med 32. Maclaren N, Lan M, Coutant R et al. Only multiple autoantibod-
1999; 341: 1749–1756. ies to islet cells (ICA), insulin, GAD65, IA-2 and IA-2 beta predict
13. Buchanan TA, Metzger BE, Freinkel N, Bergman RN. Insulin immune mediated (type 1) diabetes in relatives. J Autoimmun
sensitivity and B-cell responsiveness to glucose during late preg- 1999; 12: 279–287.
nancy in lean and moderately obese women with normal glu- 33. Mauricio D, Corcoy R, Codina M, Balsells M, Puig-Domingo
cose tolerance or mild gestational diabetes. Am J Obstet Gynecol M, Pou JM, de Leiva A. Islet cell antibodies identify a subset of
1990; 162: 1008–1014. gestational diabetic women with higher risk of developing dia-
14. Freinkel N, Metzger BE, Phelps RL et al. Gestational diabetes mel- betes mellitus shortly after pregnancy. Diab Nutr Metab 1992;
litus: Heterogeneity of maternal age, weight, insulin secretion, 5: 1–5.
HLA antigens, and islet cell antibodies and the impact of maternal 34. Kinalski M, Kretowski A, Telejko B, Kowalska I, Bingley P,
metabolism on pancreatic B-cell and somatic development in the Kinalska I. Prevalence of ICA antibodies, anti-GAD and anti-IA2
offspring. Diabetes 1985; 34(Suppl. 2): 1–7. in women with gestational diabetes treated with diet. Przegl Lek
15. Kim C, Newton KM, Knopp RH. Gestational diabetes and the inci- 1999; 56: 342–346.
dence of type 2 diabetes: A systematic review. Diabetes Care 2002; 35. Bo S, Menato G, Pinach S et al. Clinical characteristics and out-
25: 1862–1868. come of pregnancy in women with gestational hyperglycemia
16. Schranz DB, Lernmark A. Immunology in diabetes: An update. with and without antibodies to beta-cell antigens. Diabet Med
Diabetes Metab Rev 1998; 14: 3–29. 2003; 20: 64–68.
17. Palmer JP. Predicting IDDM: Use of humoral immune markers. 36. Damm P, Kuhl C, Buschard K, Kakobsen BK, Svedjaard A,
Diabetes Rev 1992; 1: 104–115. Sodoyez-Goffaux F, Shattock M, Bottazzo GF, Molsted-Pedersen
18. Bottazzo GF, Florin-Christensen A, Doniach D. Islet cell antibodies L. Prevalence and predictive value of islet cell antibodies and
in diabetes mellitus with polyendocrine autoimmune deficiencies. insulin autoantibodies in women with gestational diabetes.
Lancet 1974; 2: 1279–1282. Diabet Med 1994; 11: 558–563.
19. Baekkeskov S, Aanstoot HJ, Christgau S, Reetz A, Solimena M, 37. Lapolla A, Betterle C, Sanzari M et al. An immunological and
Cascalho M, Folli F, Richter-Olesen H, De Camilli P. Identification genetic study of patients with gestational diabetes mellitus. Acta
of the 64 autoantigen in insulin-dependent diabetes as the GABA Diabetol 1996; 33: 139–144.
synthesizing enzyme glutamic acid decarboxylase. Nature 1990; 38. Dozio N, Beretta A, Belloni C, Castiglioni M, Rosa S, Bosi E,
347: 151–156. Bonifacio E. Low prevalence of islet autoantibodies in patients
20. Kaufman DL, Erlander MG, Clare-Salzler M, Atkinson MA, with gestational diabetes mellitus. Diabetes Care 1997; 20:
Maclaren NK, Tobin AJ. Autoimmunity to two forms of glutamate 81–83.
decarboxylase in insulin-dependent diabetes mellitus. J Clin Invest 39. Ziegler AG, Hillebrand B, Rabl W, Mayrhofer M, Hummel M,
1992; 89: 283–292. Mollenhouer U, Vordemann J, Lenz A, Standl E. On the appear-
21. Atkinson MA, Maclaren NK. Islet cell autoantigens in insulin depen- ance of islet associated autoimmunity in offspring of diabetic
dent diabetes mellitus. J Clin Invest 1993; 92: 1608–1616. mothers: A prospective study from birth. Diabetologia 1993; 36:
22. Lan MS, Wassefall C, Maclaren NK, Notkins AL. IA-2, a transmem- 402–408.
brane protein of the protein tyrosine phosphatase family, as an 40. Palmer JP, Asplin CH, Clemons P. Insulin antibodies in insulin
autoantigen in insulin dependent diabetes mellitus. Proc Natl Acad dependent diabetics before insulin treatment. Science 1982; 222:
Sci USA 1996; 93: 6367–6370. 1337–1339.
23. Lu J, Li Q, Xie H, Chen ZJ, Borovitskaya AE, Maclaren NK, Notkins 41. Karjalainen J, Salmela P, Ilonen J, Surcel HM, Knip M. A compari-
AL, Lan MS. Identification of a second transmembrane protein tyro- son of childhood and adult type 1 diabetes mellitus. N Engl J Med
sine phosphatase, IA-2beta, as an autoantigen in insulin dependent 1989; 320: 881–886.
diabetes mellitus: Precursor of the 37-kDa tryptic fragment. Proc 42. Balsells M, Corcoy R, Mauricio D, Morales J, García-Patterson
Natl Acad Sci USA 1996; 93: 2307–2311. A, Carreras G, Puig-Domingo M, de Leiva A. Insulin antibody
24. Freinkel N, Metzger BE, Phelps RL, Simpson JL, Martin AO, Radvany response to a short course of human insulin therapy in women
R, Ober C, Dooley SL, Depp RO, Belton A. Gestational diabetes with gestational diabetes. Diabetes Care 1997; 20: 1172–1175.
mellitus: A syndrome with phenotypic and genotypic heterogene- 43. Jovanovic L, Ilic S, Pettitt DJ, Hugo K, Gutierrez M, Bowsher
ity. Horm Metab Res 1986; 18: 427–430. RR, Bastyr EJ. Metabolic and immunologic effects of insu-
25. Mauricio D, Balsells M, Morales J, Corcoy R, Puig-Domingo M, de lin lispro in gestational diabetes. Diabetes Care 1999; 22:
Leiva A. Islet cell autoimmunity in women with gestational diabe- 1422–1427.
tes and risk of progression to insulin-dependent diabetes mellitus. 44. Seissler J, Scherbaum WA. Autoimmune diagnostics in diabetes
Diabetes Metab Rev 1996; 12: 275–285. mellitus. Clin Chem Lab Med 2006; 44: 133–137.
26. Steel JM, Irvine WJ, Clarke BF. The significance of pancreatic islet 45. Leslie RDG, Atkinson MA, Notkins AL. Autoantigens IA-2 and
cell antibody and abnormal glucose tolerance during pregnancy. GAD in type I (insulin-dependent) diabetes. Diabetologia 1999;
J Clin Lab Immunol 1980; 4: 83–85. 42: 3–14.
27. Fallucca F, Di Mario U, Gargiulo P, Iavicoli M, Galfo C, Contreas 46. Bingley PJ, Bonifacio E, Ziegler AG, Schatz DA, Atkinson MA,
G, Pachi’ A, Andreani D. Humoral immunity in diabetic preg- Eisenbarth GS; Immunology of Diabetes Society. Proposed
nancy: Interrelationships with maternal/neonatal complica- guidelines on screening for risk of type 1 diabetes. Diabetes Care
tions and maternal metabolic control. Diabetes Metab 1985; 2001; 24: 398.
11: 387–395. 47. Palomer X, Mauricio D, Rodriguez-Espinosa J, Zapico E, Mayoral
28. Fuchtenbusch M, Ferber K, Standl E, Ziegler AG. Prediction of C, Gonzalez-Sastre F, de Leiva A, Blanco-Vaca F. Evaluation of
type 1 diabetes postpartum in patients with gestational diabe- two non-isotopic immunoassays for determination of glutamic
tes mellitus by combined islet cell autoantibody screening: A acid decarboxylase and tyrosine phosphatase autoantibodies in
prospective multicenter study. Diabetes 1997; 46: 1459–1467; serum. Clin Chem 2004; 50: 1378–1382.
Verge CF, Gianani R, Kawasaki E et al. Prediction of type I 48. Whittingham S, Byron SL, Tuomilehto J et al. Autoantibodies
diabetes in first-degree relatives using a combination of insulin, associated with presymptomatic insulin-dependent diabetes mel-
GAD, and ICA512bdc/IA-2 autoantibodies. Diabetes 1996; 45: litus in women. Diabet Med 1997; 14: 678–685.
926–933. 49. Albareda M, Caballero A, Badell G, Piquer S, Ortiz A, de Leiva A,
29. Ginsberg-Fellner F, Mark EM, Nechemias C, Hausknecht RU, Corcoy R. Diabetes and abnormal glucose tolerance in women
Rubinstein P, Dobersen NJ, Notkins AL. Islet cell antibodies in with previous gestational diabetes. Diabetes Care 2003; 26:
gestational diabetics. Lancet 1980; 2: 362–363. 1199–1205.
References 67
50. Lohmann T, Sessler J, Verlohren HJ, Schroder S, Rotger J, Dahn K, 69. Pitcher-Wilmott RW, Hindocha P, Wood CB. The placental trans-
Morgenthaler N, Scherbaum WA. Distinct genetic and immuno- fer of IgG subclasses in human pregnancy. Clin Exp Immunol
logical features in patients with onset of IDDM before and after 1980; 41: 303–308.
age 40. Diabetes Care 1997; 20: 524–529. 70. Lindberg B, Ivarsson SA, Landin-Olsson M, Sundkvist G,
51. Mayrhofer M, Rabin DU, Messenger L, Standl E, Ziegler AG. Svanberg L, Lenmark A. Islet autoantibodies in cord blood from
Value of ICA512 antibodies for prediction and diagnosis of children who developed type 1 (insulin dependent) diabetes mel-
type 1 diabetes. Exp Clin Endocrinol Diabetes 1996; 104: litus before 15 years of age. Diabetologia 1999; 42: 181–187.
228–234. 71. Schwartz R, Gruppuso PA, Petzold K, Brambilla D, Hiilesmaa V,
52. Tuomilehto J, Zimmet P, Mackay IR, Koskela P, Vidgren G, Teramo KA. Hyperinsulinemia and macrosomia in the fetus of the
Toivanen L, Tuomilehto-Wolf E, Kohtamaki K, Stengard J, Rowley diabetic mother. Diabetes Care 1994; 17: 640–648.
MJ. Antibodies to glutamic acid decarboxylase as predictors of 72. Lindsay RS, Ziegler AG, Hamilton BA, Calder AA, Johnstone
insulin-dependent diabetes mellitus before clinical onset of dis- FD, Walker JD. Scottish Multicentre Study of Diabetes and
ease. Lancet 1994; 343: 1383–1385. Pregnancy: Type 1 diabetes-related antibodies in the fetal circu-
53. Beischer NA, Wein P, Sheedy MT, Mackay IR, Rowley MJ, lation: Prevalence and influence on cord insulin and birth weight
Zimmet P. Prevalence of antibodies to glutamic acid decarbox- in offspring of mothers with type 1 diabetes. J Clin Endocrinol
ylase in women who have had gestational diabetes. Am J Obstet Metab 2004; 89: 3436–3439.
Gynecol 1995; 173: 1563–1569. 73. Bauman WA, Yalow RS. Transplacental passage of insulin com-
54. Petersen JS, Dyrberg T, Damm P, Kuhl C, Molsted-Petersen L, plexed to antibody. Proc Natl Acad Sci USA 1981; 78: 4588–4590.
Buschard K. GAD65 autoantibodies in women with gestational 74. Tamas G, Jr., Bekefi D, Gaal O. Insulin antibodies in diabetic
or insulin dependent diabetes mellitus diagnosed during preg- pregnancy. Lancet 1975; 1: 521.
nancy. Diabetologia 1996; 39: 1329–1333. 75. Di Mario U, Falluca F, Gargiulo P, Tiberti C, Scardellato A,
55. Fallucca F, Tiberti C, Torresi P, Cardellini G, Sciullo E, Arduini P. Insulin-antiinsulin complexes in diabetic women and
D’Aliberti T, Napoli A, Di Mario U. Autoimmune markers of their neonates. Diabetologia 1984; 27(Suppl.): 83–86.
diabetes in diabetic pregnancy. Ann Ist Super Sanita 1997; 33: 76. Weitgasser R, Spitzer D, Kartnig I, Zajc M, Staudach A, Sandhofer
425–428. F. Association of HELLP syndrome with autoimmune antibodies
56. Mitchell ML, Hermos RJ, Larson CA, Palomaki GE, Haddow JE. and glucose intolerance. Diabetes Care 2000; 23: 786–790.
Prevalence of GAD autoantibodies in women with gestational 77. Murata K, Toyoda N, Sugiyama Y. The effects of insulin antibod-
diabetes mellitus. Diabetes Care 2000; 23: 1705–1706. ies during diabetic pregnancy on newborn infants. Asia Oceania
57. Kousta E, Lawrence NJ, Anyaoku V, Johnston DG, McCarthy MI. J Obstet Gynaecol 1990; 16: 115–122.
Prevalence and features of pancreatic islet cell auto-immunity in 78. Balsells M, Corcoy R, García-Patterson A, Morales J, Puig M, de
women with gestational diabetes from different ethnic groups. Leiva A. Insulin antibodies and materno-fetal morbidity in gesta-
BJOG 2001; 108: 716–720. tional diabetic women. Diabetologia 1994; 37(Suppl. 1): 172A.
58. Weng J, Ekelund M, Lehto M, Li H, Ekberg G, Frid A, Aberg A, 79. Greeley SA, Katsumata M, Yu L, Eisenbarth GS, Moore DJ,
Groop LC, Berntorp K. Screening for MODY mutations, GAD Goodarzi H, Barker CF, Naji A, Noorchashm H. Elimination of
antibodies, and type 1-diabetes associated HLA genotypes in maternally transmitted autoantibodies prevents diabetes in non-
women with gestational diabetes mellitus. Diabetes Care 2002; obese diabetic mice. Nat Med 2002; 8: 399–402.
25: 68–71. 80. McKeever U, Khandekar S, Newcomb J et al. Immunization with
59. Balaji M, Shatauvere-Brameus A, Valaji V, Seshiah V, Sanjeevi soluble BDC 2.5 T cell receptor immunoglobulin chimeric pro-
CB. Women diagnosed with gestational diabetes mellitus do not tein: Antibody specificity and protection of non obese diabetic
carry antibodies against minor cell antigens. Ann N Y Acad Sci mice against adoptive transfer of diabetes by maternal immuniza-
2002; 958: 281–284. tion. J Exp Med 1996; 184: 1755–1768.
60. Füchtenbusch M, Bonifacio E, Lampasona V, Knopff A, Ziegler 81. Koczwara K, Bonifacio E, Ziegler AG. Transmission of maternal
AG. Immune responses to glutamic acid decarboxylase and insu- islet antibodies and risk of autoimmune diabetes in offspring of
lin in patients with gestational diabetes. Clin Exp Immunol 2004; mothers with type 1 diabetes. Diabetes 2004; 53: 1–4.
135: 318–321. 82. Srikanta S, Richter AT, MacCulloch DK et al. Autoimmunity to
61. Hawa MI, Fava D, Medici F, Deng YJ, Notkins AL, De Mattia G, insulin, beta-cell dysfunction and development of insulin-depen-
Leslie RD. Antibodies to IA-2 and GAD65 in type 1 and type dent diabetes mellitus. Diabetes 1986; 36: 139–142.
2 diabetes: Isotype restriction and polyclonality. Diabetes Care 83. Leslie RD, Delli Castelli M. Age-dependent influences on the
2000; 23: 228–233. origins of autoimmune diabetes: Evidence and implications.
62. Hämäläinen AM, Ronkainen MS, Akerblom HK, Knip M. Postnatal Diabetes 2004; 53: 3033–3040.
elimination of transplacentally acquired disease-associated anti- 84. Horton V, Stratton I, Bottazzo GF, Shattock M, Mackay I, Zimmet
bodies in infants born to families with type 1 diabetes: The Finnish P, Manley S, Holman R, Turner R, the UK Prospective Diabetes
TRIGR Study Group: Trial to Reduce IDDM in the Genetically at Study (UKPDS) Group. Genetic heterogeneity of autoimmune dia-
Risk. J Clin Endocrinol Metab 2000; 85: 4249–4253. betes: Age at presentation in adults is influenced by HLA DRB1 and
63. Spellacy WN, Goetz FC. Insulin antibodies in pregnancy. Lancet DQB1 genotypes (UKPDS 43). Diabetologia 1999; 42: 608–616.
1963; 2: 222–224; Gamlen TR, Aynsley-Green A, Irvine WJ, 85. Tuomi T, Groop LC, Zimmet PZ et al. Antibodies to glutamic
McCallum CJ. Immunological studies in the neonate of a mother acid decarboxylase reveal latent autoimmune diabetes melli-
with Addison’s disease and diabetes mellitus. Clin Exp Immunol tus in adults with a non-insulin dependent onset of the disease.
1977; 28: 192–195. Diabetes 1993; 42: 359–362.
64. Menon RK, Cohen RM, Sperling MA, Cutfield WS, Mimouni F, 86. Turner R, Stratton I, Horton V et al., for the UK Prospective Diabetes
Khouri JC. Transplacental passage of insulin in pregnant women Study (UKPDS) Group: UKPDS 25. Autoantibodies to islet cyto-
with insulin-dependent diabetes mellitus: Its role in fetal macro- plasm and glutamic acid decarboxylase for prediction of insulin
somia. N Engl J Med 1990; 323: 309–315. requirement in type 2 diabetes. Lancet 1997; 350: 1288–1293.
65. Mylvaganam R, Stowers JM, Steel JM, Wallace J, MacHendry JC, 87. Zimmet PZ, Tuomi T, Mackay JR et al. Latent autoimmune dia-
Wright AD. Insulin immunogenicity in pregnancy: Maternal and betes in adults (LADA): The role of antibodies to glutamic acid
fetal studies. Diabetologia 1983; 24: 19–25. decarboxylase in diagnosis and prediction of insulin depen-
66. Weiss PA, Kainer F, Purstner P, Zehetleitner G, Huttner U, Haas dency. Diabet Med 1994; 11: 299–303.
J. Anti-insulin antibodies and birth weight in pregnancies com- 88. Hawa MI, Kolb H, Schloot N et al. Adult-onset autoimmune dia-
plicated by diabetes. Early Hum Dev 1998; 53: 145–154. betes in Europe is prevalent with a broad clinical phenotype:
67. Roll U, Christie MR, Füchtenbusch M, Payton MA, Hwkes CJ, Action LADA 7. Diabetes Care 2013 April; 36(4): 908–913.
Ziegler AG. Perinatal autoimmunity in offspring of diabetic par- 89. Zimmet PZ, Tuomi T, Mackay IR, Rowley MJ, Knowles W, Cohen
ents: The German Multicenter BABY-DIAB study: Detection of M, Lang DA. The role of antibodies to glutamic acid decarbox-
humoral immune responses of islet antigens in early childhood. ylase in diagnosis and prediction of insulin dependency. Diabet
Diabetes 1996; 45: 967–973. Med 1994; 11: 299–303.
68. Omar MA, Srikanta S, Eisenbarth GS. Human islet cell antibod- 90. Desai M, Zeggini E, Horton VA et al. A association analysis of
ies: Immunoglobulin class and subclass distribution defined by the HLA gene region in latent autoimmune diabetes in adults.
monoclonal antibodies. Diabetes Res 1987; 4: 155–157. Diabetologia 2007; 50(1): 68–73.
91. Howson JMM, Rosinger S, Smyth DJ, Boehm BO, the ADBW- 105. Tiberti C, Giordano C, Locatelli M, Bosi E, Bottazzo GF, Buzzetti
END Study Group, Todd JA. Genetic analysis of adult-onset auto- R, Cucinotta D, Galluzzo A, Falorni A, Dotta F. Identification of
immune diabetes. Diabetes 2011; 60: 2645–2653. tyrosine phosphatase 2(256–760) construct as a new, sensitive
92. Padmos RC, Schloot NC, Beyan H et al., the LADA Consortium. marker for the detection of islet autoimmunity in type 2 diabetic
Distinct monocyte gene-expression profiles in autoimmune dia- patients. Diabetes 2008; 57: 1276–1283.
betes. Diabetes 2008; 57: 2768–2773. 106. Thunander M, Thorgeirsson H, Törn C, Peterson C, Landin-
93. Andersen MK, Lundgren V, Turunen JA, Forsblom C, Isomaa Olsson M. β-cell function and metabolic control in latent auto-
B, Groop PH, Groop L, Tuomi T. Latent autoimmune diabe- immune diabetes in adults with early insulin versus conventional
tes in adults differs genetically from classical type 1 diabetes treatment: A 3-year follow-up. Eur J Endocrinol 2011; 164:
diagnosed after the age of 35 years. Diabetes Care 2010; 33: 239–245.
2062–2064. 107. Mauricio D, Corcoy R, Codina M, Morales J, Balsells M, de
94. Todd JA. Genetics of type 1 diabetes. Pathol Biol 1997; 45: Leiva A. Islet cell antibodies and beta cell function in gestational
219–227. diabetic women: Comparison to first degree relatives of type
95. Redondo MJ, Kawasaki E, Mulgrew CL et al. DR- and 1 (insulin-dependent) diabetic subjects. Diabet Med 1995; 12:
DQ-associated protection from type 1A diabetes: Comparison 1009–1014.
of DRB1*1401 and DQA1*0102-DQB1*0602*. J Clin Endocrinol 108. Vauhkonen I, Niskanen L, Knip M, Ilonen J, Vanninen E,
Metab 2000; 85: 3793–3797. Kainulainen S, Uusitupa M, Laakso M. Impaired insulin secretion
96. Bell DSH, Barger BO, Go RCP et al. Risk factors for gestational in non-diabetic offspring of probands with latent autoimmune
diabetes in black population. Diabetes Care 1990; 13(Suppl. 4): diabetes in adults. Diabetologia 2000; 43: 69–78.
1196–1201. 109. Panczel P, Kulkey O, Luczay A et al. Detection of antibodies
97. Vambergue A, Fajardi I, Bianchi F et al. Gestational diabetes against pancreatic islet cells in clinical practice. Orv Hetil 1999;
mellitus and HLA class II (-DQ, -DR) association: The DIAGEST 140: 2695–2701.
Study. Eur J Immunogenet 1997; 24: 385–394. 110. Löbnner K, Knopff A, Baumgarten A, Mollenhauer U, Marienfeld
98. Ferber K, Keller E, Albert ED, Ziegler A-G. Predictive value of S, Garrido-Franco M, Bonifazio E, Ziegler AG. Predictors of post-
human leucocyte antigen Class II typing for the development of partum diabetes in women with gestational diabetes mellitus.
islet autoantibodies and insulin-dependent diabetes postpartum Diabetes 2006; 55: 792–797.
in women with gestational diabetes. J Clin Endocrinol Metab 111. Lauenborg J, Hansen T, Jensen DM, Vestergaard H, Molsted-
1999; 84: 2342–2348. Pedersen L, Hornnes P, Locht H, Pedersen O, Damm P. Increasing
99. Bartha JL, Martínez del Fresno P, Comino-Delgado R. Postpartum incidence of diabetes after gestational diabetes: A long-term
metabolism and autoantibody markers in women with gesta- follow-up in a Danish population. Diabetes Care 2004; 27:
tional diabetes mellitus diagnosed in early pregnancy. Am J 1194–1199.
Obstet Gynecol 2001; 184: 965–970. 112. Corcoy R, Mauricio D, de Leiva A. Diabetes-related antibod-
100. Murgia C, Orru M, Potoghese E et al. Autoimmunity in gesta- ies and pregnancy. In: Djelmis J, Desoye G, Ivanisevic M, eds.,
tional diabetes mellitus in Sardinia: A preliminary case-control Diabetology of Pregnancy: Frontiers in Diabetes, Vol. 17, Krager:
report. Reprod Biol Endocrinol 2008; 6: 24–29. Basel, Switzerland, 2005, pp. 195–205.
101. Palmer J, Hampe CS, Chiu H, Goel A, Brooks-Worrel BM. Is 113. Metzger B, Buchanan TA, Coustan DR et al. Summary and rec-
latent autoimmune diabetes in adult distinct from type 1 diabetes ommendations of the Fifth International-Workshop-Conference
at an older age? Diabetes 2005; 54: S62–S67. on Gestational Diabetes Mellitus. Diabetes Care 2007; 30(Suppl.
102. Mollo A, Hernández M, Marsal JR, Esquerda A, Rius F, Blanco- 2): S251–S260.
Vaca F, Verdaguer J, Pozzilli P, de Leiva A, Mauricio D. Latent 114. Bonsembiante B, Dalfrà MG, Masin M, Gallo A, Lapolla A. Adult-
autoimmune diabetes in adults is perched between type 1 and onset type 1 diabetes and pregnancy: Three case reports. Case
type 2: Evidence from adults in one región of Spain. Diabetes Rep Med 2013-dx.
Metab Res Rev 2013; 29: 446–451. 115. de Leiva A, Mauricio D, Corcoy R. Diabetes related autoanti-
103. Achenbach P, Bonifacio E, Williams JK, Ziegler AG, Gale EAM, bodies and gestational diabetes mellitus. Diabetes Care 2007;
Bingley PJ, the ENDIT Group. Autoantibodies to IA-2β improve 30(Suppl. 2): 127–133.
diabetes risk assessment in high-risk relatives. Diabetologia 2008; 116. Albareda M, Corcoy R, Piquer S et al. Prevalence and predictive
51: 488–492. value of GAD and IA2 antibodies in a group of women with ges-
104. Tiberti C, Yu L, Lucantoni F, Panimolle F, Spagnuolo I, Lenzi A, tational diabetes-mellitus related to ICA. Diabetologia 1998; 41:
Eisenbarth GS, Dotta F. Detection of four diabetes specific auto- 106–106.
antibodies in a single radioimmunoassay: An innovative high- 117. Mauricio D, de Leiva A. Autoimmune gestational diabetes melli-
throughput approach for autoimmune diabetes screening. Clin tus: A distinct clinical entity? Diabetes Metab Res Rev November–
Exp Immunol 2011; 166: 317–324. December 2001; 17(6): 422–428.
8 Epidemiology of gestational
diabetes mellitus
Yariv Yogev, Avi Ben Haroush, Moshe Hod, and Jeremy Oats
The prevalence of GDM varies in direct proportion to

Introduction the prevalence of type 2 DM in a given population or ethnic
Gestational diabetes mellitus (GDM) is defined as carbohy- group.1 The reported prevalence of GDM in the United States
drate intolerance that begins or is first recognized during ranges from 1% to 14%, with 2% to 5% being the most com-
pregnancy.1 Although it is a well-known cause of preg- mon rate.5 In an earlier report, using previous WHO criteria,
nancy complications, a systematic review of its epidemiol- on the prevalence of diabetes and impaired glucose tolerance
ogy has become more complex following the introduction (IGT) in diverse populations in women between the ages of 20
of new World Health Organization (WHO)/International and 39, the WHO Ad Hoc Diabetes Reporting Group6 noted
Association Diabetes in Pregnancy Study Group (IADPSG) lower rates of diabetes (<1%) in Bantu (Tanzania), Chinese
international criteria that are for the first time based on the rural Indian, Sri Lankan, and some Pacific populations fol-
risk of adverse perinatal outcomes.2 A further problem is the lowed (1%–3%) by Italian women and white, black, and
distinction of GDM from preexisting but undiagnosed dia- Hispanic women in the United States. Rural Fijian Indian and
betes, so that the degree of clinical surveillance may have a Aboriginal Australian women had 7% prevalence; the high-
major impact on the estimated prevalence of GDM in a given est rate was found in Pima/Papago Indians and Nauruans
population. This is especially true in high-risk populations (14%–22%). The prevalence of IGT was low (<3%) in Chinese
in which the onset of type 2 diabetes mellitus (type 2 DM) and Malays and was >10% in black and Hispanic women in
occurs at an early age.3 Further, investigators have used dif- the United States, urban Indian women in Tanzania, Pima
ferent screening programs and diagnostic criteria for GDM, and Nauruan, and some other Pacific communities. The
making comparisons among studies difficult. combined age-standardized prevalence of diabetes and IGT
In this chapter, the reported risk factors for GDM, dif- ranged from 0% to 36%, with a >10% prevalence in one-
ferences in its racial distribution, and evidence of a genetic third of the populations and a >30% prevalence in Pima and
or familial association will be discussed. These differences Nauruan. Importantly, in some populations, more than half
may well need to be further reviewed as studies using the of the cases of diabetes were undiagnosed prior to the survey.
new criteria become available. The close relationship of IGT was mostly overlooked in routine clinical practice. This,
GDM to polycystic ovarian syndrome (PCOS), the aspects of a substantial proportion of abnormal glucose tolerance in
the associations between GDM and adverse pregnancy out- pregnancy, will be undetected without screening.
comes, including congenital anomalies, and the risk of type When analyzing these older studies, it is important
2 DM will also be described. before concluding that there are necessarily marked racial
and geographic variations in the prevalence of GDM, to
make allowance for the remarkably variable approaches
Racial distribution of gestational used across different studies, including different methods
diabetes mellitus of screening, different oral glucose and intravenous glucose
loads, and different diagnostic criteria. For example, Dooley
The International Diabetes Federation recently estimated that et al.7 demonstrated that race as well as maternal age and
16.6% of live births to women in 2013 had some form of hyper- degree of obesity must be taken into account in comparing
glycemia during the pregnancy of which 16% were GDM using the prevalence of GDM in different populations. Their study
the new criteria.4 There were major regional differences rang- included 3744 consecutive pregnant women who under-
ing from 25.0% in Southeast Asian region to 10.4% in North went universal screening. The population was 39.1% white,
America and the Caribbean region. Of concern, 91.6% of cases 37.7% black, 19.8% Hispanic, and 3.4% oriental/other. Black
were in low- and middle-income countries where access to and Hispanic race, maternal age, and percentage ideal body
maternity care can be problematic. The prevalence by interna- weight had a significant independent effect on the prevalence
tional diabetes foundation (IDF) region is shown in Table 8.1. of GDM. The adjusted relative risk (RR) was higher in black
69
70 Epidemiology of gestational diabetes mellitus
Table 8.1 Hyperglycemia in pregnancy (20–49 years) by IDF Region, 2013
IDF region Cases in live births (millions) Comparative prevalence (%)

Africa 4.6 14.4
Europe 1.7 12.6
Middle east and North Africa 3.4 17.5
North America Caribbean 0.9 10.4
South and Central America 0.9 11.4
Southeast Asia 6.3 25.0
Western Pacific 3.7 11.9
Source: International Diabetes Federation, IDF Diabetes Atlas, 6th ed., IDF, Brussels, Belgium, 2013.
(1.81 95% confidence interval [CI] 11.13–2.89) and Hispanic in Beersheba, Israel, 9.9% in Barbados, West Indies, to 21% in
(2.45, 95% CI 1.48–4.04) women than in white women. The Manchester, United Kingdom, and 23.7% Cleveland, United
degree of carbohydrate intolerance was similar across racial States (Table 8.2).10
groups; nevertheless, when the 92 patients with GDM under
dietary control were analyzed separately, mean birth weight
was found to be highest in Hispanic women and was low- Risk factors for gestational diabetes
est in the blacks and Orientals. Hence, race had a significant mellitus
independent effect on birth weight, with percentage ideal
body weight a significant covariant. These findings are sup- The traditional and most often reported risk factors for GDM
ported by Gunton et al.8 who showed Asian women were are high maternal age, weight and parity, previous delivery
more likely to have GDM than Caucasian women (31.7% vs. of a macrosomic infant, and a family history of diabetes.
14%, p = 0.02), despite their lower body mass index (BMI). These and other reported risk factors are summarized in
Although the Hyperglycemia and Adverse Pregnancy Table 8.2. It is of great importance that the clinician under-
Outcome (HAPO) study9 was not a population study, when the stand and use these characteristics, along with others, such
WHO/IADPSG criteria were applied to the glucose tolerance as racial and geographic attributed risk (discussed earlier),
test (GTT) results, there were marked differences in the preva- to improve screening programs and diagnostic accuracy and
lence of GDM across the 15 field centers in 9 different coun- perhaps to design better and more cost-effective selective
tries. The overall prevalence was 16.1%, rates ranged from 8.7% screening and diagnostic tests.
Table 8.2 Frequency of gestational diabetes mellitus by hyperglycemia and adverse

pregnancy outcome field center using World Health Organization/IADPSG criteria
Field center No. participants Percent with GDM

HAPO overall 23,316 16.1
Cleveland, USA 784 23.7
Bellflower, USA 1,903 22.9
Singapore 1,695 22.4
Bangkok, Thailand 2,426 21.2
Manchester, U.K. 2,250 21.0
Chicago, USA 738 16.5
Belfast, U.K. 1,634 15.5
Toronto, Canada 1,988 14.6
Providence, USA 746 14.2
Newcastle, Australia 653 13.6
Hong Kong, China 1,620 13.4
Brisbane, Australia 1,437 12.1
Barbados, West Indies 2,034 9.9
Petah Tikva, Israel 1,798 9.2
Beersheba, Israel 1,610 8.7
Source: Sacks, D.A. et al., Diabetes Care, 35, 526, 2012.

Polycystic ovary syndrome and gestational diabetes mellitus 71
Jang et al.11 examined 3581 consecutive Korean women Is GDM a cause or an effect? A retrospective study in
and found a 2.2% prevalence of GDM. The affected women Hong Kong18 in 84 normotensive women showed that the
were older and had higher prepregnancy weights, higher progressive glucose intolerance throughout pregnancy
BMI, higher parity, and higher frequencies of known diabe- is associated with an upward shift in blood pressure in
tes in the family. The risk of diabetes was closely associated the third trimester. Hence, it is possible that blood pres-
with previous obstetric outcome, such as congenital malfor- sure changes below the diagnostic threshold for hyperten-
mation, stillbirth, and macrosomia. The number of risk fac- sive disorders of pregnancy may help identify women at
tors present in each individual increased the risk of diabetes, increased risk of GDM.
with the prevalence ranging from 0.6% in subjects without The relationship between dietary fat and glucose metabo-
any risk factors to 33% in those with four or more. Thus, it lism has been recognized for many years. Epidemiological
is possible that selective screening may be cost-effective in data in humans suggest that subject with a higher fat intake
situations where health resources are scarce and where total are more prone to disturbances in glucose metabolism.19
screening is impossible.2 Several researchers have hypothesized that polyunsatu-
Similar results were reported in a retrospective cohort rated fatty acid plays an essential role in the maintenance of
study of 2574 pregnant women, which suggested that elec- energy balance and, through regulation of gene transcrip-
tive screening programs have a true-positive yield.12 An age tion, may improve insulin resistance.20–22 A recent small
of ≥30, a family history of diabetes, obesity (BMI ≥ 27), and study reported significantly lower cord vein erythrocyte
previous fetal macrosomia were the most important risk fac- phospholipid fatty acid concentrations in 13 women with
tors. Just over half (54.2%) of the population present with one GDM compared with 12 women with normal pregnancies.23
or more risk factors. The positive predictive value of screen- Accordingly, Bo et al.24 investigated the relationship between
ing increased with the number of risk factors, from 12% for lifestyle habits and glucose abnormalities in 504 Caucasian
the women with no risk factors to 40% for those with three women with and without conventional risk factors for GDM.
or more risk factors. They identified 126 women with GDM and 84 with IGT.
In another study, Jang et al.13 demonstrated that in the These women were older and shorter than the women with
racially homogeneous population of Seoul, Korea, besides normal pregnancies and had significantly higher prepreg-
prepregnancy BMI, age, weight gain, and parental history nancy BMI, higher rates of diabetes in first-degree relatives,
of diabetes, short stature is an independent risk factor for and higher intakes of saturated fat. In a multiple logistic
GDM. Accordingly, Kousta et al.14 reported that European regression model, all of these factors were associated with
and South Asian women with previous GDM were shorter glucose abnormalities, after adjustment for gestational age.
than control women from the same ethnic groups, perhaps In the patients without conventional risk factors, only the
due to a common pathophysiological mechanism underly- percentage of saturated fats (odds ratio [OR] = 2.0, 95% CI
ing GDM and the determination of final adult height. Others 1.2–3.2) and polyunsaturated fats (OR = 0.85, 95% CI 0.77–
have reported similar results.15 0.92) were associated with gestational hyperglycemia, after
In a large retrospective cohort study in Canada, Xiong adjustment for age, gestational age, and BMI. Thus, the alleg-
et al.16 evaluated 111,563 pregnancies and detected a 2.5% edly independent role of saturated fat in the development of
prevalence of GDM. The risk factors identified were age gestational glucose abnormalities takes on greater impor-
>35 years, obesity, history of prior neonatal death, and a tance in the absence of conventional risk factors. This sug-
prior cesarean section. Interestingly, teenage mothers and gests that glucose abnormalities could be prevented in some
women who drank less alcohol were less likely to have GDM. groups of women during pregnancy.
The risk factors mentioned earlier are mainly of mater- A possible expression of the still unknown genetic link-
nal origin. However, cumulative knowledge about the long- age in GDM was reported by Lao and Ho, 25 who detected
term implications of exposure to the diabetic intrauterine GDM in 62% of 163 women with the alpha-thalassemia
environment (see Chapter X) has led to the addition of the trait compared to 14.7% out of 163 controls matched for age
mother’s fetal history to the risk factor list. Egeland et al.17 and parity.
investigated whether the mother’s own characteristics at
birth could predict her subsequent risk of GDM. Using
linked generation data from the Medical Birth Registry of Polycystic ovary syndrome and
Norway for all women born between 1967 and 1998, the gestational diabetes mellitus
authors identified 498 women aged <32 years with GDM
in one or more singleton pregnancies. They found that the PCOS is a heterogeneous disorder affecting 5%–10% of
women whose mothers had diabetes during pregnancy were women of reproductive age. It is characterized by chronic
at increased risk of GDM themselves. Significant inverse anovulation with oligo-/amenorrhea, infertility, typical
trends in diabetes were noted in relation to birth weight, sonographic appearance of the ovaries, and clinical or bio-
with an increased risk of GDM of 80%, 60%, and 40% in chemical hyperandrogenism. Insulin resistance is present in
women whose birth weights were ≤2500, 2500–2999, and 40%–50% of patients, especially in obese women.26
3000–3499 g, respectively, compared with women in the Holte et al.27 reported a higher rate of ultrasonographic,
4000–4500 g group. Similar findings were observed for cat- clinical, and endocrine signs of PCOS in 34 women who
egories of weight for gestational age. had had GDM 3–5 years before, compared to 36 matched
controls with uncomplicated pregnancies. Five of the different, suggesting a minor clinical impact. Wein et al.35
women (15%) with previous GDM had developed manifest compared the prevalence of GDM between 61,914 singleton
diabetes. The authors concluded that in women with previ- and 798 twin deliveries performed between 1971 and 1991.
ous GDM, PCOS may form a distinct subgroup from women The difference was significant only for the earlier decade
with normal ovaries and previous GDM, who may be more (5.6% vs. 7.4%, p = 0.025). However, in a follow-up program,
prone to develop features of insulin resistance syndrome. there was a trend toward higher prevalence of overt diabe-
Many other researchers reported similar results, Kousta tes in women who had a diabetic twin pregnancy (18.5%)
et al.28 found a higher prevalence of PCOS in 91 women compared to those who had a diabetic singleton pregnancy
with previous GDM compared to 73 normoglycemic control (7.4%). Whether this represents a true increased risk for dia-
women (52% vs. 27%, p = 0.002), and Anttila et al.29 reported betes is unknown.
a 44% prevalence of PCOS in women with GDM, with no dif- By contrast, using data derived from the Medical Birth
ferences in BMI before pregnancy or in weight gain during Registry of Norway, Egeland and Irgens,36 controlling for
pregnancy compared to controls. They suggested a screening other risk factors such as advance age, parity, maternal his-
program for GDM in these patients. tory of diabetes, and the woman’s own birth weight, found
Mikola et al.30 retrospectively evaluated 99 pregnancies GDM in 6.6 per 1000 multiple pregnancies (n = 9,271) and
in women with PCOS compared with an unselected control in 5.0 per 1000 singleton pregnancies (n = 640,700) (OR =
population. The average BMI and the nulliparity rate were 1.3 95% CI 1.0–1.7, p = 0.03). However, analysis stratified
higher in the PCOS group, as was the multiple pregnancy by maternal age or parity yielded no elevated risk of GDM.
rate (9.2% vs. 1.1%). GDM developed in 20% of the patients Others have also failed to demonstrate a higher prevalence of
with PCOS but in only 8.9% of the controls (p < 0.001). GDM in multiple pregnancies.37,38
A BMI >25 was the best predictor of GDM (adjusted OR =
5.1, 95% CI 3.2–8.3), and PCOS was an additional indepen-
dent predictor (adjusted OR = 1.9, 95% CI 1.0–3.5). Genetic factors
Koivunen et al.31 found that compared with 48 control
Animal studies have shown that female fetuses exposed
women, 33 women with previous GDM more often had sig-
to a diabetic intrauterine milieu have an increased risk of
nificantly abnormal oral glucose tolerance tests (OGTT);
subsequent GDM. In a family history study, Harder et al.39
higher prevalences of polycystic ovaries (39.4% vs. 16.7%,
reported a significantly greater prevalence of diabetes
p = 0.03); higher serum concentrations of cortisol, dehydro-
(mainly type 2 DM) in the mothers of women with GDM
epiandrosterone, and dehydroepiandrostenedione sulfate;
than in their fathers. A significant aggregation of type 2
and a greater area under the glucose curve.
DM was also observed in the maternal–grandmaternal line
compared to the paternal–grandpaternal line. However, in
patients with IDDM, there was no significant difference in
Multiple pregnancy and gestational the prevalence of any type of diabetes between mothers and
diabetes mellitus fathers. Therefore, a history of type 2 DM on the mother’s
side might be considered as a particular risk factor for GDM
The number of fetuses in multifetal pregnancies is expected via “intergenerational transmission of type 2 DM,” which
to influence the incidence of GDM owing to the increased might be prevented by strict avoidance of GDM.
placental mass and, thereby, the increase in diabetogenic Dorner et al.40 reported a significantly decreased familial
hormones. However, the reports are somewhat conflicting, diabetes aggregation on the maternal side in children with
probably because of the heterogeneous populations studied. type 1 DM born between 1974 and 1984 compared to those
In an interesting study of the prevalence of GDM in dizy- born between 1960 and 1973. This finding was explained
gotic (DZ) twin pregnancies with two placentas compared by the improved prevention of hyperglycemia during preg-
to monozygotic (MZ) twin pregnancies with one placenta, nancy since 1974 and particularly of GDM in women with
Hoskins32 evaluated 3458 recorded twin deliveries and found familial diabetes aggregation. These authors also noted a
a higher proportion of different sex compared with same- highly significant predominance of type 2 DM in the great-
sex twin pregnancies complicated by GDM (3.5% vs. 1.6%). grandmothers of individuals with type 1 DM compared to
The estimated risk for DZ twin pregnancies relative to MZ the paternal side. They suggested that GDM, which may rep-
pregnancies was 8.6 (95% CI 3.5–21.0). The impact of fetal resent a risk factor for diabetes transmission on the mater-
reduction on the incidence of GDM may support this theory. nal side, is often followed by “extragenerational” type 2 DM
Sivan et al.33 examined 188 consecutive triplet pregnancies at a later age. Like Harder et al., 39 these authors suggested
of which 85 were reduced to twins. The rate of GDM was that their findings were consistent with the suspected tera-
significantly higher in the triplet group than in the reduction togenic effect of GDM on diabetes susceptibility in the
group (22.3% vs. 5.8%). offspring and that this was preventable by avoiding hyper-
Similar results were reported by Schwarz et al.34 in a glycemia in pregnant women and hyperinsulinism in fetuses.
study of 29,644 deliveries. They found that GDM was sig- Histocompatibility leukocyte antigen (HLA) studies are
nificantly more frequent in the 429 twin deliveries (7.7% one way to establish a genetic linkage in certain diseases. In
vs. 4.9%, p < 0.05). However, insulin requirements were not GDM, conflicting results have been reported. Kuhl41 described
Early gestational diabetes mellitus diagnosis as a risk factor 73
similar frequencies of HLA DR2, DR3, and DR4 antigens following variables were present in the index pregnancy:
in healthy pregnant women and women with GDM and parity ≥1 (OR = 3.6), BMI ≥ 30 (OR = 3.6), GDM diagnosed
low prevalences of markers of autoimmune destruction of ≤24 weeks gestation (OR = 20.4), and insulin requirement
the beta cells in GDM pregnancies. Likewise, Vambergue (OR = 2.3). A weight gain ≥7 kg (c. 15 lb) (OR = 2.9) and an
et al.,42 in a study of 95 women with GDM, 95 with IGT, and interval between pregnancies of ≤24 months (OR = 1.6) were
95 control pregnant women, found no significant difference also associated with a recurrence of GDM. Spong et al.50
in the distribution of HLA class II polymorphism among the found a similarly high recurrence rate of 68% in 164 women
groups. However, the GDM and IGT groups presented some with GDM. Risk factors for recurrence in this study were
particular HLA patterns, pointing to a genetic heterogeneity earlier diagnosis of GDM, insulin requirement, and hospital
of glucose intolerance in pregnancy. admissions in the index pregnancy.
Lapolla et al.43 evaluated 68 women with GDM and
matched controls for the frequency of HLA A, B, C, and DR
antigens; the only significant differences were an increase Abnormal glucose tolerance test as
in Cw7 and a decrease in A10 in the GDM group. Budowle
et al.44 reported that the Bf-F allele was found significantly a risk factor for adverse pregnancy
less frequently in nonobese black women with GDM com- outcome
pared to controls and suggested similar genetic associa-
tions in nonobese black women with GDM and type 2 DM. The relationships between maternal hyperglycemia mea-
Similarly, in another study, women with GDM who required sured during a 75 g OGTT administered early in the third
insulin for glycemic control had a lower frequency of the trimester of pregnancy and adverse pregnancy outcome
Bf-F phenotype and a higher frequency of the Bf-f1 pheno- have been clearly demonstrated by the HAPO study.9 The
type; they also had a lower frequency of the type 2 allele at adjusted ORs were calculated for adverse pregnancy out-
the polymorphic locus adjacent to the insulin gene.45 comes associated with an increase in the fasting plasma
Freinkel et al.46 evaluated 199 women with GDM and glucose level of 1 SD (6.9 mg/dL [0.4 mmol/L]), an increase
148 patients with normal pregnancies and found that HLA in the 1-hour plasma glucose level of 1 SD (30.9 mg/dL
DR3 and DR4 antigens occurred significantly more often [1.7 mmol/L]), and an increase in the 2-hour plasma glucose
in black women with GDM. Ferber et al.,47 in an analysis level of 1 SD (23.5 mg/dL [1.3 mmol/L]). For birth weight
of 184 women with GDM, did not find an elevation in the above the 90th percentile, the ORs were 1.38 (95% CI, 1.32–
frequency of any HLA II alleles in the GDM patients com- 1.44), 1.46 (1.39–1.53), and 1.38 (1.32–1.44), respectively; for
pared with nondiabetic unrelated subjects. However, the cord blood serum C-peptide level above the 90th percentile,
DR3 allele was noted significantly more frequently in 43 1.55 (95% CI, 1.47–1.64), 1.46 (1.38–1.54), and 1.37 (1.30–
women with islet autoantibodies and in the 24 women who 1.44); for primary cesarean delivery, 1.11 (95% CI, 1.06–1.15),
developed type 1 DM postpartum. The cumulative risk of 1.10 (1.06–1.15), and 1.08 (1.03–1.12); and for neonatal hypo-
developing type 1 DM within 2 years after pregnancy in glycemia, 1.08 (95% CI, 0.98–1.19), 1.13 (1.03–1.26), and 1.10
the women with GDM with DR3 or DR4 was 22% and in (1.00–1.12).
the women without these alleles was 7% (p = 0.02). The risk Likewise there were significant continuous relationships
rose to 50% in the DR3- and DR4-positive women who had for the secondary outcomes, the strongest being preeclamp-
required insulin during pregnancy (p = 0.006). These results sia followed by shoulder dystocia and birth trauma where
indicate that women with GDM who have islet autoantibod- for each 1SD rise the ORs ranged from 1.21 to 1.28 and
ies at delivery or develop type 1 DM postpartum have HLA 1.20 across the three glucose measurements, respectively.
alleles typical of late-onset type 1 DM and that both HLA Premature delivery, the requirement for an increased level of
typing and islet antibodies can predict the development of neonatal care, and treatment for hyperbilirubinemia were all
type 1 DM postpartum. related to the 1- and 2-hour postload levels, but not fasting
glucose. Importantly, these relationships were independent
of maternal BMI, age, family history of diabetes, gestational
Recurrence of gestational diabetes age at OGTT, infant sex, mean arterial pressure at OGTT,
and hospitalization for complications of pregnancy.
mellitus
MacNeill et al.48 conducted a retrospective longitudinal
study of 651 women who had had a diabetic pregnancy and Early gestational diabetes mellitus
at least one other thereafter. They found a 35.6% recurrence diagnosis as a risk factor
rate of GDM. Multivariant regression models showed that
infant birth weight in the index pregnancy and maternal Early onset of GDM is a high-risk factor. Bartha et al.51 found
weight before the subsequent pregnancy were predictive of that among 3986 pregnant women, those with early-onset
recurrent GDM. GDM (n = 65) were more likely to be hypertensive (18.46% vs.
Higher recurrence rates (69% of 78 patients) were reported 5.88%, p = 0.006) and have higher glycemic levels and greater
by Major et al.49 Recurrence was more common when the needs for insulin therapy (33.85% vs. 7.06%, p < 0.001) than
those in whom diabetes developed later (n = 170). All cases Hypertensive disorders
of neonatal hypoglycemia (n = 4) and all perinatal deaths
(n = 3) were in this group. The women with early-onset Preeclampsia and gestational hypertension are apparently
GDM also had an increased risk of postpartum DM, whereas more frequent in women with GDM. As noted earlier, the
those with late-onset GDM had a minimal risk.52 The per- HAPO study demonstrated a strong association between
centages of overt diabetes and abnormal glucose tolerance increasing maternal glucose levels and preeclampsia.
were significantly higher in the early-pregnancy group A large study by Xiong et al.16 detected preeclampsia in
(n = 30) than in the late-pregnancy group (n = 72) (26.7% vs. 2.7% of 2,755 patients with GDM compared with only 1.1%
1.4% and 40% vs. 5.56% respectively). of 108,664 patients with normal pregnancies (adjusted
OR = 1.3, 95% CI 1.20–1.41). Similar results were observed
for gestational hypertension. Likewise Dukler et al. 58 stud-
Congenital malformations ied 380 primiparous women with preeclampsia and 385
primiparous control women for a total of 121,207 and 1,293
Schaefer-Graf et al. 53 in a review of 4180 pregnancies deliveries, respectively. When adjusted for confounding
complicated by GDM (n = 3764) or type 2 DM (n = 416), variables, GDM was strongly associated with the recur-
reported that the congenital anomalies in the offspring rence of preeclampsia in the second pregnancy (OR = 3.72,
affected the same organ systems described in pregnancies 95% CI 1.45–9.53).
complicated by IDDM. The risk of anomalies rose with Further evidence for the relationship between GDM
the increasing hyperglycemia at diagnosis or presenta- and preeclampsia comes from the two RCTs of treatment of
tion for care. However, most other reports had conflict- GDM—the ACHOIS59 and Landon and colleagues.60
ing findings. Bartha et al. 51 failed to find an increase in In the ACHOIS study,59 the incidence of preeclampsia was
major congenital malformations associated with GDM, as reduced from 18% to 12% from the controls to the treatment
did Kalter54 in a comprehensive review of the literature. group (adjusted treatment effect 0.70 [95% CI 0.51–0.95]).
An exception is the recent Swedish Health Registry study In the Landon study,60 the incidence of preeclampsia in
covering over 1.2 million births between 1987 and 1997. 55 the treatment group was 2.5% compared with 5.5% for the
The author identified 3864 infants born to women with controls (RR 0.46, 97% CI 0.22–0.97).
preexisting diabetes and 8688 infants born to women Go et al.,61 in an 11-year follow-up study of a cross-
with GDM. The total malformation rate in the first group sectional sample of African-American women with a his-
was 9.5% and in the second group 5.7%, similar to the tory of GDM (n = 289), reported one of the highest rates of
rate in the general population. However, the GDM group microalbuminuria (MA) of all ethnic groups. The presence
was characterized by an excess of certain malformations, of MA was not associated with insulin resistance, but it was
suggesting that a subgroup of GDM are at increased risk significantly and independently associated with glycosyl-
of diabetic embryopathy, perhaps due to preexisting but ated hemoglobin (HbA1c) levels and hypertension. Hence,
undetected type 2 DM. hypertension and glucose tolerance influence MA through
Martinez-Frias et al.56 analyzed 19,577 consecutive different mechanisms, and screening for MA should be con-
infants with malformations of unknown cause and com- sidered in this patient population.
pared those born to mothers with GDM with those of Conditions associated with increased insulin resis-
nondiabetic mothers. Their findings indicated that GDM tance, such as GDM, PCOS, and obesity, may predispose
is a significant risk factor for holoprosencephaly, upper/ patients to essential hypertension, hypertensive pregnancy,
lower spine/rib anomalies, and renal and urinary system hyperinsulinemia, hyperlipidemia, and high levels of plas-
anomalies. However, owing to the heterogeneous nature of minogen activator inhibitor-1, leptin, and tumor necrosis
GDM, which includes previously unrecognized and newly factor-alpha. These findings may also be associated with
diagnosed type 2 DM, they could not rule out the possibil- cardiovascular complications in these women.62 Joffe et al.63
ity that the teratogenic effect is related to latent type 2 DM. provided further support for the role of insulin resistance in
Nevertheless, they concluded that pregnancies compli- the pathogenesis of hypertensive disorders of pregnancy. In
cated by GDM should be considered at risk for congenital a prospective study of 4589 healthy nulliparous women, they
anomalies. found that the women with GDM had an increased RR of
By contrast, the relationship between GDM and the preeclampsia and all hypertensive disorders (RR = 1.67, 95%
development of congenital malformations was examined CI 0.92–3.05 and RR = 1.54, 95% CI 1.28–2.11 respectively).
in another population-based retrospective study using RR were not substantially reduced after further adjust-
birth certificate data for all live-born children delivered ment for race and BMI (OR = 1.41 and 1.48, respectively).
between 1984 and 1991 in Washington State.57 The preva- Furthermore, even within the normal range, multivariant
lence of congenital malformations was 7.2%, 2.8%, and 2.1% analysis demonstrated that the level of plasma glucose 1 hour
among offspring of mothers with established diabetes (n = after a 50 g oral glucose challenge was an important predic-
8869), with GDM (n = 1511) and without diabetes (n = 8934), tor of preeclampsia.
respectively. That is, the rate of congenital malformations in Innes et al.64 evaluated 54 normotensive women who
the GDM group was only slightly higher than in the control developed hypertension in pregnancy and 51 controls
group (OR = 1.3, 95% CI 1.0–1.6). with normotensive pregnancies, matched for parity. Mean
Summary 75
postload glucose levels and the total glucose area under the type 2 DM, 17% IGT), in contrast to a control group in
curve were significantly higher in the cases than the controls which none of the women had diabetes and 5.3% had IGT.
and were positively correlated with peak mean arterial pres- The independent risk factors for later development of dia-
sure. After adjustment for potential confounders, 2-hour betes were fasting glucose levels at diagnosis of GDM, deliv-
postload glucose levels remained strongly related to the risk ery >3 weeks before term and abnormal OGTT 2 months
for hypertension and to peak mean arterial pressure, as did postpartum. Low insulin secretion at diagnosis of GDM
the total area under the curve. The cases were also more was also an independent risk factor. Even the nonobese glu-
likely to have had one abnormal OGTT. Stratifying analy- cose-tolerant women with previous GDM had a metabolic
ses by case severity (preeclampsia and gestational hyperten- profile of type 2 DM, i.e., insulin resistance and impaired
sion) yielded similar results. Among all subjects, more cases insulin secretion. Thus, the first OGTT should probably be
than controls were also diagnosed with GDM (31% vs. 12%, performed 2 months postpartum to identify the women who
p = 0.008). are already diabetic and the women who are at highest risk of
later development of overt diabetes.69 Interestingly, accord-
ing to Fletcher et al.’s study, both women with a history of
Risk of type 2 diabetes mellitus GDM and their children are at risk of progressing to type 2
DM.70 Whether this effect is due to a genetic or an in utero
Women with GDM have a 17%–63% risk of type 2 DM influence has yet to be determined.
within 5–6 years.65 However, the risk varies according Overall a systematic review and meta-analysis of 20 stud-
to different parameters. For example, Greenberg et al.66 ies that included 675,455 women and 10,859 type 2 DM
reported that the most significant predictor of 6 weeks events concluded that the RR for women who had GDM of
postpartum diabetes was insulin requirement, with RR = developing subsequent type 2 DM compared with women
17.28 (95% CI 2.46–134.42) followed by poor glycemic who had a normoglycemic pregnancy was 7.43, 95% CI of
control, glucose tolerance (GT) and glucose challenge test 4.79–11.51.71
(GCT) ≥ 200 mg/dL. All of these factors probably repre-
sent the magnitude of the insulin resistance, which is the
hallmark of future diabetes and of other vascular complica- Summary
tions. Similarly, Bian et al.67 reported a diagnosis of diabe-
tes 1–5 years postpartum in 33.3% of patients with previous The IDF estimate that in 2013 diabetes affected 21.4 million
GDM (n = 45), but only 9.7% (n = 31) of those with IGT live births, which is 16.8%, of which 16% were due to gesta-
and 2.6% (n = 39) of normal controls. Two or more abnor- tional diabetes. This is set in the context of the prediction
mal OGTT values during pregnancy, a blood glucose level that the total number of cases of diabetes will rise by 55%
exceeding the maximal values at 1 and 2 hours after oral from 382 million in 2013 to 592 million in 2035.4 This should
glucose loading, and high pregnancy BMI were all useful alert physicians to the need to direct special attention to
predictors of diabetes in later life. this population, especially in developing countries.
To determine if recurrent episodes of insulin resistance The data presented in this chapter indicate that the epide-
(i.e., another pregnancy) contribute to the decline in beta- miology of GDM is characterized by several features:
cell function that leads to type 2 DM in high-risk indi-
viduals, Peters et al.68 investigated 777 Latino women with ●● Differences in screening programs and diagnostic cri-
a history of GDM. Among the 87 (13%) who completed an teria have made it difficult to compare frequencies of
additional pregnancy, the rate ration of type 2 DM increased GDM among various populations. Nevertheless, race
to 3.34 (95% CI 1.80–6.19), compared with women without has been proven to be an independent risk factor of
an additional pregnancy, after adjustment for other potential GDM, which varies in prevalence in direct proportion
diabetes risk factors during the index pregnancy (antepar- to the prevalence of type 2 DM in a given population or
tum oral glucose tolerance, high fasting glucose, gestational ethnic group.
age at diagnosis of GDM) and during follow-up (postpartum ●● Identifiable predisposing factors for GDM include high
BMI, glucose tolerance, weight change, breast-feeding, and maternal age, multiparity, obesity, previous delivery of a
months of contraceptive use). Weight gain was also inde- macrosomic infant, and a family history of diabetes.
pendently associated with an increased risk of type 2 DM; ●● In the absence of risk factors, the incidence of GDM is low.
the rate ration was 1.95 (95% CI 1.63–2.33) for each 4.5 kg Therefore, some authors suggest that selective screening
gained during follow-up after adjustment for the additional may be cost-effective, especially in view of the forecasted
pregnancy and other potential risk factors. These data show rise in the burden of GDM.
that a single pregnancy, independent of the well-known ●● PCOS is an important risk factor for GDM, with special
effect of weight gain, accelerates the development of type 2 similarity in the existence of insulin resistance.
DM in women with a high prevalence of pancreatic beta-cell ●● The genetic diathesis is not well understood.
dysfunction. ●● The recurrence rate of GDM (35%–80%) is influ-
What about milder, diet-controlled GDM? Damm69 enced by parity, BMI, early diagnosis of GDM, insulin
reported abnormal glucose tolerance in 34.4% of 241 women requirement, weight gain, and the interval between
2–11 years after a diabetic pregnancy (3.7% IDDM, 13.7% pregnancies.
●● GDM carries a significantly increased risk of an adverse considered in all patients with early diagnosis of GDM,
pregnancy outcomes relative to women with normal accompanied by appropriate patient counseling.
glucose tolerance. ●● Hypertensive disorders in pregnancy and afterward is
●● Women with an early diagnosis of GDM represent a high- more prevalent in women with GDM. One possible mech-
risk subgroup, with an increased incidence of obstetric com- anism is insulin resistance.
plication, recurrent GDM, and development of type 2 DM. ●● Women with GDM are at increased risk of developing
●● Another subgroup of GDM is characterized by an type 2 DM, especially obese patients, those who were
increased risk of diabetic embryopathy, perhaps due to diagnosed before 24 weeks gestation, and those who
preexisting but undetected type 2 DM. This should be required insulin for glycemic control.
REFERENCES
1. American College of Obstetricians and Gynecologists. 19. Lichtenstein AH, Swab US. Relationship of dietary fat to glucose
Gestational diabetes. Practice Bulletin No. 30, September 2001. metabolism. Atherosclerosis 2000; 150: 227–243.
2. World Health Organization. Diagnostic Criteria for the 20. Clarke SD. Polyunsaturated fatty acid regulation of gene tran-
Classification of Hyperglycemia First Diagnosed in Pregnancy, scription: A mechanism to improve energy balance and insulin
Geneva, Switzerland, 2013. resistance. Br J Nutr 2000; 83(Suppl. 1): s59–s66.
3. King H. Epidemiology of glucose tolerance and gestational diabe- 21. Clarke SD. Polyunsaturated fatty acid regulation of gene tran-
tes in women of childbearing age. Diabetes Care 1998; 21(Suppl. scription: A molecular mechanism to improve the metabolic syn-
2): B9–B13. drome. J Nutr 2001; 131: 1129–1132.
4. International Diabetes Federation. IDF Diabetes Atlas, 6th edn., 22. Rustan AC, Nenseter MS, Drevon CA. Omega-3 and omega-6
IDF: Brussels, Belgium, 2013. fatty acids in the insulin resistance syndrome. Lipid and lipopro-
5. Coustan DR. Gestational diabetes. Diabetes in America. In: tein metabolism and atherosclerosis. Ann N Y Acad Sci 1997;
National Institutes of Diabetes and Digestive and Kidney 827: 310–326.
Diseases, 2nd ed., NIDDK: Bethesda, MD, 1995. 23. Wijendran V, Bendel RB, Couch SC et al. Fetal erythrocyte phos-
6. WHO Ad Hoc Diabetes Reporting Group. Diabetes and impaired pholipid polyunsaturated fatty acids are altered in pregnancy
glucose tolerance in women aged 20–39 years. World Health Stat complicated with gestational diabetes mellitus. Lipids 2000; 35:
1992; 45: 321–327. 927–931.
7. Dooley SL, Metzger BE, Cho NH. Influence of race on disease 24. Bo S, Menato G, Lezo A et al. Dietary fat and gestational hyper-
prevalence and perinatal outcome in a US population. Diabetes glycaemia. Diabetologia 2001; 44: 972–978.
1991; 40: 25–29. 25. Lao TT, Ho LF. Alpha-thalassaemia trait and gestational diabetes
8. Gunton JE, Hitchman R, McElduff A. Effects of ethnicity on glu- mellitus in Hong Kong. Diabetalogia 2001; 44: 966–971.
cose tolerance, insulin resistance and beta cell function in 223 26. Franks S, Gilling-Smith C, Watson H. Insulin action in the normal
women with abnormal glucose challenge test during pregnancy. and polycystic ovary. Metab Clin North Am 1999; 28: 361–378.
Aust N Z J Obstet Gynaecol 2001; 41: 182–186. 27. Holte J, Gennarelli G, Wide L et al. High prevalence of poly-
9. The HAPO Study Cooperative Research Group. Hyperglycemia cystic ovaries and associated clinical, endocrine and metabolic
and adverse pregnancy outcomes. N Engl J Med 2008; 358: features in women with previous gestational diabetes mellitus. J
1991–2002. Clin Endocrinol Metab 1998; 83: 1143–1150.
10. Sacks DA, Hadden DR, Maresh M et al. The Hyperglycemia and 28. Kousta E, Cela E, Lawrence N et al. The prevalence of polycys-
Adverse pregnancy Outcome (HAPO) Study: Frequency of gesta- tic ovaries in women with a history of gestational diabetes. Clin
tional diabetes mellitus (GDM) at collaborating centers based on Endocrinol (Oxf) 2000; 53: 501–507.
the IADPSG Consensus Panel recommended criteria. Diabetes 29. Anntila L, Karjala K, Penttila RA et al. Polycystic ovaries in women
Care 2012; 35: 526–528. with gestational diabetes. Obstet Gynecol 1998; 92: 13–16.
11. Jang HC, Cho NH, Jung KB et al. Screening for gestational diabe- 30. Mikola M, Hiilesmaa V, Halttunen M et al. Obstetric outcome in
tes mellitus in Korea. Int J Gynecol Obstet 1995; 51: 115–122. women with polycystic ovarian syndrome. Hum Reprod 2001;
12. Jimenez-Moleon JJ, Bueno-Calvanillas A, Luna-del-Castillo JD 16: 226–229.
et al. Predictive value of a screen for gestational diabetes mellitus: 31. Koivunen RM, Juutinen J, Vauhkonen I et al. Metabolic and ste-
Influence of associated risk factors. Acta Obstet Gynecol Scand roidogenic alterations related to increased frequency of polycys-
2000; 79: 991–998. tic ovaries in women with a history of gestational diabetes. J Clin
13. Jang HC, Min HK, Lee HK et al. Short stature in Korean women: Endocrinol Metab 2001; 86: 2591–2599.
A contribution to the multifactorial predisposition to gestational 32. Hoskins RE. Zygosity as a risk factor for complications and out-
diabetes mellitus. Diabetalogia 1998; 41: 778–783. comes of twin pregnancy. Acta Genet Med Gemellol 1995; 44:
14. Kousta E, Lawrence NJ, Penny et al. Women with a history of ges- 11–23.
tational diabetes of European and South Asian origin are shorter 33. Sivan E, Maman E, Homko CJ et al. Impact of fetal reduction on the
than women with normal glucose tolerance in pregnancy. Diabet incidence of gestational diabetes. Obstet Gynecol 2002; 99: 91–94.
Med 2000; 17: 792–797. 34. Schwartz DB, Daoud Y, Zazula P et al. Gestational diabetes mel-
15. Branchtein L, Schmidt MI, Matos MC et al. Short stature and ges- litus: Metabolic and blood glucose parameters in singleton versus
tational diabetes in Brazil. Brazilian Gestational Diabetes Study twin pregnancies. Am J Obstet Gynecol 1999; 181: 912–914.
Group. Diabetalogia 2000; 43: 848–851. 35. Wein P, Warwick MM, Beischer NA. Gestational diabetes in twin
16. Xiong X, Saunders LD, Wang FL et al. Gestational diabetes mel- pregnancy: Prevalence and long-term implications. Aust N Z J
litus: Prevalence, risk factors, maternal and infant outcomes. Int J Obstet Gynaecol 1992; 32: 325–327.
Gynaecol Obstet 2001; 75: 221–228. 36. Egeland GM, Irgens LM. Is a multiple birth pregnancy a risk fac-
17. Egeland GM, Skjærven R, Irgens LM. Birth characteristics of tor for gestational diabetes? Am J Obstet Gynecol 2001; 185:
women who develop gestational diabetes: Population based 1275–1276.
study. Br Med J 2000; 321: 546–547. 37. Fitzsimmons BP, Bebbington MW, Fluker MR. Perinatal and neo-
18. Ma RM, Lao TT. Maternal mean arterial pressure and oral glu- natal outcomes in multiple gestations: Assisted reproduction ver-
cose tolerance test results. Relationship in normotensive women. sus spontaneous conception. Am J Obstet Gynecol 1998; 179:
J Reprod Med 2001; 46: 747–751. 1162–1167.
References 77
38. Henderson CE, Scarpelli S, Larosa D et al. Assessing the risk of 54. Kalter H. The non-teratogenicity of gestational diabetes. Paediatr
gestational diabetes in twin pregnancy. J Natl Med Assoc 1995; Perinat Epidemiol 1998; 12: 456–458.
87: 757–758. 55. Aberg A, Westbom L, Kallen B. Congenital malformations among
39. Harder T, Franke K, Kohlhoff R et al. Maternal and paternal family infants whose mothers had gestational diabetes or preexisting
history of diabetes in women with gestational diabetes or insu- diabetes. Early Human Dev 2001; 61: 85–95.
lin-dependent mellitus type 1. Gynecol Obstet Invest 2001; 51: 56. Martinez-Frias ML, Bermejo E, Rodriguez-Pinilla E et al.
160–164. Epidemiological analysis of outcomes of pregnancy in gestational
40. Dorner G, Plagemann A, Reinagel H. Familial diabetes aggrega- diabetic mothers. Am J Med Genet 1998; 78: 140–145.
tion in type 2 diabetics: Gestational diabetes an apparent risk 57. Janssen PA, Rothman I, Schwartz SM. Congenital malformations
factor for increased diabetes susceptibility in the offspring. Exp in newborns of women with established and gestational diabetes
Clin Endocrinol 1987; 89: 84–90. in Washington State, 1984–91. Paediatr Perinat Epidemiol 1996;
41. Kuhl C. Etiology and pathogenesis of gestational diabetes. 10: 52–63.
Diabetes Care 1998; 21(Suppl. 2): B19–B26. 58. Dukler D, Porath A, Bashiri A et al. Remote prognosis of primipa-
42. Vambergue A, Fajardy I, Bianchi F et al. Gestational diabetes mel- rous women with preeclampsia. Eur J Obstet Gynecol Reprod
litus and HLA class II (-DQ, -DR) associations. The Digest Study. Biol 2001; 96: 69–74.
Eur J Immunogenet 1997; 24: 385–394. 59. Crowther CA, Hiller JE, Moss JR et al. Effect of treatment of ges-
43. Lapolla A, Betterle C, Sanzari M et al. An immunological and tational diabetes mellitus on pregnancy outcome. N Engl J Med
genetic study of patients with gestational diabetes mellitus. Acta 2005; 352: 2477–2486.
Diabetol 1996; 33: 139–144. 60. Landon MB, Spong CY, Thom E et al. A multicentre, random-
44. Budowle B, Huddleston JF, Go RC et al. Association of HLA- ized trial of treatment for mild gestational diabetes. N Engl J Med
linked factor B with gestational diabetes mellitus in black women. 2009; 361: 1339–1348.
Am J Obstet Gynecol 1988; 159: 805–806. 61. Go RC, Desmond R, Roseman JM et al. Prevalence and risk factors
45. Bell DS, Barger BO, Go RC et al. Risk factors for gestational dia- of microscopic albuminuria in a cohort of African-American women
betes in black population. Diabetes 1990; 13: 1196–1201. with gestational diabetes. Diabetes Care 2001; 24: 1764–1769.
46. Freinel N, Metzger BE, Phelps RL et al. Gestational diabetes mel- 62. Solomon CG, Seely EW. Brief review: Hypertension in pregnancy:
litus: A syndrome with phenotypic and genotypic heterogeneity. A manifestation of the insulin resistance syndrome? Hypertension
Horm Metab Res 1986; 18: 427–430. 2001; 37: 232–239.
47. Ferber KM, Keller E, Albert ED et al. Predictive value of human 63. Joffe GM, Esterlitz JR, Levine RJ et al. The relationship between
leukocyte antigen class II typing for the development of islet cell abnormal glucose tolerance and hypertensive disorders of preg-
autoantibodies and insulin-dependent diabetes postpartum in nancy in healthy nulliparous women. Calcium for Preeclampsia
women with gestational diabetes. J Clin Endocrinol Metab 1999; Prevention (CPEP) Study Group. Am J Obstet Gynecol 1998; 179:
84: 2342–2348. 1032–1037.
48. MacNeill S, Dodds L, Hamilton DC et al. Rates and risk factors 64. Innes KE, Wimsatt JH, McDuffie R. Relative glucose tolerance
for recurrence of gestational diabetes. Diabetes Care 2001; 24: and subsequent development of hypertension in pregnancy.
659–662. Obstet Gynecol 2001; 97: 905–910.
49. Major CA, deVeciana M, Weeks J et al. Recurrence of gesta- 65. Kjos SL, Buchanan TA. Gestational diabetes mellitus. N Engl J
tional diabetes: Who is at risk? Am J Obstet Gynecol 1998; 179: Med 1999; 341: 1749–1756.
1038–1042. 66. Greenberg LR, Moore TR, Murphy H. Gestational diabetes mel-
50. Spong CY, Guillermo L, Kuboshige J et al. Recurrence of ges- litus: Antenatal variables as predictors of postpartum glucose
tational diabetes mellitus: Identification of risk factors. Am J intolerance. Obstet Gynecol 1995; 86: 96–101.
Perinatol 1998; 15: 29–33. 67. Bian X, Gao P, Xiong X et al. Risk factors for development of
51. Bartha JL, Martinez-Del-Fresno P, Comino-Delgrado R. Gesta diabetes mellitus in women with a history of gestational diabetes
tional diabetes mellitus diagnosed during early pregnancy. Am J mellitus. Chin Med J (Engl) 2000; 113: 759–762.
Obstet Gynecol 2000; 182: 346–350. 68. Peters RK, Kjos SL, Xiang A et al. Long-term diabetogenic effect
52. Bartha JL, Martinez-Del-Fresno P, Comino-Delgrado R. Post of single pregnancy in women with previous gestational diabetes
partum metabolism and autoantibody markers in women with mellitus. Lancet 1996; 347: 227–230.
gestational diabetes diagnosed in early pregnancy. Am J Obstet 69. Damm P. Gestational diabetes and subsequent development of
Gynecol 2001; 184: 965–970. overt diabetes mellitus. Dan Med Bull 1998; 45: 495–509.
53. Schaefer-Graf UM, Buchanan TA, Songster G et al. Patterns 70. Fletcher B, Gulanick M, Lamendola C. Risk factors for type 2
of congenital anomalies and relationship to initial maternal diabetes mellitus. J Cardiovasc Nurs 2002; 16: 17–23.
fasting glucose levels in pregnancies complicated by type 2 71. Bellamy L, Casa J-P, Hingorinin AD et al. Type 2 diabetes mellitus
and gestational diabetes. Am J Obstet Gynecol 2000; 182: after gestational diabetes: A systematic review and meta-analysis.
313–320. Lancet 2009; 373: 1773–1779.
9 Genetics of diabetic pregnancy
Komal Bajaj and Susan J. Gross
Introduction The foundation: Lessons learned in a

Recent discoveries due to advances in molecular technology pregenomic era
have underscored the important role genetic factors play
Prior to widespread use of advanced techniques such as
in the development, treatment, and consequences of the
next-generation sequencing and arrays, the contribution of
diabetic pregnancy. This chapter reviews the genetic etiol-
genetic factors to a disease was often characterized by obser-
ogy of maternal diabetes, factors influencing pregnancy out-
vations made through family history, the study of twins,
come, and what is known of the epigenetic consequences of
analysis of general population databases, and the character-
hyperglycemia in pregnancy.
ization of rare disorders known to run in families.
A patient’s family history may provide clues as to condi-
tions a patient is at risk for in her lifetime. The World Health
Gestational diabetes and relationship Organization and the American College of Obstetrics and
to type 1 and type 2 diabetes Gynecology both stress the importance of obtaining a thor-
ough family history as part of a patient encounter.14,15 Family
Insulin resistance is a normal physiological alteration that history alone serves as a predictor of GDM.16 When com-
occurs during pregnancy. Gestational diabetes mellitus (GDM) pared to individuals without any parental history of diabe-
may develop when this normal resistance is coupled with tes, those that reported even one parent with diabetes had a
pancreatic beta-cell insufficiency. Women diagnosed with 2.3-fold risk of developing GDM. Furthermore, women with
GDM are at much higher risk of developing type 2 diabetes a diabetic sibling had an 8.4-fold increased risk of GDM.
mellitus (T2DM) after pregnancy, ranging between 17% and Individuals who come from Hispanic, African, Native
63% within 15 years, depending upon the population studied.1 American, Asian, and Pacific Islander background are at
The connection between GDM and significantly increased risk increased risk for GDM.17 The prevalence of GDM in specific
of T2DM suggests that pregnancy serves as a stressor that often populations has been reported to be as high as 22%, whereas
catalyzes progression to diabetes in predisposed individuals.2 the prevalence in populations not considered to be high risk
GDM and T2DM share several common risk factors and is between 2% and 5%.18 Such increased prevalence suggests
underlying pathophysiology.3 Both conditions are associ- a greater hereditary susceptibility in those populations.
ated with an elevated body mass index (BMI) and history
of abnormal glucose tolerance. The hallmark of both condi-
tions is a peripheral insulin resistance coupled with a rela- Monogenic diabetes
tive insufficiency in pancreatic beta-cell insulin production. While GDM is a multifactorial disorder, with a polygenic
There are striking parallels between these two diseases, sug- heritable and environmental component, there are mono-
gesting that there is considerable overlap between the genetic genic forms of diabetes that comprise a small but important
contributors to GDM and T2DM. Both are in fact strongly cohort of patients with diabetic pregnancy.
influenced by genetic and environmental factors. Table 9.1
highlights the relative role genetic factors and obesity play in
early-onset T2DM, GDM, and late-onset T2DM. Maturity onset diabetes of the young
Unlike GDM and T2DM, type 1 diabetes mellitus The term “maturity-onset diabetes of the young (MODY)”
(T1DM) is characterized by autoimmune destruction of beta dates back to the 1960s when diabetes was classified into
cells. T1DM is typically diagnosed well before reproduction.4 two forms—“juvenile onset” and “maturity onset.” MODY
Table 9.2 outlines genes in which alterations are associated is characterized by insulin-dependent diabetes diagnosed
with susceptibility to T1DM, the majority of which relate to prior to age 25 without the presence of autoantibodies.19
immunologic function. The remainder of this chapter will MODY accounts for up to 5% of diabetes and is inherited
focus on heritable components of GDM, and, since they in an autosomal dominant fashion. The clinical features of
share similar pathophysiology, T2DM. MODY are heterogeneous and in and of themselves may not
78
Monogenic diabetes 79
Table 9.1 Comparison of the relative role of genetic factors and obesity in young-onset type 2 diabetes,
gestational diabetes, and late-onset type 2 diabetes
Early-onset type 2 diabetes Gestational diabetes Late-onset type 2 diabetes

Age of diagnosis (years) 25–45 18–40 >45
Parental history of type 2 diabetes +++ ++ +
High prevalence racial origin +++ ++ +
Obesity +++ +++ +
Source: Forbes, M. and Hattersley, A.T., The genetics of diabetic pregnancy, in: Textbook of Diabetes and Pregnancy, Hod M, Jovanovic LG,
and Di Renzo GC (Eds.), 2007, pp. 466–474. With permission.
Table 9.2 Genes associated with type 1 diabetes
Gene(s) Location
HLA class I and HLA class II genes 6p21.3
Insulin receptor (INS) 19p13.3
Cytotoxic T-lymphocyte-associated protein 4 (CTLA4) 2q33.2
Interleukin 2 receptor α (IL2Rα) 10p15.1
Interferon induced with helicase C domain 1 (IFIH1) 2q24.2
C-type lectin domain family 16 (CLEC16A) 16p13.13
Th1transcription factor (STAT4) 2q32.2-q32.3
Sources: Stankov, K. et al., Pediatrics, 132(6), 1112, 2013; C-type lectin domain family 16, available from: http://omim.org/entry/611303,
accessed on April 29, 2014; INSR insulin receptor, available from: http://www.ncbi.nlm.nih.gov/gene/3643, accessed on April
29, 2014; Cytotoxic T lymphocyte-associated 4, available from: http://omim.org/entry/123890, accessed on April 29, 2014;
Interleukin 2 receptor, alpha, available from: http://omim.org/entry/147730, accessed on April 29, 2014; Boucas, A.P. et al.,
Arq. Bras. Endocrinol. Metabol., 57(9), 667, 2013; Knudsen, N.H. and Lee, C.H., Diabetes, 62(12), 4002, 2013; HGNC,
Histocompatibility complex, available from: http://www.genenames.org/genefamily/hla.php, accessed on April 29, 2014.
raise suspicion for a monogenic syndrome unless a family GCK mutation. In fact, a 700 g birth weight difference was seen
history is clearly suggestive of an autosomal dominant pat- between infants that did not carry the mutation and those that
tern. This is further complicated by the fact that all types inherited the mutation from their mother.28,29 There are cur-
of MODY exhibit variable penetrance—not all individuals rently no standardized guidelines as to how to manage women
carrying a particular gene change will develop MODY.20 with GCK mutations during pregnancy to minimize the risk of
Patients with MODY are typically diagnosed by age 25 and maternal and fetal complications.
are often misclassified as having either T1DM or T2DM.21 While it is impractical to consider testing every woman
A clearly distinguishing feature is the lack of autoantibod- with GDM for GCK mutations, narrow fasting glucose, BMI
ies. Patients with MODY are categorized into different types levels, and family history have been suggested as parameters
depending on the gene mutation. Table 9.3 highlights the that can be utilized to select patients with GDM for genetic
gene defects associated with the various types of MODY. testing.27,30 For example, using a BMI of <25 kg/m2 and fasting
Glucokinase (GCK) is an enzyme present in the β-cells of glucose of ≥5.5 mmol/L during pregnancy as screening crite-
the pancreas that helps the cells, through phosphorylating glu- ria, 2.7 women with GDM would need to be tested to find one
cose to glucose-6-phosphate, accurately detect circulating glu- case of GCK-MODY and 66% of all cases would be detected.27
cose and adjust insulin secretion accordingly. A loss of function
mutation in the GCK gene impacts the ability to detect glucose
at lower levels, thereby causing a rise in the homeostatic set Mitochondrial mutations
point of blood glucose in that individual. Glucokinase mono- The mitochondrial tRNA leucine 3243 mutation (3243
genic diabetes (GCK-MODY) has been estimated to have a mtDNA) results in a rare form of maternally inherited
prevalence of 1.1 in 1000 primarily European population and diabetes and deafness. There is variable expressivity of the
that the prevalence in women with GDM is at least 0.9%.27 condition, though all cases have some level of aberrant insu-
Fifty percent of the fetuses of women with GCK mutations will lin secretion and sensorineural deafness. The mean age of
inherit the mutation. It is expected that an unaffected fetus of a onset of symptoms is between 30 and 40 years of age. A pro-
mother with a GCK mutation would produce excess insulin in posed mechanism of disease is that diabetes develops due
response to the increased circulating maternal blood glucose, to an inappropriate storage of triacylglycerols, resulting in
causing an increase in birth weight compared to a fetus with the deterioration of pancreatic β-cell function.31 The mutation
Table 9.3 Types of maturity-onset diabetes of the young
Type Gene defect Physiologic defect MODY cases

MODY 1 Hepatocyte nuclear factor-4-α Loss of function mutation results in decreased insulin <10%
secretion in response to glucose
MODY 2 Glucokinase gene Several different mutations result in faulty glucose detection 15%–30%
MODY 3 Hepatocyte nuclear factor-1-α Aberrant insulin secretion and altered renal excretion of 50%–60%
glucose
MODY 4 Insulin promoter factor 1 (PDX1) Decreased activation of the insulin gene in response to Rare
hyperglycemia
MODY 5 Hepatocyte nuclear factor-1-β Atrophy of the pancreas 5%–10%
MODY 6 Neurogenic differentiation factor-1 Pancreatic development Rare
MODY 7 Kruppel-like factor 11 Transcription factor Rare
MODY 8 Bile salt–dependent lipase (CEL) Pancreatic function Rare
MODY 9 PAX4 Differentiation of endocrine pancreas Rare
MODY 10 Insulin gene (INS) Aberrant insulin production Rare
MODY 11 B-lymphocyte tyrosine kinase (BLK) Modification of insulin transcription factors Rare
Sources: Naylor, R. and Philipson, L.H., Clin. Endocrinol., 75(4), 422, 2011; Shields, B.M. et al., Diabetologia, 53(12), 2504, 2010;
Thanabalasingham, G. and Owen, K.R., Br. Med. J., 343, d6044, 2011; Maturity-onset diabetes of the young, Type 7, available from:
http://omim.org/entry/610508, accessed on April 23, 2014; Maturity-onset diabetes of the young, Type 8, available from: http://omim.
org/entry/609812, accessed on April 23, 2014; Maturity-onset diabetes of the young, Type 10, available from: http://www.genetests.org/
disorders/?disid=315205, accessed on April 23, 2014; Tyrosine kinase, B-lymphocyte specific, BLK, available from: http://omim.org/
entry/191305, accessed on April 24, 2014.
has been well described in several different populations.32,33 beings and how they are distributed among diverse popula-
Since the 3243mtDNA mutation leads to an insulin secretions.34 Both these important initiatives, along with advances
tion defect, not a defect in insulin sensitivity, these patients in molecular technologies such as arrays and next-generation
initially do respond to medications that increase secretion sequencing, provided the foundation for investigators to be
such as glyburide. However, since there is a progressive loss able to evaluate many genes at one time and begin to make
of insulin secretion, the use of insulin is eventually required. meaning of changes found in genotypes of many different
individuals and populations. See further Table 9.4.
Technological advancements and Genetic predisposition to gestational

definition of terms diabetes and type 2 diabetes in a
Completed in 2003, the Human Genome Project produced genomic era
a high-quality sequence of the human genome and provided
details regarding its structure and organization. This was fol- Though genome-wide association studies (GWAS) are not
lowed by the International HapMap Project in 2005, which without their limitations, they have provided new insight
provided a catalog of common genetic variants in human into disease pathways and begun to dissect complex genetic
Table 9.4 Definitions of terms commonly utilized during the discussion of advanced molecular techniques
Term Definition
Epigenetics Any process that alters gene activity without changing the DNA sequence itself and leads to
modifications that can be transmitted to daughter cell.35
Exome Composed of all exons, which are the coding portions of genes.36
Genome The entire genetic information (genes) for an organism, which is maintained in the nucleotide
sequence of DNA; each organism has only one genome.37
Genome-wide An approach that involves rapidly scanning markers across the complete sets of DNA, or genomes, of
association study many people to find genetic variations associated with a particular disease.38
Locus (loci) A specific location of a gene or DNA sequence.
Monogenic A phenotype or disease that is the result of a change within a single gene.
Polygenic The combined action of alleles of more than one gene contribute to a particular phenotype or disease.39
Single-nucleotide Often called SNPs, these are the most common type of genetic variant in any species. Each SNP
polymorphism represents a difference in one nucleotide.40
Impact of the abnormal glycemic environment on fetus 81
Table 9.5 Genes and locations associated with type 2 diabetes mellitus
Gene Location
KCNJ11 11p15.1
HHEX/IDE 10q23.33
SLC30A8 8q24.11
KCNQ1 11p15.5-p15.4
WFS1 4p16.1
Sources: Saxena, R. et al., Science, 316(5829), 1331, 2007; Zeggini, E. et al., Science,
316(5829), 1336, 2007; Steinthorsdottir, V., Nat. Genet., 39(6), 770, 2007; Scott,
L.J. et al., Science, 316(5829), 1341, 2007; Williams, A.L. et al., Nature, 506(7486),
97, 2014; Unoki, H. et al., Nat. Genet., 40(9), 1098, 2008; Sladek, R. et al., Nature,
445(7130), 881, 2007; Hayes, M.G. et al., Diabetes, 62(9), 3282, 2013.
architecture.41 GWAS and other forms of analysis allow for function (Table 9.5). The attributable risk of T2DM associ-
the discovery of variants associated with disease in genes ated with the allelic variants within each of these genes has
not otherwise implicated in the disease process as well as been studied in several diverse populations.
the identification of variants outside coding regions of DNA. Two of these genes, KCNQ1 and WFS1, are also associated
These insights provide important clues to the pathogenesis of with other genetic syndromes. The KCNQ1 gene encodes for
common diseases, including new loci related to diabetes sus- a subunit of a potassium channel (KvLQT1). Mutations in the
ceptibility.42 The following text summarizes some examples KCNQ1 gene have been implicated in long-QT syndrome, a
of genetic susceptibility categorized by suspected gene func- group of disorders involving a characteristic cardiac arrhyth-
tion. While this area of research is constantly growing and mia. SNPs in KCNQ1 have also been shown to confer an
changing, the examples highlight the types of information increased risk of GDM in several different populations.57,58
that is being generated in the genomic era. WFS1 is a gene involved in beta-cell survival. Mutations
in this gene cause Wolfram syndrome, the features of which
Insulin action and release include diabetes, optic atrophy, and deafness.59
Single-nucleotide polymorphisms (SNPs) in the TCF7L2

gene have been associated with a significantly higher risk
Impact of the abnormal glycemic
of T2DM in several different populations.43,44 The TCF7L2 environment on fetus
gene, located on chromosome 10q25.3, is a transcription
factor that acts through regulation of pro-glucagon gene The impact of maternal hyperglycemia and short-term
expression.45 When compared to noncarriers, the relative adverse perinatal outcomes, including fetal macrosomia,
risk of T2DM was 1.45 for heterozygotes of the at-risk allele is well known.60 Over the last decade, it has become clear
and 2.41 for homozygotes of the at-risk allele.44 Individuals that exposure to a diabetic milieu in utero is associated with
with impaired glucose tolerance and the TCF7L2 gene vari- impaired glucose tolerance and defective insulin secretory
ant were more likely to progress to diabetes over 3 years than response in offspring during adulthood, independent of
those without the variant (HR 1.55, CI 1.2–2).46 Analysis in genetic predisposition to diabetes based on family history.61
GDM also suggested an increased risk of GDM in individu- When a fetus is female, in utero exposure increases her risk
als with TCF7L2 variant (OR 1.653, 95% CI 1.416–1.930).47 of developing GDM in her own pregnancies, thereby per-
ABCC8, located on 11p15.1, codes for a receptor on the pan- petuating the transgenerational effect of hyperglycemia and
creatic beta cell involved in insulin secretion. Several variants contributing to the increase in T2DM.62 Furthermore, intra-
within this gene have been associated with an increased risk uterine hyperglycemia is also associated with an increased
of GDM and/or T2DM.48 In the Han Chinese population, for risk of obesity and cardiovascular diseases as an adult.63
example, the C allele of the −3 T/C polymorphism in intron 24 There is a great deal of evidence to suggest that abnormal
and the A allele of the R1273R in exon 31 were strongly associ- nutrition while in utero (both fetal overnutrition and under-
ated with an increased susceptibility to gestational DM.49 nutrition) results in epigenetic changes to genes in important
metabolic pathways. Both animal and human models reveal
altered methylation patterns64,65 after aberrant nutritional sta-
Pancreatic beta-cell development tus in utero. Coined “the fetal origins of adult disease,”66 this
and function concept of early programming has tremendous public health
implications. Because the epigenome is vulnerable during
Some of the loci uncovered by GWAS and other applications early development, encouraging patients to avoid such expo-
of next-generation sequencing related to diabetes suscepti- sures during the preconceptual and pregnant state may have
bility are in genes involved in pancreatic development and tremendous impact on the prevention on adult diseases.63
Future frontiers The “exposome,” like modern “omics,” attempts to

characterize patterns in order to gain insight into disease.
Pharmacogenomics is an area that holds tremendous poten- In this case, the “exposome” includes all the exposures an
tial for impact in the care of GDM in the future. The ability individual encounters from conception onward. Efforts to
to develop individualized regimens based on the patient’s implement assessment tools to measure early-life exposures
genetic makeup may improve treatment outcomes. For and their impact on health outcomes are underway and may
example, glyburide (a sulfonylurea that stimulates the pan- shed light on the development and outcome of the diabetic
creas to release more insulin) has been utilized to success- pregnancy.74
fully manage hyperglycemia during pregnancy.67 Women
who fail to achieve glycemic control on glyburide must be
transitioned to insulin. It has been demonstrated that the Practice points
cytochrome P450 CYP2C9 is involved in the metabolism
of many drugs, including glyburide.68 The CYP2C9 gene ●● Monogenic forms of diabetes comprise a small but impor-
has several polymorphisms, and the degree of blood glucose tant cohort of patients with diabetic pregnancy.
reduction appears to differ depending upon the genotype.69 ●● Advances in molecular technology and understanding
Though further studies are needed, it is quite possible that of the human genome have contributed to the identifica-
in the future testing a patient’s genetic makeup to help guide tion of new loci and pathways associated with increased
treatment regimens may be a regular part of the care of the susceptibility to diabetes.
gestational diabetes. ●● Exposure to a diabetic environment in utero increases the
Inflammatory markers have been associated with the devel- risk of adult-onset chronic diseases such as obesity and
opment of diabetes.70 Abnormal metabolic disturbances at diabetes.
the level of the adipocyte have been observed in women with
GDM.71 Aberrant levels of galectin-1, a carbohydrate binding
protein, in the second and third trimester are associated with Research agenda
GDM.72 Modern “omics” approaches—genomics, metabolo-
mics, proteomics, and transcriptomics—attempt to character- ●● Deconstruct the genetic basis of the complicated meta-
ize patterns of various biological markers and translate those bolic pathways leading to diabetes in pregnancy.
patterns into deeper insights regarding normal function and ●● Further characterize the epigenetic changes related to
disease processes. The creation of a metabolic “fingerprint” exposure to a hyperglycemic environment and their
for GDM, for example, would allow for deciphering biological impact on adult-onset diseases.
processes, which translates into opportunities for prevention, ●● Establish individualized treatment regimens based on
early diagnosis, and treatment of disease.73 genetic variants and metabolic profile.
REFERENCES
1. Lobner K, Knopff A, Baumgarten A, et al. Predictors of post- 12. Knudsen NH, Lee CH. STAT4: An initiator of meta-inflammation
partum diabetes in women with gestational diabetes mellitus. in adipose tissue? Diabetes 2013; 62(12): 4002–4003.
Diabetes 2006; 55(3): 792–797. 13. HGNC. Histocompatibility complex. Available from: http://www.
2. Robitaille J, Grant AM. The genetics of gestational diabetes melli- genenames.org/genefamily/hla.php, accessed on April 29, 2014.
tus: Evidence for relationship with type 2 diabetes mellitus. Genet 14. Genetics and Health. World Health Organization. http://www.
Med 2008; 10(4): 240–250. who.int/genomics/en/, accessed on April 14, 2014.
3. Herman WH, Zimmet P. Type 2 diabetes: An epidemic requiring 15. American College of Obstetricians and Gynecologists Committee
global attention and urgent action. Diabetes Care 2012; 35(5): on Genetics. Committee Opinion No. 478: Family history as a
943–944. risk assessment tool. Obstet Gynecol 2011; 117(3): 747–750.
4. Rosenbloom AL et al. Type 2 diabetes in children and adoles- 16. Williams MA et al. Familial aggregation of type 2 diabetes and
cents. Pediatr Diabetes 2009; 10(Suppl. 12): 17–32. chronic hypertension in women with gestational diabetes mel-
5. Forbes M, Hattersley AT. The genetics of diabetic pregnancy. In: litus. J Reprod Med 2003; 48(12): 955–962.
Textbook of Diabetes and Pregnancy, Hod M, Jovanovic LG, and 17. Practice Bulletin No. 137: Gestational diabetes mellitus. Obstet
Di Renzo GC (Eds.), 2007, pp. 466–474. Gynecol 2013; 122(2 Pt. 1): 406–416.
6. Stankov K, Benc D, Draskovic D. Genetic and epigenetic factors 18. Galtier F. Definition, epidemiology, risk factors. Diabetes Metab
in etiology of diabetes mellitus type 1. Pediatrics 2013; 132(6): 2010; 36(6 Pt. 2): 628–651.
1112–1122. 19. McDonald TJ et al. Islet autoantibodies can discriminate matu-
7. C-type lectin domain family 16. Available from: http://omim.org/ rity-onset diabetes of the young (MODY) from Type 1 diabetes.
entry/611303, accessed on April 29, 2014. Diabet Med 2011; 28(9): 1028–1033.
8. INSR insulin receptor. Available from: http://www.ncbi.nlm.nih. 20. Naylor R, Philipson LH. Who should have genetic testing for
gov/gene/3643, accessed on April 29, 2014. maturity-onset diabetes of the young? Clin Endocrinol (Oxf) 2011;
9. Cytotoxic T lymphocyte-associated 4. Available from: http:// 75(4): 422–426.
omim.org/entry/123890, accessed on April 29, 2014. 21. Shields BM et al. Maturity-onset diabetes of the young (MODY):
10. Interleukin 2 receptor, alpha. Available from: http://omim.org/ How many cases are we missing? Diabetologia 2010; 53(12):
entry/147730, accessed on April 29, 2014. 2504–2508.
11. Boucas AP et al. The role of interferon induced with helicase C 22. Thanabalasingham G, Owen KR. Diagnosis and management of
domain 1 (IFIH1) in the development of type 1 diabetes mellitus. maturity onset diabetes of the young (MODY). Br Med J 2011;
Arq Bras Endocrinol Metabol 2013; 57(9): 667–676. 343: d6044.
References 83
23. Maturity-onset diabetes of the young, Type 7. Available from: 49. Niu XM et al. Study on association between gestational diabe-
http://omim.org/entry/610508, accessed on April 23, 2014. tes mellitus and sulfonylurea receptor-1 gene polymorphism.
24. Maturity-onset diabetes of the young, Type 8. Available from: Zhonghua Fu Chan Ke Za Zhi 2005; 40(3): 159–163.
http://omim.org/entry/609812, accessed on April 23, 2014. 50. Zeggini E et al. Replication of genome-wide association signals
25. Maturity-onset diabetes of the young Type 10. Available from: in UK samples reveals risk loci for type 2 diabetes. Science 2007;
http://www.genetests.org/disorders/?disid=315205, accessed on 316(5829): 1336–1341.
April 23, 2014. 51. Steinthorsdottir V et al. A variant in CDKAL1 influences insulin
26. Tyrosine kinase, B-lymphocyte specific, BLK. Available from: response and risk of type 2 diabetes. Nat Genet 2007; 39(6):
http://omim.org/entry/191305, accessed on April 24, 2014. 770–775.
27. Chakera AJ et al. The 0.1% of the population with glucokinase 52. Scott LJ et al. A genome-wide association study of type 2 diabe-
monogenic diabetes can be recognized by clinical characteris- tes in Finns detects multiple susceptibility variants. Science 2007;
tics in pregnancy: The Atlantic diabetes in pregnancy cohort. 316(5829): 1341–1345.
Diabetes Care 2014; 37: 1230–1236. 53. Williams AL et al. Sequence variants in SLC16A11 are a common
28. Spyer G et al. Pregnancy outcome in patients with raised blood risk factor for type 2 diabetes in Mexico. Nature 2014; 506(7486):
glucose due to a heterozygous glucokinase gene mutation. 97–101.
Diabet Med 2009; 26(1): 14–18. 54. Unoki H et al. SNPs in KCNQ1 are associated with susceptibility
29. Velho G, Hattersley AT, Froguel P. Maternal diabetes alters birth to type 2 diabetes in East Asian and European populations. Nat
weight in glucokinase-deficient (MODY2) kindred but has no Genet 2008; 40(9): 1098–1102.
influence on adult weight, height, insulin secretion or insulin sen- 55. Sladek R et al. A genome-wide association study identifies novel
sitivity. Diabetologia 2000; 43(8): 1060–1063. risk loci for type 2 diabetes. Nature 2007; 445(7130): 881–885.
30. Ellard S et al. A high prevalence of glucokinase mutations in gesta- 56. Hayes MG et al. Identification of HKDC1 and BACE2 as genes
tional diabetic subjects selected by clinical criteria. Diabetologia influencing glycemic traits during pregnancy through genome-
2000; 43(2): 250–253. wide association studies. Diabetes 2013; 62(9): 3282–3291.
31. Maassen JA et al. Mitochondrial diabetes and its lessons for 57. Kwak SH et al. Polymorphisms in KCNQ1 are associated with
common Type 2 diabetes. Biochem Soc Trans 2006; 34(Pt. 5): gestational diabetes in a Korean population. Horm Res Paediatr
819–823. 2010; 74(5): 333–338.
32. Kadowaki T et al. A subtype of diabetes mellitus associated with 58. Zhou Q et al. Association of KCNQ1 gene polymorphism
a mutation of mitochondrial DNA. N Engl J Med 1994; 330(14): with gestational diabetes mellitus in a Chinese population.
962–968. Diabetologia 2009; 52(11): 2466–2468.
33. Martikainen MH, Ronnemaa T, Majamaa K. Prevalence of mito- 59. WFS1-related disorders. Available from: http://www.ncbi.nlm.
chondrial diabetes in southwestern Finland: A molecular epide- nih.gov/books/NBK4144, accessed on April 14, 2014.
miological study. Acta Diabetol 2013; 50(5): 737–741. 60. Metzger BE et al. Hyperglycemia and adverse pregnancy out-
34. International HapMap Project. Available from: http://hapmap. comes. N Engl J Med 2008; 358(19): 1991–2002.
ncbi.nlm.nih.gov/, accessed on April 26, 2014. 61. Sobngwi E et al. Effect of a diabetic environment in utero on
35. Weinhold B. Epigenetics: The science of change. Environ Health predisposition to type 2 diabetes. Lancet 2003; 361(9372):
Perspect 2006; 114(3): A160–A167. 1861–1865.
36. Broad Institute. What is exome sequencing? Available from: 62. Claesson R, Aberg A, Marsal K. Abnormal fetal growth is associ-
https://www.broadinstitute.org/blog/what-exome-sequencing, ated with gestational diabetes mellitus later in life: Population-
accessed on April 17, 2014. based register study. Acta Obstet Gynecol Scand 2007; 86(6):
37. Nolan CJ, Damm P, Prentki M. Type 2 diabetes across genera- 652–656.
tions: From pathophysiology to prevention and management. 63. Lehnen H, Zechner U, Haaf T. Epigenetics of gestational diabetes
Lancet 2011; 378(9786): 169–181. mellitus and offspring health: The time for action is in early stages
38. National Human Genome Research Institute. Available from: of life. Mol Hum Reprod 2013; 19(7): 415–422.
https://www.genome.gov/20019523, accessed on April 20, 64. Bouchard L et al. Placental adiponectin gene DNA methylation
2014. levels are associated with mothers’ blood glucose concentration.
39. Genetics home reference. Available from: http://ghr.nlm.nih.gov/ Diabetes 2012; 61(5): 1272–1280.
glossary=polygenic, accessed on April 17, 2014. 65. Lillycrop KA et al. Induction of altered epigenetic regulation of
40. Genetics home reference. http://ghr.nlm.nih.gov/handbook/ the hepatic glucocorticoid receptor in the offspring of rats fed
genomicresearch/snp, accessed on April 14, 2014. a protein-restricted diet during pregnancy suggests that reduced
41. Korte A, Farlow A. The advantages and limitations of trait analysis DNA methyltransferase-1 expression is involved in impaired
with GWAS: A review. Plant Methods 2013; 9(1): 29. DNA methylation and changes in histone modifications. Br J Nutr
42. Saxena R et al. Genome-wide association analysis identifies 2007; 97(6): 1064–1073.
loci for type 2 diabetes and triglyceride levels. Science 2007; 66. Calkins K, Devaskar SU. Fetal origins of adult disease. Curr Probl
316(5829): 1331–1336. Pediatr Adolesc Health Care 2011; 41(6): 158–176.
43. Lyssenko V et al. Mechanisms by which common variants in the 67. Langer O et al. A comparison of glyburide and insulin in women
TCF7L2 gene increase risk of type 2 diabetes. J Clin Invest 2007; with gestational diabetes mellitus. N Engl J Med 2000; 343(16):
117(8): 2155–2163. 1134–1138.
44. Grant SF et al. Variant of transcription factor 7-like 2 (TCF7L2) 68. Stubbins MJ et al. Genetic analysis of the human cytochrome
gene confers risk of type 2 diabetes. Nat Genet 2006; 38(3): P450 CYP2C9 locus. Pharmacogenetics 1996; 6(5): 429–439.
320–323. 69. Yin OQ, Tomlinson B, Chow MS. CYP2C9, but not CYP2C19,
45. Yi F, Brubaker PL, Jin T. TCF-4 mediates cell type-specific regu- polymorphisms affect the pharmacokinetics and pharmacody-
lation of proglucagon gene expression by beta-catenin and namics of glyburide in Chinese subjects. Clin Pharmacol Ther
glycogen synthase kinase-3beta. J Biol Chem 2005; 280(2): 2005; 78(4): 370–377.
1457–1464. 70. Pickup JC. Inflammation and activated innate immunity in the
46. Florez JC et al. TCF7L2 polymorphisms and progression to dia- pathogenesis of type 2 diabetes. Diabetes Care 2004; 27(3):
betes in the Diabetes Prevention Program. N Engl J Med 2006; 813–823.
355(3): 241–250. 71. Miehle K, Stepan H, Fasshauer M. Leptin, adiponectin and other
47. Kang S, Xie Z, Zhang D. Association of the rs7903146 polymor- adipokines in gestational diabetes mellitus and pre-eclampsia.
phism in transcription factor 7-like 2 (TCF7L2) gene with gesta- Clin Endocrinol 2012; 76(1): 2–11.
tional diabetes mellitus: A meta-analysis. Gynecol Endocrinol 72. Blois SM et al. Getting too sweet: Galectin-1 dysregulation in ges-
2013; 29(10): 873–877. tational diabetes mellitus. Mol Hum Reprod 2014; 20: 644–649.
48. Rissanen J et al. Sulfonylurea receptor 1 gene variants are asso- 73. Dudzik D et al. Metabolic fingerprint of gestational diabetes mel-
ciated with gestational diabetes and type 2 diabetes but not litus. J Proteomics 2014; 103c: 57–71.
with altered secretion of insulin. Diabetes Care 2000; 23(1): 74. Vrijheid M et al. The Human Early-Life Exposome (HELIX): Project
70–73. rationale and design. Environ Health Perspect 2014.
10 Animal models in diabetes
and pregnancy research
Catherine Yzydorczyk, Delphine Mitanchez, and Umberto Simeoni
Maternal diabetes constitutes an unfavorable environment transported inside the cell only by the glucose transporter-2
for the development of the fetus. The concept that exposure (GLUT-2) (STZ is not recognized by the other GLUTs). Beta
in utero to a hyperglycemic environment has both a short- and cells of the islets of Langerhans express high levels of GLUT-2,
a long-term impact in the offspring is notably demonstrated therefore making them more vulnerable to STZ toxicity.8
by animal studies. Animal models for diabetes and pregnancy Various modes of injection (intravenous or intraperito-
offer various ranges of hyperglycemia, which can lead to fetal neal) and different doses (30–50 mg/kg) have been proposed.
organ alterations, congenital malformations, and micro- or Also, STZ administration can be performed during the pre-
macrosomic fetuses and affect the offspring later in life. The gestational or gestational period that leads to the develop-
purpose of this review is to give a comprehensive view of the ment of severe or mild diabetes associated with two opposite
most frequently studied animal models of diabetes and to fetus phenotypes, intrauterine growth restriction (IUGR),
demonstrate how animal models have permitted to elucidate or macrosomia. In the case of mild diabetes, mothers are
some mechanisms involved in the pathogenesis of diabetes. hyperglycemic (with blood glucose concentrations com-
prised between 120 and 300 mg/dL) and hypoinsulinemic
and deliver macrosomic fetuses, whereas in the case of severe
diabetes, mothers are insulin deficient and hyperglycemic.
Animal models This leads to small-for-gestational-age fetuses, characterizing
Animal models of pregestational type 1 diabetes IUGR, and to fetuses with congenital malformations.9
Type 1 diabetes (T1D) in humans is characterized by a spe-
cific destruction of the pancreatic beta cells, commonly Spontaneous animal models
associated with immune-mediated damage.1 The nonobese diabetic (NOD) mice and biobreeding (BB)
rats develop spontaneous pregestational T1D.
Surgical models NOD mice develop a mild form of diabetes characterized
In rodents, after puberty, partial pancreatectomy (removal of by macrosomic fetuses and adiposity. Pancreatic beta cells
95% of pancreatic weight) leads to a decrease in the number become infiltrated by mononuclear cells that lead to the devel-
of beta cells and therefore to uterine and placental defects and opment of insulitis (inflammation of the islets of Langerhans,
altered fetal glucose metabolism regulation, associated with present in NOD mice at 4–5 weeks old), followed by beta cell
mild maternal diabetes (glycemia between 150 and 200 mg/dL). destruction, and a decrease in circulating insulin concentra-
Insulin administration is not required and pregnancy rates are tions. Mild diabetes is present between 12 and 30 weeks of age.
usually good. However, postsurgery mortality is relatively ele- However, insulin-dependent diabetes develops spontaneously
vated (20%) and a delay (2–3 months) exists between the surgi- only in 9% of NOD at 12 weeks of age and in 80% of them at
cal procedure and the development of diabetic symptoms.2 30 weeks of age and is more prevalent in females.10 Contrary to
human T1D, ketoacidosis is relatively mild and affected ani-
Chemical model mals can survive for several weeks without insulin injection.
In general, human T1D is reproduced by streptozotocin BB rats, which were originally derived from a Canadian
(STZ) administration to rodents during adult life.3 STZ is a colony of outbred Wistar rats, are a good model for the study
nitrosourea derivative isolated from Streptomyces achromo- of perinatal morbidity, small-for-gestational-age fetuses, and
genes.4 It is a powerful alkylating agent that has been shown malformations.
to interfere with glucose transport5 and glucokinase func- Weight loss, polyuria, polydipsia, hyperglycemia, and insu-
tion6 and able to methylate DNA. Nevertheless, it is gener- linopenia appear around 12 weeks of age. Also, ketoacidosis
ally accepted that the cytotoxicity produced by STZ depends is severe and can be fatal if exogenous insulin is not adminis-
on DNA alkylation.7 STZ, relatively similar to glucose, is tered. The incidence is similar in males and females.11,12
84
Outcomes in animal offspring of maternal diabetes exposure 85
Animal models of type 2 diabetes fetuses at late gestation in a rat model with STZ-induced dia-
Type 2 diabetes (T2D) represents a heterogeneous group of betes.27,28 Congenital malformations are also mentioned in
disorders, which is characterized by insulin resistance (IR) fetuses from other models of diabetes such as the GK rats29
and impaired insulin secretion, and defined by a raise in and the BB rats with T1D.30 In NOD model, principally, neu-
fasting or postload blood glucose concentration. ral tube defects and skeletal alterations have been observed
in embryos from NOD female with overt diabetes.31
Spontaneously diabetic rats Also, congenital anomalies could be the result of a
The Goto-Kakizaki (GK) rat model is a model of nonobese default in gene expression involved in the control of the
T2D.13,14 This strain was developed by the successive inbreed- developmental processes. A decrease in Pax-3 (a gene that
ing of Wistar rats with the administration of highest blood encodes a transcription factor necessary for neural tube
glucose concentrations.15 GK rats develop some features closure) expression has been involved in neural tube and
that can be compared with the complications of diabetes cardiac malformations observed in mice with STZ-induced
observed in humans, as renal lesions,16 structural changes diabetes.32,33
in peripheral nerves,17 and abnormalities of the retina.18 GK
offspring exposed throughout gestation to mild diabetes
presents a severe reduction in beta cell mass. This is associ- Fetal organ alterations and long-term outcomes
ated with a lack of pancreatic reactivity to glucose that seems Maternal diabetes is also associated in the offspring with an
to be the consequence of a reduction in cell proliferation, a alteration in the development of many organs. Notably, the
defect of insulin-like growth factor signaling pathways, and heart, kidney, liver, and endocrine pancreas are the most
an increased apoptosis in the fetal pancreas.19,20 affected, and cardiometabolic diseases may develop later in life.
Chemical model Cardiovascular and renal systems

Human T2D is reproduced in rodents by administration of The heart of fetuses from rats with STZ-induced diabetes
different doses of STZ (80–100 mg/kg) in the neonatal period.21 is hypertrophic, while the offspring born from rats with
The timing of the neonatal STZ injection seems to be impor- severe maternal diabetes shows cardiovascular dysfunction.
tant in determining degrees of severity of the subsequent In early adulthood, a decreased heart rate, an impairment
diabetic state. It has been shown that the effects observed in of the vascular function characterized by a reduction of
adulthood were higher if STZ was administrated 5 days after endothelium-dependent vasodilatation capacity to acetyl-
birth22 compared to an administration on the day of birth.23 choline, and an increase in vasoconstriction to noradrena-
STZ administrated 5 days after birth leads in adulthood to line have been shown in the offspring exposed to diabetes
the development of a frank basal hyperglycemia and glucose in utero,34,35 therefore predisposing to cardiovascular dis-
intolerance, an increase in glycosylated hemoglobin, a strong eases. Fetuses from rats with STZ-induced diabetes show
decrease in the pancreatic insulin stores, a reduction (50%) nephrogenesis characterized by a decrease in kidney weight
of basal plasma insulin level, and a default of plasma insulin and nephron number,36 which leads in adulthood to hyper-
response to glucose in vivo.24 Variations in the effects accord- tension37,38 and renal dysfunction.39
ing to the timing of STZ administration could be explained by
the recovery of pancreatic insulin stores due to the regenera- Liver
tion of part of the beta cells after the STZ insult.25 The fetal and neonatal livers from STZ-induced diabetic
dams show an increase in lipid accumulation. A decrease
in insulin receptors has been observed in the fetal liver of
Animal model of gestational diabetes diabetic rats.40 Alterations in signaling pathways have also
The heterozygous leptin receptor-deficient (Lepr(db/+)) been evidenced. In female offspring (db/+), higher hepatic
mouse shows an autosomal recessive mutation in the leptin glucose-6 phosphatase activity has been found to be asso-
receptor gene and serves as an animal model with gesta- ciated with a decrease of Akt phosphorylation that leads
tional diabetes mellitus (GDM). During gestation, compared to the development of an IR in the liver at adulthood. 26
to wild-type, db/+ dams present hyperphagia, hyperlepti- Maternal diabetes interferes with fetal hepatocyte pro-
nemia, greater weight gain, abnormal insulin and glucose liferation and differentiation. It leads to hepatic damage
tolerance tests and their offsprings are heavier at birth.26 characterized by the formation of edematous blood ves-
sels, karyolitic and pyknotic hepatocyte nuclei, and dis-
tortion of blood sinusoids. Also, mitochondrial damage in
fetal hepatocytes was observed with an increase in apop-
Outcomes in animal offspring of totic cells 41 that can participate in enhancing the libera-
maternal diabetes exposure tion of reactive oxygen species (ROS) and so contribute to
the oxidative stress (OS). High levels of lipid peroxidation
Congenital malformations (thiobarbituric acid measurement) have been observed in
Maternal diabetes affects embryonic development and the fetal liver of diabetic pregnant rats associated with a
induces congenital malformations. Skeletal, facial, cardiac, decrease in vitamin E levels that can be associated with
and visceral malformations have been observed in the congenital malformations.42
86 Animal models in diabetes and pregnancy research
Pancreas islet cell proliferation, decreased islet vascularization,54

The fetal beta cells are extremely sensitive to maternal dia- and increased islet cell apoptosis.55 In basal condition, they
betes. In mothers with mild hyperglycemia, rat fetuses pres- are normoglycemic, but glucose tolerance is impaired and
ent hypertrophy of the endocrine pancreas, hyperplasia of insulin-stimulated glucose uptake is reduced54,56,57 in the
beta cells, and an increase of pancreatic and plasma insulin muscle, adipocytes, and liver.58 There is also a gender effect:
concentrations. A positive correlation has been observed The male offspring undergo an impairment in glucose tol-
between maternal glycemia and fetal weight.43–45 erance (17 months of age) and, thereafter, develop T2D and
In the case of severe maternal diabetes, fetal beta cell mass IR,59 whereas female offspring only develop hyperinsu-
has been shown to be either reduced or increased. A reduc- linemia and impaired glucose tolerance in a more advanced
tion of beta cell mass decreases the capacity of insulin secre- age (21 months).60
tion, whereas an enhancement of beta cell mass increases the
High-fat diet
capacity of insulin secretion.46,47 When the fetus is exposed
to very high concentrations of glucose, it will develop islet In rats, maternal high-fat (HF) diet intakes during gesta-
hypertrophy and beta cell hyperactivity associated with tion lead to the development of some features of the meta-
early hyperinsulinemia. On the other hand, the number bolic syndrome (MS) in male offspring characterized by
of beta cells is decreased, and they are disorganized and increased body weight, plasma levels of insulin, glucose, free
depleted in insulin granules; therefore, they are not able to fatty acids, and fatty acids and also glucose intolerance.61,62
secrete insulin in vivo and in vitro.43 This results to hypo- Furthermore, in female offspring, paternal HF diet exposure
insulinemia, a decrease in fetal glucose intakes.48 A negative leads to an early impaired insulin secretion and glucose
correlation has been mentioned between maternal glycemia tolerance that worsened with advanced age.63
and fetal weight. Long-term outcomes involve an increase
of obesity rate, an impairment of glucose tolerance, and an
increased risk of IR, contributing to the development of T2D Uteroplacental insufficiency
in adulthood.45 Uteroplacental insufficiency is the major cause of IUGR. In
rat models, IUGR can be induced by maternal bilateral uter-
ine artery ligation during late gestation (on day 19 of gesta-
tion). Their pups have LBW and decreased levels of glucose,
Animal models of diabetes: The role of insulin, and insulin-like growth factor-1. Thereafter, diabe-
intrauterine environment tes appears in adulthood at approximately 16–26 weeks of
age associated with a default in beta cell secretory capacity
An alteration of fetal environment may program the devel- and IR.64
opment of metabolic diseases later in life.
Mechanisms
Nutrition alteration
Maternal nutrition represents a major factor, and it is now Oxidative stress
well recognized that inappropriate nutrition in utero has Diabetic pregnancy and macrosomia are associated with
notable metabolic consequences in adulthood.49 Beta cell increased OS. Protein glycation and glucose auto-oxidation
development is irreversibly damaged by inadequate nutri- may generate ROS and therefore catalyzed lipid peroxida-
tion during critical periods of fetal development.50 tion.65 ROS can be detoxified by endogenous antioxidant
enzymes (superoxide dismutase [SOD], catalase, glutathione
Caloric-nutrient restriction peroxidase, and reductase) and vitamins (A, E, and C).
In rat, global restrictions (40%–50% of normal intake) dur- An increase in OS and an impairment of antioxidant
ing the last week of gestation lead to low birth weight (LBW) enzymes have been found in fetuses and neonates from rodent
in offspring and a decrease in beta cell mass that persists experimental models with mild and severe STZ-induced dia-
in adulthood with a reduction in insulin content.51,52 Also, betes.66,67 In offspring with advanced age, maternal protein
nutrient restriction during the suckling period leads to a restriction leads to an increase in OS levels, characterized
permanent reduction in beta cell mass in offspring and the by an increased lipid peroxidation (hexanoyl lysine adducts),
development of glucose intolerance with aging.53 decreased antioxidant defenses (SOD), and an increase in
islet fibrosis markers (collagen I and III), which can lead to
Protein restriction progressive beta cell loss and dysfunction.68 A poor maternal
In pregnant rats fed with a low-protein (LP) diet, the mater- nutrition leading to LBW followed by accelerated postnatal
nal plasma glucose and insulin levels are normal and fetuses growth is associated with impaired antioxidant defenses
are normoglycemic. Nevertheless, the development of the (SOD expressions) and increased markers of senescence (p21
fetal endocrine pancreas is impaired, which leads to a defi- and p16) in the beta cells of pancreatic islets.69 To confirm
cient insulin response and to fetal growth restriction. The the role of OS in these metabolic alterations, nutritional
offspring from LP dams presents a deficit in beta cell mass,54 and antioxidant therapies have been tested. Folic acid and
reduced pancreatic insulin content and islet size, altered antioxidant supplementations (α-tocopherol and ascorbate)
Mechanisms 87
have been observed to decrease congenital malformations the genes. Usually, if these islands of the promoter region
in term fetuses of rodents with STZ-induced diabetes.70,71 are unmethylated, the gene is transcribed, but when a sig-
Also, omega-3 fatty acid consumption restores the antioxi- nificant part of the CpG islands is methylated, the gene can
dant status of diabetic pregnant rats and decreases the rate no longer be transcribed and is silenced. In early life, DNA
of macrosomic pups at birth and lipid peroxidation (plasma methylation can be altered by many nutritional factors such
thiobarbituric acid–reactive substance measurement) in dia- as folate, choline, betaine, methionine, and vitamins B2, B6,
betic mothers and their offspring.72 and B12 (methyl donors).82 DNA methylation pattern can
also be affected by OS. The replacement of guanine by the
ROS radical adduct-8-hydroxyguanine affects the methyla-
Telomere shortening tion of adjacent cytosine.83 Therefore, this DNA modifica-
Telomeres are repeated sequences of DNA (TTAGGG)n at tion can affect the development and thus increase the risk
the end of chromosomes that protect them against inappro- of developing cardiometabolic diseases later in life. A global
priate DNA fusion, breaks, and shortening into coding DNA DNA hypomethylation has been observed in the liver of rats
sequences.73 Thus, they are essential for chromosome stabil- with T1D.84
ity. Telomeres have a high content in guanine and therefore LP diet is associated with impaired fetal growth and
are particularly susceptible to damage by OS,74 and a posi- development of cardiometabolic diseases in the offspring.
tive correlation has been observed between OS and telomere Many epigenetic changes have been reported in offspring
shortening.75 With each cycle of replication, the telomeres exposed to a maternal LP diet. Hypomethylation of the pro-
shorten; however, the telomerase makes additional copies of moter of the glucocorticoids receptor (GR) associated with
the hexanucleotide repeats to protect against this telomere a decrease in DNA methyltransferase-1 expression and in
loss. Data suggest that short telomeres may lead to prema- expression and binding of methyl CpG-binding protein-2
ture beta cell senescence associated with a decreased beta in the liver results in increased GR promoter activity and so
cell mass and a default in insulin secretion and glucose toler- can lead to the development of hypertension thereafter. An
ance. Using a mice model deficient in telomerase, Kuhlow increased activation of the peroxisomal proliferator–acti-
et al. have suggested that this enzyme is important to main- vated receptor-α promoter has also been observed in the
tain glucose homeostasis.76 In rat offspring, an LP diet expo- liver of these adult offspring exposed to maternal LP diet,85
sure in utero followed by an accelerated postnatal growth which can be associated with induction of dyslipidemia.86
leads to a higher number of short telomeres and fewer long LP intakes during gestation in mice lead to DNA hypo-
telomeres in pancreatic islets, reflecting an accelerated islet methylation of the leptin promoter in adipose tissues associ-
telomere shortening.69 ated with an alteration in body composition and increased
food intakes in male offspring and so could contribute to the
development of MS.87
Epigenetic regulation
Epigenetic modifications can be defined as changes in the Histone modifications
pattern of gene expression without involving changes in In the nucleus of eukaryotic cells, DNA is coiled around
the DNA sequence. The fetal environment can alter the octamers of globular proteins called histones, forming dense
epigenome in the offspring, therefore leading to different blocks of DNA and proteins named nucleosomes and that
phenotypes.77–79 Three main pathways can silence, acti- in conjunction form chromatin, the condensed form of
vate, or regulate the level and time of expression of many DNA. Histones can be modified on the tails by methylation,
genes: DNA methylation, histone modifications (acetylation, acetylation, phosphorylation, biotinylation, ubiquitination,
methylation, ubiquitination, phosphorylation, or ADP ribo- and ADP-ribosylation.79,80 These modifications facilitate or
sylation), and small noncoding RNAs, such as microRNAs obstruct the access of DNA to transcriptional factors and,
(miRNAs).80,81 In general, these three epigenetic mechanisms together with the other epigenetic factors, regulate gene
appear to work together to regulate gene expression. DNA expression.
methylation or histone modifications can alter the expres- In a rodent model of IUGR that expressed a lower level
sion of miRNAs, which can in turn regulate the expression of Pdx1 (a transcription factor critical for beta cell func-
level of DNA methyltransferases, histone methyltransferases, tion),88 epigenetic modifications characterized by the loss of
histone deacetylases, and methyl CpG-binding proteins and upstream stimulatory factor-1 binding at proximal promoter
therefore regulate the epigenetic processes of DNA methyla- of Pdx1, recruitment of histone deacetylase 1 and the core-
tion and histone modifications.80 Nowadays, it is relatively pressor Sin3A, and deacetylation of histone H3 and H4 were
well established that epigenetic alteration/modification can observed. This resulted in Pdx1 silencing, which persisted
affect several processes involved in glucose regulation and from 2 weeks to 4 months of age and was responsible for the
insulin secretion that can lead to the development of T2D. prediabetic state in rats.89
Histone modifications were also proposed to decrease
DNA methylation the expression of GLUT-4 (glycoprotein that facilitates
It is the best-known pathway. It usually leads to repression of g lucose transport that is involved in the development of glu-
transcription of the involved gene. Methylation takes place cose intolerance) in the IUGR offspring. A reduction of glucose
on CpG islands mainly located in the promoter region of transport in the muscle is a feature of IR. Such alterations
persist at adulthood and could be an adaptive response to the this IR. However, in the model with STZ-induced diabetes,
decrease of insulin pancreatic production.90 the main alteration is the beta cell destruction.
Also, histone modification and DNA methylation can be In an optimal experimental model of GDM, levels of
dependent on one another. Methylation of lysine 9 on H3 can glycemia before gestation would be normal, and then, after
promote DNA methylation, and in turn, CpG methylation midgestation, glucose intolerance and impaired insulin
can influence methylation of lysine 9 on H3.91 secretion would be observed with a perturbation in glucose
and IR in the dams and therefore affect the fetus.
Noncoding RNAs A satisfying model of GDM is db/+ mice in which dia-
miRNAs are small single-stranded RNAs that do not encode betes occurs during gestation, but the deficiency in leptin
proteins. Each miRNA binds to specific messenger RNAs receptor leads to phenotypes that differ from human GDM.
(mRNAs) resulting in the degradation of target mRNA or In STZ-induced diabetes, the destruction of beta cells during
inhibition of its translation into protein. miRNAs regulate pregnancy leads to a diabetic experimental model during the
the posttranscriptional expression level of many genes and pregnant state.
processes such as apoptosis, cell growth, and differentiation Nevertheless, in all cases, the elevated glucose and other
in a large range of tissues. Therefore, miRNAs are involved metabolic substrates in the maternal compartment reach the
in many processes including angiogenesis, cardiogenesis, fetuses and are involved in the induction of short- and long-
nephrogenesis, and related diseases.92 miRNAs are critical term outcomes of diabetes.
regulators of metabolism. It has been shown that miR-103, An alternative to the limits mentioned on rodent animal
miR-107, and miR-33 regulated insulin sensitivity and glu- models with diabetes will be the use of larger animals that
cose homeostasis.93–95 Also, an overexpression of miR-143 seem better to mimic the human diabetes.
was involved in the default of insulin pathway and glucose Treatment of rabbit with alloxan is associated with the
homeostasis.96 induction of T1D, notably on the long term. Injection of a
monodose of alloxan (100 mg/kg) is rapidly taken up by the
beta cells, which leads to the formation of free radicals and
Conclusion the fragmentation of beta cell DNA, as a consequence of the
poor antioxidant defenses from these cells. Hypoinsulinemia
The impact of an early exposure to an altered fetal environ- and hyperglycemia have been observed 48 hours after the
ment (hyperglycemia, inadequate maternal nutrition) on destruction of pancreatic b-cells.97
the offspring’s health and on the development of cardio- In cats, some similar characteristics to T2D in human are
metabolic diseases in adulthood is robustly demonstrated in reported. Cats develop T2D in middle age, associated with
animal models. These models also can offer a major contri- obesity, IR, and subsequent beta cell loss.98
bution in the understanding of the mechanisms leading to Administration of an HF diet and different doses of STZ
such alterations. However, they have some limits in mimick- in dog lead to features similar to human T2D, characterized
ing human diseases. by obesity, IR, and compensatory hyperinsulinemia followed
An ideal animal model to mimic human T1D and T2D is by a decrease in pancreatic function.99
not easy to develop. For T1D, this ideal animal model should The nonhuman primates (NHPs) present disease with
be associated with an autoimmune destruction of beta cells many similarities with the human ones like obesity, IR, dys-
at a very early stage of life to reproduce the human situation. lipidemia, and hypertension. Likewise, NHP spontaneously
However, in surgical animal models of diabetes, the decrease develop T2D with advanced age and also when a hyperca-
of beta cell number is the result of pancreatectomy, whereas loric diet is administrated.100
in STZ-induced diabetes, beta cells are destroyed due to spe- Larger animal models of diabetes share important
cific beta cell–induced death. The spontaneous animal mod- metabolic similarities with human pathology compared
els (NOD mice and BB rats) seem to be good models because to rodent models. The life span to reproduce the natural
the destruction of beta cells is the result of an autoimmune history T2D and permit to study the long-term clinical
process. In BB rat models, diabetes develops around puberty, complications observed in human. Moreover, larger ani-
whereas this occurs in aged animals in NOD mice. mal models provide a valuable resource to test the safety
The ideal animal model of T2D should be a model with and efficacy of therapeutic intervention prior to clinical
an IR and a default in pancreatic response resulting from trials.
REFERENCES
1. Dahlquist G. The aetiology of type 1 diabetes: An epidemiologi- 5. Zahner D, Ramirez R, Malaisse WJ. Kinetic behaviour of liver glu-
cal perspective. Acta Paediatr Suppl 1998; 425: 5–10. cokinase in diabetes. III. Possible role of insulinopenia. Diabetes
2. Foglia VG. Fetuses and newborns of 95 per cent pancreatecto- Res 1990; 14: 117–122.
mized female rats. Adv Metab Disord 1970; 1(Suppl. 1): 221+. 6. Zahner D, Malaisse WJ. Kinetic behaviour of liver glucokinase
3. Wu KK, Huan Y. Streptozotocin-induced diabetic models in mice in diabetes. I. Alteration in streptozotocin-diabetic rats. Diabetes
and rats. Curr Protoc Pharmacol 2008; Chapter 5: Unit 5.47. Res 1990; 14: 101–108.
4. Bono VH, Jr. Review of mechanism of action studies of the nitro- 7. Bennett RA, Pegg AE. Alkylation of DNA in rat tissues following
soureas. Cancer Treat Rep 1976; 60: 699–702. administration of streptozotocin. Cancer Res 1981; 41: 2786–2790.
References 89
8. Schnedl WJ, Ferber S, Johnson JH et al. STZ transport and 31. Otani H, Tanaka O, Tatewaki R et al. Diabetic environment and
cytotoxicity. Specific enhancement in GLUT2-expressing cells. genetic predisposition as causes of congenital malformations in
Diabetes 1994; 43: 1326–1333. NOD mouse embryos. Diabetes 1991; 40: 1245–1250.
9. Aerts L, Van Assche FA. Endocrine pancreas in the offspring of 32. Kumar SD, Dheen ST, Tay SS. Maternal diabetes induces con-
rats with experimentally induced diabetes. J Endocrinol 1981; 88: genital heart defects in mice by altering the expression of genes
81–88. involved in cardiovascular development. Cardiovasc Diabetol
10. Formby B, Schmid-Formby F, Jovanovic L et al. The offspring of 2007; 6: 34.
the female diabetic “nonobese diabetic” (NOD) mouse are large 33. Loeken MR. Advances in understanding the molecular causes
for gestational age and have elevated pancreatic insulin content: of diabetes-induced birth defects. J Soc Gynecol Investig 2006;
A new animal model of human diabetic pregnancy. Proc Soc Exp 13: 2–10.
Biol Med 1987; 184: 291–294. 34. Taylor PD, McCarthy AL, Thomas CR et al. Endothelium-
11. Mordes JP, Bortell R, Groen H et al. Autoimmune diabetes dependent relaxation and noradrenaline sensitivity in mesenteric
mellitus in the BB rat. In: Sima A, Shafrir E, eds. Frontiers in resistance arteries of streptozotocin-induced diabetic rats. Br J
animal diabetes research, a primer on animal models of dia- Pharmacol 1992; 107: 393–399.
betes. Reading, U.K.: Harwood Academic Publishers, 2001: 35. Holemans K, Gerber RT, Meurrens K et al. Streptozotocin dia-
67–78. betes in the pregnant rat induces cardiovascular dysfunction in
12. Mordes JP, Bortell R, Blankenhorn EP et al. Rat models of type adult offspring. Diabetologia 1999; 42: 81–89.
1 diabetes: Genetics, environment, and autoimmunity. ILAR J 36. Amri K, Freund N, Vilar J et al. Adverse effects of hyperglyce-
2004; 45: 278–291. mia on kidney development in rats: In vivo and in vitro studies.
13. Ling ZC, Efendic S, Wibom R et al. Glucose metabolism in Goto- Diabetes 1999; 48: 2240–2245.
Kakizaki rat islets. Endocrinology 1998; 139: 2670–2675. 37. Segar EM, Norris AW, Yao JR et al. Programming of growth, insu-
14. Galli J, Li LS, Glaser A et al. Genetic analysis of non-insulin lin resistance and vascular dysfunction in offspring of late gesta-
dependent diabetes mellitus in the GK rat. Nat Genet 1996; 12: tion diabetic rats. Clin Sci 2009; 117: 129–138.
31–37. 38. Nehiri T, Duong Van Huyen JP, Viltard M et al. Exposure to mater-
15. Goto Y, Kakizaki M, Masaki N. Production of spontaneous dia- nal diabetes induces salt-sensitive hypertension and impairs renal
betic rats by repetition of selective breeding. Tohoku J Exp Med function in adult rat offspring. Diabetes 2008; 57: 2167–2175.
1976; 119: 85–90. 39. Wichi RB, Souza SB, Casarini DE et al. Increased blood pressure
16. Janssen U, Phillips AO, Floege J. Rodent models of nephropathy in the offspring of diabetic mothers. Am J Physiol Regul Integr
associated with type II diabetes. J Nephrol 1999; 12: 159–172. Comp Physiol 2005; 288: R1129–R1133.
17. Murakawa Y, Zhang W, Pierson CR et al. Impaired glucose toler- 40. Mulay S, Philip A, Solomon S. Influence of maternal diabetes on
ance and insulinopenia in the GK-rat causes peripheral neuropa- fetal rat development: Alteration of insulin receptors in fetal liver
thy. Diabetes Metab Res Rev 2002; 18: 473–483. and lung. J Endocrinol 1983; 98: 401–410.
18. Sone H, Kawakami Y, Okuda Y et al. Ocular vascular endo- 41. El-Sayyad HI, Al-Haggar MM, El-Ghawet HA et al. Effect of
thelial growth factor levels in diabetic rats are elevated before maternal diabetes and hypercholesterolemia on fetal liver of
observable retinal proliferative changes. Diabetologia 1997; 40: albino Wistar rats. Nutrition 2014; 30: 326–336.
726–730. 42. Siman CM, Eriksson UJ. Vitamin E decreases the occurrence of
19. Calderari S, Gangnerau MN, Thibault M et al. Defective IGF2 malformations in the offspring of diabetic rats. Diabetes 1997; 46:
and IGF1R protein production in embryonic pancreas precedes 1054–1061.
beta cell mass anomaly in the Goto-Kakizaki rat model of type 2 43. Kervran A, Guillaume M, Jost A. The endocrine pancreas of the
diabetes. Diabetologia 2007; 50: 1463–1471. fetus from diabetic pregnant rat. Diabetologia 1978; 15: 387–393.
20. Miralles F, Portha B. Early development of beta-cells is impaired 44. Aerts L, Van Assche FA. Is gestational diabetes an acquired con-
in the GK rat model of type 2 diabetes. Diabetes 2001; 50(Suppl. dition? J Dev Physiol 1979; 1: 219–225.
1): S84–S88. 45. Van Assche FA, Holemans K, Aerts L. Long-term consequences
21. Blondel O, Bailbe D, Portha B. Insulin resistance in rats with non- for offspring of diabetes during pregnancy. Br Med Bull 2001; 60:
insulin-dependent diabetes induced by neonatal (5 days) strep- 173–182.
tozotocin: Evidence for reversal following phlorizin treatment. 46. Eriksson U, Andersson A, Efendic S et al. Diabetes in pregnancy:
Metabolism 1990; 39: 787–793. Effects on the foetal and newborn rat with particular regard to
22. Bonner-Weir S, Trent DF, Honey RN et al. Responses of neonatal body weight, serum insulin concentration and pancreatic con-
rat islets to streptozotocin: Limited B-cell regeneration and hyper- tents of insulin, glucagon and somatostatin. Acta Endocrinol
glycemia. Diabetes 1981; 30: 64–69. (Copenh) 1980; 94: 354–364.
23. Portha B, Levacher C, Picon L et al. Diabetogenic effect of strep- 47. Aerts L, Sodoyez-Goffaux F, Sodoyez JC et al. The diabetic intra-
tozotocin in the rat during the perinatal period. Diabetes 1974; uterine milieu has a long-lasting effect on insulin secretion by B
23: 889–895. cells and on insulin uptake by target tissues. Am J Obstet Gynecol
24. Blondel O, Bailbe D, Portha B. Relation of insulin deficiency to 1988; 159: 1287–1292.
impaired insulin action in NIDDM adult rats given streptozocin 48. Philipps AF, Rosenkrantz TS, Grunnet ML et al. Effects of fetal
as neonates. Diabetes 1989; 38: 610–617. insulin secretory deficiency on metabolism in fetal lamb.
25. Cantenys D, Portha B, Dutrillaux MC et al. Histogenesis of the Diabetes 1986; 35: 964–972.
endocrine pancreas in newborn rats after destruction by strep- 49. Yajnik C. Interactions of perturbations in intrauterine growth and
tozotocin. An immunocytochemical study. Virchows Arch B Cell growth during childhood on the risk of adult-onset disease. Proc
Pathol Incl Mol Pathol 1981; 35: 109–122. Nutr Soc 2000; 59: 257–265.
26. Yamashita H, Shao J, Qiao L et al. Effect of spontaneous gesta- 50. Cerf ME, Chapman CS, Muller CJ et al. Gestational high-fat pro-
tional diabetes on fetal and postnatal hepatic insulin resistance in gramming impairs insulin release and reduces Pdx-1 and gluco-
Lepr(db/+) mice. Pediatric Res 2003; 53: 411–418. kinase immunoreactivity in neonatal Wistar rats. Metabolism
27. Siman CM, Gittenberger-De Groot AC, Wisse B et al. 2009; 58: 1787–1792.
Malformations in offspring of diabetic rats: Morphometric analy- 51. Garofano A, Czernichow P, Breant B. In utero undernutri-
sis of neural crest-derived organs and effects of maternal vitamin tion impairs rat beta-cell development. Diabetologia 1997; 40:
E treatment. Teratology 2000; 61: 355–367. 1231–1234.
28. Roest PA, Molin DG, Schalkwijk CG et al. Specific local cardio- 52. Bertin E, Gangnerau MN, Bellon G et al. Development of beta-
vascular changes of N-epsilon-(carboxymethyl)lysine, vascular cell mass in fetuses of rats deprived of protein and/or energy
endothelial growth factor, and Smad2 in the developing embryos in last trimester of pregnancy. Am J Physiol Regul Integr Comp
coincide with maternal diabetes-induced congenital heart Physiol 2002; 283: R623–R630.
defects. Diabetes 2009; 58: 1222–1228. 53. Garofano A, Czernichow P, Breant B. Effect of ageing on beta-
29. Malaisse-Lagae F, Vanhoutte C, Rypens F et al. Anomalies of fetal cell mass and function in rats malnourished during the perinatal
development in GK rats. Acta Diabetol 1997; 34: 55–60. period. Diabetologia 1999; 42: 711–718.
30. Verhaeghe J, Peeters TL, Vandeputte M et al. Maternal and fetal 54. Dahri S, Snoeck A, Reusens-Billen B et al. Islet function in off-
endocrine pancreas in the spontaneously diabetic BB rat. Biol spring of mothers on low-protein diet during gestation. Diabetes
Neonate 1989; 55: 298–308. 1991; 40(Suppl. 2): 115–120.
55. Petrik J, Reusens B, Arany E et al. A low protein diet alters the 77. Sebert S, Sharkey D, Budge H et al. The early programming of
balance of islet cell replication and apoptosis in the fetal and metabolic health: Is epigenetic setting the missing link? Am J Clin
neonatal rat and is associated with a reduced pancreatic expres- Nutr 2011; 94: 1953S–1958S.
sion of insulin-like growth factor-II. Endocrinology 1999; 140: 78. Rogers LK, Velten M. Maternal inflammation, growth retardation,
4861–4873. and preterm birth: Insights into adult cardiovascular disease. Life
56. Bertin E, Gangnerau MN, Bailbe D et al. Glucose metabolism Sci 2011; 89: 417–421.
and beta-cell mass in adult offspring of rats protein and/or energy 79. Canani RB, Costanzo MD, Leone L et al. Epigenetic mechanisms
restricted during the last week of pregnancy. Am J Physiol 1999; elicited by nutrition in early life. Nutr Res Rev 2011; 24: 198–205.
277: E11–E17. 80. Chen M, Zhang L. Epigenetic mechanisms in developmental
57. Shepherd PR, Crowther NJ, Desai M et al. Altered adipocyte programming of adult disease. Drug Discov Today 2011; 16:
properties in the offspring of protein malnourished rats. Br J Nutr 1007–1018.
1997; 78: 121–129. 81. McKay JA, Mathers JC. Diet induced epigenetic changes and
58. Ozanne SE, Jensen CB, Tingey KJ et al. Low birthweight is asso- their implications for health. Acta Physiol 2011; 202: 103–118.
ciated with specific changes in muscle insulin-signalling protein 82. Choi SW, Friso S. Epigenetics: A new bridge between nutrition
expression. Diabetologia 2005; 48: 547–552. and health. Adv Nutr 2010; 1: 8–16.
59. Petry CJ, Dorling MW, Pawlak DB et al. Diabetes in old male 83. Franco R, Schoneveld O, Georgakilas AG et al. Oxidative
offspring of rat dams fed a reduced protein diet. Int J Exp Diabetes stress, DNA methylation and carcinogenesis. Cancer Lett 2008;
Res 2001; 2: 139–143. 266: 6–11.
60. Fernandez-Twinn DS, Wayman A, Ekizoglou S et al. Maternal 84. Williams KT, Garrow TA, Schalinske KL. Type I diabetes leads to
protein restriction leads to hyperinsulinemia and reduced insulin- tissue-specific DNA hypomethylation in male rats. J Nutr 2008;
signaling protein expression in 21-mo-old female rat offspring. 138: 2064–2069.
Am J Physiol Regul Integr Comp Physiol 2005; 288: R368–R373. 85. Lillycrop KA, Phillips ES, Jackson AA et al. Dietary protein restric-
61. Srinivasan M, Katewa SD, Palaniyappan A et al. Maternal high-fat tion of pregnant rats induces and folic acid supplementation pre-
diet consumption results in fetal malprogramming predisposing vents epigenetic modification of hepatic gene expression in the
to the onset of metabolic syndrome-like phenotype in adulthood. offspring. J Nutr 2005; 135: 1382–1386.
Am J Physiol Endocrinol Metab 2006; 291: E792–E799. 86. Burdge GC, Phillips ES, Dunn RL et al. Effect of reduced mater-
62. Nivoit P, Morens C, Van Assche FA et al. Established diet-induced nal protein consumption during pregnancy in the rat on plasma
obesity in female rats leads to offspring hyperphagia, adiposity lipid concentrations and expression of peroxisomal proliferator–
and insulin resistance. Diabetologia 2009; 52: 1133–1142. activated receptors in the liver and adipose tissue of the offspring.
63. Ng SF, Lin RC, Laybutt DR et al. Chronic high-fat diet in fathers Nutr Res 2004; 24: 8.
programs beta-cell dysfunction in female rat offspring. Nature 87. Jousse C, Parry L, Lambert-Langlais S et al. Perinatal undernutri-
2010; 467: 963–966. tion affects the methylation and expression of the leptin gene in
64. Simmons RA, Templeton LJ, Gertz SJ. Intrauterine growth retar- adults: Implication for the understanding of metabolic syndrome.
dation leads to the development of type 2 diabetes in the rat. FASEB J 2011; 25: 3271–3278.
Diabetes 2001; 50: 2279–2286. 88. Bernardo AS, Hay CW, Docherty K. Pancreatic transcription fac-
65. Hunt JV, Smith CC, Wolff SP. Autoxidative glycosylation and pos- tors and their role in the birth, life and survival of the pancreatic
sible involvement of peroxides and free radicals in LDL modifica- beta cell. Mol Cell Endocrinol 2008; 294: 1–9.
tion by glucose. Diabetes 1990; 39: 1420–1424. 89. Park JH, Stoffers DA, Nicholls RD et al. Development of type 2
66. Pustovrh MC, Jawerbaum A, Capobianco E et al. Oxidative stress diabetes following intrauterine growth retardation in rats is asso-
promotes the increase of matrix metalloproteinases-2 and -9 ciated with progressive epigenetic silencing of Pdx1. J Clin Invest
activities in the feto-placental unit of diabetic rats. Free Radic Res 2008; 118: 2316–2324.
2005; 39: 1285–1293. 90. Raychaudhuri N, Raychaudhuri S, Thamotharan M et al. Histone
67. Raza H, John A. Glutathione metabolism and oxidative stress in code modifications repress glucose transporter 4 expression in
neonatal rat tissues from streptozotocin-induced diabetic moth- the intrauterine growth-restricted offspring. J Biol Chem 2008;
ers. Diabetes Metab Res Rev 2004; 20: 72–78. 283: 13611–13626.
68. Tarry-Adkins JL, Chen JH, Jones RH et al. Poor maternal nutri- 91. Schubeler D, Lorincz MC, Cimbora DM et al. Genomic targeting
tion leads to alterations in oxidative stress, antioxidant defense of methylated DNA: Influence of methylation on transcription,
capacity, and markers of fibrosis in rat islets: Potential underlying replication, chromatin structure, and histone acetylation. Mol
mechanisms for development of the diabetic phenotype in later Cell Biol 2000; 20: 9103–9112.
life. FASEB J 2010; 24: 2762–2771. 92. Sayed D, Abdellatif M. MicroRNAs in development and disease.
69. Tarry-Adkins JL, Chen JH, Smith NS et al. Poor maternal nutri- Physiol Rev 2011; 91: 827–887.
tion followed by accelerated postnatal growth leads to telomere 93. Trajkovski M, Hausser J, Soutschek J et al. MicroRNAs 103 and
shortening and increased markers of cell senescence in rat islets. 107 regulate insulin sensitivity. Nature 2011; 474: 649–653.
FASEB J 2009; 23: 1521–1528. 94. Davalos A, Goedeke L, Smibert P et al. miR-33a/b contribute to
70. Cederberg J, Eriksson UJ. Antioxidative treatment of pregnant the regulation of fatty acid metabolism and insulin signaling. Proc
diabetic rats diminishes embryonic dysmorphogenesis. Birth Natl Acad Sci USA 2011; 108: 9232–9237.
Defects Res A Clin Mol Teratol 2005; 73: 498–505. 95. Ramirez CM, Goedeke L, Rotllan N et al. MicroRNA 33 regulates
71. Oyama K, Sugimura Y, Murase T et al. Folic acid prevents con- glucose metabolism. Mol Cell Biol 2013; 33: 2891–2902.
genital malformations in the offspring of diabetic mice. Endocr J 96. Jordan SD, Kruger M, Willmes DM et al. Obesity-induced over-
2009; 56: 29–37. expression of miRNA-143 inhibits insulin-stimulated AKT acti-
72. Yessoufou A, Soulaimann N, Merzouk SA et al. N-3 fatty acids vation and impairs glucose metabolism. Nat Cell Biol 2011; 13:
modulate antioxidant status in diabetic rats and their macrosomic 434–446.
offspring. Int J Obes 2006; 30: 739–750. 97. Wang J, Wan R, Mo Y et al. Creating a long-term diabetic rabbit
73. Blackburn EH. Structure and function of telomeres. Nature 1991; model. Exp Diab Res 2010; 2010: 289614.
350: 569–573. 98. O’Brien TD. Pathogenesis of feline diabetes mellitus. Mol Cell
74. Kawanishi S, Oikawa S. Mechanism of telomere shortening by Endocrinol 2002; 197: 213–219.
oxidative stress. Ann N Y Acad Sci 2004; 1019: 278–284. 99. Ionut V, Liu H, Mooradian V et al. Novel canine models of obese
75. Richter T, von Zglinicki T. A continuous correlation between oxi- prediabetes and mild type 2 diabetes. Am J Physiol Endocrinol
dative stress and telomere shortening in fibroblasts. Exp Gerontol Metab 2010; 298: E38–E48.
2007; 42: 1039–1042. 100. Kievit P, Halem H, Marks DL et al. Chronic treatment with a mel-
76. Kuhlow D, Florian S, von Figura G et al. Telomerase deficiency anocortin-4 receptor agonist causes weight loss, reduces insulin
impairs glucose metabolism and insulin secretion. Aging 2010; 2: resistance, and improves cardiovascular function in diet-induced
650–658. obese rhesus macaques. Diabetes 2013; 62: 490–497.
11 Pathologic abnormalities of
placental structure and function
in diabetes
Rhonda Bentley-Lewis, Maria Rosaria Raspollini, and Drucilla Roberts
intervillous space as a cover and bottom. The intervillous

Introduction space is perfused with maternal blood, which circulates
In 1892, Ballantyne wrote: “A diseased foetus without his freely and directly around the trophoblastic surface of the
placenta is an imperfect specimen, and a description of a placental villi. The villi originate from a complex, treelike
foetal malady, unless accompanied by a notice of the placen- projection of the chorionic plate into the intervillous space
tal condition, is incomplete. Deductions drawn from such a (Figure 11.1). The villous surface is surrounded by the tro-
case cannot be considered as conclusive, for in the missing phoblast and a core composed of stroma that supports the
placenta or cord may have existed the cause of the disease fetal vessels connected to the fetal circulation system via the
and death. During intrauterine life the foetus, the mem- chorionic plate and umbilical cord. At the placental mar-
branes, the cord and the placenta form an organic whole, and gin, the intervillous space is obliterated as the chorionic and
disease of any part must react upon and affect the others.”1 basal plates fuse together to form the chorion laeve.5
Indeed, a careful examination of the placenta contributes
to the determination of the causes underlying fetal growth
abnormalities, demise, or neonatal conditions. The corre- Pathophysiology of placental
lation between placental histopathology and clinical data derangements in diabetes
may augment our understanding of the sequence of patho-
logic events and enhance the management of the neonate or Maternal diabetes mellitus complicates pregnancy through
future pregnancies. a variety of growth-promoting and growth-restricting forces
Although placental specimens are not routinely sent for that disrupt the normal development of both the fetus and
pathological examination, there has been growing recog- placenta. Diabetes leads to increased inflammation, which
nition of its importance.2 Indications for placental referral inhibits implantation, similar to the effect of intrauterine
include fetal conditions requiring admission to the neona- contraceptive devices that induce local inflammation.6 In
tal intensive care unit and maternal diseases or disorders addition to inflammation, oxidative stress resulting from
related to pregnancy, such as maternal diabetes.3 At a mini- diabetes negatively impacts uteroplacental vascular func-
mum, the placenta requires gross examination after delivery tion, thereby altering the maternal intrauterine environ-
to identify any visible pathology. This macroscopic analysis ment. Because the placenta serves as the sole source of fetal
should include assessment of the length and insertion site oxygen and nutrients, metabolic derangement in the pla-
of the umbilical cord, the number of cord vessels, the state cental environment, such as changes in glucose and insulin
and color of the fetal surface and fetal membranes, the pres- levels, can affect fetal development.
ence of abnormalities on the maternal surface, the color and Maternal glucose reaches fetal circulation freely; there-
the consistency of the villous tissue, and the presence of cho- fore, when there is maternal hyperglycemia, the ensuing
rionic plate lesions. Any abnormal findings should prompt fetal hyperglycemia has several possible effects on the devel-
a complete pathologic examination regardless of clinical oping conceptus. These adverse sequelae include hyperplasia
history.4 of the islets of Langerhans, increased insulin stored in the
At term, the normal human placenta is a focal, disk- islets, and a greater probability of single umbilical artery for-
like thickening of the membranous sac that is achieved by mation.7 Several studies have shown that additional placen-
splitting the membranes into two separate sheets, the cho- tal findings in cases of maternal diabetes may be influenced
rionic plate and the basal plate. Both sheets enclose the by the level of dysglycemia or type of diabetes, possibly due
91
92 Pathologic abnormalities of placental structure and function in diabetes
The fetal–placental vasculature, developed in the second

week of pregnancy, is also negatively affected by maternal
diabetes. Under normoglycemic conditions, the gestational
sac invades the maternal decidua allowing the media and
elastica of the uterine spiral arteries to loosen. The endo-
thelium is then supported by connective tissue fibers.
However, during diabetes in pregnancy, increased oxida-
tive stress and the activation of protein kinase pathways,
particularly PK-C and mitogen-activated PK, may impair
placental vasculature formation. In addition, the response
of the fetal–placental vasculature to vasoconstrictor and
vasodilator agents in diabetic placentas is significantly
attenuated compared to that of nondiabetic placentas.16,28
Consequently, the diabetic placenta may not be able to
respond adequately to demands for altered blood flow,
thereby leading to fetal compromise.28
Figure 11.1 Hematoxylin–eosin section of normal placenta

at term showing villi that form a complex treelike projection of
the chorionic plate into the intervillous space.
Gross examination
to the different stages during pregnancy when the placenta is A recent comprehensive systematic review of the literature
exposed to hyperglycemia.2,8–11 Maternal glycemic status can provides a critical examination of the gross and histopatho-
also influence the up- or down-regulation of insulin recep- logical findings associated with dysglycemia in pregnancy.21
tors on trophoblasts.12 Because insulin does not physically On initial gross examination, placental anomalies indicative
cross the placenta into fetal circulation, the effect of insulin of chronic insult may be present, including abnormal pla-
on maternal diabetes is indirect. cental size and shape, infarcts or hemorrhages, cord defects,
Insulin has also been shown to modulate ovarian ste- and membrane anomalies. While human placentas are gen-
roidogenesis13 that, at the earliest stages of pregnancy, erally round or oval, other shapes are not uncommon. Shape
allows the development of normal endometrial receptiv- anomalies, such as a multilobulated placenta, may develop
ity.14 During the course of pregnancy, nondiabetic women from abnormalities in fetal gene expression or fetal–mater-
experience a progressive decline in insulin sensitivity nal interactions. An abnormal maternal environment, such
associated with hyperinsulinemia,15 a physiologic state as the presence of submucosal leiomyomas or a uterine scar,
required to preferentially direct maternal nutrients toward unprotected membranous fetal vessels as seen in vasa previa,
the fetoplacental unit.16 However, women with gestational or early uteroplacental vascular compromise, can also alter
diabetes mellitus have been shown to express abnormal the shape of the placenta.5
insulin secretion17 and pronounced insulin resistance.18 Placental weight, measured after removing the cord,
The resulting protein kinase C (PK-C) activation and membranes, and maternal blood clots, is a function of gesta-
increased oxidative stress lead to endothelial dysfunctional age, fetal weight, and fetal gender. The placenta weighs
tion of the decidual capillary network. This dysfunc- approximately 450–550 g at term, and the fetal/placental
tion is associated with decreased vasodilation, increased weight ratio should be approximately 6–7. In pregnancies
leukocyte–endothelial cell adhesion, and increased vascu- with appropriate-for-gestational-age infants, histological
lar permeability.19,20 differences between diabetic and nondiabetic placentas are
Type 1 diabetes during pregnancy is also associated with the most pronounced.29 Placental weight heavier than the
dysregulation of glucose and oxygen metabolic pathways, 95th or less than the 5th percentile for gestational age is often
both of which affect placental villous growth and function. a sign of chronic insult. Maternal diabetes is often associ-
Different degrees of changes in the syncytiotrophoblast, ated with large placentas heavier than the 90th percen-
cytotrophoblast, trophoblastic basement membrane, and tile although the differential diagnosis also includes fetal
fetal vessels have been described and attributed to glycemic hydrops, congenital syphilis, villous edema, and Beckwith–
status. Alteration of placental development may contribute Wiedemann syndrome.30
to increased risk of pregnancy complications such as pre- Cord length is usually between 40 and 70 cm with a diam-
eclampsia, macrosomia, fetal growth restriction, and fetal eter between 1 and 3 cm at term. A short or excessively long
hypoxia. However, the extent to which maternal glycemic cord is related to an increased risk for fetal damage. Marginal
control contributes to placental abnormalities is still being cord insertion or velamentous cord insertion, which may
considered in detail21 additionally, several studies have be secondary to aberrant blastocyst implantation, may be
identified histopathological placental abnormalities among associated with fetal or neonatal injury. Macroscopic analy-
women with well-controlled pregestational11,22–24 and gesta- sis of cord vessels can reveal histopathologic features such
tional diabetes.25–27 as thrombosis of vessels or the presence of a single artery,
Microscopic evaluation 93
(a) (b) (c)
Figure 11.2 Section of a normal placenta at 38 weeks gestational age (a), a placenta at 38 weeks gestational age complicated
by maternal insulin-dependent diabetes mellitus and chronic hypertension (b and c) (hematoxylin–eosin). In (b) the villi are
widely separated due to premature branching, and the villi have fewer capillaries and larger and more numerous syncytial knots
(arrow and arrowhead). (c) Shows a chronic abruption with the retroplacental hematoma (arrow) with an indented and infracted
parenchyma.
manifestations of vascular pathology, fetal malformation, or Microscopic evaluation

maternal diabetes.
The placental surface is red and spongy at term. A dark- Placental pathology in pregnancies complicated by maternal
red color suggests congestion of capillaries or chorangiosis diabetes varies based on physiological influences includ-
that may be associated with maternal diabetes (Figure 11.2). ing the type of diabetes, degree of dysglycemia, and insulin
Placental thickness is usually between 1.5 and 3 cm but dia- use.11,21,27 Type 1 diabetes is associated with increased cap-
betic placentas are often thicker. Macroscopic analysis of illary volume and greater degrees of vascular dysfunction,
the fetal surface of the chorionic plate should include eval- including increased branching and nonbranching angiogen-
uation of the chorionic vessels for evidence of thrombosis. esis, likely the result of increased leakiness of fetal placen-
Dilated and/or discolored vessels suggest fresh thrombo- tal vessels.21,22 In addition, studies have shown that women
sis, while tan–white or yellow fibrosed vessels are indica- with type 1 diabetes have a higher prevalence of placental
tive of an old thrombus. Thrombosis of chorionic vessels dysmaturity while women with type 2 diabetes have a higher
is more common in maternal diabetes but can also be seen prevalence of placental infarcts.21 Compared to gestational
in vascular anomalies accompanied by local trauma or sta- diabetes, pregestational diabetes results in a greater number
sis, including true knots of the umbilical cord, velamen- of histopathologic lesions, such as chorial edema and inter-
tous cord insertion, and umbilical cord entanglement.31 stitial hemorrhage,33 while gestational diabetes has been
Thrombosis can also be found in thrombophilic states, fetal associated with increased fibrinoid material, villous edema,
chromosomal disorders, toxic agents to fetal vessels, and hyperplasia, and thickening of the basement membrane.34
some viral infections.4 The literature presents conflicting views regarding the
Fetal membrane color may also be impacted by mater- degree to which maternal glycemic control affects placental
nal diabetes. Fetal membranes and the fetal surface of the histopathology. However, in a recent systematic review of the
chorionic plate are normally clean and transparent produc- topic,21 several studies have shown that both pregestational
ing a blue color on the chorionic plate and pink color on the and gestational diabetes are associated with an increased fre-
free membranes. However, in fetal acute chorioamnionitis, quency of placental villous immaturity, even with controlled
another characteristic of diabetic placentas, a diffuse yel- glycemic levels.29,35 Other studies have found that placentas
low/white–opaque feature is observed with infiltration of from women with diabetes and well-controlled glucose levels
granulocytes, lymphocytes, and other inflammatory cells. appeared normal when examined by routine light micros-
In addition, meconium staining, also seen in maternal dia- copy.36,37 A study of an aggregated population of women
betes, may produce a green coloration of the membranes or with pregestational or gestational diabetes stratified by
a brown color if the meconium is accompanied by hemo- glycemic control revealed that first-trimester “low” HbA1c
siderin deposition.32 Because several pathologies may occur levels (≤6.0%) were associated with more placental infarcts
simultaneously, the best description of membrane color is and less villous maturity compared to “high” HbA1c levels
often “normal” or “abnormal.” The macroscopic evaluation (>6.0%). When excluding the population of women with
of the maternal surface of the placenta should demonstrate preeclampsia, only the placental infarct incidence remained
the normal mosaic of 20–40 cotyledons. This mosaic can be significantly different between the two groups.11
absent in pathologic conditions, such as vascular pathologies Another well-documented effect of maternal diabetes in
causing extensive fibrin deposition. The presence of adherent pregnancy is chorangiosis, described as the presence of 10
fresh or old blood clots dissecting the placental parenchyma or more villi with 10 or more capillaries in 10 or more low-
is indicative of placental abruption. These features are also power microscopic fields. 29,38–43 Chorangiosis is believed to
observed in placentas from patients carrying thrombophilic be a response to immature villi, another effect of maternal
traits, maternal preeclampsia, or hypertension. diabetes during pregnancy. 29,44–47 Interestingly, one study
(a) (b) (c)
Figure 11.3 Hematoxylin–eosin section of a normal placenta at 38 weeks gestation (a), a placenta at 38 weeks gestation with
maternal gestational diabetes (b), and a placenta at 38 weeks gestation with insulin-dependent diabetes (c). Note the smaller villi
in (a) have 1–3 capillaries, while those in (b) have 3–5, and in (c) have 10 or more. The villous stroma in (a) is dense and pink and
the trophoblast has formed syncytial knots (arrow). The villous vessels are all peripheral (arrowhead). With diabetes, there are few
syncytial knots (b and c), and the stroma is more open and clear.
demonstrated a difference in chorangiosis prevalence by pathology includes brisk funisitis or umbilical cord vasculi-
race/ethnicity; white women with gestational diabetes tis with peripheral umbilical cord abscesses (Figure 11.4).59
had a higher rate of chorangiosis compared to nonwhite The presence of invasive hyphae within the Wharton’s jelly
women with gestational diabetes. 27 Villous immatu- of the umbilical cord is thought to represent an increased
rity, defined as inadequate or absent terminal villi, can risk for disseminated disease in the fetus, but even deeply
lead to relative hypoxia due to the fact that the centrally invasive hyphae are usually benign.3 Consequently, it
located villous capillaries will increase the distance oxy- usually does not have significant sequelae for the fetus,
gen and nutrients need to travel from the maternal to fetal although rarely fungal placentitis can lead to overwhelm-
circulation.48 ing sepsis.62,63
Chronic chorioamnionitis may also be present in the Fungal infection may also promote umbilical cord vascu-
diabetic placenta. This finding is characterized by infiltra- litis and abscess formation, also associated with acute cho-
tion of the membranes, chorionic plate, and umbilical cord rioamnionitis. The abscesses typically have easily identifiable
by mature lymphocytes, plasma cells, and histiocytes.49,50 hyphae invading the Wharton’s jelly or superficial hyphae
Although chronic chorioamnionitis may be seen in the and do not require special stains for visualization. Abscesses
setting of infectious agents, other suggested mechanisms on the surface of the umbilical cord are nearly always due
have included an imbalance of the systemic inflammatory to Candida although they can be occasionally observed in
milieu associated with insulin resistance in gestational dia- group B β-hemolytic Streptococcal infections. When vis-
betes mellitus.51 In contrast to fetal acute chorioamnionitis, ible hyphae are absent, fungal specific staining such as silver
chronic chorioamnionitis is usually associated with villi- stains should be used.
tis of unknown etiology and cannot be evaluated via gross
examination.52
One of the more common diseases associated with mater-
nal diabetes is hypertension. Whether pregnancy related,
such as preeclampsia or pregnancy-induced hypertension,
or pregestational, hypertension can cause vascular damage
to the uterus leading to poor blood flow to the placenta. The
resulting uteroplacental insufficiency causes placental isch-
emia that is associated with infarction, chronic and acute
abruption, and stunted placental growth. This impaired pla-
cental growth is associated with a small organ, weighing less
than the 10th percentile, and accelerated villous branching,
also referred to as hypermaturity (Figure 11.2).53 In addi-
tion, hypertensive diabetic placentas have more syncytial
knots, more vasculosyncytial membranes, smaller villi,
centrally located villous vessels, and decreased vessels per
villous compared to placentas from nonhypertensive dia-
betic pregnancies (Figure 11.3). Poor glucose control will
exacerbate these findings and result in placentomegaly and
Figure 11.4 Hematoxylin–eosin section of an umbilical cord
macrosomia.54 showing peripheral abscesses (arrow) that contain maternal
Fungal placentitis, a relatively uncommon finding, is white blood cells and abundant candidal forms (insert with
also increased in the setting of maternal diabetes.55–61 This silver stain).
References 95
Conclusions to be influenced by the type of diabetes and degree of dys-

glycemia, but more research is required to elucidate the roles
The placenta in maternal diabetes is heavy, large, somewhat of demographic factors, such as race/ethnicity, and clinical
immature, and characterized by chorangiosis and vascular characteristics, such as insulin use. Based on the interaction
anomalies such as decidual vasculopathy. Increased vascu- of these and additional genetic and environmental factors,
lar damage results in uteroplacental insufficiency, reduced increased risk for villitis, chorioamnionitis, and funisitis
placental growth, and villous hypermaturity, a pathologic may exist. Further research into the pathology associated
profile similar to that observed in hypertension during preg- with diabetes is warranted to enhance our understanding of
nancy. Placental pathologic characteristics have been shown clinically relevant diagnostic and management information.
REFERENCES
1. Ballantyne JW. The Disease and Deformities of the Foetus, Vol. I, 19. Feener EP, King GL. Endothelial dysfunction in diabetes mel-
Oliver and Boyd: Edinburgh, U.K., 1892. litus: Role in cardiovascular disease. Heart Fail Monit 2001; 1:
2. Redline RW. Placental pathology: Is it time to get serious? 74–82.
Editorial. Contemporary OB/GYN, Modern Medicine Network, 20. De Vriese AS, Verbeuren TJ, Van de Vorde J et al. Endothelial
February 1, 2014. Retrieved from http://contemporaryobgyn. dysfunction in diabetes. Br J Pharmacol 2000; 130: 963–974.
modernmedicine.com/contemporary-obgyn/content /tags/ 21. Huynh J, Dawson D, Roberts D, Bentley-Lewis R. A systematic
maternal-fetal-medicine/placental-pathology-it-time-get-serious. review of placental pathology in maternal diabetes mellitus.
3. Roberts DJ. The placental pathology report. In: UpToDate, Post Placenta 2015; 36(2): 101–114.
TW, eds., UpToDate: Waltham, MA. Retrieved from http://www. 22. Asmussen I. Ultrastructure of the villi and fetal capillaries of the
uptodate.com/contents/the-placental-pathology-report. placentas delivered by non-smoking diabetic women (White
4. Roberts DJ, Oliva E. Clinical significance of placental examina- group D). Acta Pathol Microbiol Immunol Scand A 1982; 90(2):
tion in perinatal medicine. J Matern Fetal Neonatal Med 2006; 95–101.
19: 255–264. 23. Boyd PA, Scott A, Keeling JW. Quantitative structural studies on
5. Benirschke K, Kaufman P, Baergen RN. The Pathology of the placentas from pregnancies complicated by diabetes mellitus. Br
Human Placenta, Springer-Verlag: Heidelberg, Germany, 2006. J Obstet Gynaecol 1986; 93(1): 31–35.
6. Ortiz ME, Croxatto HB, Bardin CW. Mechanisms of action of 24. Fox H. Pathology of the placenta in maternal diabetes mellitus.
intrauterine devices. Obstet Gynecol Surv 1996; 51: S42–S51. Obstet Gynecol 1969; 34(6): 792–798.
7. Driscoll S. Pathology of pregnancy complicated by diabetes mel- 25. Teasdale F. Histomorphometry of the placenta of the diabetic
litus. Med Clin North Am 1965; 49: 1052–1057. women: Class A diabetes mellitus. Placenta 1981; 2(3): 241–251.
8. Haust MD. Maternal diabetes mellitus-effects on the fetus and 26. Daskalakis G, Marinopoulos S, Krielesi V et al. Placental pathol-
placenta. In: Naeye RL, Kissane JM, Kaufman N, eds., Perinatal ogy in women with gestational diabetes. Acta Obstet Gynecol
Diseases, Williams & Wilkins: Baltimore, MD, 1981, pp. 201–285. Scand 2008; 87(4): 403–407.
9. Singer DB. The placenta in pregnancies complicated by diabetes 27. Bentley-Lewis R, Dawson DL, Wenger JB, Thadhani RI, Roberts
mellitus. Perspect Pediatr Pathol 1984; 8: 199–212. DJ. Placental histomorphometry in gestational diabetes mellitus:
10. Fox H. The Pathology of the Placenta, W.B. Saunders Co.: The relationship between subsequent type 2 diabetes mellitus
Philadelphia, PA, 1997. and race/ethnicity. Am J Clin Pathol 2014; 141(4): 587–592.
11. Huynh J, Yamada J, Beauharnais C et al. Type 1, type 2 and 28. Myatt L, Kossenjans W, Sahay R et al. Oxidative stress causes
gestational diabetes mellitus differentially impact placental patho- vascular dysfunction in the placenta. J Matern Fetal Med 2000;
logic characteristics of uteroplacental malperfusion. Placenta 9: 79–82.
2015; 36(10): 1161–1166. 29. Evers IM, Nikkels PG, Sikkema JM et al. Placental pathology in
12. Desoye G, Hofmann HH, Weiss PA. Insulin binding to tropho- women with type 1 diabetes and in a control group with nor-
blast plasma membranes and placental glycogen content in mal and large-for-gestational-age infants. Placenta 2003; 24:
well-controlled gestational diabetic women treated with diet or 819–825.
insulin, in well-controlled overt diabetic patients and in healthy 30. Torfs CP, van den Berg B, Oechsli FW et al. Prenatal and perina-
control subjects. Diabetologia 1992; 35(1): 45–55. tal factors in the etiology of cerebral palsy. J Pediatr 1990; 116:
13. Willis D, Mason H, Gilling-Smith C et al. Modulation by insulin 615–619.
of follicle-stimulating hormone and luteinizing hormone actions 31. Redline RW. Clinical and pathological umbilical cord abnor-
in human granulosa cells of normal and polycystic ovaries. J Clin malities in fetal thrombotic vasculopathy. Hum Pathol 2004; 35:
Endocrinol Metab 1996; 81: 302–309. 1494–1498.
14. Dominguez F, Avila S, Cervero A et al. A combined approach 32. Miller PW, Coen RW, Benirschke K. Dating the time interval from
for gene discovery identifies insulin-like growth factor-binding meconium passage to birth. Obstet Gynecol 1985; 66: 459–462.
protein related protein 1 as a new gene implicated in human 33. Rudge MV, Lima CP, Damasceno DC et al. Histopathological
endometrial receptivity. J Clin Endocrinol Metab 2003; 88: placental lesions in mild gestational hyperglycemic and diabetic
1849–1857. women. Diabetol Metab Syndr 2011; 3(1): 19.
15. Fridman JE, Ishizuka T, Shao J et al. Impaired glucose transport 34. al-Okail MS, al-Attas OS. Histological changes in placental
and insulin receptor tyrosine phosphorylation in skeletal muscle syncytiotrophoblasts of poorly controlled gestational diabetic
from obese women with gestational diabetes. Diabetes 1999; 48: patients. Endocr J 1994; 41(4): 355–360.
1807–1814. 35. Higgins M, McAuliffe FM, Mooney EE. Clinical associations with
16. Jauniaux E, Burton GJ. Villous histomorphometry and placental a placental diagnosis of delayed villous maturation: A retrospec-
bed biopsy investigation in type I diabetic pregnancies. Placenta tive study. Pediatr Dev Pathol 2011; 14(4): 273–279.
2006; 27: 468–474. 36. Mayhew TM, Jairam IC. Stereological comparison of 3D spa-
17. Buchanan TA, Metzger BE, Freinkel N et al. Insulin sensitivity and tial relationships involving villi and intervillous pores in human
beta cell responsiveness to glucose during late pregnancy in lean placentas from control and diabetic pregnancies. J Anat 2000;
and moderately obese women with normal glucose tolerance 197(Pt. 2): 263–274.
or mild gestational diabetes. Am J Obstet Gynecol 1990; 162: 37. Mooney E, Doyle E, Gearhart P, Robboy S. Placenta—Clinical
1008–1014. scenarios. In: Robboy SJ et al., eds., Pathology of the Female
18. Kautzky-Willer A, Prager R, Waldhausl W et al. Pronounced insu- Reproductive Tract, 2nd ed., Elsevier Limited: Amsterdam, the
lin resistance and inadequate β-cell secretion characterize lean Netherlands, 2009, p. 871.
gestational diabetes during and after pregnancy. Diabetes Care 38. Asmussen I. Vascular morphology in diabetic placentas. Contrib
1997; 20: 1717–1723. Gynecol Obstet 1982; 9: 76–85.
39. Altshuler G. Chorangiosis. An important placental sign of neo- 52. Baerger RN. Manual of Bernirshke and Kaufmann’s Pathology of
natal morbidity and mortality. Arch Pathol Lab Med 1984; 108: the Human Placenta, Springer: New York, 2005.
71–74. 53. Redline RW, Boyd T, Campbell V et al. Maternal vascular under-
40. Ogino S, Redline RW. Villous capillary lesions of the placenta: perfusion: Nosology and reproducibility of placental reaction
Distinctions between chorangioma, chorangiomatosis, and cho- patterns. Pediatr Dev Pathol 2004; 7: 237–249.
rangiosis. Hum Pathol 2000; 31: 945–954. 54. Setji TL, Brown AJ, Feinglos MN. Gestational diabetes mellitus.
41. De La Ossa MM, Cabello-Inchausti B, Robinson MJ. Placental Clin Diabetes 2005; 23: 17–24.
chorangiosis. Arch Pathol Lab Med 2001; 125: 1258. 55. Delprado WJ, Baird PJ, Russell P. Placental candidiasis: Report of
42. Schwartz DA. Chorangiosis and its precursors: Underdiagnosed three cases with a review of the literature. Pathology 1982; 14:
placental indicators of chronic fetal hypoxia. Obstet Gynecol 191–195.
Surv 2001; 56: 523–525. 56. Schwartz DA, Reef S. Candida albicans placentitis and funisi-
43. Gupta R, Nigam S, Arora P et al. Clinico-pathological profile of tis: Early diagnosis of congenital candidemia by histopathologic
12 cases of chorangiosis. Arch Gynecol Obstet 2006; 274: 50–53. examination of umbilical cord vessels. Pediatr Infect Dis J 1990;
44. Haust MD. Maternal diabetes mellitus—Effects on the fetus and 9: 661–665.
placenta. Monogr Pathol 1981; 22: 201–285. 57. Reed BD. Risk factors for Candida vulvovaginitis. Obstet Gynecol
45. Singer DB, Liu CT, Widness JA et al. Placental morphometric Surv 1992; 47: 551–560.
studies in diabetic pregnancies. Placenta Suppl 1981; 3: 193–202. 58. Mazziotti F, Arena V, Lo Mastro F et al. Diabetes and pregnancy:
46. Stoz F, Schuhmann RA, Haas B. Morphohistometric investigations in Prophylaxis of genital infections. Ann Ist Super Sanita 1997; 33:
placentas of gestational diabetes. J Perinat Med 1988; 16: 205–209. 343–345.
47. Younes B, Baez-Giangreco A, al-Nuaim L et al. Basement mem- 59. Qureshi F, Jacques SM, Bendon RW et al. Candida funisitis: A
brane thickening in the placentae from diabetic women. Pathol clinicopathologic study of 32 cases. Pediatr Dev Pathol 1998; 1:
Int 1996; 46: 100–104. 118–124.
48. Kingdom J, Huppertz B, Seaward G et al. Development of the 60. Kelekci S, Kelekci H, Cetin M et al. Glucose tolerance in preg-
placental villous tree and its consequences for fetal growth. Eur J nant women with vaginal candidiasis. Ann Saudi Med 2004; 24:
Obstet Gynecol Reprod Biol 2000; 92: 35–43. 350–353.
49. Salafia CM, Silberman L. Placental pathology and abnormal fetal 61. Grigoriou O, Baka S, Makrakis E et al. Prevalence of clinical vagi-
heart rate patterns in gestational diabetes. Pediatr Pathol 1989; 9: nal candidiasis in a university hospital and possible risk factors.
513–520. Eur J Obstet Gynecol Reprod Biol 2006; 126: 121–125.
50. Jacques SM, Qureshi F. Chronic chorioamnionitis: A clinicopatho- 62. Gerberding KM, Eisenhut CC, Engle WA et al. Congenital Candida
logical and immunohistochemical study. Hum Pathol 1998; 29: pneumonia and sepsis: A case report and review of the literature.
1457–1461. J Perinatol 1989; 9: 159–161.
51. Richardson A, Fulton C, Catlow D et al. Cytokine profiles in 63. Levy I, Shalit I, Birk E et al. Candida endocarditis in neonates:
pregnant women with gestational diabetes mellitus. Am J Obstet Report of five cases and review of the literature. Mycoses 2006;
Gynecol 2006; 195: S163. 49: 43–48.
12 The great obstetric syndromes:
The roots of disease
Rinat Gabbay-Benziv and Ahmet A. Baschat
is a compensatory increase in pancreatic β cells response and

Introduction insulin secretion, the opposing effects of placental hormones,
The term “great obstetrical syndromes” was originally cortisol, and progesterone on metabolism remain progressive
applied to pregnancy-related disorders such as preterm and prevalent so that the maintenance of placental and fetal
labor, preterm premature rupture of membranes, preeclamp- nutrition prevails at the expense of maternal glycemic control.
sia, spontaneous pregnancy loss, stillbirth, and abnormally It is noteworthy that this metabolic pattern is ubiquitous and
delayed or accelerated fetal growth. All of them share a independent of maternal physical characteristics but can result
placental component as part of their etiology.1 The under- in exaggerated metabolic effects in predisposed women. The
lying concept is one of multiple underlying etiologies that recognition that pregnancy is a diabetogenic state is based on
adversely interact with the maternal–fetal unit to initiate these metabolic changes that can exceed the threshold to pre-
subclinical pathology that progresses to clinical manifesta- cipitate the development of overt GDM.
tion that also results in fetal adverse outcome. This concept
is applicable to conditions such as gestational diabetes mel-
litus (GDM): a metabolic disease where the combination Maternal predisposition or
of maternal predisposition and placental factors results in
progressively impaired glucose tolerance with secondary hCG-mediated exacerbation?
maternal and fetal effects that typically manifest in the third The damaging potential of increased glucose levels is reflected
trimester.2–4 Understanding the contributory role of these by the increased rates of early pregnancy loss, placental dys-
interactions is important for prediction, screening, diagno- function, maternal hypertensive disorders, and congenital
sis, prognosis, and ultimately therapeutic interventions. birth defects in diabetic women with poor periconception gly-
This chapter aims to review the changes in maternal– cemic control.8–11 Catalano and colleagues compared longitu-
placental–fetal interface associated with normal glucose metab- dinal changes in peripheral insulin sensitivity between lean
olism and the shift to abnormal glucose metabolism up to the and obese women with subsequent normal glucose tolerance
clinical threshold of GDM with an emphasis on the periconcep- or GDM from preconception throughout pregnancy using the
tion period, early gestation, and adaptive mechanisms through- hyperinsulinemic–euglycemic clamp technique.12,13 In these
out pregnancy and their relationship to the evolution of disease. studies, all women experienced a 50%–60% decrease in insu-
lin sensitivity with advancing gestation. The greater degree
of insulin resistance that contributed to the development of
Periconception interactions GDM was attributable to preexisting insulin insensitivity and
the superimposed pregnancy-related effect. Both lean and
The placenta interfaces the maternal and fetal compartments, obese women who later on developed GDM had decrease insu-
and in the first trimester, placenta-mediated maternal adap- lin sensitivity prior to pregnancy compared to their matching
tations ensure preferential nutritional supply to the placenta controls who continued to have normal glucose tolerance in
and fetus. This is achieved through a metabolic state that is pregnancy. These studies illustrate that GDM arises when the
characterized by relative increase in insulin resistance and pregnancy-associated metabolic impacts push predisposed
anabolism,5 both are linked to the rise in human chorionic women over the threshold that constitutes normal glucose
gonadotropin (hCG) and human placental lactogen (hPL). This tolerance and that this effect is progressive with advancing
principal change in metabolic handling of nutrients results in gestation. This further implies that the degree of preexisting
higher postprandial glucose levels and increased placental glu- vulnerability to metabolic deterioration determines the onset
cose utilization contributing to a rapid fall of fasting glucose and degree of clinical disease and the potential for placental
levels accompanied by a rise in free fatty acids.6,7 While there and fetal complications.
97
98 The great obstetric syndromes
Luoto et al.14 studied the impact of first-trimester dietary or intra-arterial) results in displacement of vascular smooth
modification and physical activity on the prevention of muscle and endothelial cells and transforms the narrow spi-
GDM and delivery of a large-for-gestational-age infant. In ral arteries into a low-resistance high-capacity vascular bed.
this study, pregnant women with at least one risk factor for
GDM (age > 40, BMI > 25, prior GDM, prior macrosomia,
and family history of diabetes) were recruited at 8–12 gesta- Diabetic placenta
tional weeks. Modification of physical activity was initiated
at recruitment and dietary consult started at 16–18 gesta- The term GDM encompasses all women first discovered in
tional weeks. The study demonstrated that the interventions pregnancy and accordingly by itself does not distinguish
had an impact on infant birth weight but failed to reduce between diabetes exclusively precipitated by pregnancy from
the rate of GDM. This provides further evidence that therapy preexisting type 2 diabetes incidentally discovered during
does not prevent the development of GDM but can reduce pregnancy. In fact, in 30%–40% of women with GDM, the
the consequences of the condition. diagnosis is made prior to 20 weeks of gestation,15,16 increas-
ing the likelihood that there is an overlap with preexist-
ing type 2 diabetes. In principle, the disruption of normal
Interaction with placental placental development is most profound if it occurs early
development in gestation. Because early disruption of development occurs
at a time when structural development, especially of the
The placenta has many roles. Other than being a selective vasculature, is still incomplete, structural and functional
barrier between the maternal and fetal circulations that is placental dysfunction may ensue. In comparison, later devel-
capable of controlling nutrient and gas exchange, it also has opment of abnormal glycemic status is more likely to affect
immune and endocrine functions. The conditions that con- the functionality of the placenta. The relationship between
tribute to the development of the great obstetrical syndromes the degree of glycemic control and placental changes remains
exert their effects in part by interfering with normal placen- unclear. While some studies support the concept that the
tal development that sets the stage for the development of degree of damage is proportional to the degree of hypergly-
pregnancy-related hypertensive disease, preterm delivery, cemia, o thers have shown that even in tight glycemic control
stillbirth, abruption, and disrupted fetal growth.3 there are still histological differences in the diabetic placen-
GDM is associated with maternal and fetal short-term tas compared to the nondiabetic ones.17–19
and long-term complications. Well-recognized complica- Macroscopically, the diabetic placenta is character-
tions are the increased risk for type 2 diabetes for the mother ized by increased size and weight resulting in an increased
and accelerated growth for the fetus. Although less com- placental-to-fetal-weight ratio.20–25 These changes are pro-
mon, but not less important are fetal adverse outcomes such portional to the degree of hyperglycemia and suggest that in
as stillbirth, delivery complications attributable to macro- the diabetic environment, the placenta grows first and then
somia, and neonatal metabolic complications. Late-onset contributes to accelerated fetal growth by increased glucose
metabolic complications for the offspring are associated as well as other nutrient transfer.
with endothelial/vascular dysfunction that may later cause Microscopically, a number of lesions have been found in
obesity, hypertension, type 2 diabetes mellitus, and meta- increased frequency in placentas of diabetic women compared
bolic syndrome. At least some of these complications may be to controls. Villous fibrinoid necrosis, a condition where vil-
related to aberrant placental structure or function caused by lous stroma is replaced by fibrinoid material, and villous
the metabolic milieu of the diabetic mother. immaturity with a decreased formation of terminal villi are
An important question is how early in gestation can observed. These changes may affect maternal–fetal nutrient
placen tal changes related to diabetes be seen. Following delivery and gas exchange and accordingly may be precur-
fertilization, placentation is initiated when the trophoblast sors to otherwise unanticipated fetal complications such as
develops as a cell area that is distinct from the inner cell stillbirth.26,27 Chorangiosis and vascular hyperplasia of the
mass that will form the embryo. The progenitor cytotropho- chorionic villi are defined as the occurrence of 10 or more villi
blast cell is the stem cell of the placenta. These cells prolifer- with 10 or more capillaries in 10 or lower-power microscopic
ate throughout gestation, differentiating along two pathways fields and may be a result of increased vascular recruitment
to form villous cytotrophoblast that ultimately will become that occurs with placental overgrowth. Infarction of 10% or
either syncytiotrophoblasts (outer cellular layer) or extravil- more of the placenta and evidence of ischemia (Tenney–Parker
lous cytotrophoblasts (inner cellular layer). The syncytiotro- changes) resulting in increased maturation of branching villi
phoblast has several functions, including transport of gases, and increased numbers of fetal nucleated red blood cells as a
nutrients, and waste products and synthesis of peptide and marker of chronic fetal hypoxemia are also more commonly
steroid hormones that regulate placental, fetal, and maternal seen in the placentas of diabetic women.
systems. The extravillous trophoblast (EVT) has a prolifera- Jones et al.28 suggested that even optimal glycemic con-
tive component and an invasive component. Invasive EVT trol cannot abolish the placental effect as they have observed
that invades the decidua is called interstitial EVT, whereas various degrees of changes in the syncytiotrophoblast,
EVT that invades and remodels the spiral arteries is called cytotrophoblast, trophoblastic basement membrane, and
the endovascular EVT. Endovascular invasion (intramural fetal vessels in six out of seven women with all ranges of
References 99
diabetic severity. Madazli et al.17 compared the histology of diffusion distance of immature villi and the placental size
22 placentas from women with well-controlled GDM to 22 to perfusion mismatch may predispose the fetus to sud-
placentas from nondiabetic women that served as control. den changes in gas and nutrient exchange. Accordingly,
They found a sixfold increase in villous immaturity and the placenta “becomes its own enemy” by initiating the
chorangiosis and a threefold increase in ischemic changes sequence that leads from altered glycemic status to placen-
in the placentas of women diagnosed with GDM. These tal dysfunction. Understanding these pathways in greater
findings were confirmed by the study of Daskalakis et al.,20 detail will become important in modifying this sequence
which examined a large number of placentas by a pathologist of events.
that was blinded to the maternal glycemic status. This study
showed that GDM was associated with villous immaturity,
chorangiosis, villous fibrinoid necrosis, and nucleated red Conclusions
blood cells regardless of the maternal glycemic status.
Collectively, these studies document that the glyce- GDM can be regarded as one of the great obstetric syn-
mic control can have profound impacts on many critical dromes where early placental development coupled with
developmental and functional aspects of the placenta. maternal predisposition initiates a chain of events that
Interference with angiogenesis sets the stage for maternal affects placental, maternal, and fetal development with
hypertensive disorders. The placenta that is subjected to lasting impacts on mother and child. Understanding the
the opposing stimuli of glucose-mediated growth accel- pathways that mediate these effects to greater detail will
eration and decrease in vascular development is at risk be required to develop effective screening, prevention, and
for local ischemic changes. The increased intervillous management approaches.
REFERENCES
1. Romero R. Prenatal medicine: The child is the father of the man. 16. Meyer WJ, Carbone J, Gauthier DW et al. Early gestational glu-
Prenat Neonatal Med 1996; 1: 8–11. cose screening and gestational diabetes. J Reprod Med 1996; 41:
2. Romero R. Prenatal medicine: The child is the father of the 675–679.
man. 1996. J Matern Fetal Neonatal Med August 2009; 22(8): 17. Madazli R, Tuten A, Calay Z et al. The incidence of placental abnor-
636–639. malities, maternal and cord plasma malondialdehyde and vascular
3. Brosens I, Pijnenborg R, Vercruysse L et al. The “Great Obstetrical endothelial growth factor levels in women with gestational diabe-
Syndromes” are associated with disorders of deep placentation. tes mellitus and nondiabetic controls. Gynecol Obstet Invest 2008;
Am J Obstet Gynecol March 2011; 204(3): 193–201. 65(4): 227–232.
4. Khong Y, Brosens I. Defective deep placentation. Best Pract Res 18. Laurini RN, Visser GHA, van Ballegooie E et al. Morphological find-
Clin Obstet Gynaecol June 2011; 25(3): 301–311. ings in placentas of insulin-dependent diabetic patients treated with
5. Kühl C. Aetiology of gestational diabetes. Baillieres Clin Obstet continuous subcutaneous insulin infusion (CSII). Placenta 1987; 8:
Gynaecol June 1991; 5(2): 279–292. 153–165.
6. Metzger BE, Ravnikar V, Vileisis RA et al. “Accelerated star- 19. Desoye G, Korgun ET, Ghaffari-Tabrizi N et al. Is fetal macrosomia
vation” and the skipped breakfast in late normal pregnancy. in adequately controlled diabetic women the result of a placen-
Lancet March 13, 1982; 1(8272): 588–592. tal defect? A hypothesis. J Matern Fetal Neonatal Med 2002; 11:
7. Metzger BE, Freinkel N. Accelerated starvation in pregnancy: 258–261.
Implications for dietary treatment of obesity and gestational dia- 20. Daskalakis G, Marinopoulos S, Krielesi V et al. Placental pathology
betes mellitus. Biol Neonate 1987; 51(2): 78–85. in women with gestational diabetes. Acta Obstet Gynecol Scand
8. West NA, Kechris K, Dabelea D. Exposure to maternal diabe- 2008; 87(4): 403–407.
tes in utero and DNA methylation patterns in the offspring. 21. Gauster M, Desoye G, Tötsch M et al. The placenta and gestational
Immunometabolism 2013; 1: 1–9. diabetes mellitus. Curr Diab Rep February 2012; 12(1): 16–23.
9. El Hajj N, Pliushch G, Schneider E et al. Metabolic program- 22. Pathak S, Hook E, Hackett G et al. Cord coiling, umbilical cord
ming of MEST DNA methylation by intrauterine exposure to insertion and placental shape in an unselected cohort delivering
gestational diabetes mellitus. Diabetes 2013; 62: 1320–1328. at term: Relationship with common obstetric outcomes. Placenta
10. Ruchat SM, Houde AA, Voisin G et al. Gestational diabetes November 2010; 31(11): 963–968.
mellitus epigenetically affects genes predominantly involved in 23. Taricco E, Radaelli T, Rossi G et al. Effects of gestational diabetes
metabolic diseases. Epigenetics September 2013; 8(9): 935–943. on fetal oxygen and glucose levels in vivo. BJOG December 2009;
11. Cheng X, Chapple SJ, Patel B et al. Gestational diabetes impairs 116(13): 1729–1735.
Nrf2-mediated adaptive antioxidant defenses and redox signal- 24. Taricco E, Radaelli T, Nobile de Santis MS et al. Foetal and pla-
ing in fetal endothelial cells in utero. Diabetes December 2013; cental weights in relation to maternal characteristics in gesta-
62(12): 4088–4097. tional diabetes. Placenta April 2003; 24(4): 343–347.
12. Catalano PM, Tyzbir ED, Wolfe RR et al. Carbohydrate metabo- 25. Lao TT, Lee CP, Wong WM. Placental weight to birthweight ratio
lism during pregnancy in control subjects and women with ges- is increased in mild gestational glucose intolerance. Placenta
tational diabetes. Am Physiol Soc 1993; 264: E60–E67. March–April 1997; 18(2–3): 227–230.
13. Catalano PM, Huston L, Amini SB et al. Longitudinal changes 26. Evers IM, Nikkels PG, Sikkema JM et al. Placental pathology in
in glucose metabolism during pregnancy in obese women with women with type 1 diabetes and in a control group with nor-
normal glucose tolerance and gestational diabetes mellitus. Am J mal and large-for-gestational-age infants. Placenta September–
Obstet Gynecol 1999; 180: 903–916. October 2003; 24(8–9): 819–825.
14. Luoto R, Kinnunen TI, Aittasalo M et al. Primary prevention of ges- 27. Björk O, Persson B. Villous structure in different parts of the
tational diabetes mellitus and large-for-gestational-age newborns cotyledon in placentas of insulin-dependent diabetic women. A
by lifestyle counseling: A cluster-randomized controlled trial. PLoS morphometric study. Acta Obstet Gynecol Scand 1984; 63(1):
Med May 2011; 8(5): e1001036. 37–43.
15. Bartha JL, Martinez-Del-Fresno P, Comino-Delgado R. Gestational 28. Jones CJ, Fox H. Placental changes in gestational diabetes. An
diabetes mellitus diagnosed during early pregnancy. Am J Obstet ultrastructural study. Obstet Gynecol September 1976; 48(3):
Gynecol 2000; 182: 346–350. 274–280.
13 Placental origins of diabesity
and the origin of preeclampsia
Gernot Desoye and Berthold Huppertz
changes, which are essential for determining the fetal

Introduction phenotype in maternal diabesity, while intrinsic pla-
The placenta is a fetal organ that early in gestation rap- cental changes may determine the maternal phenotype
idly develops through a series of essential steps. These (preeclampsia) and/or the fetal phenotype (FGR). All of
encompass these changes will have long-term consequences for the
fetus and childhood development as well as maternal
●● Anchoring of the placenta to the maternal decidua health later in life.
●● Opening of uterine glands to enable nutrient supply of
the embryo during the first trimester of pregnancy, prior
to onset of maternal blood flow Diabesity
●● Subsequent opening of spiral arteries to increase oxygen
and nutrient supply to the fetus starting with the begin- Diabetes and obesity share several features such as insulin
ning of the second trimester resistance, dyslipidemia, and inflammation. The term “dia-
●● Establishing villous structures and the first primi- besity” was coined to reflect these commonalities. The prev-
tive vascular plexus within the placenta independent of alence of diabesity is increasing dramatically.1,2 Especially
vasculogenesis within the embryo proper alarming is that the percentage of overweight or obese chil-
dren is now above 20% in Europe.3 Therefore, childhood
Given the necessity to rapidly establish the uteroplacen- obesity is one of the greatest current challenges.4 It leads to
tal vascular connection as well as the vasculature between an increasing number of individuals developing type 2 dia-
embryo and placenta and because of the complex nature betes mellitus (T2DM) and other metabolic diseases early in
of the biological processes driving implantation, placenta- life5 and therefore has become of considerable public health
tion, and embryo development, early gestation is a period concern.3 Childhood obesity in girls increases their risk to
of high vulnerability of both the placenta and embryo. develop diabesity prior to or within their pregnancy giv-
Although tightly controlled by a variety of mechanisms, ing birth to offspring with again a higher risk for develop-
each of these processes can malfunction, which in cer- ing diabesity later in their life, creating a self-perpetuating
tain situations may lead to impaired placental function situation.
later in pregnancy, which could manifest in pregnancy The rise in the prevalence and severity of early-life dia-
conditions such as preeclampsia and/or fetal growth besity has a multifactorial background, where complex
restriction (FGR). interactions between genetic and environmental factors
Its localization between the maternal and fetal blood contribute.6 These interactions are dynamic with a strong
systems makes obvious the essential contribution of the effect during critical periods of development in prenatal and
placenta to the fetal environment and thus fetal growth and early postnatal life. The first mechanistic principle of this
development by mediating transfer of maternal nutrients. environment-dependent programming with a long-term
As the placenta is exposed to changes in both the maternal impact on health later in life was postulated in the 1970s of
and the fetal circulation, one can expect that placental struc- the last century.7 Since then, more detailed concepts have
ture and function are altered in pregnancy conditions asso- been developed trying to explain the association between
ciated with changes in the maternal environment such as in later disease risk and fetal overnutrition (“fuel-mediated
maternal diabesity. in utero hypothesis”), fetal undernutrition and postna-
This chapter will discuss the intense interplay tal overnutrition (“mismatch hypothesis”), and postnatal
between the mother and fetus and the role of the placenta overnutrition (“accelerated postnatal growth hypothesis”).8
as mediator. Maternal changes may lead to placental Recently, a unifying concept has been proposed.9 All these
100
Transplacental transport of nutrients in diabesity 101
concepts ultimately explain why and how diabetes and/or Transplacental transport of nutrients
obesity may develop later in life.
A key role is attributed to the intrauterine environment, in diabesity
in which the fetus grows and develops. This intrauterine
Insulin secretagogues
environment is characterized and determined by the con-
centrations of principle nutrients, cytokines, growth factors, The placenta is richly endowed with glucose transporters
and hormones. foremost with GLUT1,36 which are dynamically regulated
across gestation.37 These confer a high glucose transport
capacity, which only becomes saturated at a maternal–fetal
glucose concentration difference of 20–25 mM d-glucose.38
Placenta in diabesity This overcapacity to transport glucose renders the placenta
How the placenta integrates maternal and fetal exposures10 relatively insensitive to glucose transporter changes as they
has been the subject of several reviews.11–24 It is unclear may occur in diabesity39 induced by hyperglycemia40,41 and
whether any of these placental changes are causally linked other components of the intrauterine environment.42 In fact,
with the fetal and neonatal phenotype of pregnancies in direct measurements of maternal-to-fetal glucose transfer
diabesity. Therefore, we will focus on the specific con- in placentas from normal and gestational diabetic pregnan-
cept that fetal and neonatal insulin plays a central role in cies did not find differences.43–45 Thus the placenta per se
determining the neonatal phenotype25 and represents the does neither limit nor promote glucose flux at least at the
mechanistic link with childhood diabesity and associated end of gestation. The amount of glucose reaching the fetal
comorbidities.26 By emphasizing on the maternal and fetal circulation is thus dictated by its maternal–fetal concentra-
glucose/insulin axis and its effects on placenta and fetus, we tion gradient with some contribution of uteroplacental and
will explain the fetal phenotype and some aspects of intra- fetoplacental blood flow.46
uterine programming of childhood obesity in the wake of Arginine uptake into various placental cells has been stud-
maternal diabesity. ied, because of its precursor role for NO biosynthesis. Glucose
as well as insulin is capable of stimulating arginine uptake
into a trophoblast model cell line,47 which may suggest altera-
The fetus/neonate in maternal diabesity tions associated with maternal diabetes mellitus. However,
trophoblast uptake does not necessarily correspond with
A growing body of evidence shows that the major difference
overall transfer. There are no data available on transplacental
in the neonatal phenotype between normal and diabetic/
arginine transport and potential changes in maternal diabetes
obese pregnancies is the fetal body composition indepen-
mellitus. Arginine concentrations in the maternal circulation
dent of birth weight. While the fat-free mass is unaltered,
are higher in diabetes mellitus as compared to nondiabetic
neonates of diabesity pregnancies contain more fat. These
pregnancies, and this likely leads to an increase in the fetus.
differences are found even when born in the appropri-
This increase must be transient, because fetal steady-state
ate-for-gestational-age range (between the 5th/10th and
concentrations of arginine are unchanged in diabetes.48 The
90th birth weight percentile).27–30 Interestingly, gestational
transient rise in fetal arginine elicits a pancreatic response as
diabetes mellitus (GDM) and maternal obesity are inde-
reflected by elevated insulin levels.49 This response is, how-
pendent risk factors for neonatal body fat and contribute
ever, limited to diabetic pregnancies, which is an indication
additively. 31
of the accelerated β-cell maturation under the conditions of
The relative body fat of the neonate associates with that
permanent overstimulation by glucose and perhaps arginine.
of the offspring at age 9 years, whereas total body weights
do not associate. 32 This reinforces the concept that the
number of adipocytes for a human being is determined
very early in the life cycle if not already in utero. 33 The tra- Lipids
jectory of adipocyte number cannot be changed not even The augmented fat accretion in offspring born to diabetes
when weight, i.e., fat, is lost. Thus, once the growing fetus or obesity pregnancies has sparked interest in the question
has developed more adipocytes in the wake of the diabesity whether there is an increased transfer of lipids across the pla-
environment, then the neonate may be strongly disposed centa under these conditions to contribute to the buildup of
to remain on this trajectory and have more adipocytes also fetal fat depots.50,51 Lipids are a complex class of molecules
in childhood. most of which contain fatty acids as esters such as triglyc-
According to the well-established Pedersen hypothesis, erides (TGs), phospholipids, and cholesterol esters. These
maternal hyperglycemia will lead to fetal hyperglycemia. As in turn are building blocks for lipoproteins, in which more
a result, fetal islets will become hypertrophic and hyperse- than 95% of fatty acids are bound.
crete insulin.34 The lipoproteins have to be hydrolyzed by lipases to release
The hyperinsulinism may not only be the result of fetal fatty acids, which is the lipid moiety transported to the fetus.
hyperglycemia, because some amino acids such as arginine For topological reasons, these lipases have to be located on
can also stimulate insulin secretion, especially from the the microvillous membrane of the syncytiotrophoblast. For
overactive fetal pancreas as in maternal diabetes.35 most fatty acids, a maternal–fetal concentration gradient
102 Placental origins of diabesity and the origin of preeclampsia
exists, and hence, free fatty acids may traverse the placenta is also unknown whether overall transfer of the lipoprotein-
by simple diffusion, but an intermediate esterification of free borne fatty acid is altered in diabesity.
fatty acids to TGs within the placenta may occur, followed
by lipolysis and release into the fetal circulation. Lipid stores
in the placenta are elevated in diabesity.52,53 It is unknown Effect of fetal hyperinsulinism
whether this affects overall fatty acid transfer.
Endothelial lipase (EL), a close relative of lipopro- Fetal hyperinsulinism as it often occurs in diabesity has sev-
tein lipase (LPL), was recently identified on the microvil- eral effects, which may not only help sustain fetal develop-
lous syncytiotrophoblast membrane at the end of gestation ment and growth but also mediate long-term consequences
(Figure 13.1), whereas LPL was found in cells subjacent to associated with intrauterine exposure to the diabesity envi-
the fetoplacental endothelium.54 EL is upregulated in placen- ronment (Figure 13.2).
tas from obese women with GDM, but unchanged in placentas
from lean GDM cases.55 This suggests that metabolic inflam-
mation together with diabetic conditions accounted for Fetal metabolism and uptake of glucose and fatty acids
dysregulation of EL in pregnancies complicated by obesity Glucose uptake into fetal tissues is stimulated by hyperin-
and GDM. TNF-α and leptin, both inflammatory cytokines sulinism as reflected by a higher glucose disappearance rate
often elevated in diabesity, are putative key regulators of EL. (“k-value”) in newborns of diabetic as compared to nondia-
After hydrolysis and release of fatty acids, they have to betic pregnancies.34 The insulin-stimulated glucose uptake
be transported across the placenta to ultimately reach the will transiently decrease glucose levels in the fetal circulation,
fetal circulation. The transport will be accomplished by a thereby steepening the maternal–fetal glucose concentration
range of fatty acid transporters present on the surface and gradient with ensuing increased glucose flux across the pla-
of fatty acid–binding proteins in the cytoplasm of placental centa until steady-state concentrations will be reached. At
cells.56–58 Each of these has a different preference for the vari- this stage, the amount of glucose passing from mother to
ous individual fatty acids and may be subject to regulation fetus is independent of the maternal glucose level and dic-
by distinct changes in the diabesity environment.59,60 tated by the fetus. Whether a similar mechanism exists also
Notwithstanding the presumed key role of EL in hydro- for fatty acids is unknown. When fatty acid profiles were
lyzing maternal lipoproteins on the placental surface and measured in the cord blood of newborn from gestational
the fatty acid transporters for transplacental movement of diabetic and nondiabetic pregnancies, then lower proportion
fatty acids, it is unclear whether total placental EL activity or of polyunsaturated fatty acids were found in gestational dia-
the transporter activity is the limiting and therefore deter- betic newborn but only in the arterial and not in the venous
mining factor for overall fatty acid transport. Moreover, it cord blood. All other fatty acids were similar.61 This would
Compartments
Maternal Stroma Fetal
Arginine Arginine
+
+
Glucose Glucose Insulin
Liver +
Lipogenesis
TG + +
Lipogenesis LPL
HDL White
Lipoproteins adipose
TG
EL tissue +
FFA FFA
Growth
Syncytiotrophoblast Endothelium
Figure 13.1 The Pedersen–Freinkel concept expanded. Maternal glucose and arginine cross the placenta and stimulate fetal insu-
lin secretion. Lipoprotein–triglycerides (TGs) and phospholipids will be hydrolyzed by lipases such as endothelial lipase present
on the microvillous membrane of the syncytiotrophoblast to release fatty acids, which then traverse the placenta to contribute to
the fetal fatty acid pool. Fetal glucose and insulin enhance hepatic TG formation. Insulin stimulates growth of adipocytes in white
adipose tissue as well as lipoprotein lipase (LPL) transcription. LPL hydrolyzes TGs released from the liver and provides free fatty
acids for uptake into adipocytes and reesterification under insulin promotion to ultimately form TGs.
Conclusions 103
Compartments
Maternal Stroma Fetal Hypothalamus Adipocytes

+ Orexigenic + Growth
– Anorexigenic + Fat deposition
neurons
Glucose Glucose Insulin
Vascularization
Liver
Glucose uptake
Oxygen
Metabolism
Syncytiotrophoblast Endothelium
Figure 13.2 Fetal hyperinsulinism leads to multiple changes. High insulin levels promote hepatic glucose metabolism and indi-
rectly glucose uptake into hepatocytes and other insulin-sensitive peripheral tissues. This will increase fetal oxygen demand for
aerobic metabolism. In a situation of inadequate oxygen supply, this results in fetal hypoxia. Insulin directly and hypoxia indirectly
stimulates angiogenesis-promoting factors, which are potent stimulators of placental angiogenesis serving to ultimately increase the
degree of placental vascularization. Insulin also acts on the adipocytes with its dual role as growth factor to stimulate adipocyte
growth and as hormone to promote triglyceride formation and fat deposition. The central role of insulin is reflected by stimulating
orexigenic and reducing anorexigenic neurons, which may contribute to neonatal overfeeding in such situations.
indicate that the newborn takes up polyunsaturated fatty end of gestation, when its levels in the umbilical cord are
acids, which are needed in particular for the development of associated with the capillary surface area.20,67 In vitro exper-
the retina and the brain, and this uptake may be under the iments using primary human fetoplacental endothelial cells
regulation of fetal insulin. Insulin also stimulates TG forma- also demonstrated insulin stimulation of several pathways
tion in the fetal liver and in the white adipocytes. involved in angiogenesis.68,69
Fetal hypoxia
Fetal hypothalamus
The insulin-stimulated glucose uptake into peripheral tissues
The human brain contains insulin receptors in various
and, in the presence of hyperglycemia, also into the liver is
regions.70 During fetal development, insulin binding to brain
accompanied by enhanced aerobic glucose metabolism gen-
receptors increases as the fetus grows, i.e., with advancing ges-
erating an increased demand for oxygen. At the same time,
tational age.71 One of the sites in the brain susceptible to insu-
maternal and placental oxygen supply is impaired, because
lin action is the hypothalamus. Insulin exerts neurotrophic
of more maternal hemoglobin being present in the glyco-
functions and contributes to the organization of neuronal
sylated form and thickening of the basement membrane of
networks during critical periods of hypothalamic develop-
the syncytiotrophoblast. Occasionally, this is aggravated by
ment.72–74 Thus, fetal hyperinsulinism is likely to affect the
reduced uteroplacental and fetoplacental blood flow. Overall
development of hypothalamus. Indeed in rat experiments,
this will lead to an imbalance of oxygen demand and supply
fetal hyperinsulinism was accompanied by increased insu-
and chronic fetal hypoxia. A reduction in umbilical oxygen
lin concentrations in the immature brain75 with subsequent
saturation and oxygen content,62 enhanced erythropoiesis,63
permanent dysplasia and neuronal hypotrophy of the ventro-
and elevated cord blood erythropoietin levels64 are often
medial hypothalamic nucleus.9 The prolonged intrauterine
manifestations of inadequate oxygen supply.
exposure to high insulin levels ultimately induces hypotha-
lamic insulin resistance. Since insulin also serves as satiety
Placental vascularization signal, this may contribute to neonatal overfeeding observed
The placenta as a fetal tissue receives fetal signals and adapts in some mother–offspring pairs. Most recently, maternal
its development and function according to fetal needs. This hyperglycemia during an oral glucose tolerance test (oGTT)
explains why the placenta is often hypervascularized in was associated with slower fetal brain responses.76
maternal diabetes.13 Corresponding data for maternal obe-
sity are missing. Mechanistically, low oxygen levels may
upregulate hypoxia-sensitive proangiogenic factors such as Conclusions
fibroblast growth factor 2, which promotes sprouting angio-
genesis and is increased in the placenta and fetal cord blood Collectively, fetal hyperinsulinism associated with mater-
in diabetic pregnancies.65,66 Fetal insulin may also be capable nal diabesity not only modifies the classical peripheral
of directly stimulating placental angiogenesis at least at the insulin-sensitive tissues but also has a broad spectrum of
central effects (Figure 13.2). These effects have been postu- 10%–15% of maternal deaths, 12% of infants born SGA (small
lated to contribute to the permanently increased disposition for gestational age), and up to 25% of stillbirths and neonatal
to adiposity and associated comorbidities such as T2DM.9 mortality rates.79 Although preeclampsia is pregnancy spe-
Conceptually, this links the placenta and its absence of cific, it has been shown to have long-term consequences for
mechanism to protect the fetus from nutritional overload those women who experienced preeclampsia. Such women
associated with diabesity, with the medium- to long-term are at increased risk of a variety of diseases including chronic
offspring risk for diabesity, and may help explain the trans- hypertension, diabetes mellitus, ischemic heart disease, cere-
generational transmission of diabesity risk. brovascular disease, kidney disease, thromboembolism,
hypothyroidism, and even impaired memory.80
Origin of preeclampsia Stratification of preeclampsia

Definition of preeclampsia During the last decade, efforts have been undertaken to strat-
ify preeclampsia into various subtypes and trying to define
Since the cause, development, and etiology of preeclampsia are
groups that are more homogeneous in terms of severity,
still far from being clear in a variety of aspects, the syndrome
marker expression, and onset of symptoms. One approach
of preeclampsia remains a topic that generates new hypoth-
used blood pressure values to differentiate between a mild
eses and new research data on a weekly basis. Although the
form of preeclampsia (systolic blood pressure between 140
real origin of preeclampsia is still a mystery, it is generally
and 160 mmHg and/or diastolic blood pressure between 90
accepted that the placenta rather than the fetus is mandatory
and 110 mmHg) and a severe form of the syndrome (systolic
for this syndrome to develop. Hence, preeclampsia is based
blood pressure ≥160 mmHg and/or diastolic blood pressure
on a dysregulation of the placenta, which in turn leads to the
≥110 mmHg). Another approach has used gestational age to
maternal symptoms that define preeclampsia. Here, the inter-
define an early-onset type (before 34 weeks of gestation) and
play between factors released from a dysfunctioning placenta
a late-onset type (after 34 weeks of gestation).81
on one hand and the maternal response on the other hand is
The latter approach in combination with other studies has
based on the susceptibility of the mother to placenta-derived
led to the idea that the early-onset type of preeclampsia is
factors. This complex interplay finally defines whether or
completely different to the late-onset type of the syndrome.
not a pregnant woman develops preeclampsia. Furthermore,
At the same time, it has to be stressed that the maternal symp-
time of onset, severity, and progression of preeclampsia into
toms of both types are identical and that mild and severe cases
eclampsia all depend on this interaction between the mother
of preeclampsia can be found in both types. The major differ-
and placenta. Thus, even if the origin of preeclampsia lies
ence between the two types—besides the different gestational
within the placenta, the interactions between the fetus, pla-
age at onset—is the large overlap with FGR in the early-onset
centa, and mother need to be taken into account to define the
group82 (Figure 13.3). Hence, the features of preeclampsia per
factors crucial for the onset of maternal symptoms.
se are quite similar in both groups—if the other conditions
Preeclampsia is specific for pregnancy, defined as occur-
that cluster around the diagnosis of preeclampsia and that
ring after 20 weeks of gestation and is characterized by the
lead to adverse outcomes are not taken into account.83 Today,
onset of proteinuria and hypertension in a woman with
the notion whether or not the two subtypes are different in
so far normal blood pressure.77 Proteinuria is defined as
terms of their origin is still discussed controversially.
≥0.3 g protein in 24-hour urine, or 0.1 g/L (>2+ on dipstick),
while hypertension is defined as systolic blood pressure of
≥140 mmHg and/or diastolic blood pressure of ≥90 mmHg.77 Late-onset type of preeclampsia
Additionally, multiple features on top of proteinuria and Worldwide most of the preeclampsia cases (80% and more84)
hypertension may occur and are mostly restricted to the can be attributed to the late-onset type. The onset of symp-
more severe types of the syndrome, including cerebral or toms of such cases occurs at a gestational age around the
visual disturbances, headache, nausea and vomiting, epi- normal time of delivery (≥37 weeks of gestation), with most
gastric or right upper quadrant pain, pulmonary edema of the neonates showing normal growth without any signs
or cyanosis, oliguria (less than 500 mL urine in 24 hours), of growth restriction (Figure 13.3). Also the placenta mostly
impaired liver function, and thrombocytopenia.77 does not show any gross morphological changes and displays
The only cure for preeclampsia is still the delivery of the normal values for volume and weight. Blood flow from the
baby (and hence the placenta). So far there is no effective mother toward the placenta as assessed by uterine artery
interventional treatment to prevent or even treat preeclamp- blood flow and blood flow from the fetus toward the placen-
sia. Hence, there is the danger for a woman suffering from tas as assessed by umbilical artery blood flow are mostly nor-
preeclampsia to progress into eclampsia, a life-threatening mal in those cases of late-onset preeclampsia as well.
exaggeration associated with tonic–clonic seizures, stroke, Hence, it seems as if the late-onset type of preeclampsia—
and death. although originating from factors derived from the placenta—
Preeclampsia is one of the major reasons for maternal, fetal, is a syndrome that only affects the mother rather than the
and neonatal mortality and morbidity, and worldwide pre- fetus. Also, pregnancies with an enlarged mass and/or surface
eclampsia affects about 2%–8% of all pregnancies, especially of the placenta, such as in diabetes mellitus, multiple preg-
in developing countries.77,78 Here, preeclampsia accounts for nancies, anemia, or high altitude, tend to have an increased
Origin of preeclampsia 105
Early and late onset preeclampsia
Normal mother Normal mother Predisposed mother Predisposed mother

+ + + +
Normal placenta Abnormal placenta Normal placenta Abnormal placenta
Normal outcome Mild or severe Mild or severe Severe

Preeclampsia Preeclampsia Preeclampsia
Normal risk of woman later in life Higher risk of woman later in life
Figure 13.3 Origin of preeclampsia. Representation of a very simplified scheme showing the players and combinations leading to
preeclampsia. While the combination of a normal mother with a normal placenta should end in a normal outcome, all of the other
combinations may result in preeclampsia. If one player is normal, the syndrome may be mild or severe, while the combination of
both being abnormal or predisposed mostly results in severe cases of preeclampsia. In these scenarios, a predisposed mother will
also carry a higher risk for, e.g., cardiovascular diseases later in life.
risk to develop the late-onset type of preeclampsia.85 Here However, features thought to be specific for early-onset pre-
the question arises whether there needs to be a dysregula- eclampsia such as changes of blood flow within the maternal
tion of placental development at all to cause this type of uterine arteries or even abnormal blood flow in the umbilical
preeclampsia. There may well be cases where an increased arteries are regular findings in cases with pure FGR.
maternal susceptibility in combination with an inadequate Hence, a dispute arose whether these findings are directly
maternal response to placenta-derived factors could lead to related to early-onset preeclampsia or whether there is an over-
preeclampsia without any dysfunctional placental deteriora- lap of two independent syndromes, FGR and preeclampsia.82
tion.82 However, respective data are missing. From a clinical point of view, this dispute is of no relevance.
In a clinical setting, it is important to do the best for both,
Early-onset type of preeclampsia mother and child. Here it does not play a role whether there
are two independent syndromes or just one with features of
Different to the late-onset type of preeclampsia, the early-
both. However, scientifically it indeed plays an important
onset type of the syndrome only develops in a small propor-
role since only a clear demarcation between FGR and pre-
tion of all preeclampsia cases, in 5%–20%84 (Figure 13.4).
eclampsia will help to identify the origins of both syndromes.
Early-onset preeclampsia only affects 1 in 200 pregnancies;
however, at the same time, it is responsible for the majority of
fetal and maternal morbidity and mortality in the developed Definition of FGR
world that are related to preeclampsia.86 FGR affects 5% of all pregnancies and is one of the leading
It is not only the gestational age at the onset of symptoms causes of perinatal mortality and morbidity.87–89 In litera-
that makes the early-onset type different to the late-onset ture, several terms are used in parallel, leading to massive
type. The early-onset cases are mostly associated with growth confusion. The terms fetal growth restriction and intrauter-
restriction of the fetus. This is why the early-onset type is rec- ine growth restriction (IUGR) are synonymous and can be
ognized as showing features not present in the late-onset type. used to describe the same issue: a fetus who has not been
Preeclampsia
Weeks of gestation 0 20 34 40
Start of clinical symptoms

33% 80%
20% 66%
Early onset preeclampsia
Early onset FGR PE FGR Overlap of
PE and FGR
Late onset preeclampsia
Late onset FGR/SGA PE FGR/ Overlap of
SGA PE and FGR/SGA
Figure 13.4 Early-onset versus late-onset preeclampsia. Schematic representation of the length of pregnancy in weeks (0–40) with
two specific time points: 20 weeks of gestation as the start of clinical symptoms of preeclampsia (per definition) and 34 weeks to
discriminate between early-onset (prior to 34 weeks) and late-onset (after 34 weeks) preeclampsia. The two types of preeclampsia and
fetal growth restriction (FGR) are further characterized by their occurrence rates in all preeclampsia and FGR cases (% values) and
also by their massive overlap in early-onset cases and their negligible overlap in late-onset cases. In late-onset cases, it is still very dif-
ficult to achieve a clear distinction between FGR and small-for-gestational-age cases without growth restriction.
able to achieve its growth potential.90 These two terms (fetal Doppler ultrasound between 11 and 14 weeks of gestation.
growth restriction and intrauterine growth restriction) are These authors were able to calculate a sensitivity to predict
regularly mixed with the term small for gestational age and early-onset preeclampsia of only 33.3%, while the sensitiv-
are even used as synonyms. However, fetuses who are SGA ity to predict all cases of preeclampsia (early and late onset)
are defined being born below a certain growth percentile further decreased to 21%. At the same time, the sensitivity
(e.g., below the 10th percentile). Not all of such fetuses are to predict early-onset FGR was 100% using the same pool
also growth restricted since there are fetuses with a normal of women.98
growth trajectory but who are constitutionally small while
otherwise normal.89,90 Ott has calculated that of 100 SGA
babies with a growth below the 10th percentile, only 30 are Etiology of preeclampsia
growth restricted and hence are FGR.91 The etiology of preeclampsia is still unclear, but as described
Furthermore, FGR can also appear with a baby being born earlier, the placenta is essential. Without the placenta, pre-
above the 10th percentile. In such cases, the growth trajec- eclampsia does not develop. Only the presence of at least one
tory of a fetus did not remain constant over the duration of of the two, defective placenta and/or a predisposed mother,
pregnancy but rather massively declined (e.g., from the 70th may result in preeclampsia.82
percentile to the 30% percentile). Also such newborn babies Even today it is more than difficult not only to exactly
are growth restricted without being SGA.90,92 define the syndrome of preeclampsia but also to define the
individual that is responsible for the development of the
syndrome (mother or fetus). Thus, preeclampsia may derive
from at least three different combinations of mother and
Features of preeclampsia and FGR fetus/placenta (Figure 13.5).
The clinically relevant features of early-onset preeclamp-
sia (excluding hypertension and proteinuria of the mother) Cases with a normal placenta and a predisposed mother
may be caused by preeclampsia or may be the features of the Here the mother suffers from an inadequate response system
overlying syndrome FGR.82 This deleterious combination of and may simply be overloaded by the normal release of fac-
early-onset FGR and early-onset preeclampsia dramatically tors from the placenta. Today, a number of factors leading to
increases the risk for both, mother and baby. This is why predisposition are defined as risk factors to increase the risk to
these cases are highly relevant for clinical intervention. develop preeclampsia including chronic hypertension or renal
The idea that there are two overlapping syndromes rather disease.
than a single syndrome with both features is supported by a
variety of studies and data:
Preeclampsia and FGR
1. In cases with preeclampsia, the placenta does not dis- Normal or predisposed mother + Abnormal placenta
play any gross morphological changes and also shows Villous trophoblast Extravillous trophoblast
normal values for volume and weight. Of course, a Affected Normal Normal Affected
macroscopically normal placenta does not necessar-
ily reflect a histologically and functionally normal pla-
centa. Looking into the morphology of the placenta,
also here preeclampsia has only modest effects, if at all.
There have been a number of variables that have been Preeclampsia Preeclampsia and FGR FGR
quantified using stereological techniques, including
placental weight and volume, length, diameter, volume Normal growth Reduced growth of fetus
and surface area of villi, villous components and vascu- of fetus (no FGR) (FGR)
lature,93–95 and villous domains and intervillous pores.96
Normal levels of Higher levels of
All of the aforementioned measures were not different angiogenic factors angiogenic factors
between healthy control placentas and those from pre- (due to absence of FGR) (due to presence of FGR)
eclampsia cases (without FGR). In the presence of FGR,
these measures were significantly different, whether or Higher risk of woman Higher risk of newborn
not such cases were further suffering from preeclampsia. later in life later in life
2. Studies predicting preeclampsia and/or FGR using Doppler (if predisposed)
ultrasound to assess blood flow in uterine arteries showed Figure 13.5 Combinations of fetal growth restriction (FGR)
that again it is FGR rather than preeclampsia that is the and preeclampsia. As described earlier,82 preeclampsia could
effective component here. Nicolaides et al.97 have assessed be derived from a dysfunctional villous trophoblast, while FGR
uterine artery blood flow between 11 +0 and 13 +6 weeks could be derived from a maldeveloped extravillous tropho-
by Doppler ultrasound. These authors were able to calcu- blast. The combinations of these defects in a single placenta
determine whether only preeclampsia, only FGR, or the com-
late a detection rate for early-onset preeclampsia of only bination of both develops. This may have direct effects on fetal
40% at a 10% false-positive rate.97 Pilalis et al.98 used a simi- growth and outcome, changes in biomarkers such as angio-
lar approach and assessed uterine artery blood flow using genic factors, and long-term effects on the mother or offspring.
References 107
The overload of the maternal system may develop slowly symptoms will appear early in pregnancy (early-onset pre-
resulting in normally grown babies (no FGR), clinical eclampsia) with no effect on the growth of the fetus (no FGR)
symptoms appearing mostly late in pregnancy (late-onset and the absence of changes in predictive angiogenic markers.
preeclampsia), and the absence of changes of predictive However, a larger proportion of these maldeveloped pla-
angiogenic markers such as PlGF and sFlt-1 prior to the centas may also show defects in extravillous trophoblast
onset of symptoms.99 development and hence may additionally result in impaired
Such women may well be the ones who suffer from a trophoblast invasion.105 In those cases, the villous syncy-
higher risk to develop cardiovascular diseases later in life. tiotrophoblast is affected and releases factors by necrotic
and aponecrotic shedding, while the maldevelopment of
Cases with a dysfunctioning placenta and a normal the extravillous trophoblast will alter maternal blood flow
mother toward the placenta106 subsequently impairing fetal growth.
Here, it seems as if mostly the villous syncytiotrophoblast is Again, the overload of the maternal system will develop early
affected, releasing factors systemically harming the mater- in pregnancy (early-onset preeclampsia), growth of the fetus
nal vascular system during pregnancy finally leading to the will be impaired as well (FGR), and changes in predictive
clinical symptoms of preeclampsia. The syncytiotrophoblast angiogenic markers will be quite striking.107
does no longer release most of the factors by controlled secre- Again, such women may well suffer from a higher risk to
tion and apoptotic shedding. Rather, necrosis and aponecro- develop cardiovascular diseases and other morbidities later
sis82,100,101 result in the release of subcellular fragments and in life. In cases with FGR, the babies may show fetal pro-
micro- and nanoparticles102,103 systemically affecting the gramming as well.108,109
maternal endothelium.104
The normal maternal system may resist any deleteri-
ous changes for a longer time period, and hence clinical Conclusions
symptoms may develop only later in pregnancy (late-onset
preeclampsia) and in the absence of changes of predictive Preeclampsia and FGR are different syndromes that have
angiogenic markers such as PlGF and sFlt-1.99 Also the different etiologies. Both may start as impaired development
fetus may develop into a normally grown baby (no FGR). of the placenta affecting different cell populations and tis-
The mother may recover completely after delivery without sues within the placenta. FGR may develop on its own only
any further long-term effects. affecting the fetus without any disturbance found in the
mother. The same is true for preeclampsia that only affects
Cases with a maldeveloped and dysfunctioning placenta the mother without any changes of the fetus. It has to be con-
and a predisposed mother sidered that preeclampsia may be solely derived from a pre-
These are the most severe cases since the combination of disposed mother who cannot response to normally released
both defects will definitively increase the severity of the placental factors in a proper way and hence develops pre-
symptoms of the mother. eclampsia in the presence of a normal placenta.
A small proportion of these maldeveloped placentas may The different origins and pathways of the etiology of
only show alterations of the syncytiotrophoblast. In those preeclampsia in combination with the simultaneous occur-
cases, the villous syncytiotrophoblast is affected and releases rence of FGR clearly hinder the further progress in identify-
factors by necrotic and aponecrotic shedding, subsequently ing the origins of the syndrome. It would be ideal to have
affecting a maternal system that is already impaired. The over- markers in hand distinguishing between the different ori-
load of the maternal system will develop fast, and hence clinical gins and combinations of preeclampsia.
REFERENCES
1. Zimmet P, Thomas CR. Genotype, obesity and cardiovascular 8. Koletzko B, Symonds ME, Olsen SF et al. Programming research:
disease—Has technical and social advancement outstripped Where are we and where do we go from here? Am J Clin Nutr
evolution? J Intern Med 2003; 254(2): 114–125. 2011; 94(6 Suppl.): 2036S–2043S.
2. Zimmet PZ, Magliano DJ, Herman WH et al. Diabetes: A 21st 9. Plagemann A. Toward a unifying concept on perinatal program-
century challenge. Lancet Diabetes Endocrinol 2014; 2(1): 56–64. ming: Vegetative imprinting by environment-dependent biocy-
3. Wijnhoven TM, van Raaij JM, Spinelli A et al. WHO European bernetogenesis. In: Plagemann A, ed., Perinatal Programming:
Childhood Obesity Surveillance Initiative 2008: Weight, height The State of the Art, Walter DeGruyter: Berlin, Germany, 2012,
and body mass index in 6–9-year-old children. Pediatr Obes pp. 242–282.
2013; 8(2): 79–97. 10. Lewis RM, Demmelmair H, Gaillard R et al. The placental expo-
4. Katzmarzyk PT, Shen W, Baxter-Jones A et al. Adiposity in chil- some: Placental determinants of fetal adiposity and postnatal
dren and adolescents: Correlates and clinical consequences body composition. Ann Nutr Metab 2013; 63(3): 208–215.
of fat stored in specific body depots. Pediatr Obes 2012; 7(5): 11. Desoye G, Hauguel-de Mouzon S, Shafrir E. The placenta in dia-
e42–e61. betic pregnancy. In: Hod M, Jovanovic L, DiRenzo GC et al., eds.,
5. Amed S, Daneman D, Mahmud FH et al. Type 2 diabetes in children Textbook of Diabetes and Pregnancy, Martin Dunitz: London,
and adolescents. Expert Rev Cardiovasc Ther 2010; 8(3): 393–406. U.K., 2003, pp. 126–147.
6. Biro FM, Wien M. Childhood obesity and adult morbidities. Am J 12. Desoye G, Myatt L. The placenta. In: Reece EA, Coustan DR,
Clin Nutr 2010; 91(5): 1499S–1505S. Gabbe SG, eds., Diabetes Woman—Adolescence, Pregnancy
7. Dorner G. Environment-dependent brain differentiation and funda- and Menopause, 3rd ed., Williams & Wilkins: Philadelphia, PA,
mental processes of life. Acta Biol Med Ger 1974; 33(2): 129–148. 2004, pp. 147–157.
13. Desoye G, Shafrir E. The human placenta in diabetic pregnancy. 38. Hauguel S, Desmaizieres V, Challier JC. Glucose uptake, utiliza-
Diabetes Rev 1996; 4: 70–89. tion, and transfer by the human placenta as functions of maternal
14. Desoye G, Shafrir E. Placental metabolism and its regulation in glucose concentration. Pediatr Res 1986; 20(3): 269–273.
health and diabetes. Mol Aspects Med 1994; 15(6): 505–682. 39. Gaither K, Quraishi AN, Illsley NP. Diabetes alters the expression
15. Hiden U, Froehlich J, Desoye G. Diabetes and the placenta. and activity of the human placental GLUT1 glucose transporter.
In: Kay H, Neslon DM, Wang Y, eds., The Placenta: From J Clin Endocrinol Metab 1999; 84(2): 695–701.
Development to Disease, 1st ed., Wiley-Blackwell: Oxford, U.K., 40. Hahn T, Barth S, Weiss U et al. Sustained hyperglycemia in vitro
2011, pp. 228–236. down-regulates the GLUT1 glucose transport system of cultured
16. Hiden U, Desoye G. Insulin and the placenta in GDM. In: Kim C, human term placental trophoblast: A mechanism to protect fetal
Ferrara A, eds., Diabetes in and after Pregnancy, Springer: New development? FASEB J 1998; 12(12): 1221–1231.
York, 2010, pp. 97–111. 41. Hahn T, Hahn D, Blaschitz A et al. Hyperglycaemia-induced sub-
17. Hiden U, Desoye G. The placenta in diabetic pregnancy. In: cellular redistribution of GLUT1 glucose transporters in cultured
McCance DR, Maresh M, Sacks D, eds., Practical Manual of human term placental trophoblast cells. Diabetologia 2000;
Diabetes in Pregnancy, Wiley Blackwell: Oxford, U.K., 2010, 43(2): 173–180.
pp. 26–33. 42. Hahn T, Barth S, Graf R et al. Placental glucose transporter
18. Desoye G, Shafrir E, Hauguel de Mouzon S. The placenta in diabetic expression is regulated by glucocorticoids. J Clin Endocrinol
pregnancy: Placental transfer of nutrients. In: Hod M, Jovanovic L, Metab 1999; 84(4): 1445–1452.
DiRenzo GC et al., eds., Textbook of Diabetes in Pregnancy, 2nd 43. King RG, Osmond DT, Brennecke SP et al. Effect of fetal macro-
ed., Informa Healthcare: London, U.K., 2008, pp. 47–56. somia on human placental glucose transport and utilization in
19. Desoye G, Hauguel-de Mouzon S. The human placenta in ges- insulin-treated gestational diabetes. J Perinatal Med 2003; 31(6):
tational diabetes mellitus. The insulin and cytokine network. 475–483.
Diabetes Care 2007; 30(Suppl. 2): S120–S126. 44. Osmond DT, King RG, Brennecke SP et al. Placental glucose
20. Gauster M, Desoye G, Totsch M et al. The placenta and gesta- transport and utilisation is altered at term in insulin-treated, ges-
tional diabetes mellitus. Curr Diab Rep 2012; 12(1): 16–23. tational-diabetic patients. Diabetologia 2001; 44(9): 1133–1139.
21. Hiden U, Desoye G. The placenta in a diabetic pregnancy. 45. Osmond DT, Nolan CJ, King RG et al. Effects of gestational diabe-
J Reprod Med Endocrinol 2010; 7: 27–33. tes on human placental glucose uptake, transfer, and utilisation.
22. Lappas M, Hiden U, Desoye G et al. The role of oxidative stress Diabetologia 2000; 43(5): 576–582.
in the pathophysiology of gestational diabetes mellitus. Antioxid 46. Illsley NP, Hall S, Stacey T. The modulation of glucose transfer
Redox Signal 2011; 15(12): 3061–3100. across the human placenta by intervillous flow rates: An in vitro
23. Higgins L, Greenwood SL, Wareing M et al. Obesity and the pla- perfusion study. Trophoblast Res 1987; 2: 535–544.
centa: A consideration of nutrient exchange mechanisms in rela- 47. Eaton BM, Sooranna SR. In vitro modulation of l-arginine trans-
tion to aberrant fetal growth. Placenta 2011; 32(1): 1–7. port in trophoblast cells by glucose. Eur J Clin Invest 1998; 28(12):
24. Hauguel-de Mouzon S, Guerre-Millo M. The placenta cyto- 1006–1010.
kine network and inflammatory signals. Placenta 2006; 27(8): 48. Cetin I, de Santis MS, Taricco E et al. Maternal and fetal amino
794–798. acid concentrations in normal pregnancies and in pregnancies
25. Freinkel N. Banting Lecture 1980. Of pregnancy and progeny. with gestational diabetes mellitus. Am J Obstet Gynecol 2005;
Diabetes 1980; 29(12): 1023–1035. 192(2): 610–617.
26. Dorner G, Mohnike A, Honigmann G et al. Possible significance 49. Fallucca F, Sabbatini A, Di Biase N et al. Fetal pancreatic function
of prenatal hyperinsulinism for postnatal development of diabe- in infants of diabetic and rhesus-isoimmunized women. Obstet
tes mellitus. Endokrinologie 1973; 61(3): 430–432. Gynecol 2000; 95(2): 195–198.
27. Durnwald C, Huston-Presley L, Amini S et al. Evaluation of body 50. Cetin I, Parisi F, Berti C et al. Placental fatty acid transport in
composition of large-for-gestational-age infants of women with maternal obesity. Dev Orig Health Dis 2012; 3: 409–414.
gestational diabetes mellitus compared with women with nor- 51. Desoye G, Gauster M, Wadsack C. Placental transport in
mal glucose tolerance levels. Am J Obstet Gynecol 2004; 191(3): pregnancy pathologies. Am J Clin Nutr 2011; 94(6 Suppl.):
804–808. 1896S–1902S.
28. Catalano PM, Thomas A, Huston-Presley L et al. Phenotype of 52. Diamant YZ, Metzger BE, Freinkel N et al. Placental lipid and
infants of mothers with gestational diabetes. Diabetes Care 2007; glycogen content in human and experimental diabetes mellitus.
30(Suppl. 2): S156–S160. Am J Obstet Gynecol 1982; 144(1): 5–11.
29. Ortega-Senovilla H, Schaefer-Graf U, Meitzner K et al. Decreased 53. Saben J, Lindsey F, Zhong Y et al. Maternal obesity is associated
concentrations of the lipoprotein lipase inhibitor angiopoietin-like with a lipotoxic placental environment. Placenta 2014; 35(3):
protein 4 and increased serum triacylglycerol are associated with 171–177.
increased neonatal fat mass in pregnant women with gestational 54. Gauster M, Hiden U, Blaschitz A et al. Dysregulation of placental
diabetes mellitus. J Clin Endocrinol Metab 2013; 98(8): 3430–3437. endothelial lipase and lipoprotein lipase in intrauterine growth-
30. Sewell MF, Huston-Presley L, Super DM et al. Increased neonatal restricted pregnancies. J Clin Endocrinol Metab 2007; 92(6):
fat mass, not lean body mass, is associated with maternal obesity. 2256–2263.
Am J Obstet Gynecol 2006; 195(4): 1100–1103. 55. Gauster M, Hiden U, van Poppel M et al. Dysregulation of pla-
31. Catalano PM, McIntyre HD, Cruickshank JK et al. The hyper- cental endothelial lipase in obese women with gestational diabe-
glycemia and adverse pregnancy outcome study: Associations tes mellitus. Diabetes 2011; 60(10): 2457–2464.
of GDM and obesity with pregnancy outcomes. Diabetes Care 56. Larque E, Demmelmair H, Klingler M et al. Expression pattern
2012; 35(4): 780–786. of fatty acid transport protein-1 (FATP-1), FATP-4 and heart-fatty
32. Catalano PM, Farrell K, Thomas A et al. Perinatal risk factors for acid binding protein (H-FABP) genes in human term placenta.
childhood obesity and metabolic dysregulation. Am J Clin Nutr Early Hum Dev 2006; 82(10): 697–701.
2009; 90(5): 1303–1313. 57. Gil-Sanchez A, Koletzko B, Larque E. Current understanding of
33. Spalding KL, Arner E, Westermark PO et al. Dynamics of fat cell placental fatty acid transport. Curr Opin Clin Nutr Metab Care
turnover in humans. Nature 2008; 453(7196): 783–787. 2012; 15(3): 265–272.
34. Pedersen J. Diabetes mellitus and pregnancy: Present status of 58. Duttaroy AK. Transport of fatty acids across the human placenta:
the hyperglycaemia-hyperinsulinism theory and the weight of the A review. Prog Lipid Res 2009; 48(1): 52–61.
newborn baby. Postgrad Med J 1971; (Suppl.): 66–67. 59. Haggarty P. Placental regulation of fatty acid delivery and its
35. King KC, Adam PA, Yamaguchi K et al. Insulin response to argi- effect on fetal growth—A review. Placenta 2002; 23(Suppl. A):
nine in normal newborn infants and infants of diabetic mothers. S28–S38.
Diabetes 1974; 23(10): 816–820. 60. Pagan A, Prieto-Sanchez MT, Blanco-Carnero JE et al. Materno-
36. Illsley NP. Glucose transporters in the human placenta. Placenta fetal transfer of docosahexaenoic acid is impaired by gestational
2000; 21(1): 14–22. diabetes mellitus. Am J Physiol Endocrinol Metab 2013; 305(7):
37. Novakovic B, Gordon L, Robinson WP et al. Glucose as a fetal E826–E833.
nutrient: Dynamic regulation of several glucose transporter genes 61. Ortega-Senovilla H, Alvino G, Taricco E et al. Gestational diabe-
by DNA methylation in the human placenta across gestation. tes mellitus upsets the proportion of fatty acids in umbilical arte-
J Nutr Biochem 2013; 24(1): 282–288. rial but not venous plasma. Diabetes Care 2009; 32(1): 120–122.
References 109
62. Taricco E, Radaelli T, Rossi G et al. Effects of gestational diabetes 87. Wolfe HM, Gross TL. Increased risk to the growth retarded fetus.
on fetal oxygen and glucose levels in vivo. BJOG 2009; 116(13): In: Gross TL, Sokol RJ, eds., Intrauterine Growth Retardation:
1729–1735. A Practical Approach, Year Book Medical Publishers: Chicago,
63. Yeruchimovich M, Dollberg S, Green DW et al. Nucleated red IL, 1989, pp. 111–124.
blood cells in infants of smoking mothers. Obstet Gynecol 1999; 88. Neerhof MG. Causes of intrauterine growth restriction. Clin
93(3): 403–406. Perinatol 1995; 22(2): 375–385.
64. Teramo KA, Widness JA. Increased fetal plasma and amniotic 89. Bernstein I, Gabbe SG. Intrauterine growth restriction. In:
fluid erythropoietin concentrations: Markers of intrauterine Gabbe SG, Niebyl JR, Simpson JL et al., eds., Obstetrics:
hypoxia. Neonatology 2009; 95(2): 105–116. Normal and Problem Pregnancies, Churchill Livingstone: New
65. Arany E, Hill DJ. Fibroblast growth factor-2 and fibroblast growth York, 1996, pp. 863–886.
factor receptor-1 mRNA expression and peptide localization in 90. Mayer C, Joseph KS. Fetal growth: A review of terms, concepts
placentae from normal and diabetic pregnancies. Placenta 1998; and issues relevant to obstetrics. Ultrasound Obstet Gynecol
19(2–3): 133–142. 2013; 41(2): 136–145.
66. Grissa O, Yessoufou A, Mrisak I et al. Growth factor concentra- 91. Ott WJ. The diagnosis of altered fetal growth. Obstet Gynecol
tions and their placental mRNA expression are modulated in ges- Clin North Am 1988; 15(2): 237–263.
tational diabetes mellitus: Possible interactions with macrosomia. 92. Cetin I, Alvino G. Intrauterine growth restriction: Implications
BMC Pregnancy Childbirth 2010; 10: 7. for placental metabolism and transport. A review. Placenta
67. Nelson SM, Coan PM, Burton GJ et al. Placental structure in type 2009; 30(Suppl. A): S77–S82.
1 diabetes: Relation to fetal insulin, leptin, and IGF-I. Diabetes 93. Mayhew TM, Ohadike C, Baker PN et al. Stereological investi-
2009; 58(11): 2634–2641. gation of placental morphology in pregnancies complicated by
68. Lassance L, Miedl H, Absenger M et al. Hyperinsulinemia stimu- pre-eclampsia with and without intrauterine growth restriction.
lates angiogenesis of human fetoplacental endothelial cells: A Placenta 2003; 24(2–3): 219–226.
possible role of insulin in placental hypervascularization in dia- 94. Mayhew TM, Wijesekara J, Baker PN et al. Morphometric evi-
betes mellitus. J Clin Endocrinol Metab 2013; 98(9): E1438–E1447. dence that villous development and fetoplacental angiogenesis
69. Hiden U, Lassance L, Tabrizi NG et al. Fetal insulin and IGF-II are compromised by intrauterine growth restriction but not by
contribute to gestational diabetes mellitus (GDM)-associated up- pre-eclampsia. Placenta 2004; 25(10): 829–833.
regulation of membrane-type matrix metalloproteinase 1 (MT1- 95. Egbor M, Ansari T, Morris N et al. Pre-eclampsia and fetal
MMP) in the human feto-placental endothelium. J Clin Endocrinol growth restriction: How morphometrically different is the pla-
Metab 2012; 97(10): 3613–3621. centa? Placenta 2006; 27(6–7): 727–734.
70. Werner H, Leroith D. Insulin and insulin-like growth factor recep- 96. Rainey A, Mayhew TM. Volumes and numbers of intervil-
tors in the brain: Physiological and pathological aspects. Eur lous pores and villous domains in placentas associated with
Neuropsychopharmacol 2014; 24(12): 1947–1953. intrauterine growth restriction and/or pre-eclampsia. Placenta
71. Sara VR, Hall K, Misaki M et al. Ontogenesis of somatomedin 2010; 31(7): 602–606.
and insulin receptors in the human fetus. J Clin Invest 1983; 71(5): 97. Nicolaides KH, Bindra R, Turan OM et al. A novel approach
1084–1094. to first-trimester screening for early pre-eclampsia combin-
72. Steculorum SM, Vogt MC, Bruning JC. Perinatal programming of ing serum PP-13 and Doppler ultrasound. Ultrasound Obstet
metabolic diseases: Role of insulin in the development of hypo- Gynecol 2006; 27(1): 13–17.
thalamic neurocircuits. Endocrinol Metab Clin North Am 2013; 98. Pilalis A, Souka AP, Antsaklis P et al. Screening for pre-eclamp-
42(1): 149–164. sia and small for gestational age fetuses at the 11–14 weeks
73. Plagemann A. A matter of insulin: Developmental programming scan by uterine artery Dopplers. Acta Obstet Gynecol Scand
of body weight regulation. J Matern Fetal Neonatal Med 2008; 2007; 86(5): 530–534.
21(3): 143–148. 99. Ohkuchi A, Hirashima C, Matsubara S et al. Alterations in pla-
74. Davidowa H, Plagemann A. Inhibition by insulin of hypothalamic cental growth factor levels before and after the onset of pre-
VMN neurons in rats overweight due to postnatal overfeeding. eclampsia are more pronounced in women with early onset
Neuroreport 2001; 12(15): 3201–3204. severe preeclampsia. Hypertension Res 2007; 30(2): 151–159.
75. Plagemann A, Harder T, Rake A et al. Hypothalamic insulin and 100. Huppertz B, Kadyrov M, Kingdom JC. Apoptosis and its role in
neuropeptide Y in the offspring of gestational diabetic mother the trophoblast. Am J Obstet Gynecol 2006; 195(1): 29–39.
rats. Neuroreport 1998; 9(18): 4069–4073. 101. Huppertz B. IFPA Award in Placentology Lecture: Biology of the
76. Linder K, Schleger F, Ketterer C et al. Maternal insulin sensitiv- placental syncytiotrophoblast—Myths and facts. Placenta 2010;
ity is associated with oral glucose-induced changes in fetal brain 31(Suppl.): S75–S81.
activity. Diabetologia 2014; 57(6): 1192–1198. 102. Johansen M, Redman CW, Wilkins T et al. Trophoblast depor-
77. ACOG Committee on Obstetric Practice. Diagnosis and manage- tation in human pregnancy—Its relevance for pre-eclampsia.
ment of preeclampsia and eclampsia. Int J Gynaecol Obstet 2002; Placenta 1999; 20(7): 531–539.
77: 67–75. 103. Redman CW, Sargent IL. Placental debris, oxidative stress and
78. WHO. Make every mother and child count, Geneva, Switzerland, pre-eclampsia. Placenta 2000; 21(7): 597–602.
2005. 104. Goswami D, Tannetta DS, Magee LA et al. Excess syncytiotro-
79. Duley L. The global impact of pre-eclampsia and eclampsia. phoblast microparticle shedding is a feature of early-onset pre-
Semin Perinatol 2009; 33(3): 130–137. eclampsia, but not normotensive intrauterine growth restriction.
80. Williams D. Long-term complications of preeclampsia. Semin Placenta 2006; 27(1): 56–61.
Nephrol 2011; 31(1): 111–122. 105. Huppertz B. Placental pathology in pregnancy complications.
81. von Dadelszen P, Magee LA, Roberts JM. Subclassification of pre- Thrombosis Res 2011; 127(Suppl. 3): S96–S99.
eclampsia. Hypertens Pregnancy 2003; 22(2): 143–148. 106. Burton GJ, Woods AW, Jauniaux E et al. Rheological and physio-
82. Huppertz B. Placental origins of preeclampsia: Challenging the logical consequences of conversion of the maternal spiral arteries
current hypothesis. Hypertension 2008; 51(4): 970–975. for uteroplacental blood flow during human pregnancy. Placenta
83. von Dadelszen P, Magee L. What matters in preeclampsia are 2009; 30(6): 473–482.
the associated adverse outcomes: The view from Canada. Curr 107. Schaarschmidt W, Rana S, Stepan H. The course of angiogenic
Opin Obstet Gynecol 2008; 20(2): 110–115. factors in early- vs. late-onset preeclampsia and HELLP syn-
84. Lisonkova S, Joseph KS. Incidence of preeclampsia: Risk factors drome. J Perinatal Med 2013; 41(5): 511–516.
and outcomes associated with early- versus late-onset disease. 108. Longtine MS, Nelson DM. Placental dysfunction and fetal pro-
Am J Obstet Gynecol 2013; 209(6): 544.e1–544.e12. gramming: The importance of placental size, shape, histopathol-
85. Zhong Y, Wang Y, Chen S. Circulating fetal DNA in maternal ogy, and molecular composition. Semin Reprod Med 2011; 29(3):
plasma is increased in pregnancies at high altitude and is further 187–196.
enhanced by preeclampsia. Clin Chem 2004; 50: 2403–2405. 109. Hogg K, Blair JD, McFadden DE et al. Early onset pre-eclampsia
86. Chappell LC, Enye S, Seed P et al. Adverse perinatal outcomes and is associated with altered DNA methylation of cortisol-signalling
risk factors for preeclampsia in women with chronic hyperten- and steroidogenic genes in the placenta. PLOS ONE 2013; 8(5):
sion: A prospective study. Hypertension 2008; 51(4): 1002–1009. e62969.
14 Diagnosis of gestational diabetes
mellitus
Donald R. Coustan and Boyd E. Metzger
In a subsequent publication, 17–23 years of follow-up deter-

History mined that almost two-thirds had developed diabetes within
Gestational diabetes mellitus (GDM) is generally consid- that time period.6 The 1.9% prevalence of GDM identified
ered to be carbohydrate intolerance with variable severity by these cutoffs was similar to the 2% prevalence of diabetes
with onset or first recognition during pregnancy. The con- in a nonpregnant population in the 1940s.7 If one standard
cept that pregnancy may bring about alterations of carbohy- deviation had been used as a cutoff, then 16% of pregnant
drate metabolism goes back more than 140 years. In 1882, women would have been diagnosed with GDM. O’Sullivan
J. Mathews Duncan described what would later be called and Mahan expressed concern about the potential adverse
GDM, indicating that “diabetes may occur only during preg- psychological effect of overdiagnosis if the lower thresholds
nancy being absent at other times or may cease with the ter- were applied and noted that the predictive value for future
mination of pregnancy recurring sometime afterwards.”1 In diabetes, using the recommended thresholds, was similar to
1946, Miller observed that perinatal mortality was increased the predictive value of a steroid-modified OGTT, which was
in the previous pregnancies of women with diabetes,2 and in in common use at the time.
1952 Jackson noted a high likelihood of previous stillbirth O’Sullivan and Mahan decided to recommend that at
and fetal macrosomia in women with diabetes.3 A few years least two values needed to be elevated in order to diagnose
later, Carrington coined the term “gestational diabetes.”4 In GDM, as was also the case for the United States Public Health
the middle of the twentieth century, one commonly used set Service (USPHS) criteria for diabetes in nonpregnant indi-
of diagnostic criteria for diabetes in the United States was viduals. The authors stated “It was considered expedient…
that put forth by the U.S. Public Health Service (USPHS)5: a to require two or more values to be met or exceeded. In this
100 g, 3-hour oral glucose tolerance test (OGTT) was admin- way misclassification due to a laboratory error, or occasional
istered. Diabetes was diagnosed if both the fasting and single high peaks resulting from unusually rapid absorption
3-hour whole blood glucose values were ≥130 mg/dL or if one of glucose, could be avoided.”8 It was thus determined that
of these two was elevated and, in addition, either the 1-hour the 100 g, 3-hour OGTT would be used to diagnose GDM
value was ≥195 mg/dL or the 2-hour value was ≥140 mg/dL. and that two or more elevated values would be required for
In 1964, O’Sullivan and Mahan published their classic this diagnosis. These two requirements have persisted in the
treatise recommending pregnancy-specific glucose toler- United States well into the twenty-first century.
ance test criteria. Noting that pregnancy changes carbohy- Although the O’Sullivan and Mahan criteria were
drate metabolism, such that glucose tolerance test norms accepted by the American College of Obstetricians and
might be altered, they described the results of 100 g, 3-hour Gynecologists in the late 1970s,9 changes in laboratory meth-
OGTTs administered to 752 unselected pregnant women, odology necessitated modification of the cutoffs, which were
predominantly in the late second or early third trimester. originally based on the Somogyi–Nelson method of mea-
They considered potential diagnostic cutoffs at one, two, and suring glucose, in venous whole blood samples, rounded
three standard deviations above the means for each of the to the nearest 5 mg/dL for easier memorization. In 1979,
four blood glucose determinations. The three sets of thresh- the National Diabetes Data Group (NDDG) in the United
olds were then applied retrospectively to a second dataset States recommended using the O’Sullivan criteria, modi-
of OGTTs in 1333 previous pregnancies, with the outcome fied to account for the use of plasma rather than whole blood
variable being the development of diabetes within 8 years by adding 15% to the already rounded values, then round-
of the index pregnancies. These subjects had been followed ing again to the nearest 5 mg/dL.10 In 1982, Carpenter and
with periodic OGTTs. Cutoffs at two standard deviations Coustan published a second modification of the O’Sullivan
above the means were chosen because 22% of gravidas meet- and Mahan criteria, in which 5 mg/dL was subtracted from
ing those thresholds developed diabetes within 8 years. each of the original unrounded values, to account for the
110
Hyperglycemia and Adverse Pregnancy Outcome study 111
Table 14.1 Various conversions of criteria for diagnosing gestational diabetes with 100 g, 3-hour oral
glucose tolerance test
O’Sullivan6 mmol/L (mg/dL)a

National Diabetes Data Group10 Carpenter/Coustan11 mmol/L
Test interval Unrounded Rounded mmol/L (mg/dL)b (mg/dL)c
Fasting 5.0 (90) 5.0 (90) 5.8 (105) 5.3 (95)
1 hour 9.2 (165) 9.2 (165) 10.6 (190) 10 (180)
2 hours 7.9 (143) 8.0 (145) 9.2 (165) 8.6 (155)
3 hours 7.1 (127) 7.0 (125) 8.0 (145) 7.8 (140)
Note: Two or more elevated values are needed for diagnosis.
a Venous whole blood, Somogyi–Nelson technique.
b Corrected for conversion from whole blood to plasma.
c Corrected for conversion of whole blood to plasma and Somogyi–Nelson to enzymatic glucose measurement.
change from less specific Somogyi–Nelson methodology evaluating a 7.8 mmol/L (140 mg/dL) threshold reported
to more specific enzymatic glucose measurement, and the that 70%–88% of GDM cases would be identified, whereas
resulting values were increased by 14% because of the change a threshold of 7.2 mmol/L (130 mg/dL) identified 88%–99%
from whole blood to plasma.11 These various conversions of cases. At a threshold of 7.8 mmol/L (140 mg/dL), 11%–31%
of the original O’Sullivan and Mahan criteria are depicted of unaffected gravidas would require an OGTT, while at
in Table 14.1. When Sacks et al.12 reconstructed the origi- 7.2 mmol/L (130 mg/dL), 23%–34% require further testing.
nal methodology of O’Sullivan and Mahan and ran parallel
samples of whole blood and plasma, using Somogyi–Nelson
versus enzymatic methodology, only the Carpenter and The need for new criteria
Coustan conversions were within 95% confidence limits
of the original O’Sullivan and Mahan values. By 2002, the In 1991, the summary and recommendations of the
American Diabetes Association (ADA) had endorsed the Third International Workshop-Conference on Gestational
Carpenter and Coustan conversions.13 Diabetes18 noted that there was a lack of international agree-
While the 100 g, 3-hour OGTT became the standard ment on testing for GDM, with various countries and pro-
approach to diagnosing GDM in the United States, the fessional organizations promoting 50, 75, and 100 g glucose
75 g, 2-hour OGTT was in wide use throughout the rest of challenges as well as various sets of diagnostic thresholds. It
the world. The World Health Organization (WHO) criteria was predicted that the 2-hour, 75 g OGTT eventually would
for diagnosing diabetes and impaired glucose tolerance in become universally employed, just as it was for diagnosing
nonpregnant individuals were commonly applied to preg- diabetes outside of pregnancy. The report noted that none
nancy, such that if the fasting plasma glucose was below of the commonly used criteria were based on pregnancy
7 mmol/L (126 mg/dL) and the 2-hour value (after a 75 g glu- outcomes. The summary pointed out that “With respect to
cose challenge) was 7.8–11 mmol/L (140–199 mg/dL), GDM GDM, the heart of the matter is the relationship between
was diagnosed.14 A number of other tests and criteria are in the degree of glucose intolerance and/or hyperglycemia and
use in various locations. the risk of adverse maternal, fetal and neonatal outcomes…”
Another development in the United States was the use The report concluded that “The highest priority should be
of a two-step protocol for screening and diagnosis of GDM. given to efforts to develop international consensus on meth-
In 1973, O’Sullivan and colleagues described the glucose ods and definitions.” A 1992 international workshop spon-
challenge test (GCT) based on data from 752 unselected sored by the Eunice Kennedy Shriver National Institute of
gravidas who underwent the 50 g, 1-hour test and the 100 g, Child Health and Human Development (NICHD) added
3-hour OGTT.15 A whole blood, Somogyi–Nelson glucose further support to the need for carefully designed multina-
of 7.2 mmol/L (130 mg/dL) was identified in 79% of indi- tional studies to measure adverse perinatal effects of GDM.19
viduals whose OGTT indicated GDM. This cutoff would be
equivalent to 7.9 mmol/L (143 mg/dL) if plasma and an enzy-
matic method of analysis were used. A subsequent study16 Hyperglycemia and Adverse
demonstrated that a plasma glucose cutoff of 7.2 mmol/L Pregnancy Outcome study
(130 mg/dL), using an enzymatic method of glucose analysis,
would identify nearly 100% of GDMs, while a 7.8 mmol/L The Hyperglycemia and Adverse Pregnancy Outcome
(140 mg/dL) threshold identified 90% of cases. A recent sys- (HAPO) study20 was based upon the rationale that overt
tematic review of screening tests for GDM,17 performed for diabetes clearly increases the risk of adverse pregnancy out-
the U.S. Preventive Services Task Force, found that studies come, while controversy exists regarding the potential risk of
112 Diagnosis of gestational diabetes mellitus
glucose intolerance during pregnancy less severe than overt held, with mild attenuation, when adjusted for multiple
diabetes. This blinded observational study of over 23,000 potential prespecified confounders, including maternal
gravidas in 15 field centers in 9 different countries through- age, body mass index (BMI), smoking status, alcohol use,
out the world was designed to determine the association the presence or absence of a family history of diabetes,
between various levels of glycemia on a 2-hour, 75 g OGTT gestational age at the oral glucose tolerance test, sex of the
(performed between 24 and 32 weeks gestation) and adverse infant, parity (0, 1, or ≥2, except for primary cesarean deliv-
pregnancy outcomes. Primary outcomes included neonatal eries), mean arterial pressure, and the presence or absence
macrosomia (birth weight >90th percentile for the infant’s of hospitalization before delivery (except for preeclampsia),
sex, gestational age, race or ethnic group, field center, and the presence or absence of a family history of hypertension,
maternal parity), primary cesarean section, clinical neonatal and maternal urinary tract infection (for analysis of pre-
hypoglycemia, and cord blood C-peptide (a marker for fetal eclampsia only). Height was also included as a potential con-
insulin level). Prespecified secondary outcomes included founder, on the basis of post hoc findings of an association
delivery before 37 weeks of gestation, shoulder dystocia or with birth weight greater than the 90th percentile. The three
birth injury, the need for intensive neonatal care, hyperbili- glucose tolerance test values were not highly cross-correlated
rubinemia, and preeclampsia. Figure 14.1 depicts the rela- and were independently predictive of adverse outcomes.
tionships between each of the three glucose tolerance test The association between maternal glucose intolerance and
measurements and each of the four primary outcomes. Each macrosomia as well as cord C-peptide levels supports the
relationship was significant but continuous, increasing from 60-year-old Pedersen hypothesis that maternal hypergly-
the lowest to highest glucose levels, with no obvious inflec- cemia leads to fetal hypoglycemia, which in turn leads to
tion point. Secondary outcomes were similarly continuously fetal hyperinsulinemia and hence excessive fetal growth.21
related to glucose tolerance test values, particularly with A subsequent analysis of HAPO data 22 demonstrated that
respect to shoulder dystocia/birth injury and preeclampsia. neonatal adiposity was strongly correlated with both mater-
With the exception of clinical neonatal hypoglycemia (a rare nal OGTT values and cord C-peptide, lending further sup-
outcome occurring in only 2.1% of cases), the relationships port to the aforementioned mechanism. Another analysis of
Cord C-peptide >90th percentile Birth weight >90th percentile

35 30
Fasting Fasting
30 25
1 hour 1 hour
25
2 hour 20 2 hour
Frequency (%)
Frequency (%)
20
15
15
10
10
5 5
0 0
1 2 3 4 5 6 7 1 2 3 4 5 6 7
(a) Glucose categories (b) Glucose categories
Primary C-section Clinical hypoglycemia

35 5
Fasting
30
4 1 hour
25 2 hour
Frequency (%)
Frequency (%)
20 3
15 2
10 Fasting
1 hour 1
5
2 hour
0 0
1 2 3 4 5 6 7 1 2 3 4 5 6 7
(c) Glucose categories (d) Glucose categories
Figure 14.1 (a through d) Hyperglycemia and Adverse Pregnancy Outcome study: frequency of primary outcomes across glucose
categories. (Reprinted from HAPO Study Cooperative Research Group, N. Engl. J. Med., 358, 1991. Copyright 2008. With permis-
sion from Massachusetts Medical Society.)
International Association of Diabetes and Pregnancy Study Groups recommendations 113
HAPO data23 revealed that maternal BMI and glycemia were concentrations below levels that are diagnostic of diabetes.25–29
independently predictive of adverse pregnancy outcomes, However, because procedures for selection and enrollment of
refuting the suggestion that GDM is simply a marker for study participants, laboratory analyses, data collection, blind-
maternal obesity. ing study participants and caregivers to OGTT results, and
analysis of results were all well standardized, HAPO study
data were used by IADPSG as the basis to identify diagnostic
International Association of Diabetes thresholds for GDM.
Excess fetal growth, adiposity, and insulin secretion
and Pregnancy Study Groups are integral features of diabetic fetopathy. The IADPSG-
recommendations recommended diagnostic thresholds are the average glu-
cose values at which odds ratios (ORs) for birth weight,
Because there were no clear inflection points in the rela- cord C-peptide, and percent body fat, each >90th percentile,
tionships between OGTT glucose values and the various reached 1.75 times the estimated ORs of these outcomes at
adverse outcomes, the HAPO research group did not make mean fasting, 1-hour, and 2-hour OGTT glucose values, based
any recommendations for GDM diagnostic criteria, pre- on fully adjusted logistic regression models. The resulting fast-
ferring to leave such recommendations to a wider group ing, 1-hour, and 2-hour post 75 g glucose load thresholds are
of experts so as to foster better international ownership 92, 180, and 152 mg/dL, respectively (5.1, 10.0, 8.5 mmol/L).
and acceptance. The International Association of Diabetes Risk ratios for the HAPO study outcomes in those defined
and Pregnancy Study Groups (IADPSG) sponsored a 2008 as GDM compared with non-GDM are similar using glucose
Workshop-Conference on Gestational Diabetes Diagnosis and thresholds at ORs of 1.5, 1.75, or 2.0 relative to the mean glu-
Classification attended by more than 225 conferees from 40 cose levels. Thus, in the final analysis, the choice of thresholds
countries. Published and unpublished HAPO results and for associations that are continuous and linear is arbitrary.
other work on associations of maternal glycemia with peri- Preexisting diabetes requires more intensive evalua-
natal and long-term outcomes in offspring were reviewed. tion and treatment during early pregnancy. Because of the
A consensus panel of more than 50 individuals representing increasing prevalence of obesity and of diabetes in the gen-
the member organizations of IADPSG and other groups car- eral population and during pregnancy, the IADPSG also
ried out further review and analysis of HAPO results, held made recommendations for testing in early pregnancy to
a second face-to-face meeting of panel members, and then identify women with preexisting, untreated overt diabetes
published recommendations for the diagnosis and classifica- mellitus (Table 14.2).24 The IADPSG panel concluded that
tion of hyperglycemia in pregnancy.24 there are insufficient data available to determine whether
Results of a number of other epidemiological studies there was benefit to be derived from testing for GDM prior
are consistent with the HAPO study findings indicating to 24–28 weeks gestation. Women diagnosed with diabetes
that associations between maternal glycemia and adverse or GDM during pregnancy should undergo postpartum
outcomes are continuous across the range of glucose glucose testing to detect persistent diabetes or prediabetes.
Table 14.2 International Association of Diabetes and Pregnancy Study Groups recommendations for detecting
hyperglycemia in pregnancy
First prenatal visit

Measure fasting plasma glucose, A1c, or random plasma glucose on all gravidas or only
high-risk gravidas:
• If diabetes (fasting plasma glucose [FPG] ≥ 7 mmol/L [126 mg/dL] or A1c ≥ 6.5% or
random glucose ≥ 11.1 mmol/L [200 mg/dL], confirmed by one of the other tests), treat as
preexisting diabetes.
• If no diagnosis of diabetes and fasting plasma glucose ≥5.1 mmol/L (92 mg/dL) but <7.0
mmol/L (126 mg/dL), diagnose GDM.
• If no diagnosis of diabetes and fasting plasma glucose <5.1 mmol/L (92 mg/dL), test for
GDM at 24–28 weeks.
24–28 weeks gestation
Perform 2-hour 75 g OGTT after overnight fast in women not previously diagnosed with
diabetes or GDM earlier in pregnancy:
• Diabetes if fasting plasma glucose ≥7 mmol/L (126 mg/dL)
• GDM if one or more values equals or exceeds the following thresholds:
• Fasting 7.0 mmol/L (92 mg/dL)
• 1 hour 10 mmol/L (180 mg/dL)
• 2 hour 7.9 mmol/L (143 mg/dL)
• Normal if all values below the above thresholds
Source: International Association of Diabetes and Pregnancy Study Groups Consensus Panel, Diabetes Care, 33, 676, 2010.
Case against adopting IADPSG rather than separate strategies for pregnant and non-
pregnant individuals, (2) diagnostic criteria based on
recommendations pregnancy outcomes, (3) comparability of data among
the various regions of the world, and (4) simplification of
Since the IADPSG recommendations were published, there
the testing protocol and elimination of the dilemma sur-
has been considerable controversy, with many authors
rounding the management of gravidas with a single ele-
expressing the view that they are unproven and it is not yet
vated value on the 3-hour OGTT. In the paragraphs that
time for their universal adoption30–33 and others expressing
follow, we shall deal with each of the concerns expressed
opposing views.34–36 In the spring of 2013, the U.S. National
by the consensus panel:
Institutes of Health (NIH) convened a consensus panel to
consider the available data and make recommendations.37
The panel consisted of a number of individuals from various 1. Increased healthcare costs and lack of data on cost-effec-
specialties who were considered unbiased, in that they have tiveness. The IADPSG recommendations, when applied
never published anything having to do with GDM. Experts to the data in the HAPO study, identify 17.8% of par-
in the field, on both sides of the issue being considered, pre- ticipants as having GDM, when subjects unblinded
sented available data for the panel to consider. In addition, a because of GTT results above the unblinding criteria are
number of evidence reviews were conducted by the Agency included.38 At the 15 field centers, rates of GDM varied
for Healthcare Research and Quality. Recommendations from 9.3% to 25.5%. Although it is not currently possible
from the consensus panel, which are not official recommen- to compare rates of GDM around the world because of
dations of the NIH or any government agency, were pub- the variability of diagnostic criteria, the International
lished in the summer of 2013.37 The panel acknowledged Diabetes Federation recently estimated that, based
that there would be clear benefits to international standard- upon available data that were then adjusted to reflect
ization of the diagnostic approach to GDM. However, the the IADPSG criteria, the global prevalence of “hyper-
panel opined that there is not sufficient evidence to adopt glycemia in pregnancy” is 16.9%, ranging from 10.4% in
a one-step approach and recommended continuation of the the North America and Caribbean region to 25% in the
current two-step approach for the time being. With regard Southeast Asia region.39 For comparison, rates of GDM
to the two-step approach, the panel recommended that a single (current criteria), based upon the 2008 hospital dis-
standard for the screening test cutoff and for the diagnostic charge data, from 23 states in the United States, ranged
thresholds be adopted by appropriate professional organiza- from 3.5% to 7.2%.40 Hospital discharge data likely
tions. The rationale for the recommendation not to adopt the underestimate GDM rates due to incomplete testing of
IADPS one-step approach can be summarized as follows: the population. Much of the difference among rates was
attributable to differences in age, race/ethnicity, and
1. The twofold to threefold increased frequency of the insurance status. Institution of the IADPSG one-step
diagnosis of GDM that would result from adoption of approach would likely result in a twofold to threefold
the one-step approach will generate increased healthcare increase in GDM, although the increment might be more
costs; it is not clear that this would be cost-effective. or less dramatic in particular locations. Thus, it is highly
2. It is not clear whether the additional women who would likely that healthcare costs will be increased. However,
be diagnosed with milder GDM, and their offspring, the current epidemic of obesity and diabetes is not lim-
would benefit from identification and treatment. ited to pregnant women; 4.5% of U.S. women aged 18–44
3. Labeling of these additional women may have unintended have diabetes41 and 26% have prediabetes.42 Since the
consequences, possibly including increased cesarean sec- IADPSG thresholds for GDM are not too different from
tions and adverse emotional impact on the mothers. the current diagnostic criteria for prediabetes, an 18%
4. The variability inherent in the OGTT is known to be con- incidence of GDM does not represent over diagnosis. As
siderable; the use of a one-step process is likely to result in diabetes has reached epidemic proportions throughout
more “false-positives” than a two-step test. the world, we have not raised the diagnostic criteria to
lower its prevalence; rather, healthcare providers and
Among its recommendations for future research, the panel systems are attempting to meet the challenges posed. The
suggested evaluation of the use of diagnostic thresholds same should be true for the “epidemic” of GDM; newer
based on ORs for adverse outcomes of 2.0 rather than 1.75 and more efficient ways to deliver care to these patients
as recommended by IADPSG. will need to be developed. One potential cost-efficient
approach may utilize group prenatal counseling and
care. It is not clear whether patients with milder forms of
GDM will require the same intensity of antepartum fetal
The case for implementing IADPSG testing and monitoring as those with more severe GDM.
recommendations Self-monitoring of blood glucose may be performed less
frequently with mild GDM without adversely impacting
There is no question that the IADPSG recommendations on the time of initiation of treatment.43 In two random-
offer numerous benefits, including but not limited to (1) ized trials of diagnosis and treatment of mild GDM,
the use of a uniform 75 g, 2-hour OGTT in all situations insulin was required in only 8%–20% of patients.44,45
The case for implementing IADPSG recommendations 115
A recent cost-effectiveness analysis46 compared the of LGA >90th percentile (13% vs. 22%), macrosomia >4 kg
one-step IADPSG approach with the current two-step (10% vs. 21%), and maternal preeclampsia (12% vs. 18%). It
approach. It did not assume any increase in efficiency should be noted that the GDM 2-hour OGTT diagnostic
leading to decreased costs of care as outlined p
reviously. criterion of 7.8–11 mmol/L (140–199 mg/dL) is lower than
The IADPSG approach was more expensive but was more the IADPSG recommended 2-hour criterion of 8.5 mmol/L
effective in improving maternal and neonatal outcomes. (153 mg/dL). While the upper limit of fasting plasma
In another study, the one-step approach was more cost- glucose in ACHOIS was 7.8 mmol/L (140 mg/dL), the
effective but only if women with GDM received postpar- average fasting plasma glucose of ACHOIS subjects was
tum testing and counseling to prevent the subsequent 4.8 mmol/L (86 mg/dL). ACHOIS demonstrated benefit
development of diabetes.47 This finding highlights the to identifying and treating GDM at diagnostic thresholds
opportunity for improvement in maternal long-term even lower than the IADPSG recommendations. In sum,
health that is presented by the diagnosis of GDM. While there is evidence that identification and treatment of mild
pregnancy outcome is the focus of efforts to test for GDM is worthwhile.
GDM, there is additional benefit if future diabetes can be 3. Will labeling of additional patients with GDM have adverse
prevented. In a subset analysis of data from the Diabetes consequences? The NIH consensus panel37 noted studies
Prevention Program, subjects with previous GDM and demonstrating negative emotional impact of screening
current prediabetes who were randomized to intensive and diagnosis of GDM. However, when the ACHOIS
lifestyle intervention or metformin treatment developed RCT investigators measured emotional well-being of
type 2 diabetes at an approximate rate of 7.5% per year, subjects and controls at 3 months postpartum,44 women
compared to 15% per year in those randomized to the diagnosed and treated for GDM did significantly better
placebo arm of the study.48 The numbers needed to treat than those in the control group who received routine
with metformin or lifestyle change to prevent one case pregnancy care. In another study, women with the
of diabetes over 3 years were 6 and 5, respectively. The diagnosis of GDM, including those who required insulin,
diagnosis of GDM offers an important opportunity to manifested no difference in psychological status when
dampen the epidemic of diabetes. compared to nondiabetic controls.49 An apparent effect
2 . Will identification and treatment of mild GDM provide of labeling a patient with GDM was demonstrated in the
benefit? Two large randomized trials of identification Toronto Tri-Hospital Study, 50 in which identification
and treatment of mild GDM were completed around and treatment of GDM (NDDG criteria) resulted in
the time of the IADPSG recommendations and have a relative normalization of macrosomia rates but a
relevance to the aforementioned question. The Maternal- persistent increase in the cesarean rate compared to
Fetal Medicine Units Network (MFMU) Network individuals with normal glucose tolerance; cesarean
study45 recruited women whose 50 g, 1-hour GCT was deliveries were not related to fetal macrosomia. An
7.5–11 mmol/L (135–199 mg/dL) to undergo a blinded intermediate group with mild GDM (Carpenter and
OGTT. The 958 subjects with mild GDM (Carpenter Coustan criteria), whose caregivers were blinded to
and Coustan criteria with two or more elevated values the OGTT results, demonstrated an increased rate of
but fasting plasma glucose <5.3 mmol/L [95 mg/dL]) both macrosomia and cesarean section, with cesareans
were randomized to identification and treatment versus being related to macrosomic babies. These findings
blinded OGTT results and standard prenatal care. The suggest that obstetricians caring for the patients with
latter group also included a number of women with known GDM were more likely to intervene despite the
all OGTT results normal to ensure blinding of the treatment being successful in preventing macrosomia,
caregivers. Identification and treatment of mild GDM presumably because of concern over shoulder dystocia
resulted in reductions in preeclampsia, macrosomia, when GDM was present. These findings are in contrast to
and shoulder dystocia of 50% or more, as well as a those of the MFMU Network trial45 in which diagnosis
20% reduction in cesarean sections. The Australian and treatment of mild GDM resulted in a significant
Carbohydrate Intolerance Study of Pregnant Women reduction in primary cesareans. Overall the evidence is
(ACHOIS) trial44 enrolled gravidas who had risk factors mixed regarding unintended adverse consequences of
for GDM or who had a 50 g GCT of 7.8 mmol/L (140 mg/ diagnosing GDM.
dL) or greater to undergo a blinded 75 g, 2-hour OGTT. 4. Will the inherent variability of the OGTT increase the rate
Subjects who had GDM, defined as a 2-hour plasma of false-positive diagnoses of GDM? The NIH consensus
glucose of 7.8–11.0 mmol/L (140–199 mg/dL) and fasting panel report37 noted that the results of OGTTs “may differ
plasma glucose below 7.8 mmol/L (140 mg/dL), were then in as many as 25% of women if performed at different
randomized to identification and treatment of GDM or times.” It is true that OGTTs are known to manifest
continued blinding of subjects and caregivers and routine variability; this is true for both pregnant and nonpregnant
prenatal care. Treated GDMs had a significantly lower populations. However, the 75 g, 2-hour OGTT is the
incidence (1% vs. 4%) of the composite outcome (perinatal standard method of diagnosis of diabetes throughout the
death, shoulder dystocia, bone fracture, or nerve palsy). world despite its instability. While the two-step approach
In addition, identification and treatment of GDM was introduces two OGTTs, with the second (diagnostic) test
associated with significantly lower birth weight and rates presumably uncovering any false-positive result from the
screening test, the same is not true for a false-negative

Table 14.3 Rates of adverse outcomes in
screening test. The systematic review cited earlier in this
Hyperglycemia and Adverse Pregnancy Outcome
chapter17 reported that with a 50 g, 1-hour GCT cutoff of
subjects without and with gestational diabetes
7.8 mmol/L (140 mg/dL), approximately 10% of GDMs
mellitus by International Association of Diabetes
would be missed, when compared to a cutoff of 7.2 mmol/L
and Pregnancy Study Groups criteria
(130 mg/dL). A similar result was reported in two earlier
studies16,51 not included in the systematic review. Thus,
the two-step process leads to both false-positives and Normal
false-negatives, and the false-negative results remain glucose
undetected. If our goal is to minimize the diagnosis of tolerancea GDMb
(%) (%)
GDM, this approach would make sense. However, if our
goal is to identify as many patients with GDM as possible LGA (birth weight >90th 8.3 16.2
so as to maximize the chance of preventing adverse percentile)
outcomes, then the two-step process is not optimal. In Fetal hyperinsulinemia (cord 6.7 17.5
C-peptide >90th percentile)
contrast, the IADPSG-recommended one-step process
produces no false-negatives or false-positives, since the Neonatal adiposity (percent 8.5 16.6
body fat >90th percentile)
75 g, 2-hour OGTT is a diagnostic test and no screening
Preeclampsia 4.5 9.1
test is involved. It could be argued that the diagnosis of
Shoulder dystocia/birth injury 1.3 1.8
GDM is really a “screening test” for adverse pregnancy
outcomes, but if that were the case, then every GDM Source: Based on International Association of Diabetes and
who does not have an adverse outcome would be termed Pregnancy Study Groups Consensus Panel, Diabetes
Care, 33, 676, 2010.
a “false-positive,” whether diagnosed with the one-step a All three 2-hour, 75 g OGTT values below IADPSG threshold
or the two-step procedure. Such an argument defies values.
logic, since there is no “disease” in which all diagnosed b IADPSG proposed criteria. All differences significant at p < 0.01
individuals suffer the same end results. or better.
Finally, the suggestion that ORs of 2.0 be evaluated for GDM

cutoffs, rather than the 1.75 utilized by IADPSG, misses the
point that the 1.75 ORs referred to the comparison of adverse 257 with this approach and to prevent one LGA baby was 117;
outcomes in gravidas for whom one or more glucose val- both of these were fewer than the number needed to screen
ues exceeded the proposed thresholds to individuals at the with the previous WHO criteria. At around the same time,
mean for each of the three glucose measurements. When ACOG confirmed its support for the two-step approach,54
individuals in HAPO who met the IADPSG thresholds were and in early 2015, the ADA revised its 2011 recommendation
compared with all individuals who had no glucose values for the one-step approach and declared that either method
exceeding thresholds, the likelihood of adverse outcomes is acceptable, but stating that the one-step approach may be
was generally doubled or more (Table 14.3). preferable.55 The Endocrine Society endorsed the IADPSG
criteria and approach in 2013.56 In 2015 the International
Federation of Gynecology and Obstetrics endorsed the
Current status IADPSG one-step approach.57 As of the writing of this chap-
ter, there is not agreement in the United States as to which
As noted earlier, the IADPSG recommendations were pub- diagnostic approach is preferable. Around the world, vari-
lished in 2010.24 In 2011, the ADA endorsed a modified ver- ous countries and professional organizations are debating
sion of the IADPSG recommendations.52 In 2013, the NIH which diagnostic criteria to adopt. It is to be hoped that
consensus panel in the United States made its recommenda- in the not-too-distant future more evidence will be available
tion to maintain the status quo of the two-step, 100 g, 2-hour to determine the costs, benefits, and risks of the IADPSG
OGTT,37 and in the summer of 2013, WHO recommended recommendations and that there will ultimately be universal
the one-step IADPSG approach53 and noted that the num- agreement on how to diagnose GDM, similar to the agree-
ber needed to screen to prevent one case of preeclampsia was ment on the diagnosis of diabetes outside of pregnancy.
REFERENCES
1. Duncan JM. On puerperal diabetes. Trans Obstet Soc Lond 1882; 5. Eastman NJ, Hellman LM. Williams’ Obstetrics, 13th ed.,
24: 256–285. Appleton-Century-Crofts: New York, 1966, pp. 779–785.
2. Miller HC. The effect of diabetic and prediabetic pregnancies on 6. O’Sullivan JB. Subsequent morbidity among gestational diabetic
the fetus and newborn infant. J Pediatr 1946; 26: 455–461. women. In: Sutherland HW, Stowers JM, eds., Carbohydrate
3. Jackson WPU. Studies in pre-diabetes. Br Med J 1952; 3: Metabolism in Pregnancy and the Newborn, New York: Churchill
690–696. Livingstone, 1984, pp. 174–180.
4. Carrington ER, Shuman CR, Reardon HS. Evaluation of the predi- 7. Wilkerson HLC, Krall LP. Diabetes in a New England town. J Am
abetic state during pregnancy. Obstet Gynecol 1957; 9: 664–669. Med Assoc 1947; 135: 209–214.
References 117
8. O’Sullivan JB, Mahan CM. Criteria for the oral glucose tolerance 29. Schmidt MI, Duncan BB, Reichelt AJ et al. Gestational diabetes
test in pregnancy. Diabetes 1964; 13: 278–285. mellitus diagnosed with a 2-h 75-g oral glucose tolerance test
9. American College of Obstetricians and Gynecologists. and adverse pregnancy outcomes. Diabetes Care 2001; 24(7):
Management of diabetes mellitus in pregnancy. ACOG Technical 1151–1155.
Bulletin No. 48. Washington, DC, April 1978. 30. Ryan EA. Diagnosing gestational diabetes. Diabetologia 2011; 54:
10. National Diabetes Data group. Classification and diagnosis of 480–486.
diabetes mellitus and other categories of glucose intolerance. 31. Long H. Diagnosing gestational diabetes: Can expert opinions
Diabetes 1979; 28: 1039–1057. replace scientific evidence? Diabetologia 2011; 54: 2211–2213.
11. Carpenter MW, Coustan DR. Criteria for screening tests for gesta- 32. Cundy T. Proposed new diagnostic criteria for gestational
tional diabetes. Am J Obstet Gynecol 1982; 144: 768–773. diabetes—A pause for thought? Diabetic Med 2012; 29: 176–180.
12. Sacks DA, Abu-Fadil S, Greenspoon JS, Fotheringham N. Do the 33. Long H, Cundy T. Establishing consensus in the diagnosis of
current standards for glucose tolerance testing in pregnancy rep- gestational diabetes following HAPO: Where do we stand? Curr
resent a valid conversion of O’Sullivan’s original criteria? Am J Diab Rep 2013; 13(1): 43–50.
Obstet Gynecol 1989; 161: 638–641. 34. Coustan DR, Lowe LP, Metzger BE. The Hyperglycemia and
13. American Diabetes Association. Report of the Expert Committee Adverse Pregnancy Outcome (HAPO) Study: Paving the way for
on the Diagnosis and Classification of Diabetes Mellitus. Diabetes new diagnostic criteria for gestational diabetes mellitus. Am J
Care 2002 January; 25(Suppl. 1): s5–s20. Obstet Gynecol 2010; 202: 654.e1–654.e6.
14. World Health Organization. About diabetes. WHO web site. 35. Coustan DR. Point: The ADA and the IADPSG recommendations
Available from: http://www.who.int/diabetes/action_online/basics/ for diagnosing gestational diabetes should be used worldwide.
en/index1.html, accessed on February 9, 2014. Clin Chem 2012; 58: 1094–1097.
15. O’Sullivan JB, Mahan CM, Charles D, Dandrow RV. Screening 36. McIntyre HD, Metzger BE, Coustan DR, Dyer DR, Hadden DR,
criteria for high-risk gestational diabetic patients. Am J Obstet Hod M, Lowe LP, Oats JJN, Persson B. Counterpoint: Establishing
Gynecol 1973; 116: 895–900. consensus in the diagnosis of GDM following the HAPO Study.
16. Coustan DR, Nelson C, Carpenter MW, Carr SR, Rotondo L, Curr Diab Rep 2014; 14: 497.
Widness JA. Maternal age and screening for gestational dia- 37. National Institutes of Health Consensus Development Conference
betes: A population-based study. Obstet Gynecol 1989; 73: Panel. National Institutes of Health Consensus Development
557–561. Conference Statement: Diagnosing gestational diabetes mellitus,
17. Donovan L, Hartling L, Muise M, Guthrie A, Vandermeer B, March 4–6, 2013. Obstet Gynecol 2013; 122: 358–369.
Dryden DM. Screening tests for gestational diabetes: A system- 38. Sacks DA, Hadden DR, Maresh M et al., for the HAPO Study
atic review for the US Preventive Services Task Force. Ann Intern Cooperative Research Group. Frequency of gestational diabetes
Med 2013; 159: 115–122. mellitus at collaborating centers based on IADPSG consensus
18. Metzger BE and the Organizing Committee. Summary and rec- panel-recommended criteria: The Hyperglycemia and Adverse
ommendations of the Third International Workshop-Conference Pregnancy Outcome (HAPO) Study. Diabetes Care 2012; 35:
on Gestational Diabetes Mellitus. Diabetes 1991; 40(Suppl. 2): 526–528.
197–201. 39. Guariguata L, Lennenkamp U, Beagley J et al. Global estimates
19. Blank A, Grave GD, Metzger BE. Effects of gestational diabetes of the prevalence of hyperglycaemia in pregnancy for 2013 for
on perinatal morbidity reassessed. Report of the International the IDF Diabetes Atlas. Diabetes Res Clin Pract 2014; 103(2):
Workshop on Adverse Perinatal Outcomes of Gestational 176–185.
Diabetes Mellitus, December 3–4, 1992. Diabetes Care 1995; 40. Bardenheier BH, Elixhauser A, Imperatore G et al. Variation
18: 127–129. in prevalence of gestational diabetes mellitus among hospital
20. HAPO Study Cooperative Research Group. Hyperglycemia discharges for obstetric delivery across 23 states in the United
and adverse pregnancy outcomes. N Engl J Med 2008; 358: States. Diabetes Care 2013; 29: 1209–1214.
1991–2002. 41. Metzger BE and Buchanan TA. Gestational diabetes. In: Textbook
21. Pedersen J. Diabetes and pregnancy: Blood sugar of new- of Diabetes and Pregnancy, 2nd edn. National Diabetes Data
born infants. PhD thesis, Danish Science Press: Copenhagen, Group, National Institute of Diabetes and Digestive and Kidney
Denmark, 1952, p. 230. Diseases, National Institutes of Health, Bethesda, MD.
22. The HAPO Study Cooperative Research Group. Hyper 42. Bullard KM, Saydah SH, Imperatore G, Cowie CC, Gregg EW,
glycemia and Adverse Pregnancy Outcome (HAPO) Study: Geiss LS, Cheng YJ, Rolka DB, Williams DE, Caspersen CJ.
Associations with neonatal anthropometrics. Diabetes 2009; Secular changes in US prediabetes prevalence defined by Hb A1c
58: 453–459. and fasting plasma glucose. Diabetes Care 2013; 36: 2286–2293.
23. Catalano PM, McIntyre HD, Cruickshank JK et al.; for the HAPO 43. Mendez-Figueroa H, Daley J, Lopes VV, Coustan DR. Comparing
Study Cooperative Research Group. The Hyperglycemia and daily versus less frequent blood glucose monitoring in patients
Adverse Pregnancy Outcome (HAPO) Study: Associations of with mild gestational diabetes. J Matern Fetal Neonatal Med
GDM and obesity with pregnancy outcomes. Diabetes Care 2013; 26: 1268–1272.
2012; 35: 780–786. 44. Crowther CA, Hiller JE, Moss JR et al.; for the Australian
24. International Association of Diabetes and Pregnancy Study Carbohydrate Intolerance Study of Pregnant Women (ACHOIS)
Groups Consensus Panel. International Association of Diabetes Study Group. Effect of treatment of gestational diabetes mellitus
and Pregnancy Study Groups recommendations on the diagnosis on pregnancy outcomes. N Engl J Med 2005; 352: 2477–2486.
and classification of hyperglycemia in pregnancy. Diabetes Care 45. Landon MB, Spong CY, Thom E et al.; for the Eunice Kennedy
2010; 33: 676–682. Shriver National Institute of Child Health and Human
25. Pettitt DJ, Bennett PH, Saad MF, Charles MA, Nelson RG, Development Maternal-Fetal Medicine Units Network. A multi-
Knowler WC. Abnormal glucose tolerance during pregnancy in center randomized trial of treatment for mild gestational diabetes.
Pima Indian women. Long-term effects on offspring. Diabetes N Engl J Med 2009; 361: 1339–1348.
1991; 40(Suppl. 2): 126–130. 46. Mission JF, Ohno MS, Cheng YW, Caughey AB. Gestational
26. Sacks DA, Greenspoon JS, Abu-Fadil S, Henry HM, Wolde- diabetes screening with the new IADPSG guidelines: A
Tsadik G, Yao JF. Toward universal criteria for gestational dia- cost-effectiveness analysis. Am J Obstet Gynecol 2012; 207:
betes: The 75-gram glucose tolerance test in pregnancy. Am J 326.e1–326.e9.
Obstet Gynecol 1995; 172(2 Pt. 1): 607–614. 47. Werner EF, Pettker CM, Zuckerwise L et al. Screening for ges-
27. Sermer M, Naylor CD, Gare DJ et al. Impact of increasing car- tational diabetes mellitus: Are the criteria proposed by the
bohydrate intolerance on maternal-fetal outcomes in 3637 International Association of the Diabetes and Pregnancy Study
women without gestational diabetes. The Toronto Tri-Hospital Groups cost-effective? Diabetes Care 2012; 35: 529–535.
Gestational Diabetes Project. Am J Obstet Gynecol 1995; 173(1): 48. Ratner RE, Christophi CA, Metzger BE et al. Prevention of diabetes
146–156. in women with a history of GDM: Effects of metformin and life-
28. Jensen DM, Damm P, Sørensen B et al. Clinical impact of mild style interventions. J Clin Endocrinol Metab 2008; 93: 4774–4779.
carbohydrate intolerance in pregnancy: A study of 2904 non- 49. Spirito A, Williams C, Ruggiero L et al. Psychological impact of
diabetic Danish women with risk factors for gestational diabe- the diagnosis of gestational diabetes. Obstet Gynecol 1989; 73:
tes mellitus. Am J Obstet Gynecol 2001; 185(2): 413–419. 562–566.
50. Sermer M, Naylor CD, Farine D et al. The Toronto Tri-Hospital 54. American College of Obstetricians and Gynecologists Practice
Gestational Diabetes Project. A preliminary review. Diabetes Bulletin No. 137. Gestational diabetes mellitus. Obstet Gynecol
Care 1998; 21(Suppl. 2): B33–B42. 2013; 122: 406–416.
51. Dooley SL, Keller JD, Metzger BE et al. Screening for gestational 55. American Diabetes Association. Standards of care 2015. Diabetes
diabetes: Is the 140 mg/dL threshold appropriate? In: Society of Care 2015; 8(Suppl. 1): s8–s16.
Perinatal Obstetricians Annual Scientific Meeting, New Orleans, 56. Blumer I, Hadar E, Hadden DR et al. Diabetes and pregnancy:
LA, February 1989, Abstract 52. An endocrine society clinical practice guideline. J Clin Endocrinol
52. American Diabetes Association. Standards of care 2011. Diabetes Metab 2013; 98: 4227–4249.
Care 2011; 34(Suppl. 1): s15. 57. Hod M, Kapur A, Sacks DA et al. The International Federation
53. World Health Organization. Diagnostic criteria and classification of Gynecology and Obstetrics (FIGO) Initiative on gestational
of hyperglycaemia first detected in pregnancy. Available from: diabetes mellitus: A pragmatic guide for diagnosis, management,
http://www.who.int/diabetes/publications/Hyperglycaemia_In_ and care. International Journal of Gynecology and Obstetrics
Pregnancy/en/, accessed on February 5, 2014. 2015; 131(S3): S173–S211.
15 Cost-effectiveness of screening
and management programs for
gestational diabetes mellitus
Louise K. Weile, James G. Kahn, Elliot Marseille, and Nicolai Lohse
these include up to a 3.5-fold higher risk of macrosomia7–10

Introduction and elevated risk of shoulder dystocia and hypoglycemia.7–8
Whether gestational diabetes mellitus (GDM) screening and For the women, these include increased risk of cesarean
management by different criteria and in different settings section, polyhydramnios, gestational hypertension, and

is cost-effective has been debated for more than a decade. preeclampsia.7–12 The recently published IADPSG diagnos-
A 2009 report from The Lancet’s Stillbirth Series steering tic criteria,5 which are based on data from a large multi-
committee put detection and management of GDM on a pri- center and multinational prospective study,7 are lower than
ority list of interventions to reduce stillbirth,1 and the poten- some commonly used diagnostic criteria13–16 and therefore
tial key role of GDM on the long-term health of the mother lead to a two- to threefold higher prevalence of GDM.17,18
and her offspring is being discussed at the highest political The increase in GDM risk factors and the changed diag-
level in the United Nations (UN) General Assembly and nostic criteria are both pushing GDM prevalence upward.
other UN bodies.2–4 The suggestion of new screening criteria The World Health Organization very recently approved the
by the International Association of Diabetes and Pregnancy IADPSG criteria.19
Study Groups (IADPSG)5 in 2010 has added to the debate. Less visible are the long-term risks that potentially have
These extensive deliberations notwithstanding, we are only a much larger effect on population health. GDM is increas-
beginning to understand which interventions might be ingly mentioned as a possible contributor to the type 2 dia-
effective at reducing the burden of GDM, and at what cost.6 betes mellitus (T2DM) epidemic.20–23 Women with GDM
To institute appropriate policies, decision makers at national have a more than sevenfold increased risk of T2DM (pooled
and global levels need additional reliable information on the relative risk = 7.43),24 with cumulative incidence up to 70% in
cost and cost-effectiveness of GDM screening and treatment. 3 years in some populations.25 The risk of T2DM is most pro-
This chapter aims to equip the reader with the concep- nounced for women with high glucose values at the diagnos-
tual tools and substantive background needed to understand tic OGTT during pregnancy or postpartum continuation of
and evaluate cost-effectiveness assessments of GDM and the glucose intolerance,26 greater body mass index, use of insu-
implications of global policy from a health economic per- lin during pregnancy (as a marker of severity of GDM), and
spective. We start by a short background section. We then earlier gestational age at diagnosis of GDM.27 The offspring
introduce and explain basic concepts of health economics as of mothers with hyperglycemia during pregnancy also carry
they relate to this issue, including resources and cost, health an increased risk of prediabetes and T2DM later in life, four
metrics, and different types of health economic evaluations. to eight times that of offspring born to mothers with nor-
We review the current health economic evaluations on GDM mal glucose tolerance during pregnancy,28,29 with a 20-year
screening and management, while elaborating these concepts. cumulative incidence of up to 21%.29 A modeling study from
Finally, we provide a brief summary of cost-effectiveness Canada estimated GDM as a forerunner of as many as 30%
results and discuss future directions for research. of T2DM cases in a high-prevalence diabetes population.30,31
This link between GDM and the long-term health of the
pregnant woman and her offspring has only recently come to
Background the attention of the GDM community.
GDM and T2DM share risk factors such as obesity and
Current diagnostic and therapeutic recommendations for other metabolic disturbances, and the two epidemics are
GDM are founded on the growing body of evidence of the likely to move in tandem.4,32–34 Since the prevalence of
short-term risks associated with GDM. For the offspring, GDM is increasing globally, identifying and providing
119
120 Cost-effectiveness of screening and management programs for gestational diabetes mellitus
Table 15.1 Key types of cost-effectiveness analysis in health
Definition Outcome metric Use to decide what?

Cost-effectiveness analysis Net cost per added unit of ICER (i.e., cost/T2DM case What’s the incremental cost
health averted or cost/brachial per added health benefit?
plexus injury averted) Is it worth doing versus a
threshold?
Cost-effectiveness analysis Net cost per added ICER (net costs/QALY What’s the incremental cost
(CUA)—with a standardized standardized unit of gained or DALY averted) per added QALY or
health metric health (a ratio) averted DALY?
Is it worth doing versus a
threshold?
Cost–benefit analysis Cost of intervention minus Net costs (benefits−costs) Does the program produce
savings (a difference, not Typically includes benefits more value than it costs?
a ratio) beyond health Is it worth doing?
Abbreviations: DALY, disability-adjusted life year, a measure of disease burden that sums premature mortality (LY) plus disabling morbidity
(DA); ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life year, a measure of health that tallies years alive
adjusted for health status of those years.
good care for women with GDM could have substantial costs in monetary units, whereas they differ in the assess-
impact on population health in both high- and low-income ment of the consequences (Table 15.1).35
countries.
Cost-effectiveness and cost–utility analysis
Within a GDM screening context, the outcome of a CEA
Basic health economic theory could be found cases of GDM or prevented cases of neona-
tal morbidity depending on the time horizon used. As seen,
The benchmark of health economics is that resources such the former outcome focuses on the effectiveness of screen-
as labor, supplies, facilities, and equipment are scarce. ing, whereas the latter gives the opportunity of incorporat-
The use of resources must therefore be based on priority. ing both effectiveness of GDM screening and management.
The focus of many economic evaluations both within the However, both outcomes are specific to GDM and the obstet-
healthcare sector and elsewhere is on whether the com- ric field, and using these outcomes will narrow the scope of
mitments of resources in one area yield as good a return the economic evaluation. For decision making, these CEAs
as would have been obtained if resources were committed can only be used for comparison within an obstetric budget,
elsewhere. An essential component in any such analysis which limits their use for broader priority setting.
is the marginal cost, understood as the cost of produc- A specific type of CEA combines both mortality and
ing one additional unit of output such as a unit of health morbidity effects into one standardized single metric of
benefit. Costs and consequences are the input and output, change in health. This metric is either “disability-adjusted
respectively, of given health interventions. Drummond life years” (DALYs) or “quality-adjusted life years” (QALYs).
et al. define economic evaluation as “(…) the compara- DALYs measure the disease burden as it both represents
tive analysis of alternative courses of action in terms of premature mortality and disability due to morbidity. 36
both their costs and consequences.”35 Economic evalua- Imagine an individual who would normally expect to live
tions therefore aim at identifying, measuring, valuing, and another 10 years. He is now diagnosed with T2DM and
comparing the costs and consequences of certain health T2DM-related complications that give him a 10% disabil-
alternatives. Thus, economic evaluations offer input for ity compromise while alive, and he dies 8 years later. He
decision makers once issues of efficacy, effectiveness, and will have a DALY burden of 2 (lost life years) + 0.1 (disabil-
availability are settled. 35 ity) ∗ 8 (years living with that disability) = 2.8 DALYs. The
QALY is a measure of health and essentially the negative
of the DALY. Thus, an illness that shortens life by 2 years
Evaluation types and lowers “health status utility” by 10% for 8 years would
Only so-called full analyses can address the question of cost- decrease QALYs by 2.8. Since QALY incorporates the util-
effectiveness. Cost-effectiveness analysis (CEA), comprising ity of different health states into the outcome measure,
cost–utility analysis (CUA), and cost–benefit analysis (CBA) analyses using this outcome are often classified as cost–
are full analyses, because they compare costs and conse- utility analysis.
quences between two and more alternatives. Partial analyses A CEA compares two different health interventions and
focus solely on costs or consequences, or they neglect com- expresses the net cost with the units of health benefit gained
parison between alternatives. CEAs, CUAs, and CBAs assess in ratio form, for example, as the “net cost per death averted”
Basic health economic theory 121
or the “net cost per QALY gained.” This ratio is called the can be linked to endpoints by extending the time horizon.
“incremental cost-effectiveness ratio” (ICER), because both These models are suitable for decision making under uncer-
the costs and health benefits are incremental.35 Thus, the tainty, and they also allow effectiveness to be estimated on
ICER expresses the additional cost incurred to gain one multiple empirical studies. In the latter case, inclusion of
incremental health unit. The ICER has no useful meaning studies should be based on a systematic literature search
when programs result in net savings as there is no trade-off and include potential weighing and pooling of the studies.
between costs and benefits. Instead, such intervention is clas- Decision models are often structured as decision trees or
sified as “dominant” (both cheaper and better) and as such Markov models that can both be classified as cohort models.
should be a “no brainer” to health planners. Conventionally, Decision trees represent the possible prognosis of an indi-
the absolute magnitude of savings and health gains, with no vidual with or without an intervention. Events can be linked
ratio, is then reported.36 to costs and consequences, whereby the expected values of
DALYs and QALYs are generic health metrics and there- course of action within the competing alternatives can be
fore not restricted to a certain health field. Thus, analyses calculated.35
using one of these metrics are often sufficient for decision
makers comparing interventions broadly within the health- Confronting uncertainty
care sector. The models constructed to calculate cost-effectiveness are
only representations of the factors that determine costs and
Cost–benefit analysis
consequences and of the way those factors interact. The use-
Though the concept of “cost–benefit” is frequently used as a fulness of a model is therefore limited by the fidelity with
layman’s term, CBA is actually a rarely used type of analy- which it represents the underlying reality. Fidelity in turn
sis in healthcare. The power of CBA is that it permits com- depends importantly on the accuracy and precision of the
parisons of widely disparate funding alternatives. In the estimates of the models’ input values such as disease preva-
CBA framework, the value of spending on sewerage work lence, the cost of screening, or the effectiveness of the inter-
could be compared with spending on a maternal and neo- ventions being evaluated. An integral part of almost any
natal health initiative. Hence, in CBA, all health outcomes economic evaluation is therefore the systematic and trans-
including human life are assigned a monetary value,36 i.e., parent portrayal of the level of uncertainty and the confi-
the value of longer life and lower morbidity due to T2DM or dence that should therefore be placed in the outputs of the
other GDM-associated complications might be expressed as model. The rigorous assessment of the magnitude and effects
the sum of medical costs averted and economic productivity of uncertainty consists of a set of techniques known as “sen-
added compared with the cost of delivering GDM screening sitivity analyses” since they quantify how “sensitive” results
and management. are to changes in input values or in the way those inputs
Due to the use of a monetary outcome, CBA enables gen- interact. Sensitivity analyses should not be considered an
eral priority setting within all sectors, and this is a large part optional feature of CEAs, but rather an essential part of the
of its power and appeal. However, the monetary valuation analysis, without which it is perilous to draw any conclusions.
of health outcomes often places great demand on data and Figure 15.1 represents a set of so-called one-way sensi-
analytic technique. Furthermore, placing a value on human tivity analyses arrayed in a “tornado diagram.” It shows the
health raises ethical issues, and for these reasons, CBA effect on the outcome, in this case the cost-effectiveness of
is rarely used in the evaluation of clinical or public health GDM screening in Chennai, India,37 of an array of model
programs.36 inputs. The input with the greatest influence is at the top,
and each successively less important one is stacked beneath
Designing health economic evaluations it. The input with the greatest influence on cost-effectiveness
Health economic evaluations typically combine patient-level is the cost of postpartum care. At 50%–150% of the $2846
data with decision analytic modeling. Use of patient-specific base case value, the ICER ranges from $887 to $2385 per
cost and consequence data, often obtained from clinical tri- DALY averted. These are called “one-way” sensitivity analy-
als, may increase the precision of the analysis, but the results ses because they show the effect of changing one input at a
may not be applicable outside of the trial setting. Clinical tri- time, holding all of the other inputs constant. Other sensi-
als measure efficacy, whereas the economic evaluation wants tivity analyses techniques can document the effect of simul-
to estimate the effectiveness of a health intervention. Thus, taneously varying two, three, or more variables, including
using data collected with strict inclusion criteria will reduce stochastic variations.
the external validity of the economic evaluation. An eco-
nomic evaluation of GDM screening and management based Costing methods
on data from a clinical trial will have difficulties estimating Costs are traditionally classified as direct, indirect, or intangi-
the effect on persons outside of the trial, and the often short ble costs, but these can also be classified according to context:
duration of clinical trials may impede the very important consumed within the healthcare sector and other sectors, by
estimation of long-term cost-effectiveness. patients or relatives, or if a loss of production is seen. In order
Decision analytic modeling allows all relevant alterna- to take account of inflation due to different timing of the
tives to be taken into account, and intermediary outcomes costs, these are typically valued after a given currency from a
Postpartum care cost

T2DM incidence—Mother
T2DM reduction—Mother
Discount rate
DALYs—T2DM—Mother
Perinatal AE costs—Intervention
T2DM incidence—Child
T2DM reduction—Child
Prevalence
DALYs—T2DM—Child
Antenatal care cost
DALYs—AEs—No Intervention
Initial screen cost
T2DM cost—Mother
DALYs—AEs—Intervention
T2DM cost—Child
$800 $1000 $1200 $1400 $1600 $1800 $2000 $2200 $2400 $2600
Net cost per DALY averted (International dollars)
Figure 15.1 Sensitivity analyses arrayed in a tornado diagram. (From Marseille, E. et al., J. Matern. Fetal. Neonatal Med., 26(8),
802, 2013. With permission.)
certain price year. Sources and methods for cost valuation are In total, 12 publications were identified. Over time, the
essential for the validity of the costing. As the health market published evaluations moved from analyses of measured
is flawed by several attributes of the free market (i.e., third- patient outcomes to decision analytic modeling, which
party payment, regulations, and monopolies), costs should allowed long-term outcomes to be considered. Furthermore,
optimally be valued as the marginal cost—understood as the the most recently published decision models were more
value of those benefits that cannot be gained if the resource sophisticated, including input values and model design
is not available.35 However, the concept of marginal costs will (Table 15.2).
not be further elaborated in this chapter.
Discounting is based on the existence of “time preference”
that is a preference of having resources now as opposed to Main findings
later from the assumption that one can benefit from them in In general, large variations in the reported cost-effectiveness
the interim. Both costs and consequences can be discounted were identified. It was therefore impossible to draw broad
according to a chosen discount rate per year. Discounting is conclusions about the cost-effectiveness of GDM screening
a standard technique, but is sometimes considered contro- from this dataset. First, many of the results (Table 15.3) were
versial.35 In relation to GDM screening and management, for derived from one study,38 which used natural outcomes such
example, discounting future saved costs and health gains of the as “averted cases of perinatal mortality” or “averted cases of
offspring will imply inequality between mothers and offspring. hypertension disorder.” In these analyses, total costs were
divided by each of the outcomes separately, thus preventing
Comprehensiveness and scope an assessment of the cost of the combined results. Second,
A comprehensive assessment of cost-effectiveness relies on even where broad measures such as QALYs gained or DALYS
inclusion of all relevant health program alternatives, meaning averted were used, ICERs were calculated against different
that a no-intervention alternative is essential if the full relevant comparators and included different beneficiaries (i.e., some
range of options are to be assessed. The scope of an analysis is women only, some offspring only, and some both). Finally,
defined by the perspective (of the patient, healthcare system, one study37 included averted cases of T2DM in both women
or society) and the applied time horizon. Ideally, a societal and their offspring. This required the estimation of both
perspective and lifelong horizon should be applied, and there the cost of postnatal interventions to avert T2DM and the
must be consistency between stated perspective, time horizon, discounted future benefits of averted case of T2DM. Since
and the design of the analysis and included data.35 averted T2DM is a potentially large component of the ben-
efit of GDM screening, studies that do not take these into
account cannot be compared directly with those that do.
Few assessments have compared IADPSG criteria to other
Review of current literature on accepted screening criteria, and results point toward the
GDM cost-effectiveness IADPSG criteria being cost-effective, but with great varia-
tions between studies.49
The following review includes all health economic evalua- The heterogeneity of the cost-effectiveness ratios could
tions published from 2002 to 2014.37–48 partly be explained by the different methods for calculation.
Review of current literature on GDM cost-effectiveness 123
Table 15.2 Identification, measurement, and valuation of cost and consequences
Consequence
sources/data Key input for
Study Key cost components Cost sources/data handling handling decision modeling
Poncet et al.38 Screening tests Prospective study (n = 120) Clinical GDM prevalence
Obstetrical care Hospital in Rhône-Alpes, effectiveness: Screening refusal
Management of GDM France published rate
Delivery care French key-letter costing literature Risk factors
Sick leave system (≈DRG), Test effectiveness
literature, and expert GDM treatment
opinions success
Discounting (NS) Decision modeling
Discounting (NS)
Larijani et al.39 Test materials Source for cost valuation Clinical
Time for testing (staff) (NS) effectiveness:
cohort (n =
2416), four
university
hospitals in
Teheran, Iran,
period or retro-/
prospective
design: NS
Discounting (NS) Prevalence estimate
on patient-level
data
Discounting (NS)
Nicholson et al.40 Treatment of maternal/ The 2003 Medicare Clinical GDM prevalence
neonatal resource-based relative effectiveness and Test effectiveness
complications (direct value units/The Maryland utilities: Hypertensive
and indirect costs) HealthCare Commission published disease
Database/The Bureau of literature Polyhydramnios
Labor Statistics Life expectancy: Cesarean/vaginal
the National delivery and
Center for Health associated
Statistics complications
Market prices adjusted Decision modeling Neonatal
using cost-to-charge Discounting at 3% hypoglycemia
ratios rate per year Macrosomia
Inflated into 2003 US$ Shoulder dystocia
using consumer price Morbidity/infant
index death
Discounted at 3% rate per
year
Ayach et al.41 Laboratory costs (ND) Public reimbursements Clinical
effectiveness:
prospective
cohort (n = 341)
in university
hospital in Mato
Grosso do Sul,
Brazil, July
1997–December
1999
Discounting (NS) Prevalence estimate
on patient-level
data
Discounting (NS)
(Continued)
Table 15.2 (Continued) Identification, measurement, and valuation of cost and consequences
Consequence
Thung et al.42 Identified costs (ND) Units/quantity/sources (NS) Clinical Perinatal death
effectiveness, Shoulder dystocia
lifetime
expectancy, and
utilities (NS)
Discounting (NS)
NCCWCH (United Testing (ND) National Health Service/ Clinical Test effectiveness/
Kingdom)43 Treatment of GDM literature effectiveness and acceptability
(staff’s time, materials, utilities: GDM treatment
medicine) published Induction of labor/
Treatment of GDM- literature cesarean section
associated Lifetime expectancy Stillbirth/neonatal
complications (ND) (NS) death
Inflated into 2006 U.K. Decision modeling Shoulder dystocia
pounds using retail price Discounting (NS) Bone fracture
index Nerve palsy
Discounting (NS) Jaundice
Hypoglycemia
Lee et al.44 Treatment of GDM Units/quantity/sources (NS) Clinical Maternal death due
Treatment long-term effectiveness: to preeclampsia
child complications published Permanent brachial
Preeclampsia literature plexus injury
management Lifetime expectancy
and utilities (NS)
Discounting (NS)
Meltzer et al.45 Test materials Statistics Canada/Canada Clinical
Time (staff/women) Revenue Agency effectiveness:
RCT (n = 1594)
in university
hospital in
Montreal,
Quebec,
Canada, January
2001–2002
Inflated into 2002 Prevalence estimate
Canadian $ on patient-level
Discounting (NS) data
Discounting (NS)
Round et al.46 Test materials Literature/National Health Clinical Test effectiveness/
Time for testing (staff) Service data effectiveness and acceptability
Treatment of GDM utilities: GDM treatment
Treatment of GDM- published Preeclampsia
associated literature Induction of labor/
complications Lifetime expectancy: cesarean section
80 years Neonatal jaundice
(reference NS) Admission to
Inflated to 2009 U.K. pounds Decision modeling neonatal nursery
Discounting at 3.5% rate Discounted at 3.5%
per year rate per year
(Continued)
Table 15.2 (Continued) Identification, measurement, and valuation of cost and consequences
Consequence
Mission et al.47 Treatment/diagnosis of Published literature Clinical Test effectiveness
GDM (direct and effectiveness, GDM treatment
indirect costs) utilities, lifetime Preeclampsia
Admission for GDM- expectancy: Mode of delivery
related conditions published Shoulder dystocia
omitted literature Macrosomia
Inflated into 2012 US$ Decision modeling Brachial plexus
using consumer price Discounting at 3% injury
index rate per year Hypoglycemia
Discounting (NS) Hyperbilirubinemia
NICU admission
Maternal/neonatal
death
Werner et al.48 Testing (ND) Published literature/ Clinical Test effectiveness
Prenatal care Medicare effectiveness, Preeclampsia
Treatment of GDM- reimbursements utilities and Shoulder dystocia
associated lifetime Preterm birth
complications (short expectancy: Cesarean delivery
and long term) published Stillbirth
Prevention of GDM literature NICU admission
Inflated into 2011 US$ Decision modeling Maternal diabetes
using consumer price Discounting at 3% postpartum
index rate per year
Discounting at 3% rate per (sensitivity
year (sensitivity analyses: analyses: 1%–5%
1%–5% rate per year) rate per year)
Marseille et al.37 Testing (materials, staff) Personal communications/ Clinical GDM prevalence
GDM treatment and published literature effectiveness and Test performance
intervention utilities: Effectiveness of
Treatment of T2DM published treatment
literature Full set of perinatal
Lifetime adverse events
expectancy: Maternal/offspring
country-specific T2DM
WHO life tables
from 2009
Converted into 2011 Decision modeling
International $ Discounting at 3%
Discounting at 3% rate per rate per year
year
Source: Weile, L.K. et al., Best Pract. Res. Clin. Obstet. Gynaecol., 29(2), 206, 2015.
Abbreviations: DRG, diagnosis-related group; GDM, gestational diabetes mellitus; ND, not defined; NS, not stated; RCT, randomized controlled
trial; T2DM, type 2 diabetes mellitus.
It was questionable whether the separately calculated ICERs GCT ± 100 g OGTT were most frequently found to be cost-
performed in two studies38,40 clinically made sense. As the effective. However, it was questionable if the studies includ-
screening alternatives found most cost-effective in both these ing these two screening types were comparable, as not only
studies, it did not differ within strata. As the two studies44,47 the criteria used for diagnosis but also the definition of the
did not include a no-screening alternative, the possibility of screened population varied widely among the evaluations.
the alternatives in these evaluations actually not being cost- Thus, some studies relied on universal screening, whereas
effective compared with no-screening could not be excluded. others relied on selected screening. Furthermore, within
One-step 2-hour 75 g OGTT and two-step 1-hour 50 g the selective screening options, different definitions of risk
Table 15.3 Estimates of cost-effectiveness
Screening Incremental Incremental

Studya alternatives Reference cost cost (2014 US$) Per Comment
Poncet et al.38 Two-step No screening €21,069.52 40,331 Averted case of
1-hour 50 g macrosomia
OGTT
€9,953.24 19,055 Averted case of
±100 g
prematurity
3-hour
OGTT, on €7,871.55 15,071 Averted case of
high-risk perinatal
pregnant mortality
women €28,674.90 54,898 Averted case of
hypertension
disorder
Two-step €23,135.36 44,292 Averted case of
OGTT €10,965.20 20,992 Averted case of
±100 g prematurity
3-hour
€8,663.83 16,587 Averted case of
OGTT,
perinatal
universal
mortality
screening
€31,898.74 61,070 Averted case of
hypertension
disorder
One-step €68,933.79 13,1974 Averted case of
OGTT, €37,320.89 71,451 Averted case of
universal prematurity
screening
€29,444.16 56,370 Averted case of
perinatal
mortality
€94,506.04 18,0931 Averted case of
hypertension
disorder
Nicholson Two-step Referent QALYs saved Maternal
et al.40 1-hour 50 g consequences
GCT considered
±3-hour
100 g
OGTT
One-step Two-step US$32,374 41,147
3-hour 100 g 1-hour 50 g
OGTT GCT
±3-hour
100 g
OGTT
Two-step Referent Neonatal
1-hour 50 g consequences
GCT considered
±3-hour
100 g
OGTT
OGTT GCT
±3-hour
100
OGTT
(Continued)
Table 15.3 (Continued) Estimates of cost-effectiveness

Thung et al. 42 Two-step No screening US$12,269 14,232 QALYs gained
1-hour 50 g
GCT (BG >
7.8 1 hour),
±3-hour
100 g
OGTT
Two-step Two-step US$14,961 17,355
GCT (BG GCT (BG >
>7.2 1 7.8 1 hour),
hour), ±3-hour
±3-hour 100 g
100 g OGTT
OGTT
NCCWCH One-step No screening UK£3,677 5,843 QALYs gained
(United 2-hour 75 g
Kingdom)43 OGTT +
ADA risk
criteriaa
One-step One-step UK£21,738 34,543
OGTT + OGTT +
high-risk ADA risk
ethnicity criteriab
Lee et al.44 One-step Referent QALYs gained
2-hour 75 g
OGTT (BG
> 7.8
mmol/L 2
hour)
One-step One-step US$76,000 85,804
OGTT (BG OGTT
>7.0 (BG >7.8
mmol/L 2 2 hour)
hour)
Round et al.46 No screening Distribution Maximum Maximum QALYs gained Woman’s
of the willingness willingness to individual risk
most to pay pay threshold of GDM <1%
Two-step cost- threshold at at US$34,063 Woman’s
FBG ±2-hour effective UK£20,000 individual risk
75 g OGTT screening of GDM
alternative 1.0%–4.2%
One-step Woman’s
2-hour 75 g individual risk
OGTT of GDM >4.2%
Mission et al.47 Two-step Referent QALYs gained

1-hour 50 g
GCT
±3-hour
100 g OGTT
OGTT GCT
±3-hour
100 g OGTT
(Continued)
Table 15.3 (Continued) Estimates of cost-effectiveness

Werner et al. 48 Two-step No screening US$543,119 564,844 QALYs gained If diagnosis
1-hour 50 g provides no
GCT long-term
±3-hour maternal
100 g OGTT health benefit
FBG ±2-hour 1-hour 50 g
75 g OGTT GCT
±3-hour
100 g OGTT
Two-step No screening US$16,689 17,351 If diagnosis
1-hour 50 g provides
GCT long-term
±3-hour maternal
100 g OGTT health benefit
FBG ±2-hour 1-hour 50 g
75 g OGTT GCT
±3-hour
100 g OGTT
Marseille et al.37 One-step No screening International 580 Averted DALY India
2-hour 75 g $1,626
OGTT International 2,070 Israel
$1,830
Source: Weile, L.K. et al., Best Pract. Res. Clin. Obstet. Gynaecol., 29(2), 206, 2015.
Abbreviations: BMI, body mass index; DALY, disability-adjusted life years; QALY, quality-adjusted life years; 1-hour 50 g GCT, 1-hour 50 g
glucose challenge test; 2-hour 75 g OGTT, 2-hour 75 g oral glucose tolerance test; 3-hour 100 g OGTT, 3-hour 100 g oral
glucose tolerance test.
a Results from studies not performing incremental analyses were excluded in the synthesis of cost-effectiveness estimates.
b ADA risk criteria: American Diabetes Association definition of risk is having more of the following risk factors: >25 years, BMI > 27 kg/m2;
family history of diabetes, and high-risk ethnic group.
factors were identified. As these small variations might have supports a comprehensive review of feasible options and
a significant clinical effect, none of the evaluations could be helps place other alternatives in context. Moreover, the
said to compare the same combination of alternatives. Thus, potential of preventing effects on T2DM and other long-term
no summary conclusions could be made on these findings.49 complications for both mothers and their offspring should
be incorporated in the analysis.
To produce results as relevant as possible to decision
making, we encourage researchers to challenge their results
Future directions for health policies by addressing ethical issues in the discussion of the results.
and health economic evaluations If no screening is found to be the most cost-effective, is
this actually a politically viable option within the relevant
As stated earlier, making general assessments of the cost- setting? Similarly, if the cost per health gain is high (i.e.,
effectiveness of GDM screening and management based on poor cost-effectiveness), how will implementation of such
the current evidence is difficult. We recommend that future alternatives affect other health initiatives depending on
research addressing this topic from a health economic per- the same budget? As resources are scarce, priority is indis-
spective focus on national or regional policy alternatives pensable and trade-offs should be quantified. This is the
instead of global conditions: a relevant assessment at least heart of economic assessment and the main reason to per-
includes the national or regional existing screening program form CEAs. Hence, discussions of, for instance, whether
and screening alternatives that from a clinical perspective is GDM screening and management should take precedence
judged worthy of recommendation. Currently, the screen- over other interventions in pregnancy (i.e., screening
ing program and criteria recommended by IADPSG there- for preeclampsia or general health promotion) are most
fore should be included as one alternative. Furthermore, a needed. In theory, the discussion could even be expanded
no-screening alternative should always be included, as this to compare GDM interventions with other nonobstetric
References 129
health interventions. The complexity of the full range of and bureaucratic actors at any given time. By helping to
such possible comparisons is overwhelming. In practice, guide us to options that yield high value for money, high-
this complexity is greatly constrained by far smaller num- quality health economic evaluations can thus help address
ber of the actual policy-relevant choices facing political issues of broad importance to society.
REFERENCES
1. Pattinson R, Kerber K, Buchmann E et al. Stillbirths: How can 19. Colagiuri S, Falavigna M, Agarwal MM et al. Strategies for imple-
health systems deliver for mothers and babies? Lancet 2011; menting the WHO diagnostic criteria and classification of hyper-
377(9777): 1610–1623. glycaemia first detected in pregnancy. Diabetes Res Clin Pract
2. United Nations. Prevention and control of non-communicable March 2014; 103(3): 364–372.
diseases, Report of the United Nations Secretary General Ban 20. Dabelea D. The predisposition to obesity and diabetes in off-
Ki-Moon, UN, 2011. Available from: http://www.un.org/en/ spring of diabetic mothers. Diabetes Care 2007; 30(Suppl. 2):
ga/search/view_doc.asp?symbol=A/66/83&Lang=E (Accessed S169–S174.
November 10, 2015.) 21. Dabelea D, Hanson RL, Bennett PH et al. Increasing prevalence
3. United Nations. Political declaration of the high-level meeting of Type II diabetes in American Indian children. Diabetologia
of the general assembly on the prevention and control of non- 1998; 41(8): 904–910.
communicable diseases, United Nations General Assembly 22. Cheung NW, Byth K. Population health significance of gesta-
2011. Available from: http://www.un.org/ga/search/view_doc. tional diabetes. Diabetes Care 2003; 26(7): 2005–2009.
asp?symbol=A/66/L.1 (Accessed November 10, 2015.) 23. Dabelea D, Snell-Bergeon JK, Hartsfield CL et al. Increasing
4. United Nations Procurement Division. Discussion Paper prevalence of gestational diabetes mellitus (GDM) over time and
Addressing the Social Determinants of Noncommunicable by birth cohort: Kaiser Permanente of Colorado GDM screening
Diseases, UNDP 2013. Available from: http://www.undp. program. Diabetes Care 2005; 28(3): 579–584.
org /content /dam/undp/library/hivaids/English/Discussion_ 24. Bellamy L, Casas JP, Hingorani AD, Williams D. Type 2 diabetes
Paper_ Addressing _the_ Social_ Determinant s _of_ NCDs _ mellitus after gestational diabetes: A systematic review and meta-
UNDP_2013.pdf (Accessed November 10, 2015.) analysis. Lancet 2009; 373(9677): 1773–1779.
5. International Association of Diabetes and Pregnancy Study 25. Kim C, Newton KM, Knopp RH. Gestational diabetes and the
Groups Consensus Panel et al. International Association of incidence of type 2 diabetes: A systematic review. Diabetes Care
Diabetes and Pregnancy Study Groups recommendations on 2002; 25(10): 1862–1868.
the diagnosis and classification of hyperglycemia in pregnancy. 26. Lauenborg J, Hansen T, Jensen DM et al. Increasing incidence
Diabetes Care March 2010; 33(3): 676–682. of diabetes after gestational diabetes: A long-term follow-up in a
6. Jiwani A, Marseille E, Lohse N et al. Gestational diabetes mel- Danish population. Diabetes Care 2004; 27(5): 1194–1199.
litus: Results from a survey of country prevalence and practices. 27. Baptiste-Roberts K, Barone BB, Gary TL et al. Risk factors for type
J Matern Fetal Neonatal Med June 2012; 25(6): 600–610. 2 diabetes among women with gestational diabetes: A systematic
7. Metzger BE, Lowe LP, Dyer AR et al. Hyperglycemia and review. Am J Med 2009; 122(3): 207–214.e1–e4.
adverse pregnancy outcomes. N Engl J Med 2008; 358(19): 28. Dabelea D, Hanson RL, Lindsay RS et al. Intrauterine exposure to
1991–2002. diabetes conveys risks for type 2 diabetes and obesity: A study of
8. Langer O, Yogev Y, Most O, Xenakis EM. Gestational diabetes: discordant sibships. Diabetes 2000; 49(12): 2208–2211.
The consequences of not treating. Am J Obstet Gynecol 2005; 29. Clausen TD, Mathiesen ER, Hansen T et al. High prevalence of
192(4): 989–997. type 2 diabetes and pre-diabetes in adult offspring of women
9. Keshavarz M, Cheung NW, Babaee GR et al. Gestational dia- with gestational diabetes mellitus or type 1 diabetes: The role
betes in Iran: Incidence, risk factors and pregnancy outcomes. of intrauterine hyperglycemia. Diabetes Care 2008; 31(2):
Diabetes Res Clin Pract 2005; 69(3): 279–286. 340–346.
10. Yang X, Hsu-Hage B, Zhang H et al. Women with impaired glu- 30. Pettitt DJ, Jovanovic L. The vicious cycle of diabetes and preg-
cose tolerance during pregnancy have significantly poor pregnancy. Curr Diab Rep 2007; 7(4): 295–297.
nancy outcomes. Diabetes Care 2002; 25(9): 1619–1624. 31. Osgood ND, Dyck RF, Grassmann WK. The inter- and intra-
11. Siribaddana SH, Deshabandu R, Rajapakse D et al. The prev- generational impact of gestational diabetes on the epidemic
alence of gestational diabetes in a Sri Lankan antenatal clinic. of type 2 diabetes. Am J Public Health January 2011; 101(1):
Ceylon Med J 1998; 43(2): 88–91. 173–179.
12. Conde-Agudelo A, Belizan JM. Risk factors for pre-eclampsia in 32. Ferrara A, Kahn HS, Quesenberry CP et al. An increase in the
a large cohort of Latin American and Caribbean women. BJOG incidence of gestational diabetes mellitus: Northern California,
2000; 107(1): 75–83. 1991–2000. Obstet Gynecol 2004; 03(3): 526–533.
13. American Diabetes Association. Diagnosis and classification of 33. Aljohani N, Rempel BM, Ludwig S et al. Gestational diabetes in
diabetes mellitus. Diabetes Care 2010; 33(Suppl. 1): S62–S69. Manitoba during a twenty-year period. Clin Invest Med 2008;
14. National Diabetes Data Group. Classification and diagnosis of 31(3): E131–E137.
diabetes mellitus and other categories of glucose intolerance. 34. Anna V, van der Ploeg HP, Cheung NW et al. Sociodemographic
Diabetes 1979; 28(12): 1039–1057. correlates of the increasing trend in prevalence of gestational dia-
15. World Health Organization, International Diabetes Federation. betes mellitus in a large population of women between 1995 and
Definition and diagnosis of diabetes mellitus and intermediate 2005. Diabetes Care 2008; 31(12): 2288–2293.
hyperglycemia: Report of a WHO/IDF consultation, WHO Press: 35. Drummond MF, Sculpher MJ, Torrance GW et al. Methods for
Geneva, Switzerland, 2006. Available from http://www.who.int/ the Economic Evaluation of Health Care Programs, 3rd edn.,
topics/global_burden_of_disease/en/, accessed on January 15, 2011. Oxford University Press: Oxford, U.K., 2005.
16. Carpenter MW, Coustan DR. Criteria for screening tests for gesta- 36. Kahn JG, Marseille E, Bennett R et al. Cost-effectiveness of anti-
tional diabetes. Am J Obstet Gynecol 1982; 44(7): 768–773. retroviral therapy for prevention. Curr HIV Res September 2011;
17. Agarwal MM, Dhatt GS, Shah SM. Gestational diabetes mel- 9(6): 405–415.
litus: Simplifying the international association of diabetes and 37. Marseille E, Lohse N, Jiwani A et al. The cost-effectiveness of
pregnancy diagnostic algorithm using fasting plasma glucose. gestational diabetes screening including prevention of type 2 dia-
Diabetes Care 2010; 33(9): 2018–2020. betes: Application of a new model in India and Israel. J Matern
18. Morikawa M, Yamada T, Akaishi R et al. Change in the number of Fetal Neonatal Med May 2013; 26(8): 802–810.
patients after the adoption of IADPSG criteria for hyperglycemia 38. Poncet B, Touzet S, Rocher L et al. Cost-effectiveness of gesta-
during pregnancy in Japanese women. Diabetes Res Clin Pract tional diabetes mellitus screening in France. Eur J Obstet Gynecol
December 2010; 90(3): 339–342. Reprod Biol July 10, 2002; 103(2): 122–129.
39. Larijani B, Hossein-nezhad A, Rizvi SW et al. Cost analysis of 45. Meltzer SJ, Snyder J, Penrod JR et al. Gestational diabetes mellitus
different screening strategies for gestational diabetes mellitus. screening and diagnosis: A prospective randomized controlled
Endocr Pract November–December 2003; 9(6): 504–509. trial comparing costs of one-step and two-step methods. BJOG
40. Nicholson WK, Fleisher LA, Fox HE, Powe NR. Screening for gesta- March 2010; 117(4): 407–415.
tional diabetes mellitus: A decision and cost-effectiveness analysis 46. Round JA, Jacklin P, Fraser RB et al. Screening for gestational dia-
of four strategies. Diabetes Care June 2005; 28(6): 1482–1484. betes mellitus: Cost-utility of different screening strategies based
41. Ayach W, Costa RA, Calderon Ide M, Rudge MV. Comparison on a woman’s individual risk of disease. Diabetologia February
between 100-g glucose test and two other screening tests for gesta- 2011; 54(2): 256–263.
tional diabetes: Combined fasting glucose with risk factors and 50-g 47. Mission JF, Ohno MS, Cheng YW, Caughey AB. Gestational dia-
glucose tolerance test. Sao Paulo Med J January 5, 2006; 124(1): 4–9. betes screening with the new IADPSG guidelines: A cost-effec-
42. Thung S, Pettker C, Funai E. Screening for gestational diabetes. Is tiveness analysis. Am J Obstet Gynecol October 2012; 207(4):
a 130 mg/dl or 140 mg/dl glucose challenge test threshold more 326.e1–326.e9.
cost-effective? Am J Obstet Gynecol 2007; 197(6 Suppl.): S109. 48. Werner EF, Pettker CM, Zuckerwise L et al. Screening for ges-
43. National Collaborating Centre for Women’s and Children’s Health tational diabetes mellitus: Are the criteria proposed by the
(United Kingdom). Cost-effectiveness of screening, diagnosis and international association of the Diabetes and Pregnancy Study
treatment for gestational diabetes. Appendix D. In: Diabetes in Groups cost-effective? Diabetes Care March 2012; 35(3):
Pregnancy: Management of Diabetes and Its Complications from 529–535.
Preconception to the Postnatal Period, RCOG Press: London, 49. Weile LK, Kahn JG, Marseille E, Jensen DM, Damm P, Lohse N.
U.K., 2008, pp. 165–190. Global cost-effectiveness of GDM screening and management:
44. Lee S, Pettker C, Funai E et al. Is lowering the diagnostic thresh- Current knowledge and future needs. Best Pract Res Clin Obstet
old for gestational diabetes (GDM) cost-effective? Implications Gynaecol 2015; 29(2): 206–224.
from the hyperglycemia and adverse pregnancy outcomes
(HAPO) trial. Am J Obstet Gynecol 2008; 199(6 Suppl. A): S199.
16 Changing health policy: From
study to national policy
Ofra Kalter-Leibovici, Nicky Lieberman, Ronni Gamzu, and Moshe Hod
cohort included 25,505 pregnant women recruited at 15

Background centers in 9 countries. Two HAPO centers were located in
Gestational diabetes mellitus (GDM) is associated with Israel, including 3345 participants. All participants had a
greater risk for immediate and late maternal and fetal adverse 2-hour 75 g OGTT between 24 and 32 weeks of gestation.
outcomes. Immediate adverse outcomes include fetal mac- Caregivers remained blinded with respect to the fasting and
rosomia, shoulder dystocia and birth injury, preeclampsia, postload plasma glucose levels if these levels were 105 mg/dL
and cesarean section.1 Late adverse outcomes include greater or less and 200 mg/dL or less, respectively. Participants were
maternal risk for type 2 diabetes mellitus (T2DM)1,2 and followed till birth and information on maternal and fetal
higher prevalence of childhood obesity and insulin resis- outcomes was collected. The study results showed a linear
tance in offspring.3,4 relationship between fasting and postload plasma glucose
The risk of adverse outcomes in pregnancies complicated levels and adverse pregnancy outcomes, even in the range
with GDM correlates with the level of maternal hyperglyce- previously considered normal.13
mia,5 and interventions aimed to decrease hyperglycemia The results of the HAPO study motivated the Inter
were reported to lower the risk of some of these outcomes (i.e., national Association of Diabetes and Pregnancy Study Group
fetal macrosomia, shoulder dystocia, and preeclampsia).6–8 (IADPSG) to recommend universal screening for GDM
Clinical guidelines advocate screening for GDM between with 2-hour 75 g glucose OGTT performed between 24 and
24 and 28 weeks of gestation, and most countries have des- 28 weeks of gestation. The IADPSG diagnostic criteria for
ignated screening programs that involve the testing of fast- GDM were determined according to the fasting and postload
ing and post-oral glucose load plasma glucose levels. In some plasma glucose levels that were associated with odds ratios of
countries, including Israel, a two-step screening program 1.75 for having >90th percentiles of birth weight (large-for-
is employed, where pregnant women have 50 g oral glucose gestational-age [LGA]), cord C-peptide, and neonatal percent
challenge test (OGCT), and those found with 1-hour postchal- body fat, as compared to the corresponding mean values of
lenge plasma glucose level of 140 mg/dL or higher have 3-hour the entire HAPO cohort. The recommended plasma glucose
100 g oral glucose tolerance test (OGTT).9 The performance thresholds are 92, 180, and 153 mg/dL for fasting, 1-hour, and
of OGCT has been recently evaluated in a systematic review 2-hour post oral glucose load state, respectively. A woman is
of cohort studies. When GDM was diagnosed according to diagnosed with GDM if she has at least one plasma glucose
the Carpenter and Coustan criteria, a postload plasma glu- level equal to or higher than the threshold.14
cose threshold of 140 mg/dL or higher provided 85% sensitiv-
ity and 86% specificity for GDM diagnosis, with positive and
negative likelihood ratios of 5.9 and 0.18, respectively.10 Over Making local health policy
the years, however, single- and two-step screening approaches
have been proposed, using various oral glucose load doses and The IADPSG recommendation evoked international discus-
blood glucose thresholds for the diagnosis of GDM, with no sion as to whether GDM screening and diagnosis practices
consensus over a preferred method.11 Furthermore, the fasting should be changed or not. The main concern raised was
and postload plasma glucose thresholds used for diagnosis of the need of healthcare systems to cope with a significant
GDM have been set according to the maternal risk of develop- increase in the number of pregnancies previously consid-
ing T2DM during follow-up, rather than the risk of immediate ered normal that will now be classified and treated as GDM
adverse pregnancy outcomes.12 pregnancies, while there is no evidence from randomized
The Hyperglycemia and Adverse Pregnancy Outcomes trials showing that implementation of the new diagnostic
(HAPO) study addressed the ongoing debate on the best criteria results in better pregnancy outcomes compared to
screening and diagnostic practices for GDM. The study those currently employed.15
131
132 Changing health policy
We describe in this chapter the rationale and process of To estimate the expected yield of implementing the
changing health policy of GDM screening and diagnosis in IADPSG recommendation, we calculated the number needed
Israel, including the following steps: to screen (NNS) and the number needed to treat (NNT) to
avoid one case of LGA or preeclampsia.
1. Provision of locally relevant information for health pol- We developed a risk stratification tool for adverse preg-
icy decision making, based on the analysis of the Israeli nancy outcomes among IADPSG-positive cases.
HAPO study cohort data We also explored two alternative screening strategies,
2. Cost-effective assessment of implementation of the either with fasting plasma glucose (FPG) or body mass index
IADPSG recommendation, based on the data from Clalit (BMI) and calculated FPG and BMI thresholds that provide
Health Services (CHS), the largest health maintenance a similar number of cases as the number of cases identified
organization in Israel by the IADPSG criteria.
3. Changing current health policy of GDM screening, diag- Finally, we explored a two-step screening approach, using
nosis, and care in CHS FPG testing for the first step, and an OGTT with the IADPSG
thresholds for the second step.
We found that the Israeli cohort significantly differed
Step I: Data analysis from the rest of the HAPO study cohort. Israeli par-
The analytical approach has been previously described.16 In ticipants were younger, less heavy, and more frequently
brief, we compared the baseline characteristics and rates multiparous and had lower FPG and postload plasma glu-
of pregnancy adverse outcomes in the Israeli HAPO study cose levels. Adverse pregnancy outcomes were also less
cohort (n = 3,345) and the rest of the HAPO study partici- common in the Israeli HAPO cohort compared to the
pants (n = 19,971). Based on the Israeli HAPO cohort data, rest of the HAPO study cohort (Table 16.1). These find-
we calculated the expected number of women diagnosed ings emphasize the need to base health policy decisions,
with GDM according to the IADPSG diagnostic criteria where possible, on local data. Nevertheless, the associa-
and compared this number with the number of women tion between FPG and postload plasma glucose levels and
diagnosed with GDM in Israel, under the current two-step the rate of adverse pregnancy outcomes were similar in
screening and diagnostic policy. the two HAPO groups.
Table 16.1 Comparisons of baseline characteristics and pregnancy outcomes between Israeli and other
Hyperglycemia and Adverse Pregnancy Outcomes study participants
HAPO study participants

Israeli Others
n = 3,345 n = 19,971 P
Age, years, mean (SD) 27.8 (5.4) 29.4 (6.0) <0.001
BMI, kg/m2, mean (SD) 26.9 (4.4) 27.8 (5.4) <0.001
Current cigarette smoking, % 5.9 6.9 0.024
Any alcohol consumption, % 0.5 8.0 <0.001
Multiparous, % 60.6 51.1 <0.001
Fasting plasma glucose, mg/dL, mean (SD) 78.4 (6.7) 81.3 (7.1) <0.001
1-hour post-oral glucose load, mg/dL, mean (SD) 124.5 (30.4) 135.7 (30.7) <0.001
2-hour post-oral glucose load, mean (SD) 101.8 (21.7) 112.5 (23.6) <0.001
Cumulative percent positive by threshold <0.001
Fasting ≥92 mg/dL 2.8 8.3
1-hour post-oral glucose load ≥180 mg/dL 6.9 14.0
2-hour post-oral glucose load ≥153 mg/dL 8.3 16.1
Unblindeda 9.0 17.8
Fetal macrosomia, % 8.8 9.7 0.11
Preeclampsia/eclampsia, % 1.3 5.4 <0.001
Primary cesarean section, % 10.0 17.0 <0.001
Birth injury/shoulder dystocia, % 0.3 1.5 <0.001
Preterm delivery, % 5.3 7.2 <0.001
Admission to neonatal intensive care unit, % 7.0 8.1 0.034
a Women with fasting plasma glucose > 105 mg/dL, plasma glucose at 2-hour post-oral glucose load >200 mg/dL, or a random plasma
glucose level > 160 mg/dL. Information on the plasma glucose levels of these women was disclosed to their caregivers.
Making local health policy 133
Currently, about 7500 women, 4.3% of pregnancies in reducing or preventing maternal obesity are associated with
Israel are diagnosed with GDM, annually.17 Following lower risk of fetal macrosomia and preeclampsia.25 Thus, the
the implementation of the IADPSG recommendation, the assessment of maternal BMI and provision of effective life-
number of women diagnosed with GDM will double, adding style advice should be pursued regardless of the evaluation of
extra 8000 women each year. Similar results were reported maternal glucose metabolism.
in other populations. For example, the implementation of Screening with FPG may be an attractive alternative to
the IADPSG recommendation is expected to cause a two- OGTT, especially in developing countries. Agarwal et al.26
to threefold increase in the number of women diagnosed reported that screening with FPG ≥ 88 mg/dL has reason-
with GDM in the United States.15 Such increases require ably good performance in terms of sensitivity (85%), speci-
reorganization of the prenatal services and assessment of ficity (71%), and area under the curve (AUC) (0.88) when
extra resources needed. To meet these needs, we developed tested against the American Diabetes Association diag-
an LGA risk score, based on maternal height, parity, and nostic criteria of GDM. FPG of ≥90 mg/dL was reported
BMI at the time of screening. Using this score, we identified to have 80% sensitivity and 91% specificity, with likelihood
a subgroup consisting of about one-third of women diag- ratios of positive and negative test results of 8.84 and 0.21,
nosed with GDM according to the IADPSG criteria, who had respectively, when compared to the IADPSG diagnostic
a lower rate of LGA than the rest of the IADPSG-positive criteria for GDM.27 Screening with FPG using a threshold
cohort (9.8% vs. 19.7%, respectively). The rate of LGA off- of 85 mg/dL was previously reported to be associated with
spring in this subgroup was closer to the rate of LGA in the better sensitivity and specificity as well as positive and neg-
IADPSG-negative group (8.1%). This subgroup, considered ative predictive values compared to OGCT.28
to have mild GDM, would probably benefit from less inten- It should be noted, however, that FPG and IADPSG diag-
sive perinatal care setup, focusing mainly on healthy lifestyle nostic criteria do not necessarily detect the same population
advice. Thus, the implementation of the IADPSG recommen- of positive cases.
dation with further risk stratification may promote efficient Screening with FPG exclusively can miss cases that may
utilization of available healthcare resources and allow the still have high risk of adverse pregnancy outcomes despite
healthcare system to cope with the excess number of women having lower levels of FPG. We therefore developed a risk
diagnosed with GDM under the new diagnostic criteria. score for LGA offspring for women who have FPG < 89 mg/dL.
Among IADPSG-positive women in the Israeli cohort, the Using this risk score, we could identify a subgroup of the
rates of LGA offspring and preeclampsia were 16.4% and 1.8%, women with FPG < 89 mg/dL who have 17.5% risk for LGA
respectively. According to Landon et al.18 intensive treatment offspring, similar to the rate in women with FPG ≥ 89 mg/dL
of mild GDM is associated with a mean reduction of 51% in (18.8%). To avoid missing these cases, a two-step screening
the risk of LGA offspring and a mean reduction of 54% in the for GDM may be offered; first, all women will have a FPG
risk of preeclampsia. Thus, the NNS and NNT to avoid one test. Women with FPG ≥ 89 mg/dL will be diagnosed with
case of LGA are 145 (95% CI: 90, 232) and 12 (95% CI: 7.5, GDM. Women with FPG < 89 mg/dL, who have high LGA
19), respectively. The corresponding numbers for maternal risk score (about 20% of the women with FPG < 89 mg/dL),
preeclampsia are 1242 (95% CI: 431, 3584) and 103 (36, 292). will have OGTT and will be diagnosed with GDM according
These numbers are well within the range for other interven- to the IADPSG criteria. This two-step screening approach
tions in obstetrics, e.g., magnesium treatment for prevention reduces the risk of missing LGA cases while still avoiding
of eclampsia19 and for neuroprotection,20,21 or progesterone OGTT in more than 80% of pregnant women.
treatment for the prevention of premature birth.22 Göbl et al.29 recently reported a similar two-step screen-
To explore alternative screening methods for detecting ing approach, using the FPG threshold recommended by the
women at high risk of adverse pregnancy outcomes, we cal- IADPSG (≥92 mg/dL) for first-step screening and OGTT for
culated the BMI and FPG thresholds at 24–32 weeks of ges- second-step screening among women with FPG <92 mg/dL
tation providing rates of positive cases similar to the rate of and high LGA risk score.
positive cases defined by the IADPSG criteria. These thresh-
olds (at 24–32 weeks of gestation) were 33.5 kg/m2 for BMI
and 89 mg/dL for FPG. Using these thresholds, screening Step II: Cost-effectiveness assessment
with FPG or BMI produced similar rates of LGA cases, as Economic analysis of data from the Australian Carbohydrate
did screening according to the IADPSG recommendations Intolerance Study in Pregnant Women trial (ACHOIS) ran-
(18.8% for FPG ≥ 89 mg/dL, 17.3% for BMI ≥ 33.5 kg/m2, domized trial showed that treatment of mild GDM is associ-
and 16.4% for IADPSG-positive women). Similarly, screen- ated with incremental cost of $27,503 per serious perinatal
ing with FPG or BMI will detect a similar or even greater complication prevented and $2,988 per discounted life year
proportion of pregnancies complicated with preeclampsia gained.30 Altogether, the intervention was judged as highly
(2.2% and 4.3%, respectively), as compared to the screening cost-effective. However, the costs of screening and diagnosis
with the IADPSG criteria (1.8%). were not included in the analysis.7
Maternal overweight is associated with increased risk for Data from randomized controlled trials on the effect of
adverse pregnancy outcomes (e.g., fetal macrosomia, pre- GDM treatment on the risk of long-term adverse maternal
eclampsia, cesarean section, and shoulder dystocia), irre- and offspring adverse outcomes are scarce. Although insulin
spective of blood glucose levels.23,24 Interventions aimed at treatment in women with GDM was associated with reduction
134 Changing health policy
in perinatal mortality, it did not affect the risk of maternal overrepresentation of ethnically diverse and underprivileged
T2DM after 16 years of follow-up.31 Similarly, active treatment population subgroups. Although people can move from one
of GDM was associated with lower rate of fetal macrosomia health plan to another every 2 months, less than 1% chose
compared to routine care in the ACHOIS trial; treatment had to do so each year. The low attrition rate provides an incen-
no effect on offspring BMI at the age of 4–5 years.32 tive for the health plans to invest in preventive medicine and
The costs of one-step screening methods for GDM were in disease management programs. CHS was one of the first
studied in a Canadian randomized trial. The results showed health plans worldwide to develop and implement a desig-
that a two-step screening process (OGCT followed by OGTT nated diabetes prevention program for people with pre-
in positive cases) was associated with lower costs compared diabetes. The Israeli healthcare system is almost completely
to one-step screening.33 computerized, using electronic medical records with alerts
Werner et al.34 analyzed the cost-effectiveness of the pro- and reminder generator and embedded decision support
posed IADPSG criteria for GDM screening and diagnosis, systems, allowing data sharing between different providers.
using a decision tree model. The investigators found that CHS runs a comprehensive diabetes community pro-
a two-step screening for GDM (using FPG and OGTT for gram since 1997. Most people with diabetes are managed by
the first and second step, respectively), implementing the their primary practitioners and other community health-
IADPSG recommended glucose thresholds, was cost-effec- care providers, including nurses, pharmacists, dietitians, and
tive providing that post-delivery lifestyle intervention will physiotherapists. Patients are referred to consultant diabetes
reduce the incidence of T2DM in mothers. specialist in case of uncontrolled diabetes or severe diabetic
Lohse et al.35 developed a computer simulation model complications. CHS developed its own diabetes practice
to assess the cost-effectiveness of GDM screening and life- guidelines, endorsing international standards, and being
style change to prevent T2DM. The model was used to regularly reviewed and revised every 2–3 years. The diabetes
assess the cost-effectiveness of one-step GDM screening disease management program within the primary care setup
with 2-hour 75 g OGTT compared to no screening, using in CHS addresses also other comorbidities and risk factors
disability-adjusted life years (DALYs) for effectiveness mea- that are associated with diabetes, e.g., hypertension, dyslip-
sure.36 The investigators used CHS database for information idemia, and cardiovascular disease. The comprehensive care
on the prevalence of GDM and costs. The model assumed provided is culturally sensitive, addressing the ethnic diver-
that 80% of the women will receive postnatal care, includ- sity of the insured population. As a result, more than 55% of
ing lifestyle advice. The effect of the lifestyle intervention people with diabetes have HbA1C level of less than 7%, and
was assumed to decrease the relative risk for T2DM both in more than 70% have their HbA1c level within their personally
mothers and children by 40%, resulting in 4 cases of T2DM adapted target; more than 60% of the diabetic patients have
averted among mothers and 2 cases among children per 1000 LDL cholesterol levels lower than 100 mg/dL, and more than
women screened for GDM. The net incremental costs (in 75% of the patients with diabetes have LDL cholesterol level of
international dollars) compared to no screening was $76,102 less than 130 mg/dL. As a result of the comprehensive diabetes
per 1,000 women screened. The cost per DALY averted was care provided to people with diabetes, there has been a signifi-
$1830, and 95% of the DALYs averted were due to T2DM pre- cant decline in the rate of nontraumatic amputations, invasive
vention. In sensitivity analyses, lower uptake of postpartum procedures for coronary reperfusion and blindness due to dia-
care (50% instead of 80%) would raise the cost per DALY betic retinopathy. As a result, the costs of the person with dia-
averted to $2895. Assuming lower effectiveness of lifestyle betes decreased by about 40%, from the provider perspective.
advice (20% instead of 40% relative risk reduction of T2DM) Given the general policy that promote disease prevention,
would increase the cost per DALY averted to $7242. With no the high risk of T2DM in women with GDM, and the favor-
benefit in reducing T2DM, the cost per DALY averted rises able result of the cost–benefit analysis, CHS decided to adopt
to $182,750. the IADPSG recommendation and to adapt the care model
It should be noted that the estimated proportion of according to the increased number of patients diagnosed
pregnancies complicated with GDM in this analysis was with GDM.
2.6%, which is significantly lower than the rate of GDM CHS will adopt the risk-stratification approach described
according to the IADPSG criteria reported for the two par- earlier, stratifying the women diagnosed with GDM accord-
ticipating Israeli centers in the HAPO study.37 ing to the new diagnostic criteria into high-, intermediate-,
and low-risk subgroups. Women considered being in the
low-risk group will be provided with lifestyle advice, includ-
Step III: Changing current health policy ing dietary counseling and physical exercise, and follow-up
In order to understand CHS decision to change the cur- by their primary practitioners and consultant gynecologists.
rent policy, the reader should get acquainted with the Israeli Counseling will be provided by nurse educators, dietitians,
health system and the basic concepts of CHS’s policy. and occasionally diabetes specialists. Healthcare for women
According to the National Health Insurance Law enacted in the intermediate-risk subgroup will be provided in com-
in 1995, all Israeli residents are insured by one of the four munity, Women Health Centers of CHS, by multidisciplinary
health plans, currently providing healthcare services to over teams, which consists of lifestyle advice and hypoglycemic
eight million people. CHS, the largest health insurer in Israel, drug therapy. High-risk patients will be referred to high-risk
insures more than half of the total Israeli population, with pregnancy units in the regional hospitals.
References 135
REFERENCES
1. Brody SC, Harris R, Lohr K. Screening for gestational diabetes: A 20. Costantine MM, Weiner SJ, for the Eunice Kennedy Shriver
summary of the evidence for the U.S. Preventive Services Task National Institute of Child Health and Human Development
Force. Obstet Gynecol 2003; 101: 380–392. (NICHD) Maternal-Fetal Medicine Units Network (MFMU).
2. Bellamy L, Casas JP, Hingorani AD, Williams D. Type 2 diabetes Effects of antenatal exposure to magnesium sulfate on neuropro-
mellitus after gestational diabetes: A systematic review and meta- tection and mortality in preterm infants: A meta-analysis. Obstet
analysis. Lancet 2009; 373: 1773–1779. Gynecol 2009; 114(2 Pt. 1): 354–364.
3. Boerschmann H, Pflüger M, Henneberger L, Ziegler AG, 21. Doyle LW, Crowther CA, Middleton P, Marret S. Antenatal mag-
Hummel S. Prevalence and predictors of overweight and insulin nesium sulfate and neurologic outcome in preterm infants: A sys-
resistance in offspring of mothers with gestational diabetes mel- tematic review. Obstet Gynecol 2009; 113: 1327–1333.
litus. Diabetes Care 2010; 33: 1845–1849. 22. Dodd JM, Flenady VJ, Cincotta R, Crowther CA. Progesterone
4. Vääräsmäki M, Pouta A, Elliot P et al. Adolescent manifestation for the prevention of preterm birth. A systematic review. Obstet
of metabolic syndrome among children born to women with Gynecol 2008; 112: 127–134.
gestational diabetes in a general-population birth cohort. Am J 23. O’Brien TE, Ray JG, Chan WS. Maternal body mass index and the
Epidemiol 2009; 169: 1209–1215. risk of preeclampsia: A systematic overview. Epidemiology 2003;
5. Langer O, Mazze R. The relationship between large-for-gesta- 14: 368–374.
tional-age infants and glycemic control in women with gesta- 24. Usha Kiran TS, Hemmadi S, Bethel J, Evans J. Outcome of preg-
tional diabetes. Am J Obstet Gynecol 1988; 159: 1478–1483. nancy in a woman with an increased body mass index. BJOG
6. Horvath K, Koch K, Jeitler K et al. Effects of treatment in women 2005; 112: 768–772.
with gestational diabetes mellitus: Systematic review and meta- 25. Thangaratinam S, Rogozińska E, Jolly K et al. Interventions to
analysis. Br Med J 2010; 340: c1395. reduce or prevent obesity in pregnant women: A systematic
7. Waugh N, Royle P, Clar C, Henderson R, Cummins E, Hadden D, review. Health Technol Assess 2012; 16(31): 1–191.
Lindsay R, Pearson D. Screening for hyperglycemia in pregnancy: 26. Agarwal MM, Dhatt GS, Punnose J. Gestational diabetes: Utility
A rapid update for the National Screening Committee. Health of fasting plasma glucose as a screening test depends on the diag-
Technol Assess 2010; 14: 1–183. nostic criteria. Diabet Med 2006; 23: 1319–1326.
8. Hartling L, Dryden DM, Guthrie A, Muise M, Vandermeer B, 27. Agarwal MM, Dhatt GS, Shah SM. Gestational diabetes mel-
Donovan L. Benefits and harms of treating gestational diabetes litus: Simplifying the International Association of Diabetes and
mellitus. Ann Intern Med 2013; 159: 123–129. Pregnancy diagnostic algorithm using fasting plasma glucose.
9. American College of Obstetricians and Gynecologists. Diabetes Care 2010; 33: 2018–2020.
Gestational diabetes. ACOG Practice Bulletin No. 137. Obstet 28. Poomalar GK, Rangaswamy V. A comparison of fasting plasma
Gynecol 2013; 122: 406–412. glucose and glucose challenge test for screening of gestational
10. Donovan L, Hartling L, Muise M, Guthrie A, Vandermeer B, diabetes mellitus. J Obstet Gynaecol 2013; 33: 447–450.
Dryden DM. Screening tests for gestational diabetes: A system- 29. Göbl CS, Bozkurt L, Rivic P et al. A two-step screening algorithm
atic review for the U.S. Preventive Services Task Force. Ann including fasting plasma glucose measurement and a risk estima-
Intern Med 2013; 159: 115–122. tion model is an accurate strategy for detecting gestational diabe-
11. Sacks DB. Diagnosis of gestational diabetes mellitus: It is time for tes mellitus. Diabetologia 2012; 55: 3173–3181.
international consensus. Clin Chem 2014; 60: 141–143. 30. Moss JR, Crowther CA, Hiller JE, Willson KJ, Robinson JS, for the
12. Carpenter MW, Coustan DR. Criteria for screening tests for gesta- Australian Carbohydrate Intolerance Study in Pregnant Women
tional diabetes. Am J Obstet Gynecol 1982; 144: 768–773. (ARCHOIS) Trial Group. Costs and consequences of treatment for
13. The HAPO Study Cooperative Research Group. Hyperglycemia mild gestational diabetes mellitus—Evaluation from the ACHOIS
and adverse pregnancy outcomes. N Engl J Med 2008; 358: randomized trial. BMC Pregnancy Childbirth 2007; 7: 27.
1991–2002. 31. O’Sullivan JB, Mahan CM. Insulin treatment and high risk groups.
14. International Association of Diabetes and Pregnancy Study Diabetes Care 1980; 3: 482–485.
Groups Consensus Panel. International Association of Diabetes 32. Gillman MW, Oakey H, Baghurst PA, Volkmer RE, Robinson JS,
and Pregnancy Study Groups recommendations on the diagnosis Crowther CA. Effect of treatment of gestational diabetes melli-
and classification of hyperglycemia in pregnancy. Diabetes Care tus on obesity in the next generation. Diabetes Care 2010; 33:
2010; 33: 676–682. 964–968.
15. National Institutes of Health Consensus Development Conference 33. Meltzer SJ, Snyder J, Penrod JR, Nudi M, Morin L. Gestational
Statement. Diagnosing gestational diabetes mellitus, March 4–6, diabetes mellitus screening and diagnosis: A prospective ran-
2013. Obstet Gynecol 2013; 122: 358–369. domized controlled trial comparing costs of one-step and two-
16. Kalter-Leibovici O, Freedman LS, Olmer L, Liebermann N, step methods. BJOG 2010; 117: 407–415.
Heymann A, Tal O, Lerner-Geva L, Melamed N, Hod M. 34. Werner EF, Pettker CM, Zuckerwise L, Reel M, Funai E,
Screening and diagnosis of gestational diabetes mellitus: Critical Henderson J, Thung SF. Screening for gestational diabetes mel-
appraisal of the new International Association of Diabetes in litus: Are the criteria proposed by the International Association
Pregnancy Study Group recommendations on a national level. of the Diabetes and Pregnancy Study Groups cost-effective?
Diabetes Care 2012; 35: 1894–1896. Diabetes Care 2012; 35: 529–535.
17. Sella T, Shalev V, Elchalal U, Choval-Sella A, Chodick G. 35. Lohse N, Marseille E, Kahn JG. Development of a model to assess
Screening for gestational diabetes in the 21st century: A popu- the cost-effectiveness of gestational diabetes mellitus screening
lation-based cohort study in Israel. J Matern Fetal Neonatal Med and lifestyle change for the prevention of type 2 diabetes mel-
2013; 26: 412–416. litus. Int J Gynecol Obstet 2011; 115(Suppl. 1): S20–S25.
18. Landon MB, Spong CY, Thom E et al. A multicenter, random- 36. Marseille E, Lohse N, Jiwani A et al. The cost-effectiveness of
ized trial of treatment for mild gestational diabetes. N Engl J Med gestational diabetes screening including prevention of type 2 dia-
2009; 361: 1339–1348. betes: Application of a new model in India and Israel. J Matern
19. Duley L, Gülmezoglu AM, Henderson-Smart DJ, Chou D. Fetal Neonatal Med 2013; 26: 802–810.
Magnesium sulphate and other anticonvulsants for women with 37. Sacks DA, Hadden DR, Maresh M et al. Frequency of gestational
pre-eclampsia. Cochrane Database Syst Rev 2010; (Issue 11) Art. diabetes mellitus at collaborating centers based on IADPSG con-
No.: CD000025. sensus panel-recommended criteria. Diabetes Care 2012; 35: 526.
17 Ideal weight gain in diabetic
pregnancy
Gerard H.A. Visser and Harold W. de Valk
The guidelines are shown in Table 17.1, according to prepreg-

Introduction nancy BMI. In women with type 1 diabetes, it has recently
Weight gain during pregnancy is related to pregnancy out- been shown that weight gain above the recommended value,
comes that include preterm delivery, fetal macrosomia, and as defined before, may result in a striking increase in fetal
intrauterine growth restriction. The former two are common macrosomia: 42% of LGA as compared to 8% in women
in women with diabetes; the latter one has become rare with with the recommend weight gain.5 In women with gesta-
better care and lower incidence of vascular complications tional diabetes mellitus (GDM), most of the weight gain
in these women.1,2 Optimal weight gain differs according to occurs prior to GDM diagnosis (for references, see Most and
the maternal body mass index (BMI) and is lower in women Langer6). In a study in GDM, the lowest incidence of LGA
with a high BMI. Weight loss during pregnancy is usually occurred in normal weight women when weight gain was in
associated with an impaired outcome. In this chapter, opti- accordance with the Institute of Medicine guidelines. In case
mal weight gain during pregnancy in women with diabetes of overweight (BMI 25–30) or obesity (BMI > 30), the low-
is discussed. est LGA incidence occurred in women with a lower weight
gain than recommended in the guidelines (<8 and <4.5 kg,
respectively).6
The guidelines on optimal weight gain do not contain
Optimal weight gain subdivision according to the three obesity classes (I, BMI
Most data on optimal weight gain concern the general preg- 30–35; II, 35–40; III, >40), since data were not sufficient to
nant population, with specific data on maternal diabetes do so. However, in severely obese nondiabetic women with
generally lacking. Weight gain should especially be restricted a BMI > 40, there is evidence that weight gain during preg-
in obese women, since excessive weight gain hampers an nancy should be below 5 kg for an optimal outcome with
already impaired outcome. This has been nicely shown in a low preterm delivery rate.7 In a large study in 6500 obese
a study in 481 obese glucose-tolerant women: a gestational women in California, it was found that the lowest risks for
weight gain of 5–10 kg was associated with a doubling of the preterm birth, small for gestational age (SGA), and LGA
incidence of maternal hypertensive disease, cesarean deliv- occurred with a maternal weight gain of 9–13.5 kg in the
ery, induction of labor, and large-for-gestational-age (LGA) Class I group, 5–9 kg in the Class II, and 2–5 kg in Class 3. In
infant as compared to a weight gain of less than 5 kg.3 In case Class III, African-American women’s best outcome occurred
of a weight gain of 10–15 kg and more than 15 kg, these risks in the case of weight gain below 2 kg.7 In the case of weight
were on average threefold and fourfold higher, respectively. loss, there was an increase in preterm delivery in all three
In 2009, new guidelines on optimal weight gain were pub- obesity classes. A large German study also has shown that
lished by the Institute of Medicine.4 These guidelines were gestational weight loss was associated with an increase in
developed preterm delivery and SGA, apart from obesity class III.8 The
potential hazards of gestational weight loss should not be
to minimize the negative health consequences for both underestimated and weight loss cannot be recommended.9
mother and fetus of both inadequate or excessive gain.
This was assessed primarily by considering the outcomes
most consistently and plausible associated with gesta- Discussion
tional weight gain: infants born either small- or large-for-
gestational (SGA and LGA, respectively) age, the mother’s Fetal macrosomia in women with diabetes depends on
risk for an unplanned cesarean delivery and excessive many factors, such as genetics and epigenetics, pre-
(>5 kg) postpartum weight retention. pregnancy BMI, glucose control during pregnancy, and
136
References 137
interventions in overweight and women obese has shown

Table 17.1 Recommended weight gain in pregnancy
a 2 kg lower maternal weight gain in the treatment arm with
according to the Institute of Medicine Guidelines
a trend toward a reduction in the prevalence of gestational
diabetes.12 There were no differences in the incidence of
Recommended weight gain
LGA infants. All studies were assessed as being of low to
BMI kg lb medium quality.
<18.5 12–18 28–40 In women with preexisting diabetes or GDM, the inci-
dence of fetal macrosomia is high, and macrosomia is one
18.5–25 11.5–16 25–35
of the key factors important for childhood and later obesity
25–30 6.8–11.4 15–25
and metabolic syndrome. Adequate counseling and dietary
>30 5–9 11–20
advices are, therefore, particularly important in these
Source: Rasmussen, K.M. et al., Curr. Opin. Obstet. Gynecol., 21, women. This is facilitated by the fact that these women will
521, 2009. be frequently seen during pregnancy by a dedicated endo-
crinology/obstetrics team, with their weight be measured
during each visit. In the second trimester of pregnancy, it
gestational weight gain. Good glucose control in early appears possible to identify women at risk of gaining weight
pregnancy is associated with an increase in macrosomia, outside of the guidelines, which may facilitate dedicated
probably due to a better placentation, whereas a good con- action.13 However, in the case of GDM, most weight gain has
trol late in pregnancy decreases the risk (see Chapter 49). already occurred before diagnosis. One study has indicated
Disappointingly, many of these risk factors cannot be influ- that optimal weight gain in overweight and obese women
enced although weight gain in pregnancy may be modifiable with GDM should be lower than that recommended for
through dietary and behavioral interventions. However, the general population. Weight gain in pregnancy should
even with intensive guidance, the results seem to be only regularly be assessed, and in the case of weight gain below
moderate.10,11 A recent meta-analysis of randomized and or above recommended values, appropriate dietary advices
nonrandomized clinical trials on dietary and lifestyle should be given.
REFERENCES
1. Evers IM, de Valk HW, Visser GHA. Risk of complications of 8. Beyerlein A, Schiessl B, Lack N, von Kries R. Associations of ges-
pregnancy in women with type 1 diabetes: Nationwide prospec- tational weight loss with birth-related outcome: A retrospective
tive study in the Netherlands. Br Med J 2004; 328: 915. cohort study. Br J Obstet Gynaecol 2011; 118: 55–61.
2. Persson M, Pasupathy D, Hanson U, Norman M. Birth size 9. Dodd JM, Grivell RM, Crowther JS, Robinson JS. Antenatal inter-
distribution in 3,705 infants born to mothers with type 1 dia- ventions for overweight or obese pregnant women: A systematic
betes: A population-based study. Diabetes Care 2011; 34(5): review of randomized trials. Br J Obstet Gynaecol 2010; 117:
1145–1149. 1316–1326.
3. Jensen DM, Ovesen P, Beck-Nielsen H, Mølsted-Pedersen L, 10. Dodd JM, Robinson JS. Gestational weight loss in overweight and
Sørensen B, Vinter C, Damm P. Gestational weight gain and preg- obese women is associated with an increased risk of small for
nancy outcomes in 481 obese glucose-tolerant women. Diabetes gestational age infants. Evid Based Med 2011; 16: 125–126.
Care 2005; 28(9): 2118–2122. 11. Phelan S, Phipps MG, Abrams B, Darroch F, Schaffner A, Wing
4. Rasmussen KM, Catalano PM, Yaktine Al. New guidelines for RR. Randomized trial of a behavioral intervention to prevent
weight gain during pregnancy: What obstetrician/gynecologists excessive weight gain; the Fit for Delivery Study. Am J Clin Nutr
should know. Curr Opin Obstet Gynecol 2009; 21: 521–526. 2011; 93: 772–779.
5. Scifres CM, Feghali MN, Althouse AD et al. Effect of excess 12. Oteng-Ntim E, Varma R, Croker H, Poston L, Pat Doyle P. Lifestyle
gestational weight gain on pregnancy outcomes in women with intervention for overweight and obese women to improve preg-
type 1 diabetes. Obstet Gynecol 2014; 123: 1295–1302. nancy outcome: Systemic review and meta-analysis. BMC Med
6. Most O, Langer O. Gestational diabetes: Maternal weight gain 2012; 10: 47.
in relation to fetal growth, treatment modality, BMI and glyce- 13. Chmitorz A, von Kries R, Rasmussen KM, Nehring I, Ensenauer R.
mic control. J Matern Fetal Neonatal Med 2012; 25: 2458–2463. Do trimester-specific cutoffs predict whether women ultimately
7. Bodnar LM, Siega-Riz AM, Simhan HN et al. Severe obesity, ges- stay within the Institute of Medicine/National Research Council
tational weight gain and adverse birth outcomes. Am J Clin Nutr guidelines for gestational weight gain? Findings of a retrospective
2010; 91: 1642–1648. cohort study. Am J Clin Nutr 2012; 95: 1432–1437.
18 Medical nutritional therapy for
Lois Jovanovic
women.5 Lipid metabolism is also altered in hepatic and

Introduction adipose tissue. Early-pregnancy hormones, estrogen, pro-
Pregnancy causes a multitude of metabolic changes within gesterone, and insulin, promote the storage of lipids within
a woman’s body in order to provide the proper nutrients maternal tissues.
to the developing fetus. In women with type 1 and type 2 Therefore, in early pregnancy, initially there is a
diabetes and gestational diabetes mellitus (GDM), these decrease in serum triacylglycerols, fatty acids, cholesterol,
metabolic perturbations must be treated distinctly and lipoproteins, and phospholipids. However, as estrogen
aggressively to optimize fetal development and health. and insulin resistance impacts the mother, lipolysis and
Pregestational diabetes (either type 1 or type 2) has the hypertriglyceridemia ensues. 5 For example, at week 12,
potential to subject the developing fetus to abnormal estrogen causes cholesterol, specifically HDL, to be uti-
maternal glucose levels resulting in problems with organo- lized as a major metabolic fuel for the placenta through-
genesis producing congenital abnormalities or spontaneous out the remaining weeks of pregnancy; thus serum
abortion. Furthermore, GDM presents after organogenesis concentrations increase. Furthermore, VLDL is altered in
in the second part of pregnancy; therefore, the major risk the second and third trimesters due to a change in adipose
for the fetus is macrosomia. Although the goal for dietary and hepatic enzyme activity, specifically the decrease of
therapy for each of these disorders is the same, which is lipoprotein lipase (LPL). Notably, when maternal glucose
euglycemia, the means to achieve it are very different and levels decrease as can be seen during the fasting state,
somewhat controversial. In the case of GDM, the main- hepatic LPL activity increases allowing the mobilization
stay of therapy is medical nutritional therapy, whereas in of lipids and ketones for fetal nutrition. Human chorionic
i nsulin-requiring diabetes, dietary therapy is compen-
gonadotropin, prolactin, and glucagon also contribute
sated with premeal insulin injections. In this chapter, the by stimulating lipolysis in late pregnancy. This serves to
metabolic changes in normal pregnancy will be presented preferentially send glucose and protein to the fetus, while
followed by the general guidelines for pregnancy. Fetal the maternal tissues rely more on fatty acid oxidation and
complications associated with inadequate nutrition or ketogenesis to meet their energy requirement. Finally,
metabolic perturbation will be briefly explored, followed by during the third trimester, a change in hepatic gluconeo-
issues and treatment for GDM, with emphasis on specific genesis can be seen. Due to a 10–15 mg/dL decrease in
dietary therapies for GDM. the basal rate of glucose and an insulin concentration of
two times the concentration seen in nonpregnant women,
the liver must compensate by secreting 16%–30% more
glucose to meet the energy needs of both the mother and
Metabolic changes in normal fetus. 3,6,7 Assel et al.7 found that the maternal hepatic glu-
pregnancy cose production is dependent upon the maternal body
weight in a linear fashion. Late pregnancy also is charac-
During pregnancy, metabolism increases by 15%–26% to terized by a rise in the carbohydrate contribution to the
support both the mother and developing fetus.1 Early preg- overall oxidative metabolism as energy.
nancy is characterized by normal glucose tolerance, normal In normal pregnancy, metabolic changes occur to shunt
hepatic gluconeogenesis, and normal or improved insu- nutrients from the mother to the fetus, allowing for opti-
lin sensitivity.1–5 As pregnancy progresses, carbohydrate mal development and growth; however, as we explore the
metabolism becomes altered due to an increase in insulin following sections of this chapter, it will become apparent
secretion and decreased insulin sensitivity. Thus, some that alteration based on input and dysfunction can alter the
insulin resistance occurs, by late pregnancy, overall insulin body’s natural plan and cause distress to both the mother
action is decreased 50%–70% as compared to nonpregnant and fetus.
138
Fetal complications of maternal hyperglycemia 139
General nutritional guidelines for the prominent characteristic.12 In the United States, GDM
affects 2%–14% of the pregnant population per year depend-
pregnancy ing on the ethnicity of the population studied.13,14 A women
who is most likely to be affected by GDM is one who is obese,
Normal pregnancy nutritional guidelines focus on several
is >25 years of age, has a family history of diabetes especially
dietary elements. Major topics include caloric intake, mac-
in first-degree relatives, and has a past medical history for
ronutrient proportion, vitamins and minerals, and alco-
glucose intolerance or metabolism problems and/or mis-
hol consumption. The energy requirements of the fetus
carriages or other obstetric problems. Additionally, Latino,
must be met to ensure proper development and provide
African, Native American, South or East Asian, and Pacific
postpartum lactation without causing excessive maternal
Islanders are at a higher risk for developing GDM.15 The
weight gain. The energy standard to support a pregnancy
Santa Barbara County Health Care Services Program Study16
has been debated heavily and will be explored in the sec-
found that women who meet several of these criteria, at the
tion “GDM and nutritional therapy.” The American College
greatest risk, should be tested as early as feasible, while those
of Obstetricians and Gynecologists advocates several basic
with average risk should be tested at 24–28 weeks gestation.
concepts for a balanced diet for pregnant women. They sug-
However, GDM can easily appear in low-risk women; there-
gest eating three to four servings of fruits and vegetables,
fore, universal screening would be ideal.
nine servings of whole grains for energy, three servings of
dairy for calcium, and three servings of meat to reach daily
protein requirements. Vitamin supplementation to achieve
daily nutrients, as an adjunct to a healthy diet, is encour-
Screening methods for GDM
aged when recommended by the woman’s physician. Certain The screening methods for GDM are very controversial.
foods should be avoided in pregnancy due to fetal develop- The two sets of criteria endorsed by the World Health
mental harm. These include deli meat, certain preparations Organization (WHO) or the U.S. National Diabetes Data
of smoked fish, soft cheeses, unpasteurized milk, refriger- Group are each distinct and are primarily based on statisti-
ated pate, raw meat, and raw eggs, which have been associ- cal standard deviations without regarding the level of clinical
ated with bacterial infections such as Salmonella, Listeria, outcome achieved. An attempt to set international standards
and Escherichia coli. Toxoplasma gondii, the protozoan that and identify those at risk for developing GDM is currently
causes toxoplasmosis, has also been found as contaminant in under way in a 5-year, prospective, observational, and mul-
unwashed vegetables and raw meat. Fish containing mercury ticenter study, the Hyperglycemia and Adverse Pregnancy
and raw shellfish should be avoided. Caffeine has been asso- Outcomes study. It involves 25,000 women in 10 countries
ciated with miscarriage, premature birth, low birth weight, and specifically will look at the clinical outcomes with
and withdrawal symptoms in the neonate when consumed respect to cesarean delivery, fetal hyperinsulinemia, mac-
in large amounts in pregnancy. However, other studies have rosomia, and neonatal morbidity in correlation to maternal
implicated caffeine intake in modest levels to be nondetri- glycemic levels.17
mental in pregnancy. Until further studies can evaluate the
effects of caffeine, it is recommended to be avoided alto-
gether.8,9 Alcohol should not be used in any amount during Fetal complications of maternal
pregnancy. In utero exposure has been linked to develop-
mental disorders such as fetal alcohol syndrome. Also, alco- hyperglycemia
hol should be avoided postpartum while breast-feeding.10
Uncontrolled hyperglycemia primarily affects fetal growth
on both extremes of the normal growth curve. In those
diabetic mothers that have advanced vascular disease, fetal
GDM growth deceleration may occur due to placental insuf-
ficiency. Fetal growth deceleration is defined as those in
As discussed in the previous section “Metabolic changes in lower fifth percentile on a growth curve adjusted for gesta-
normal pregnancy,” the nondiabetic woman undergoes dras- tional age.18,19 Macrosomia is defined as an absolute birth
tic and dynamic metabolic changes to provide glucose as the weight of greater than 4000 g or greater than the 90th per-
preferential energy source to the developing fetus. Although centile (adjusted for gender, ethnicity, and gestational age).
the pathophysiological mechanism behind GDM remains Cesarean sections often must be performed when the baby
unknown, some current theories include a predisposition to is at term to reduce the risk of birth trauma such as Erb’s
future type 2 diabetes triggered by the changes in metabo- palsy or Klumpke’s paralysis.20 Cesarean sections also adds
lism that normally accompany pregnancy or an increased risk to the mother’s health. As explained by the Pederson
response by the woman’s body to normal metabolic changes hypothesis,21 the effects of an intrauterine environment of
of pregnancy. GDM has been defined as “glucose intoler- hyperglycemia and hyperinsulinemia include hypoglycemia,
ance of variable severity with onset or first recognized dur- organ developmental problems (especially gastrointestinal),
ing pregnancy.”11 It is important to note that GDM does erythrocytosis, iron redistribution, calcium and magnesium
not cause a malfunction in insulin secretion or improper deficiencies, respiratory problems (respiratory distress syn-
proinsulin or glucagon activity: insulin resistance remains drome), cardiac problems (intraventricular hypertrophy and
140 Medical nutritional therapy for gestational diabetes mellitus
cardiomyopathy or heart failure), hyperbilirubinemia, and adjustments based upon those levels. Women in the routine
neurological sequelae. care group received care from blinded clinicians as to the
A multitude of metabolic problems occur that not only status of the previously performed oral glucose tolerance
affect the immediate future of the neonate, but as these chil- test (OGTT). However, if the clinician felt the patient was
dren mature, they are predisposed to future metabolic prob- experiencing glucose intolerance, assessment and treatment
lems, such as type 2 diabetes and metabolic syndrome.22,23 could be instituted at his or her discretion. Otherwise, pre-
Subsequently, this pattern of metabolic disease takes a cyclic natal care that was specific to the center visited was given.
course affecting future generations. Primary outcomes included one or more perinatal complica-
tions defined as death, shoulder dystocia, bone fracture, and
nerve palsy. Admission to the neonatal intensive care unit
and jaundice that required phototherapy was also assessed.
Does treatment of GDM make a Secondary outcomes included a consideration of the primary
difference in pregnancy outcome? outcome, gestational age at delivery, overall birth weight and
birth weight adjusted for gestational age, and the presence of
In June 2005, Crowther and colleagues24 published the macrosomia or fetal decelerated growth. Maternal outcomes
results of a 10-year multicenter randomized clinical trial were assessed on the basis of general and mental health
in Australia and the United Kingdom called the Australian including depression, anxiety, gestational age at birth, mode
Carbohydrate Intolerance Study in Pregnant Women of birth, weight gain during pregnancy, hospital admis-
(ACHOIS) (Table 18.1). The purpose of the ACHOIS was sions and prenatal visits, and common complications, such
to determine whether medical nutritional therapy, glucose as pregnancy-induced hypertension. The ACHOIS study24
monitoring, and insulin therapy were superior to routine established several significant results in both the neonatal
prenatal care with regard to reducing the risk of perina- and maternal outcomes. The rate of serious adverse perinatal
tal complications and postpartum maternal health status. outcomes was significantly different between the interven-
A total of 1000 women participated in this trial, 490 in the tional and routine care groups at 1% versus 4%, respectively
intervention group and 510 in the routine care group with (p < 0.01). This rate established the number needed to treat
eligibility based upon the presence of one or more risk fac- at 34 to reduce the incidence of perinatal complication.
tors for GDM or a positive 50 g glucose challenge test (GCT) Newborns in the interventional group were admitted to the
who did not have an indication of pregestational diabetes, neonatal intensive care units more often, at a rate of 71% ver-
history of GDM, or an active chronic disease. The WHO sus 61% admission rate in the routine group (p < 0.01). No sig-
criteria were used to identify those with GDM, and women nificant difference was seen when considering the length of
with severe glucose impairments were excluded. Therapies stay in the NICU or use of phototherapy due to jaundice. The
provided to the women in the interventional group consisted percentage of mothers who were induced into labor was 39%
of dietary counseling with consideration to prepregnancy in the interventional group versus 29% in the routine care
weight, activity level, normal dietary intake, and weight group (p < 0.001). However, rates and the reason for cesarean
gain. Women were asked to self-monitor their glucose levels sections were similar between the groups.24 With regard to
four times a day and proceed with insulin therapy dosage the secondary outcomes, the interventional group neonates
Table 18.1 Summary of the Australian Carbohydrate Intolerance Study in Pregnant Women study
Maternal and neonate primary, secondary Intervention group, n = 490 Routine group, n = 510
outcomes evaluated mothers, 506 neonates mothers, 524 neonates
Perinatal complications (p = 0.01) 1% 4%
NICU admission rate (p = 0.01) 71% 61%
Length of stay and phototherapy Median 1 day Median 1 day
implementation (NSD) Interquartile range, 1–2 days Interquartile range, 1–3 days
Labor induction (p < 0.001) 39% 29%
Rate of cesarean section (NSD) 31% 32%
Mean birth weight (p < 0.001) 3335 ± 551 g 3482 ± 660 g
Large for gestational age (p < 0.001) 13% 22%
Macrosomia (p < 0.001) 10% 21%
Small for gestational age (NSD) 7% 7%
Maternal Edinburgh postnatal depression 8% 17%
Score > 12 (p = 0.001)
Source: Crowther, C.A. et al., N. Engl. J. Med., 352, 2477, 2005.
Abbreviation: NSD, no significant difference.
Caloric restriction 141
had significantly lower mean birth weights (p < 0.01) and to fetal detriment. However, in studies conducted by Rizzo
were born earlier presumably due to higher labor induc- and colleagues, hyperketonemia during pregnancy, which
tion rates. However, when adjusting for gestational age with results from maternal diabetes (hyperglycemia), has been
respect to birth weight, there were significantly fewer neo- implicated to affect the fetus’s intellectual and behavioral
nates born to interventional mothers that qualified in the development as measured by the Bayley Scales of Infant
large-for-gestational-age category. Additionally, macrosomia Development and the Stanford–Binet Intelligence Scale,
occurred significantly less often, but the rate of infants in the which were administered at ages 2 and 3, 4, and 5 years,
small-for-gestational-age group did not differ between the respectively. Hence, it was suggested that ketonemia be
interventional and routine care groups. Maternal outcomes avoided in all pregnant women.28 Buchanan and colleagues
regarding maternal perception at 3 months postpartum of contrasted the metabolic response in normal pregnant
health showed an improved quality of life in the interven- women without GDM, to those who were obese with GDM.
tional mothers specifically with a reduction in the incidence He subjected both groups to an overnight fast and then an
depression, 8% versus 17% as measured by the Edinburgh extended 18-hour total fast. Obese women with GDM had
Postnatal Depression Scale24 (Table 18.1). The ACHOIS study a greater decrease in plasma glucose levels and were not
clearly shows the benefits of using a multifaceted approach more prone to develop ketonemia than the normal pregnant
toward managing GDM to the neonate and the mother. Not women. This result would support the use of decreasing the
only does it advocate dietary and insulin therapies, but also frequency of meals in order to achieve lower preprandial glu-
when one considers the population chosen and the method cose levels in obese women with GDM.29 In a study of type 1
of randomization, it supports the use of universal screening. diabetic women, Jovanovič and coworkers showed that those
Based on the study design, the interventional group would infants whose mothers had the lowest betahydroxybutyrate
be equivalent to universal screening practices and the rou- levels had the largest infants because mothers’ postprandial
tine group would represent those in which GDM screen- glucose concentrations were higher due to the increased
ing is not routine. Therefore, by applying the results of the caloric intake prescribed to avoid the ketonemia.30
study in the women receiving routine care, only 34 women
would need to be treated with interventional therapies to
produce one improved neonatal outcome. Given the severity Caloric restriction
of an adverse neonatal outcome, this finding would support
expanding the GDM screening population to include women Caloric restriction in pregnant women with GDM is another
who would routinely not be screened for GDM. aspect of medical nutritional therapy that needs to be addressed.
When women who are classified as obese or overweight prior
to pregnancy, the amount of weight gain in pregnancy differs
GDM and nutritional therapy from those who are at a normal or underweight prior to preg-
nancy. The National Academy of Science has recommended
As demonstrated earlier, management of GDM is multifac- that for women greater than 150% of ideal body weight, no
eted. Insulin therapy, exercise, and diet are all vital components more than 15 lb should be gained with pregnancy. Optimal
toward reducing the incidence of maternal hyperglycemia and infant birth weight was achieved when less than 3 kg or no
ultimately fetal complications. The remainder of this chapter weight was gained in these women.31 Hypocaloric diets have
will focus on GDM nutritional therapy. Currently, there is been explored in women with GDM based upon a 2400 kcal/
no universally accepted medical nutritional therapy for the day diet. Investigators32 compared a 2400 kcal/day diet to a
treatment of GDM. The American Dietetic Association advo- 1200 kcal/day diet and achieved significant differences in aver-
cates the standard medical nutritional therapy for a GDM age glucose and fasting insulin levels, but not in fasting or post-
mother to be the standard therapy advocated in nonpregnant GCTs. Those in the 1200 kcal/day group developed ketonemia
adults with the carbohydrate content standard being <60% and ketonuria; therefore, the study was discontinued due to the
carbohydrate per meal. However, when these standards were controversial association of ketones with fetal developmental
followed, an increase in insulin therapy was seen in more harm. Subsequently, in another study conducted by the same
than 50% of women with GDM.25 Additional studies have investigators, within the first week, when compared with the
supported lower carbohydrate percentages. For example, in 2400 kcal/day diet, a 1600 kcal/day diet improved fasting and
a study involving obese women with GDM, when the car- mean daily glucose values without the development of keto-
bohydrate was restricted to 33%, the infants were all within nemia. Further studies have advocated a 1500–1800 kcal/day
normal birth weight ranges and there was no evidence of diet for obese women with GDM with similar results.26 The
maternal ketonemia. standards for energy requirements for pregnant women with
GDM, as supported by the American College of Obstetrics and
Gynecology, determine the amount of energy needed to main-
Ketonemia and ketonuria tain pregnancy based upon the pregravid weight. For women
with GDM, who are 1.5 times their ideal body weight, the caloric
Following an overnight fast, 10%–20% of all pregnant intake is 12–15 kcal/kg of the current pregnant weight, while
women have ketones in their blood.26,27 This fasting keto- those at less than 0.8 of their ideal body weight are to increase
nemia or “starvation ketonemia” has not been associated their caloric intake to 35–40 kcal/kg current pregnant weight.
For those at 0.8–1.2 times their ideal body weight, 30 kcal/kg, GDM who have greater than 130% of their ideal body weight
and those at 1.2–1.5 times ideal body weight, 24 kcal/kg cur- at 32–36 weeks’ gestation. The goal was to achieve a post-
rent pregnant weight is the standard.15 The “euglycemic diet” prandial of 120 mg/dL at 1 hour. None of the patients were
advocated the lower range of the spectrum set by the American on an insulin regimen and a caloric restriction of 24 kcal/
College of Obstetrics and Gynecology. kg/day was established. Patients kept a diary of glucose lev-
els taken four times a day and food intake. The carbohydrate
parameters of the diet were as follows: 12.5% of the total daily
Carbohydrate restriction carbohydrate at breakfast, 28% at lunch and dinner, with the
remainder in three snacks disturbed throughout the day.
The Pederson hypothesis attributes fetal macrosomia due The postprandial glucoses recorded by the women correlated
to hyperinsulinemia caused by maternal hyperglycemia. to the carbohydrate intake. From this study, the author has
Several studies have shown that when maternal glucose lev- adapted this diet to achieve optimal control of glycemic levels
els are well controlled, the incidence of macrosomia, fetopa- in her patients. Most women with GDM are very compliant
thy, and cesarean sections decreases.21,33–35 Currently, there and want to do what is necessary to have a healthy baby. By
is no set standard for pre- and postprandial levels in women having patients take an active role in their medical care, they
with GDM. Optimally, the therapeutic target of glucose lev- can significantly reduce their risk for fetal macrosomia. In a
els in a woman with GDM would be the same as those who study conducted by de Veciana et al.36 when patients moni-
are pregnant without diabetes. Normoglycemia in the preg- tored their pre- and postprandial glucose levels, the risk of
nant, nondiabetic, nonobese woman was demonstrated in fetal macrosomia decreased from 42% to 12%. Additionally,
studies conducted in 2001, by Parretti and colleagues.34 In these patients also had lower hemoglobin A1c levels, there-
this study, the postprandial mean glucose during the third fore supporting that they maintained lower glycemic levels.
trimester did not exceed 105.2 mg/dL (5.8 ± 0.27 mmol/L). Hemoglobin A1c is an effective clinical tool for accessing gly-
At 28 weeks gestation, the daily mean glucose levels was cemic control and can be performed every 2 weeks to chart
71.9 ± 5.7 mg/dL, and at 38 weeks, it increased to 78.3 ± management because the turnover rate of the red blood cells
5.4 mg/dL, which would coincide with the normal insu- during pregnancy is only 90 days as compared with 120 days
lin intolerance increase, respectively. Parretti et al.34 also in the nonpregnant state. Thus a significant improvement
accessed the clinical outcome of these pregnancies based on in glucose control is manifest by a significantly improved
fetal growth. They found that 1-hour postprandial glucose hemoglobin A1c level, although the steady state has not been
levels at 28 weeks through the third trimester had a posi- achieved until after 6 weeks. Patients monitor their pre- and
tive correlation to fetal abdominal growth. Furthermore, the postprandial glucose levels and only proceed with a meal
results are supportive in attributing the postprandial 1-hour when their preprandial glucose levels are 90 mg/dL or less;
glucose levels as a predictor of infant birth weight, fetal otherwise, insulin is initiated. A preprandial glucose of
macrosomia, fetal hyperinsulinemia, and fetal abdominal 90 mg/dL and a postprandial of 120 mg/dL may seem con-
circumference in nondiabetic pregnancies. Therefore, one troversial or strict; however, given the risk of macrosomia
may consider the level of insulin resistance as a spectrum, in and the positive outcomes that have been obtained clinically,
which those with GDM are affected in the same way as non- the authors of this chapter advocate these glycemic goals for
diabetic pregnant women but to a greater extent. Hence, the medical nutritional therapy.37
levels of glycemia achieved in nondiabetic pregnant women It is imperative that patients learn which foods have high
to decrease incidence of fetal complications and growth carbohydrate content, so educational lessons and nutri-
would be applicable to those with GDM. Another study to tional food label reading is essential for the success of any
support the importance of postprandial glucose levels is the therapy that is instituted. Patients are given a list of recom-
Diabetes in Early Pregnancy Study, which was conducted mended high-carbohydrate foods in order to remind them
with type 1 diabetic mothers. When postprandial glucose what needs to be portioned. For example, the “big 5” are
levels increased, there was an increased risk of macrosomia. potatoes, rice, pasta, bread/tortillas, and cereal (Figure 18.1).
The threshold for the marked increase was seen when post- By teaching patients to adhere to a euglycemic diet, not only
prandial glucose levels reached 120 mg/dL.35 Thus, a dietary are they able to control their glucose levels effectively, but
therapy, “the euglycemic diet” was developed on the basis of also they are able to modify their diet postpartum facilitat-
this study. ing weight loss.
The euglycemic diet takes into account the metabolic A simple teaching tool that is used in Santa Barbara
changes that occur within the pregnant woman as she goes County is shown in Figure 18.2. The one-page handout
throughout her day. In the morning, a surge of cortisol is identifies the foods to avoid, foods to eat with caution, and
seen (“the dawn phenomenon”), which causes the release of foods that may be eaten that minimally impact postpran-
glucose from stored sources and hepatic gluconeogenesis; dial glucose concentrations and thus can be eaten liberally.
thus the blood glucose is higher to begin with. Therefore, a Ideally, a breakfast of less than 33% of the daily carbohydrate
decreased amount of carbohydrate is needed in the breakfast intake, lunch at 45%, and a dinner at 55% are suggested to
meal. A small study (n = 14) was conducted in women with maintain a postprandial glucose level of 120 mg/dL.38
Role of protein 143
Goal: Postprandial 1-hour blood glucose level ≤120 mg/dL
10:30 a.m.: Snack

5% of daily CA
Limit portions
8:00 a.m.: Breakfast of the “Big 5”
12:00 p.m.: Lunch
dawn phenomenon:
if blood glucose
if blood glucose
≤90 mg/dL
≤90 mg/dL
eat 30% daily CA
eat 10% of daily CA
Bread
Potatoes Rice Pasta Cereal
tortillas
11:00 p.m.: Snack 3:00 p.m.: Snack

10% of daily CA 10% of daily CA
Calories as: Daily total (%)
8:00 p.m.: Snack 5:00 p.m.: Dinner Fat ≥40

5% of daily CA if blood glucose
≤90 mg/dL Carbohydrate <40
eat 30% daily CA
Protein 20
CA, carbohydrate allowance
Figure 18.1 The euglycemic diet.
Role of fats in GDM therapy daily caloric intake. Increased satiety has also been cor-
related with meals that are high in protein content.40,41
Fat content in the American Diabetic Association’s diet con- Thus, this aspect could help morbidly obese patients
sists of less than 25% of the total caloric intake, whereas the manage their overall caloric intake especially when
euglycemic diet is composed of 40% of the total daily caloric moderate caloric restriction therapy is being used. Low-
intake. The role of saturated and monounsaturated fats in carbohydrate/high-protein diets in normal pregnant
women with GDM is different with respect to the uptake of women have been explored notably in the Motherwell
glucose postprandially. In a study comparing these two types studies running from 1938 to 1977. The Motherwell stud-
of fats, 1-hour postprandial glucose levels are approximately ies suggested a link between increased protein content and
equal; however, the duration of the elevated glucose levels dif- low birth weight.42 Recent studies have expanded the ini-
fers. In women with GDM who consumed monounsaturated tial Motherwell studies by looking at the offspring of these
fat, the glucose levels remained elevated at a longer period studies as adults. It has been hypothesized that increased
of time, and thus insulin dosage had to be adjusted to coun- protein intake can stimulate maternal cortisol production
teract the maintained elevated glycemia. Conversely, meals and expose the fetus to high levels of cortisol which may
consumed containing saturated fat had a shorter duration of facilitate lifelong hypersecretion of cortisol. Herrick and
elevated glucose levels, making them preferential with regard to colleagues found a correlation to increased plasma corti-
glycemic control of postprandial glucose levels. Furthermore, sol levels consequently causing hypertension in the adult
lower postprandial durations decrease the risk of macroso- Motherwell offspring.43 However, they postulated that the
mia and the need for increased insulin doses.39 Advocating type of protein consumed and the type of carbohydrate
saturated fats over monounsaturated fats is understandably paired with it may factor into the physiologic affects seen.
controversial due to the correlation that has been made with Cofactors needed for protein metabolism such as folate and
saturated fats and heart disease in nonpregnant individuals. vitamin B6 may be excluded when certain types of carbo-
Further studies are needed to answer whether eating a higher hydrate are avoided, such as bread and potatoes and green
proportion of saturated fat during medical dietary therapy for leafy vegetables, as was done in the Motherwell studies.
GDM at approximately gestation weeks 24–40 is a significant “In mothers with a limited capacity to synthesize nones-
time period to have adverse long-term effects on the mother sential amino acids, maternal amino acid oxidation could
versus the benefit of controlling p ostprandial glucose level impair fetal growth as a result of reduced availability of
duration, which decreases the risk of fetal complication. nonessential amino acids.”42 Therefore, the physiologic effect
of low birth weight on the fetus would be caused multi-
Role of protein factorially and not just due to high protein consumption.
Additional studies will be necessary to determine the role
Protein content in the American Diabetes Association of high-protein diets in pregnant women and, specifically, in
(ADA) diet and euglycemic diet makes up 20% of the total women with GDM.
Nutrition jump start...

for improved blood sugar control
STOP Foods to avoid
• Cake • Cold cereal

• Oranges • Grapes
• Cookies • Bread
• Bananas • Squash
• Candy • Tortilla
• Melon • Juice
• Ice cream • Rice/pasta
• Peaches • Mango
• Pie • Soda
• Plums • Dried fruit
• Chips • Pastries
• Potato • Ketchup
• Sugar • Dressing
Foods to limit (1–2 servings per meal)
• ½ cup cooked peas/corn • ½ cup cooked oatmeal

• ½ cup cooked beans • 4–6 whole-wheat crackers
• ½ cup cooked lentils • 3 cups plain air-popped popcorn
Foods you can enjoy

Proteins Vegetables
• Chicken • Lettuce/spinach
• Fish • Carrots
• Beef Fruits • Celery
Fats
• Pork • Lemons • Mushrooms
• Olive oil • Green beans
• Cheese • Limes
• Avocado • Cucumber
• Eggs • Tomatoes
• Sour cream • Broccoli
• Plain yogurt • Apples
• Butter • Asparagus
• All nuts • Strawberries
• Cottage cheese • Cabbage
• Nopal
Note: Not an exhaustive list
Figure 18.2 The one-page handout used in Santa Barbara County.
Conclusion be directly linked to the level of neurological functioning

of the child, and these children are more prone to develop
GDM is a period of glucose intolerance that manifests at metabolic syndromes such as type 2 diabetes. This would
the beginning of the third trimester. Metabolic changes in affect generations to come as well. The management of
the normal pregnant women also have a degree of insulin GDM is based upon the synergistic effects of medical nutri-
resistance that shunts glucose preferentially to the fetus. tional therapy, exercise, and an insulin regimen when nec-
To maintain blood glucose levels within a tight range, the essary. Therefore, the identification and treatment of GDM
normal pregnant woman must increase her insulin secre- is crucial. Screening tests using at-risk formulations and
tion up to fourfold. When the pancreas is not able to com- OGTTs remain a point of controversy. Universal screen-
pensate for the increased insulin needs of pregnancy, GDM ing would be optimal to identify those with GDM. The
occurs resulting in hyperglycemia and hyperinsulinemia research clearly shows the benefit of expanding screening
in both the mother and fetus. These increased glucose and providing medical nutritional therapy, glucose moni-
and insulin levels manifest a multitude of fetal and mater- toring, and insulin therapy to all women who manifest even
nal complications, the most prevalent being macrosomia. minor elevations of glycemia and thereby decrease perinatal
Other complications include hypoglycemia, erythrocytosis, complications, such as macrosomia. Maternal postpartum
hypocalcemia and hypomagnesemia, hyperbilirubinemia, depression rates may also be improved with improved care
iron redistribution, respiratory distress, and neurologi- during pregnancy. Multiple studies have correlated fetal
cal effects. Poor gestational metabolic management can complications such as macrosomia to 1-hour postprandial
References 145
glucose levels. By restricting carbohydrate concentration in different with respect to detriment to the fetus. Fat content
the euglycemic diet and modifying the caloric intake based also remains controversial although studies have shown that
on pregravid weight, success has been achieved in reducing meals with saturated fat as compared to monounsaturated
large-for-gestational-age and macrosomic infants. The eug- fat result in the same-hour postprandial glucose level, but
lycemic diet targets a preprandial glucose of 90 mg/dL or less the duration of the level is shorter facilitating lower insulin
and a 1-hour postprandial of 120 mg/dL. Optimal glucose dosages. High-protein/low-carbohydrate diets are also con-
levels have been heavily debated and there is not currently troversial and in normal pregnant women have been cor-
a universal standard. However, research has shown that related to lower birth weights and adult offspring increased
normal pregnant women in the third trimester have lower cortisol levels. However, satiety is also important and pro-
preprandial and postprandial glucose concentrations than tein malnutrition should be avoided in pregnancy. More
nonpregnant women. This would support advocating lower research is necessary to determine the effect of these macro-
standards of ≤90 mg/dL preprandially and 120 mg/dL post- molecules on normal pregnant individuals and those with
prandially for women with GDM. Hypocaloric diets have GDM. Overall, medical nutritional therapy is one of the
been explored and appear safe for the obese woman with staples of GDM management. Women with GDM are very
GDM. The American College of Obstetrics and Gynecology compliant and most are willing to make dietary changes in
advocate consideration of the mother’s pregravid weight their lives for the benefit of their babies. The successful triad
when considering the caloric needs per kilogram per day. of medical nutritional therapy, exercise, and insulin therapy
The presence of maternal ketonemia and ketonuria is con- for GDM is essential not only to achieve healthy babies but
troversial with respect to fetal development, and the mecha- also to assure that generations to come will begin life with
nisms and outcomes associated with ketonemia resulting normal metabolism, and thus future metabolic aberrancy is
from uncontrolled glucose levels and starvation may be reduced in the offspring.44–47
REFERENCES
1. Butte NJ, Hopkinson JMM, Mehta N et al. Adjustments in energy 15. American College of Obstetrics and Gynecology. Gestational
expenditure and substrate utilization during late pregnancy and diabetes. Practice Bulletin No. 30. Obstet Gynecol 2001; 98:
lactation. Am J Clin Nutr 1999; 69: 299–307. 525–538.
2. Catalano PM, Tyzbir ED, Roman NM. Longitudinal changes in 16. Jovanovič L, Bevier W, Peterson CM, for the Santa Barbara County
insulin release and insulin resistance in non-obese pregnant Health Care Services Program. Birth weight change concomitant
woman. Am J Obstet Gynecol 1991; 165: 1667–1672. with screening for and treatment of glucose-intolerance of preg-
3. Catalano PM, Tyzbir ED, Wolfe RR et al. Longitudinal changes in nancy: A potential cost-effective intervention. Am J Perinatol
basal hepatic glucose production and suppression during insulin 1997; 14: 221–228.
infusion in normal pregnant women. Am J Obstet Gynecol 1992; 17. HAPO Study Cooperative Research Group. The Hyperglycemia
167: 913–919. and Adverse Pregnancy Outcome (HAPO) Study. Int J Gynecol
4. Catalano PM, Tyzbir ED, Wolfe RR et al. Carbohydrate metabo- Obstet 2002; 78: 69–77.
lism during pregnancy in control subject and women with gesta- 18. Georgieff MK. Therapy of infants of diabetic mothers. In: Burg FD,
tional diabetes. Am J Physiol 1993; 264: E60–E67. Ingelfinger JR, Wald ER, Polin RA, eds., Current Pediatric Therapy,
5. Butte NJ. Carbohydrate and lipid metabolism in pregnancy: 15th ed., WB Saunders: Philadelphia, PA, 1995, pp. 793–803.
Normal compared with gestational diabetes mellitus [Review]. 19. Creasy RK, Resnik R. Intrauterine growth restriction. In: Creasy
Am J Clin Nutr 2000; 71(Suppl.): 1256S–1261S. RK, Resnick R, eds., Maternal–Fetal Medicine, 4th ed., WB
6. Kalhan SC, D’Angelo LJ, Savin SM, Adam PAJ. Glucose produc- Saunders: Philadelphia, PA, 1999, pp. 569–589.
tion in pregnant women at term gestation. Sources of glucose for 20. Nold JL, Georgieff MK. Infants of diabetic mothers. Pediatr Clin
human fetus. J Clin Invest 1979; 63: 388–394. N Am 2004; 51: 619–637; Schwartz RP. Neonatal hypoglycemia:
7. Assel B, Rossi K, Kalhan D. Glucose metabolism during fasting How low is too low? J Pediatr 1997; 131: 171–173.
through human pregnancy. J Clin Invest 1997; 100: 1775–1781. 21. Pederson J. The Pregnant Diabetic and Her Newborn, 2nd edn.,
8. Narod SA, De Sanjose S, Victoria C. Coffee during pregnancy: Williams & Wilkins: Baltimore, MD, 1977.
A reproductive hazard? Am J Obstet Gynecol 1991; 164: 1109–1114. 22. Weiss PA, Scholz HS, Haas J et al. Long term follow-up of infants
9. Leviton A, Cowan L. A review of the literature relating caffeine of mothers with Type I diabetes: Evidence for hereditary and
consumption by women to their risk of reproductive hazards non-hereditary transmission of diabetes and precursors. Diabetes
[Review]. Food Chem Toxicol 2002; 40: 1271–1310. Care 2000; 23: 905–911.
10. Abel EL. Prenatal effects of alcohol [Review]. Drug Alcohol 23. Boney CM, Verma A, Tucker R, Vohr BR. Metabolic syndrome in
Depend 1984; 14: 1–10. childhood: Association with birth weight, maternal obesity, and
11. Report of the expert committee on the diagnosis and classifi- gestational diabetes mellitus. Pediatrics 2005; 115: e290–e296.
cation of diabetes mellitus. Diabetes Care 2003; 26(Suppl. 1): 24. Crowther CA, Hiller JE, Moss JR et al. Effect of treatment of ges-
S5–S20. tational diabetes mellitus on pregnancy outcomes. N Engl J Med
12. Buchanan TM, Metzger BE, Freinkel N. Insulin sensitivity and 2005; 352: 2477–2486.
beta cell responsiveness to glucose during late pregnancy in lean 25. Kitzmiller JL, Gavin LA, Gin GD et al. Preconception care of
and moderately obese women with normal glucose tolerance or diabetes: Glycemic control prevents congenital anomalies. J Am
gestational diabetes mellitus. Am J Obstet Gynecol 1990; 162: Med Assoc 1991; 265: 726–731.
1008–1014. 26. Knopp RH, Magee M, Raisys V. Hypocaloric diets and ketogen-
13. Coustan DR. Gestational diabetes. In: Harris MI, Cowie CC, Stern esis in the management of obese gestational diabetic women.
MP et al., eds., Diabetes in America, 2nd edn., National Institutes J Am Coll Nutr 1991; 10: 649–667.
of Health: Baltimore, MD, 1995, pp. 703–717. 27. Mills JL, Simpson JL, Driscoll SG et al. Incidence of spontaneous
14. Miller E, Hare JW, Clogerty JP et al. Elevated maternal hemoglo- abortion among normal women and insulin dependent diabetic
bin A1c in early pregnancy and major congenital anomalies in women whose pregnancies were identified within 21 days of
infants of diabetic mothers. N Engl J Med 1981; 304: 1331–1335. conception. N Engl J Med 1988; 319: 1617.
28. Rizzo T, Metzger BE, Urns WJ et al. Correlations between ante- 39. Ilic S, Jovanovič L, Pettit DJ. Comparison of the effect of saturated
partum maternal metabolism and intelligence of offspring. N Engl and monounsaturated fat on postprandial plasma glucose and
J Med 1991; 325: 911–916. insulin concentration in women with gestational diabetes mel-
29. Buchanan TA, Metzger BE, Freinkel N. Accelerated starvation in litus. Am J Perinatal 2000; 16: 489–495.
late pregnancy: A comparison between obese women with and 40. Hill AJ, Blundell JE. Macronutrients and satiety: The effects of
without gestational diabetes mellitus. Am J Obstet Gynecol 1990; high protein or high carbohydrate meal on subjective motiva-
162: 1015–1020. tion to eat and food preferences. Nutr Behav 1986; 3: 133–144.
30. Jovanovič L, Metztger B, Knopp RH. Beta hydroxybutyrate levels 41. Westerterp-Plantenga MS. The significance of protein in food
in type 1 diabetic pregnancy compared with normal pregnancy. intake and body weight regulation [Review]. Curr Opin Clin
Diabetes Care 1998; 21: 1–5. Nutr Metab Care 2003; 6: 635–638.
31. King J, Allen L. Nutrition during Pregnancy, National Academy 42. Kerr JF, Campbell-Brown BM, Johnstone FD. A study on the
Press: Washington, DC, 1990. effect of high protein low carbohydrate diet on birthweight on
32. Magee MS, Knopp RH, Benedetti, TJ. Metabolic effects of an obstetric population. In: Sutherland HW, Stowers JM, eds.,
1200 kcal diet in obese pregnant women with gestational diabe- Carbohydrate Metabolism in Pregnancy and the Newborn
tes. Diabetes 1990; 39: 234–240. 1978, Springer-Verlag: Berlin, Germany, 1979, pp. 518–534.
33. Combs CA, Gunderson E, Kitzmiller JL et al. Relationship of fetal 43. Herrick K, Phillips DI, Haselden S et al. Maternal consump-
macrosomia to maternal postprandial glucose control during tion of a high-meat, low-carbohydrate diet in late pregnancy:
pregnancy. Diabetes Care 1992; 15: 1251–1257. Relation to adult cortisol concentrations in the offspring. J Clin
34. Parretti E, Mecacci F, Papini M et al. Third-trimester maternal Endocrinol Metab 2003; 88: 3554–3560.
glucose levels from diurnal profiles in nondiabetic pregnancies. 44. Jovanovič L. Endocrine Society Guidelines for care for the
Diabetes Care 2001; 24: 1319–1327. person with diabetes. J Clin Endocrinol Metab 2012; 96:
35. Jovanovič L, Peterson CM, Reed GF et al. Maternal postpran- 1270–1285.
dial glucose levels and infant birth weight: The Diabetes in Early 45. Blumer I, Hadar E, Hadden DR, Jovanovič L, Mestman JH,
Pregnancy study. The National Institute of Child Health and Murad MH, Yogev Y. Diabetes and pregnancy: An Endocrine
Human Development—Diabetes in Early Pregnancy Study. Am Society Clinical Practice Guideline. J Clin Endocrinol Metab
J Obstet Gynecol 1991; 164: 103–111. 2013; 98: 4227–4249.
36. de Veciana M, Major CA, Morgan MA et al. Postprandial versus 46. Jovanovič L. Using meal-based self-monitoring of blood glu-
preprandial blood glucose monitoring in women with gestational cose as a tool to improve outcomes in pregnancy complicated
diabetes mellitus requiring insulin therapy. N Engl J Med 1995; by diabetes. Endocr Pract 2008; 14: 239–247.
333: 1237–1241. 47. Jovanovič L, Savas H, Mehta M, Trujillo A, Pettitt D. Frequent
37. Jovanovič L, Pettitt DJ. Contempo updates. Linking evidence and monitoring of A1C during pregnancy as a treatment tool to
experience: Gestational diabetes mellitus. J Am Med Assoc 2001; guide therapy. Diabetes Care 2011; 34(1): 53–54.
286: 2516–2518.
38. Peterson CM, Jovanovič L. Percentage of carbohydrate and gly-
cemia response to breakfast, lunch, and dinner in women with
gestational diabetes. Diabetes 1991; 40(Suppl. 2): 172–174.
19 Pharmacologic treatment of
gestational diabetes mellitus:
When to start and what agent
to use
Celeste P. Durnwald and Mark B. Landon
The mainstay of gestational diabetes mellitus (GDM) manage- to use. For years, the gold standard for treating women with
ment is dietary therapy, exercise, and self-monitoring of blood GDM who fail to achieve euglycemia on dietary manage-
glucose. Nutritional management focuses on a carbohydrate- ment alone has been insulin injections. Women with GDM
restricted diet consisting of 35%–40% of calories from car- have endogenous insulin production but are unable to either
bohydrates and the remainder of calories divided between mount the twofold to threefold increase in insulin secre-
protein (20%) and fat (40%). The general recommendation for tion or to overcome the insulin resistance of pregnancy in
daily glucose monitoring is four times daily performed as fast- order to maintain euglycemia. Therefore, oral hypoglycemic
ing and either 1 or 2 hours after the beginning of each meal. agents can also be considered as options for treatment. In
The American Congress of Obstetricians and Gynecologists recent years, investigations have focused on examining the
endorses glycemic targets of a fasting glucose ≤95 mg/dL, use of both glyburide and metformin.
1 hour postprandial ≤140 mg/dL, and 2 hour postprandial
≤120 mg/dL as measures of adequate control.1 The American
Diabetes Association endorses similar pre- and postprandial Oral hypoglycemic agents
levels.2 These thresholds have been extrapolated from recom-
mendations for pregnant women with pregestational diabetes. The rationale for oral hypoglycemic use in pregnancy is appeal-
Institution of strict glycemic control has been demonstrated ing to both the woman diagnosed with GDM and the clinician.
to reduce neonatal morbidity and mortality associated Advantages of oral hypoglycemic agents include less patient
with the diabetic pregnancy by decreasing the incidence of discomfort, fewer supplies, and less office infrastructure needed
stillbirth and macrosomia.3,4 Infants born to women with well- that may lead to increased patient satisfaction and compliance
controlled diabetes also have fewer neonatal complications with pharmacologic therapy. In addition, prescribing glyburide
such as respiratory distress syndrome, hypoglycemia, and has been shown to be more cost-effective compared with insu-
hyperbilirubinemia.5 Treating GDM has also shown mater- lin injections.9 The average cost savings per patient based on
nal benefit including a decrease in hypertensive disorders of wholesale drug costs and hospital costs was $165.84. When
pregnancy, and in one trial, a reduction in cesarean delivery comparing new pharmacologic treatments with the previously
was observed.6 Approximately, 50% of women will meet these accepted standard for management, one must consider the
glycemic target values within the first 2 weeks of dietary ther- maternal–fetal unit in its entirety. Therefore, clinical efficacy in
apy, but only an additional 10% will achieve euglycemia by the regard to maternal glycemic control must be addressed, as well
fourth week.7 Most clinicians use this time frame to determine as transplacental passage and perinatal outcomes.
dietary failure and to initiate pharmacologic therapy. There is
no conclusive evidence for the threshold number of blood glu-
coses above target or the absolute blood glucose value at which Glyburide
a clinician should initiate pharmacologic therapy.8 In our Glyburide is a second-generation secretagogue that increases
practice, if more than a third of finger-stick glucoses are above insulin secretion and sensitivity in peripheral tissues. It also
the target range, pharmacologic therapy is initiated. reduces hepatic insulin clearance by binding to pancreatic
Once the decision to initiate medication treatment is beta cell ATP-sensitive calcium channel receptors. Therefore,
made, the clinician must decide which pharmacologic agent it is a reasonable option for the treatment of GDM in women
147
148 Pharmacologic treatment of gestational diabetes mellitus
who have underlying beta cell dysfunction. Glyburide typi- hypoglycemia were similar between groups. Glyburide was
cally lowers blood glucose by approximately 20% and works not detected in the cord serum of any infant with the mean
best in those individuals who are normal in weight to slightly time from the last dose of glyburide to cord blood sampling
overweight. It is well absorbed and not dependent on food of 8 ± 4 hours. Additionally, in 12 randomly selected women
intake with a primary side effect of hypoglycemia, although in the glyburide group, glyburide was measured at the same
fairly low. A typical starting dose is 2.5 mg with a maximum time in maternal and cord serum. Maternal serum concentra-
dose of 20 mg, usually dosed once or twice daily. Although tions ranged from 50 to 150 ng/mL, whereas glyburide was
the majority of studies published using glyburide for the undetectable in cord serum. In a secondary analysis of the
treatment of GDM in pregnancy utilize this once or twice parent study, the clinical efficacy of glyburide was comparable
daily dosing, based on the Obstetric-Fetal Pharmacology to insulin regardless of GDM severity as defined by the fast-
Research Units Network findings, there is new insight into ing value on 3 hour oral glucose tolerance test (OGTT).12 The
timing and number of doses.10 First, glyburide concen- majority of women with a fasting glucose value ≤100 mg/dL
trations begin to increase within 30–60 minutes, peak in were able to achieve glucose control with low-dose glyburide
2–3 hours, and return to baseline by 8 hours. Based on these (≤10 mg). In contrast, for those with higher fasting values
findings, it appears glyburide would work best if dosing regi- ≥106 mg/dL, the number of women requiring low and high
mens are adjusted to approximately 1 hour prior to the meal dose (>10 mg/dL) were similar. There was no significant differ-
to allow for maximum coverage of the drug for postprandial ence in the rates of LGA, macrosomia, or metabolic complica-
glycemic excursion. In addition, a predinner dose may be tions in the neonates of the treatment groups.
necessary in some women. These practical points regarding Since the publication of the RCT, several prospective
administration of the medication should be discussed at the cohort studies and retrospective reviews have evaluated the
time of initiation. Glyburide is primarily metabolized by the use of glyburide outside the constraints of a large RCT.13–17
liver. The oral clearance of the drug from plasma is at double In these studies, failure rates are closer to 20% with rates
the rate of oral clearance in nonpregnant individuals, most of macrosomia similar to those seen with insulin therapy
likely related to the metabolism of glyburide primarily from (Table 19.1). These smaller studies have identified predic-
the cytochrome P450 system that is upregulated in pregnancy. tors of glyburide failure to include both a diagnosis of GDM
Therefore, plasma drug levels in pregnancy are 50% lower earlier in pregnancy and a need for pharmacologic therapy
compared with nonpregnant individuals.10 earlier in pregnancy, higher mean fasting and 1-hour post-
prandial glucoses in the week prior to initiating therapy, and
Maternal glycemia and perinatal outcomes higher values on the 3-hour OGTT. Specifically, Conway
The landmark randomized clinical trial (RCT) comparing identified that women with fasting glucose >110 on 3-hour
glyburide and insulin for clinical efficacy in women with OGTT were 50% more likely to fail glyburide therapy.14 In
GDM was conducted by Langer and colleagues.11 Women Kahn’s study, women with GDM diagnosis <25 weeks of ges-
diagnosed with GDM between 11 and 33 weeks who failed tation had an eightfold risk of glyburide failure after control-
dietary management were randomly assigned to glyburide or ling for fasting blood glucose, gravidity, and parity.16
insulin. In this predominantly Hispanic population, there was The most common side effect of pharmacologic therapy
no difference in glycemia between groups as seen with fasting, in GDM management is maternal hypoglycemia. In the
preprandial, and postprandial values. A 4% failure rate was original RCT by Langer, the rate of hypoglycemia in women
identified in which women were unable to achieve adequate treated with glyburide (2%) was significantly less than their
glycemic control on maximum dosing of glyburide neces- counterparts treated with insulin (20%), p < 0.03.11 Yogev
sitating a switch to insulin therapy. Although the study was also examined the prevalence of undiagnosed asymptomatic
not adequately powered to evaluate neonatal outcomes, rates hypoglycemia in women with GDM treated with insulin and
of large for gestational age (LGA), macrosomia, and neonatal glyburide after at least 2 weeks of therapy.18 Using continuous
Table 19.1 Studies evaluating rates of macrosomia, failure rate of glyburide, and clinical
characteristics associated with glyburide failure
n Failure rate (%) Macrosomia (%) Predictors of failure

Chmait et al.13 69 19 17 Earlier gestational age at diagnosis
Higher glucose levels
Conway et al.14 75 16 11 Gestational age treatment initiated
Higher values on OGTT
Kremer and Duff15 73 19 19 Not evaluated
Kahn et al.16 95 19 7 Earlier gestational age at diagnosis
Yogev et al.17 124 25 6.5 Higher weight gain
GCT > 200 mg/dL
Oral hypoglycemic agents 149
glucose sensing device for 72 hours, the 25 women treated in significant amounts.21 In a follow-up study by the same
with glyburide had an average of half the number of asymp- group, they compared placental transfer of glyburide to
tomatic hypoglycemic episodes each day defined as a blood first-generation sulfonylureas, showing that glyburide’s high
sugar less than 50 for 30 minutes compared with the 30 molecular weight, high plasma protein binding, and short
women treated with insulin. In women treated with insu- half-life were favorable characteristics in limiting its trans-
lin, 84% of hypoglycemic episodes were nocturnal com- placental passage.22 More recently, both Kraemer and Hebert
pared with glyburide treated women in which the episodes have published conflicting results. Kraemer’s placental perfu-
were equally divided between night and day. Brustman and sion studies documented substantial fetal-to-maternal trans-
colleagues examined if the rate of hypoglycemia was dose fer of glyburide against a concentration gradient strongly
dependent.19 In this study, 33% of glyburide-treated women implying active transport.23 Hebert, in work conducted
had at least one episode of documented hypoglycemia through the Obstetric-Fetal Pharmacology Research Unit
<50 mg/dL, but no episodes of symptomatic or severe hypo- Network, reported that glyburide crosses the placenta with
glycemia. There were no significant hypoglycemic episodes an average cord-to-maternal plasma concentration ratio of
in either group. Often, side effect profiles of drugs can be a 0.7 demonstrating that umbilical cord concentrations aver-
rate limiting factor to patient acceptability and compliance aged 70% of maternal concentrations.24 This relationship
with treatment. These findings confirm the favorable side persisted over various time intervals from the last maternal
effect profile with glyburide treatment. dose, although both concentrations were quite low due to the
In 2005, Jacobsen reported on the experience of imple- elapsed time since last dosing before delivery. The difference
menting glyburide use for the treatment of GDM in a large in results between the two groups is attributed to the different
managed care setting.20 This was a retrospective review of assay methodologies used, in which the high-performance liq-
women treated with insulin and glyburide after a diagno- uid chromatography assay used by Hebert is a more sensitive
sis of GDM (by the National Diabetes Data Group), with technique than the UV detection used in the older study by
exclusion for women with a fasting blood glucose >140 on Elliott. The clinical significance, if any, of placental transfer of
3-hour OGTT. Because of its retrospective nature, treatment low to moderate levels of the drug and the potential long-term
groups were not matched. The insulin group (n = 268) was effect of this fetal exposure remain unknown.
treated from 1999 to 2000 and the glyburide group (n = 236)
was treated from 2001 to 2002. The insulin group had a 4%
higher mean body mass index (BMI) and a 4% higher fasting Metformin
OGTT. Glyburide was dosed starting with 2.5 mg once daily The other hypoglycemic agent studied for the treatment of
up to 10 mg, and split dosing was employed if >10 mg was GDM is metformin, a biguanide. It inhibits hepatic gluco-
required. Maternal glycemic targets were a fasting BS ≤ 100, neogenesis, increases intestinal glucose absorption, and
1-hour postprandial ≤155, and 2-hour postprandial ≤130. stimulates glucose uptake in the liver and peripheral tissues.
In evaluating maternal glycemic targets, women treated Metformin does not stimulate insulin secretion or release
with glyburide were more likely to meet target blood sugar and, therefore, does not cause hypoglycemia. It reaches peak
goals, with both lower fasting and 1-hour postprandial glu- concentrations in 3–4 hours and has a half-life of 2–5 hours.
coses, which remained significant after adjustment for BMI, Ninety percent of the drug is excreted within 12 hours of
ethnicity, degree of glucose intolerance, and gestational age at administration. Typical dosing regimens include a starting
diagnosis. Interestingly, women treated with glyburide had a dose of 500 mg once or twice daily with a slow incremental
higher rate of hypoglycemia and preeclampsia. The combined dose increase over 2 weeks to achieve desired glycemia with
failure and discontinuation rate was similar to other reports a maximum dose of 2500 mg. This slow increase limits the
ranging from 12% to 17%. Although this study was also not unattractive gastrointestinal side effects, such as diarrhea
adequately powered to assess neonatal outcomes, there was and flatulence in addition to its most concerning side effect
no difference in gestational age at delivery, birth weight, fetal of lactic acidosis, which is quite rare.
overgrowth (LGA and macrosomia), or neonatal hypoglyce-
mia. Neonates in the glyburide group were less likely to be Maternal glycemia and perinatal outcomes
admitted to the neonatal intensive care unit (NICU), although The RCT by Rowan evaluated metformin compared with
they had a longer length of stay if admission was warranted. insulin for the treatment of GDM (the MiG Trial).25 In this
Neonates born to women treated with glyburide also were RCT, women diagnosed with GDM after a 2-hour 75 g oral
more likely to need phototherapy. In assessing perinatal out- glucose tolerance test between 20 and 33 weeks were assigned
comes reported, these findings suggest that glyburide can be to metformin versus insulin therapy. In the MiG Trial, the
an effective treatment in a large, diverse managed care popu- primary outcome of the study was a composite neonatal out-
lation achieving similar results to insulin therapy. come including neonatal hypoglycemia, respiratory distress,
need for phototherapy, birth trauma, 5 minutes Apgar score
Transplacental passage less than 7, or prematurity. The study was powered to detect
The initial studies by Elliott using an in vitro cotyledon model a 33% difference between treatment groups. Secondary out-
that infused the maternal compartment with glyburide lev- comes studied included neonatal anthropometry, m aternal
els at higher concentrations than therapeutic human dos- glycemic control, hypertensive disorders of pregnancy, post-
ing demonstrated that glyburide did not cross the placenta partum glucose tolerance, and patient acceptability.
The rate of the primary composite outcome was similar in GDM diagnosis and the time of dietary failure. Women
those treated with metformin compared to insulin (32% vs. randomized to metformin had less weight gain during the
32.2%, respectively).25 Only preterm birth was higher in the pregnancy, but this finding was not consistently reported.
metformin group (12.1% vs. 7.6%, p = 0.04). This increase in Rates of fetal overgrowth were similar between those treated
preterm birth was related to an increased number of sponta- with metformin and insulin.
neous preterm births, which may be due to chance or some
yet to be determined effect of metformin. There was no sig- Transplacental passage
nificant difference in birth weight, rates of fetal overgrowth, Likely the biggest drawback of metformin use in pregnancy
or neonatal anthropometric measurements. Umbilical cord is that it freely crosses the placenta due to its low molecular
serum insulin concentrations were also similar between weight, hydrophilic properties, and lack of protein binding.
groups. Although this finding should be interpreted with Umbilical cord concentrations of the drug have been shown
caution as only half of the infants had cord blood levels to be up to twice those seen in maternal serum.30 The fetal
assessed. effects, whether beneficial or detrimental as an insulin sen-
In regard to maternal glycemia, both fasting and 2-hour sitizer, remains largely unknown. It has been proposed that
postprandial blood sugars were similar between groups. in utero exposure to metformin may enhance the action of
Although metformin had great patient acceptability, with fetal insulin and prevent adverse fetal effects or alternatively,
92.6% of women who were randomized to the drug continu- exposure could increase the likelihood of diabetic fetopathy.
ing therapy, 46.3% of women required supplemental insu-
lin. Women who required supplemental insulin had a higher
BMI and higher blood glucoses prior to treatment. The rates Glyburide versus metformin
of the primary outcome did not differ between those treated Two small RCTs have compared glyburide and metformin
with metformin and supplemental insulin (34.5%) compared for the treatment of GDM. In the first randomized trial, gly-
with women on metformin monotherapy (29.7%). Maternal cemic efficacies of 74 women with GDM treated with gly-
weight gain from randomization to 36–37 weeks was sig- buride and 75 women with GDM treated with metformin
nificantly less in those randomized to metformin compared were analyzed.31 In those women who achieved adequate
with insulin. This may play an important role in patient control, mean fasting and 2-hour postprandial glucoses were
counseling and the advantages of using metformin for the similar between treatment groups. However, the failure rate
treatment of GDM. was 2.1 times higher in those receiving metformin compared
Since the seminal RCT by Rowan, a few smaller ran- to glyburide (34.7% vs. 16.2%, p = 0.01, respectively). Despite
domized trials have been published (Table 19.2).26–29 These this failure rate, infants born to women treated with met-
studies are limited by small sample size and the differences formin weighed significantly less than those born to women
in diagnostic criteria for both GDM and dietary failure, in treated with glyburide (3103 ± 600 vs. 3330 ± 334 g), although
addition to the primary outcomes studied. Both Niromanesh the overall rate of LGA was similar between groups.
and Spaulonci focused on maternal glycemic control as the In the report by Silva and colleagues, 96 women with
primary outcome, while Ijas and Tertti focused on neona- GDM were randomized to glyburide and 104 women
tal birth weight and rates of macrosomia. In these studies, with GDM to metformin.32 During treatment, maternal fast-
the rates of need for supplemental insulin were between ing and postprandial glucose levels were similar between
14% and 32%, which is less than 46% reported in the MiG oral hypoglycemic groups. In contrast, the failure rate was
Trial. Similar to the large RCT, women who required insulin similar in both women treated with glyburide and metfor-
supplementation were more likely to have a higher BMI and min (29% vs. 21%). Similar to Moore’s findings, infants born
higher pretreatment glucose values. Other patient character- to women treated with metformin had significantly lower
istics associated with supplemental insulin usage included birth weights (3193 ± 521 vs. 3387 ± 512 g, p = 0.01, respec-
higher fasting values on the diagnostic oral glucose toler- tively), yet the rate of LGA was similar between groups.
ance test and an earlier gestational age at both the time of Interestingly, the ponderal index of neonates born to women
Table 19.2 Randomized controlled trials comparing metformin and insulin use in
the treatment of gestational diabetes
Macrosomia (%)*
Supplemental
n insulin (%) Metformin versus insulin
Ijas et al.26 50 metformin/50 insulin 32 19.1 versus 22
Niromanesh et al.27 80 metformin/80 insulin 14 3.8 versus 10
Spaulonci et al.28 47 metformin/47 insulin 26 0 versus 6.5
Tertti et al.29 110 metformin/107 insulin 21 4.5 versus 0.9
* All p values <0.05.

Insulin therapy 151
treated with metformin was significantly less compared The majority of data published on insulin analogue use in
with those born to women treated with glyburide (2.87 vs. pregnancy pertains to the treatment of women with preex-
2.96, p = 0.05). Rates of neonatal hypoglycemia, defined as a isting diabetes, most specifically type 1 diabetes. However,
blood glucose <40 mg/dL, were also similar between groups the tenets of therapy and clinical experience with these
in both studies. newer insulin formulations can be extrapolated for the
These two trials are limited by small sample sizes and treatment of women with GDM. Certain aspects of treat-
inconsistent results. It is thus difficult to make any definitive ment, such as congenital malformation risk, are not as
conclusions regarding the choice of oral hypoglycemic based pertinent to GDM because the diagnosis occurs beyond
on the results. The greatest limiting factor to using metfor- the period of organogenesis. However, transplacental pas-
min preferentially over glyburide remains the concern for sage, immunogenicity, and clinical efficacy, in addition to
the unknown effects of transplacental passage. However, it maternal and neonatal outcomes, will be similar regard-
should be noted that the large RCT conducted by Rowan and less of diabetes type.
colleagues was adequately powered to assess a composite
neonatal outcome. This is in contrast to the RCT evaluat- Prandial insulin
ing glyburide use in pregnancy in which the primary out-
Prior to the introduction of insulin analogues, women with
come was maternal glycemic efficacy. Further investigations
GDM were treated with the standard insulins, neutral prot-
would be helpful to better elucidate both short- and long-
amine hagedorn (NPH) and regular insulin. In regard to
term effects of fetal exposure to oral hypoglycemic agents.
prandial insulin, regular insulin has almost exclusively been
replaced by the rapid-acting insulin analogues (RAIAs),
insulin lispro and aspart. RAIAs are the preferred choice
Insulin therapy for short-acting insulin due to their superior pharmacologic
profiles, which leads to greater flexibility and convenience of
Insulin replacement is designed to mimic the normal physi- dosing. This has resulted in greater patient satisfaction and
ologic release of insulin by the pancreas and is typically improved quality of life that could lead to a higher patient
divided into basal and prandial insulin. Basal insulin is acceptability of insulin as a therapy.
designed to restrain hepatic glucose production between
meals and in the fasting state. Prandial insulin reduces glu- Rapid-acting insulin analogues
cose excursions associated with feeding. Unlike women with RAIAs are the result of recombinant DNA technology.
preexisting diabetes, insulin therapy for a woman with GDM Unlike regular insulin, which is a hexameric molecule,
is often tailored to address the individual glucose profile of RAIAs quickly dissociate into monomers in the subcutane-
the woman with GDM. The prescribed insulin regimen may ous tissue, which allows for a shorter onset of action, peak
include as few as one and as many as four insulin injections. action, and duration of action. The duration of action of
In those women with isolated fasting glucose elevations, a standard insulins and insulin analogues is outlined in Table
single nighttime injection of basal insulin may be adequate. 19.3. The pharmacokinetic and pharmacodynamic proper-
In women with both fasting and postprandial elevations, ties of RAIA result in twice the maximum concentration
a multiple insulin injection regimen is necessary. In these of insulin in approximately half the time compared with
cases, the typical starting dose would be 0.8–1.0 units/kg regular insulin (Figure 19.1), leading to less mean glucose
daily, in divided doses using both basal and prandial insu- excursion in response to a food bolus and less hypoglycemia
lins depending on gestational age. between meals. This superior drug profile allows maximum
Over the past few years, clinical experience with insu- flexibility and convenience in dosing resulting in greater
lin analogues in pregnancy has increased dramatically. patient satisfaction and improved quality of life.33
Table 19.3 Pharmacologic profiles of standard insulins and insulin analogues
Onset of action Peak action Duration of action

Standard
Regular 30–60 minutes 2–3 hours 8–10 hours
NPH 2–4 hours 4–10 hours 12–18 hours
Rapid-acting analogues
Lispro 5–15 minutes 30–90 minutes 4–6 hours
Aspart 5–15 minutes 30–90 minutes 4–6 hours
Glulisine 5–15 minutes 30–90 minutes 4–6 hours
Long-acting analogues
Glargine 2–4 hours None 20–24 hours
Detemir 3–4 hours None 20 hours
Aspart, lispro, glulisine
Regular
Plasma insulin levels

NPH
Detemir
Glargine
0 2 4 6 8 10 12 14 16 18 20 22 24
Hours
Figure 19.1 Pharmacodynamic profile of standard insulin and insulin analogues.
The RAIA include insulin lispro, insulin aspart, and insu- lower in women treated with insulin lispro.39 Yet the rate of
lin glulisine. Of the rapid-acting analogues, insulin lispro is increase in insulin between trimesters was similar between
the most studied in pregnancy. A limited number of studies groups. Predelivery hemoglobin A1c was improved in those
have also evaluated insulin aspart use in pregnancy, and no treated with insulin lispro compared with regular insulin
clinical studies to date have been published on insulin gluli- (5.9 ± 1.0 vs. 6.7 ± 1.3 mg/dL, p = 0.02). In contrast, a recent
sine in pregnancy. Due to the similar pharmacologic action, retrospective, multicenter Italian study reported a significant
the clinical experience related to transplacental passage, gly- improvement in hemoglobin A1c levels in the first-trimester
cemic efficacy, and perinatal outcomes would not likely to be women treated with insulin lispro compared with regular
different based on the specific rapid-acting analogue used to insulin (6.7% ± 1.3% vs. 7.3% ± 1.4%, p < 0.001), but similar
treat women with GDM. A1c levels in both treatment groups in the third trimester.40
Rates of severe hypoglycemic episodes were similar.
Maternal glycemia and perinatal outcomes Two recent studies have evaluated maternal obstetrical
In the first study of women with GDM treated with insu- outcomes for women with preexisting diabetes undergo-
lin lispro, areas under the curve for glucose, insulin, and ing lispro treatment in pregnancy. Both studies have shown
C-peptide were significantly lower in those women treated similar rates of preeclampsia, preterm birth, and cesarean
with insulin lispro compared to the regular insulin group, delivery in those treated with insulin lispro compared with
despite a similar hemoglobin A1c level.34 Women with GDM regular insulin.39,40 In regard to neonatal outcomes, rates
treated with insulin lispro experienced fewer hypoglycemic of shoulder dystocia, NICU admission, respiratory distress
episodes defined as blood glucose <55 mg/dL. In a retrospec- syndrome, and neonatal hypoglycemia were comparable.40
tive review of 53 women with GDM and 33 pregestational Insulin lispro has a great homology with insulin growth fac-
women, mean hemoglobin A1c during pregnancy was simi- tor 1 (IGF-1) raising concern for the possibility of accelerated
lar in those treated with insulin lispro compared to regu- fetal growth with in utero exposure to this RAIA. Table 19.4
lar insulin.35 However, when the 53 women with GDM were shows the rates of fetal overgrowth and mean maternal glyce-
evaluated separately, those treated with insulin lispro had mia as measured by hemoglobin A1c in pregnancies treated
lower hemoglobin A1c levels compared with those treated with insulin lispro, the majority of which are in women with
with regular insulin (5.25 ± 0.8 vs. 6.5 ± 2.0, p = 0.013). pregestational diabetes. Similar rates of macrosomia have
The majority of studies published on insulin lispro use been reported in pregnancies treated with insulin lispro
during pregnancy have focused on women with pregesta- and regular insulin.35,36,39–42 From the culmination of these
tional diabetes. These analyses have demonstrated conflict- reports, insulin lispro does not appear to be associated with
ing results with some studies showing no improvement in higher rates of fetal overgrowth than would be expected in
hemoglobin A1c with the use of insulin lispro36,37 and oth- diabetic pregnancy.
ers showing significantly lower predelivery A1c, without To date, only one study has evaluated anthropomet-
an increase in hypoglycemic episodes.33 The largest report ric characteristics of newborns exposed to insulin lispro in
of insulin lispro use in pregestational gravidas showed an utero.43 This study enrolled 49 pregnant women with GDM,
improvement in hemoglobin A1c values throughout preg- randomly assigned to treatment with insulin lispro or regu-
nancy. Hemoglobin A1c of women treated with insulin lis- lar insulin, and 50 pregnant women with a normal glucose
pro decreased from a mean of 8.9 ± 4.2 mg/dL in the first challenge test, matched for age, parity, prepregnancy weight,
trimester to 6.2 ± 2.4 mg/dL in the third trimester without and BMI. Rates of LGA and small for gestational age (SGA)
an increase in hypoglycemia compared to those treated with infants were similar between groups. However, the rate of
regular insulin.38 In a prospective observational study of infants with a cranial–thoracic circumference ratio between
58 pregestational women treated with insulin lispro and 49 the 10th and 25th percentile was significantly less in women
with regular insulin of similar age, BMI, and vascular com- treated with insulin lispro compared with regular insulin
plications, total insulin requirements in each trimester were (12% vs. 37.5%). This rate was comparable to the rate seen
Insulin therapy 153
Table 19.4 Maternal glycemia and rates of fetal overgrowth with insulin lispro
use in pregnancy
Author Year n Mean A1ca (mg/dL) Rate of LGA (%)

Aydin et al.35 2008 27 lispro 6.3 ± 2.2 7.4
59 regular 7.1 ± 2.1 13.6
Lapolla et al.40 2008 72 lispro 7.4 ± 1.7 55.1
298 regular 7.4 ± 1.5 39.2
Durnwald and Landon39 2008 58 lispro 7.1 ± 2.2 32.8
49 regular 8.3 ± 2.6 20.4
Cypryk et al.36 2004 25 lispro 7.8 ± 1.4 43.5
46 regular 7.5 ± 1.5 30.3
Masson et al.42 2003 71 lispro 7.4 ± 1.7 35
Garg et al.41 2003 61 lispro 7.2 ± 0.2 24
Persson et al.37 2002 16 lispro 6.5 No difference
17 regular 6.6
a First trimester or overall, if first trimester not reported.
in those with normal glucose tolerance (14%). In this study, The largest evaluation of insulin aspart use in pregnancy
1-hour postprandial glucoses in the lispro group were close to was an open-label, randomized, parallel group study of 322
the physiologic levels as demonstrated by women with normal women with type 1 diabetes conducted at 63 sites in 18 coun-
glucose tolerance. In the regular insulin group, 1-hour post- tries.46 Study entry required a hemoglobin A1c of ≤8% at
prandial glucoses were higher than both women with normal pregnancy confirmation. In this study, the mean of the dif-
glucose tolerance and women with lispro-treated GDM. The ference between preprandial and postprandial plasma glu-
level of glycemia may have contributed to the aberration in coses were lower for those women receiving insulin aspart,
fetal growth pattern seen in this group. Although interesting, despite comparable hemoglobin A1c levels. Despite similar
further investigation is needed on a larger scale before any mean total amount of insulin in both groups, mean daily
conclusions can be made whether the difference in anthro- requirement of prandial insulin was significantly lower in
pometry is an effect of treatment with insulin lispro. the insulin aspart group. There was no difference in hypo-
Pettitt and colleagues were the first to study the clinical glycemic episodes between groups. Maternal outcomes such
efficacy of insulin aspart compared to regular insulin in 15 as rates of preeclampsia, preterm labor, and cesarean section
women with GDM.44 In this study, peak insulin and glu- were similar between groups. In contrast, women treated
cose concentrations were measured after eating a breakfast with insulin aspart reported significantly greater satisfaction
meal for three consecutive days: the first when no insulin with their assigned treatment as measured by the Diabetes
was given, the second with a random assignment to either Treatment Satisfaction Questionnaire. In a second report,
regular insulin or insulin aspart, and the third when the 137 infants born to women treated with insulin aspart and
other insulin preparation was administered. Peak insulin 131 infants born to women treated with regular insulin were
concentrations were higher, and peak glucose concentra- evaluated. Rates of fetal overgrowth, neonatal hypoglycemia,
tions were lower with both insulins compared with no and perinatal mortality were similar between groups.47
insulin. Glycemic control, as measured by the glucose con-
centration under the curve above baseline, was significantly Transplacental passage
improved for insulin aspart compared with no insulin. In Transplacental passage of insulin can occur when insulin
a follow-up study of 27 women with GDM randomized to complexes with immunoglobulins to form an antigen–
treatment with insulin aspart versus regular insulin, there antibody complex. Previous studies have linked transpla-
was a greater reduction in average glucose values from cental passage of insulin complexes with fetal overgrowth.
baseline in those treated with apart indicative of improved Jovanovic first reported on the immunologic effects and pla-
postprandial glycemic control.45 There were more hypogly- cental transfer of insulin lispro in women with GDM.34 In
cemic episodes (71%) associated with aspart use, although this study, anti-insulin antibody levels were similar in both
no severe hypoglycemic event were noted. Treatment with groups, both at enrollment and delivery. Insulin lispro was
insulin aspart was also associated with significantly lower undetectable in the umbilical cord blood, including four
levels of C-peptide compared with regular insulin, reflect- women who received continuous intravenous administra-
ing a lower demand on beta cell function. This is likely tion of the drug during labor. There were no cases of fetal
attributed to the higher peak insulin concentrations that anomaly, macrosomia, or neonatal hypoglycemia. Similarly,
were achieved in this treatment group. Insulin-specific in a subset of 95 women enrolled in a randomized trial of
antibody binding was low in both groups (<1.5%). aspart and regular insulin, analysis of maternal and cord
blood insulin antibody levels demonstrated low levels of Four studies have evaluated women diagnosed with GDM
insulin-specific antibodies in both insulin groups at baseline treated with insulin glargine.52–55 All studies included a mixed
and delivery.48 In this analysis, insulin aspart was undetect- population of women with pregestational diabetes and GDM.
able in all cord blood samples evaluated. In a prospective cohort study of 82 women with GDM ran-
Two in vitro perfusion studies have evaluated the trans- domly assigned to either insulin glargine or NPH, the total
fer of insulin lispro across the human placenta.49,50 In the dose of basal insulin was less in those women treated with
first study, the placental transfer of a range of insulin lispro glargine. Fasting glucose levels in the 36th week of gesta-
concentrations was examined. The maternal side of the pla- tion were also lower in the glargine group. Rates of maternal
centa was perfused with constant concentrations of the drug hypoglycemia and hypertensive disorders of pregnancy were
and the fetal circulation was closed. No placental transfer of higher in those women treated with NPH. In regard to neona-
lispro was detected during perfusion with 100–200 μU/mL. tal outcomes, jaundice was more common in the NPH-treated
Although there was a possibility of transfer with concentra- group, but no differences were noted in the rate of macro-
tions mimicking a single subcutaneous dose of over 50 units somia or LGA, neonatal hypoglycemia, respiratory distress
that was maintained for greater than 60 minutes, there was syndrome, or 5-minute Apgar <7.52 In a smaller retrospec-
no placental transfer at single standard doses.49 Given the tive review of a mixed population of 112 women with preges-
half-life of insulin lispro, it is unlikely that this physiologic tational diabetes and GDM treated with glargine compared
state could be reproduced in the clinical setting, and there- with NPH, there was no significant difference in maternal
fore, insulin lispro does not appear to cross the placenta complications of preeclampsia, hypoglycemia, or cesarean
in any significant amount. The lack of transplacental pas- delivery between treatment groups in the 59 women with
sage was also confirmed on a smaller scale by Holcberg.50 GDM studied.53 Rate of LGA infants, Apgar scores, neona-
No in vitro perfusion studies have evaluated the transfer of tal hypoglycemia, and hyperbilirubinemia, as well as NICU
insulin aspart across the human placenta. Given the simi- admission, were also similar between groups. However, only
lar biochemical characteristics, it is likely that the findings 15 of the 59 women with GDM were treated with glargine.53 In
reported with insulin lispro would be similar to those of two additional retrospective reviews including a small num-
insulin aspart if such studies were conducted. ber of women with GDM, perinatal outcomes were similar in
women treated with NPH insulin and insulin glargine.54,56
Basal insulin Clinical experience with insulin detemir in pregnancy is
In the majority of women with GDM, the basal insulin pri- limited to case reports and one RCT of women with type 1
marily prescribed continues to be the intermediate acting diabetes.57–59 In this RCT, the efficacy of insulin detemir com-
insulin, NPH insulin. In most women, the greater insulin pared with NPH insulin (both with prandial insulin aspart)
resistance associated with GDM and subsequent higher was evaluated. Of 310 women with type 1 diabetes, 152 were
insulin requirements cannot be adequately treated with a randomly assigned to insulin detemir and 158 to NPH up
single dose of a long-acting insulin analogue (LAIA) such as to 1 year prior to pregnancy or during the 8th–12th week of
glargine or detemir, which are peakless. If LAIAs are used, pregnancy. Maternal hemoglobin A1c at the 36th week was
it is not uncommon for women with GDM to require twice similar between groups. However, fasting plasma glucoses
daily dosing. There is limited experience in pregnancy with were lower in women treated with insulin detemir at both
LAIA. Similar to the RAIA, the reports published focus on 24 weeks and 36 weeks of gestation. There was no difference
the treatment of women with preexisting diabetes, most spe- in maternal hypoglycemic episodes. In a follow-up study of
cifically type 1. the neonates born to the women in the RCT, neonatal out-
comes including congenital malformations, fetal growth
Long-acting insulin analogues disturbance (SGA and LGA), preterm delivery, and neonatal
The biochemical properties of the LAIA allows for a slow, con- hypoglycemia were similar between treatment groups.60
stant release of insulin without a pronounced peak in insulin There has been concern raised about the increased affin-
concentrations. The relatively flat activity profile is beneficial ity to IGF-1 that is demonstrated by insulin glargine com-
for women at significant risk for nocturnal hypoglycemia. One pared to human insulin with a potential to stimulate fetal
of the biggest advantages of the LAIA is less hypoglycemia. growth. A retrospective review in which 77% of the study
The once-daily dosing regimens also may enhance patient population was diagnosed with GDM evaluated the use of
acceptance and satisfaction. In studies of nonpregnant diabet- insulin glargine on select neonatal outcomes.55 This analysis
ics, insulin detemir has been shown to have more consistent reported a mean birth weight of 3142 ± 606 g for the study
insulin absorption compared with both insulin glargine and population, with a 2% incidence of macrosomia in GDM
NPH insulin.51 Compared with insulin glargine, detemir is women. In addition, none of the studies published to date
similarly peakless but exhibits a slightly shorter duration of demonstrate an increase in the rate of LGA infants or fetal
action and therefore is dosed every 12 hours in most diabetics. macrosomia with glargine treatment. In contrast to glargine,
insulin detemir has a low affinity for the IGF-1 receptor
Maternal glycemia and perinatal outcomes (approximately 1/10 that of human insulin). Although a
Studies evaluating the use of glargine in pregnancy have strictly GDM population has not be studied, the risks of fetal
primarily been limited to retrospective reviews and observa- overgrowth and macrosomia are not unique to abnormal
tional cohorts of women with type 1 diabetes. glucose metabolism of the GDM cohort.
References 155
Transplacental passage published data on the use of insulin detemir in GDM as well,
Results from human placental cotyledon models of uncom- the results are more convincing given the RCT of women
plicated term pregnancies have shown that there is no trans- with preexisting diabetes.
placental passage at maternal therapeutic concentrations of
insulin glargine.61 Transport across the placenta was dem-
onstrated at concentrations 1000-fold higher than clinically Conclusion
therapeutic levels, yet there was a significant difference in the
rate of disappearance from the maternal compartment com- Pharmacologic treatment of GDM should be designed to
pared with the rate of appearance in the fetal compartment. achieve euglycemia that has been linked to both maternal
This suggests that the placenta may sequester some amount and neonatal benefit. Diagnosis of dietary failure should be
of insulin glargine at concentrations at these high levels. prompt and initiation of pharmacologic therapy expedited
Published studies evaluating glargine use in pregnancy in those women unable to achieve glycemic targets on nutri-
are limited by small sample size, a mixed population of tion therapy alone. Both insulin therapy and oral hypoglyce-
GDM, and pregestational women studied and do not show mic agents are viable options. Treatment strategies should be
a consistent clinical benefit for the use of insulin glargine individualized based on a woman’s glycemic profile, clinical
in the pregnant women. Given these findings, there is little characteristics, and fetal growth projection. Ideally, thera-
justification to choose glargine over NPH for basal insulin pies should be aimed at lowering the demand on beta cell
for the treatment of GDM. Although there is also limited function during this time of physiologic stress.
REFERENCES
1. Committee on Practice Bulletins–Obstetrics. Practice Bulletin 16. Kahn BF, Davies JK, Lynch AM, Reynolds RM, Barbour LA.
No. 137: Gestational diabetes mellitus. Obstet Gynecol 2013; Predictors of glyburide failure in the treatment of gestational dia-
122(2 Pt. 1): 406–416. betes. Obstet Gynecol 2006; 107(6): 1303–1309.
2. American Diabetes Association. Standards of medical care in 17. Yogev Y, Melamed N, Chen R, Nassie D, Pardo J, Hod M.
diabetes–2014. Diabetes Care 2014; 37(Suppl. 1): S14–S80. Glyburide in gestational diabetes—Prediction of treatment fail-
3. Jovanovic-Peterson L, Peterson CM, Reed GF et al. Maternal ure. J Matern Fetal Neonatal Med 2011; 24(6): 842–846.
postprandial glucose levels and infant birth weight: The Diabetes 18. Yogev Y, Ben-Haroush A, Chen R, Rosenn B, Hod M, Langer O.
in Early Pregnancy Study. The National Institute of Child Health Undiagnosed asymptomatic hypoglycemia: Diet, insulin, and
and Human Development—Diabetes in Early Pregnancy Study. glyburide for gestational diabetic pregnancy. Obstet Gynecol
Am J Obstet Gynecol 1991; 164(1 Pt. 1): 103–111. 2004; 104(1): 88–93.
4. Langer O, Mazze R. The relationship between large-for-gesta- 19. Brustman L, Langer O, Scarpelli S, El Daouk M, Fuchs A, Rosenn
tional-age infants and glycemic control in women with gesta- B. Hypoglycemia in glyburide-treated gestational diabetes: Is it
tional diabetes. Am J Obstet Gynecol 1988; 159(6): 1478–1483. dose-dependent? Obstet Gynecol 2011; 117(2 Pt. 1): 349–353.
5. Landon MB, Gabbe SG, Piana R, Mennuti MT, Main EK. 20. Jacobson GF, Ramos GA, Ching JY, Kirby RS, Ferrara A, Field
Neonatal morbidity in pregnancy complicated by diabetes mel- DR. Comparison of glyburide and insulin for the management of
litus: Predictive value of maternal glycemic profiles. Am J Obstet gestational diabetes in a large managed care organization. Am J
Gynecol 1987; 156(5): 1089–1095. Obstet Gynecol 2005; 193(1): 118–124.
6. Landon MB, Spong CY, Thom E et al. A multicenter, random- 21. Elliott BD, Langer O, Schenker S, Johnson RF. Insignificant trans-
ized trial of treatment for mild gestational diabetes. N Engl J Med fer of glyburide occurs across the human placenta. Am J Obstet
2009; 361(14): 1339–1348. Gynecol 1991; 165(4 Pt. 1): 807–812.
7. McFarland MB, Langer O, Conway DL, Berkus MD. Dietary ther- 22. Elliott BD, Schenker S, Langer O, Johnson R, Prihoda T.
apy for gestational diabetes: How long is long enough? Obstet Comparative placental transport of oral hypoglycemic agents in
Gynecol 1999; 93(6): 978–982. humans: A model of human placental drug transfer. Am J Obstet
8. Nicholson WK, Wilson LM, Witkop CT et al. Therapeutic man- Gynecol 1994; 171(3): 653–660.
agement, delivery, and postpartum risk assessment and screening 23. Kraemer J, Klein J, Lubetsky A, Koren G. Perfusion studies of gly-
in gestational diabetes. Evidence Report/Technology Assessment buride transfer across the human placenta: Implications for fetal
No. 162, 2008, pp. 1–96. safety. Am J Obstet Gynecol 2006; 195(1): 270–274.
9. Goetzl L, Wilkins I. Glyburide compared to insulin for the treat- 24. Hebert MF, Ma X, Naraharisetti SB et al. Are we optimizing ges-
ment of gestational diabetes mellitus: A cost analysis. J Perinatol tational diabetes treatment with glyburide? The pharmacologic
2002; 22(5): 403–406. basis for better clinical practice. Clin Pharmacol Ther 2009; 85(6):
10. Caritis SN, Hebert MF. A pharmacologic approach to the 607–614.
use of glyburide in pregnancy. Obstet Gynecol 2013; 121(6): 25. Rowan JA, Hague WM, Gao W, Battin MR, Moore MP. Metformin
1309–1312. versus insulin for the treatment of gestational diabetes. N Engl J
11. Langer O, Conway DL, Berkus MD, Xenakis EM, Gonzales O. Med 2008; 358(19): 2003–2015.
A comparison of glyburide and insulin in women with gestational 26. Ijas H, Vaarasmaki M, Morin-Papunen L et al. Metformin should
diabetes mellitus. N Engl J Med 2000; 343(16): 1134–1138. be considered in the treatment of gestational diabetes: A pro-
12. Langer O, Yogev Y, Xenakis EM, Rosenn B. Insulin and gly- spective randomised study. BJOG 2011; 118(7): 880–885.
buride therapy: Dosage, severity level of gestational diabetes, 27. Niromanesh S, Alavi A, Sharbaf FR, Amjadi N, Moosavi S, Akbari
and pregnancy outcome. Am J Obstet Gynecol 2005; 192(1): S. Metformin compared with insulin in the management of gesta-
134–139. tional diabetes mellitus: A randomized clinical trial. Diabetes Res
13. Chmait R, Dinise T, Moore T. Prospective observational study to Clin Pract 2012; 98(3): 422–429.
establish predictors of glyburide success in women with gesta- 28. Spaulonci CP, Bernardes LS, Trindade TC, Zugaib M, Francisco RP.
tional diabetes mellitus. J Perinatol 2004; 24(10): 617–622. Randomized trial of metformin vs insulin in the management of ges-
14. Conway DL, Gonzales O, Skiver D. Use of glyburide for the tational diabetes. Am J Obstet Gynecol 2013; 209(1): 34.e1–34.e7.
treatment of gestational diabetes: The San Antonio experience. 29. Tertti K, Ekblad U, Koskinen P, Vahlberg T, Ronnemaa T.
J Matern-Fetal Neonatal Med 2004; 15(1): 51–55. Metformin vs. insulin in gestational diabetes. A randomized study
15. Kremer CJ, Duff P. Glyburide for the treatment of gestational dia- characterizing metformin patients needing additional insulin.
betes. Am J Obstet Gynecol 2004; 190(5): 1438–1439. Diabetes Obes Metab 2013; 15(3): 246–251.
30. Vanky E, Zahlsen K, Spigset O, Carlsen SM. Placental passage of 46. Mathiesen ER, Kinsley B, Amiel SA et al. Maternal glycemic con-
metformin in women with polycystic ovary syndrome. Fertil Steril trol and hypoglycemia in type 1 diabetic pregnancy: A random-
2005; 83(5): 1575–1578. ized trial of insulin aspart versus human insulin in 322 pregnant
31. Moore LE, Clokey D, Rappaport VJ, Curet LB. Metformin com- women. Diabetes Care 2007; 30(4): 771–776.
pared with glyburide in gestational diabetes: A randomized con- 47. Hod M, Damm P, Kaaja R et al. Fetal and perinatal outcomes
trolled trial. Obstet Gynecol 2010; 115(1): 55–59. in type 1 diabetes pregnancy: A randomized study comparing
32. Silva JC, Fachin DR, Coral ML, Bertini AM. Perinatal impact of insulin aspart with human insulin in 322 subjects. Am J Obstet
the use of metformin and glyburide for the treatment of gesta- Gynecol 2008; 198(2): 186.e1–186.e7.
tional diabetes mellitus. J Perinat Med 2012; 40(3): 225–228. 48. McCance DR, Damm P, Mathiesen ER et al. Evaluation of insu-
33. Bhattacharyya A, Brown S, Hughes S, Vice PA. Insulin lispro and lin antibodies and placental transfer of insulin aspart in pregnant
regular insulin in pregnancy. Q J Med 2001; 94(5): 255–260. women with type 1 diabetes mellitus. Diabetologia 2008; 51(11):
34. Jovanovic L, Ilic S, Pettitt DJ et al. Metabolic and immunologic 2141–2143.
effects of insulin lispro in gestational diabetes. Diabetes Care 49. Boskovic R, Feig DS, Derewlany L, Knie B, Portnoi G, Koren G.
1999; 22(9): 1422–1427. Transfer of insulin lispro across the human placenta: In vitro per-
35. Aydin Y, Berker D, Direktor N et al. Is insulin lispro safe in preg- fusion studies. Diabetes Care 2003; 26(5): 1390–1394.
nant women: Does it cause any adverse outcomes on infants or 50. Holcberg G, Tsadkin-Tamir M, Sapir O et al. Transfer of insulin
mothers? Diabetes Res Clin Pract 2008; 80(3): 444–448. lispro across the human placenta. Eur J Obstet Gynecol Reprod
36. Cypryk K, Sobczak M, Pertynska-Marczewska M et al. Pregnancy Biol 2004; 115(1): 117–118.
complications and perinatal outcome in diabetic women treated 51. Heise T, Nosek L, Ronn BB et al. Lower within-subject variability of
with Humalog (insulin lispro) or regular human insulin during insulin detemir in comparison to NPH insulin and insulin glargine
pregnancy. Med Sci Monit 2004; 10(2): PI29–PI32. in people with type 1 diabetes. Diabetes 2004; 53(6): 1614–1620.
37. Persson B, Swahn ML, Hjertberg R et al. Insulin lispro therapy in 52. Negrato CA, Rafacho A, Negrato G et al. Glargine vs. NPH insu-
pregnancies complicated by type 1 diabetes mellitus. Diabetes lin therapy in pregnancies complicated by diabetes: An observa-
Res Clin Pract 2002; 58(2): 115–121. tional cohort study. Diabetes Res Clin Pract 2010; 89(1): 46–51.
38. Wyatt JW, Frias JL, Hoyme HE et al. Congenital anomaly rate in 53. Fang YM, MacKeen D, Egan JF, Zelop CM. Insulin glargine com-
offspring of mothers with diabetes treated with insulin lispro dur- pared with Neutral Protamine Hagedorn insulin in the treatment
ing pregnancy. Diabet Med 2005; 22(6): 803–807. of pregnant diabetics. J Matern Fetal Neonatal Med 2009; 22(3):
39. Durnwald CP, Landon MB. A comparison of lispro and regular 249–253.
insulin for the management of type 1 and type 2 diabetes in preg- 54. Egerman RS, Ramsey RD, Kao LW et al. Perinatal outcomes
nancy. J Matern Fetal Neonatal Med 2008; 21(5): 309–313. in pregnancies managed with antenatal insulin glargine. Am J
40. Lapolla A, Dalfra MG, Spezia R et al. Outcome of pregnancy Perinatol 2009; 26(8): 591–595.
in type 1 diabetic patients treated with insulin lispro or regular 55. Henderson CE, Machupalli S, Marcano-Vasquez H, Kerr P, Reilly
insulin: An Italian experience. Acta Diabetol 2008; 45(1): 61–66. KD. A retrospective review of glargine use in pregnancy. J Reprod
41. Garg SK, Frias JP, Anil S, Gottlieb PA, MacKenzie T, Jackson Med 2009; 54(4): 208–210.
WE. Insulin lispro therapy in pregnancies complicated by type 56. Smith JG, Manuck TA, White J, Merrill DC. Insulin glargine versus
1 diabetes: Glycemic control and maternal and fetal outcomes. neutral protamine Hagedorn insulin for treatment of diabetes in
Endocr Pract 2003; 9(3): 187–193. pregnancy. Am J Perinatol 2009; 26(1): 57–62.
42. Masson EA, Patmore JE, Brash PD et al. Pregnancy outcome in 57. Lapolla A, Di Cianni G, Bruttomesso D et al. Use of insulin
Type 1 diabetes mellitus treated with insulin lispro (Humalog). detemir in pregnancy: A report on 10 Type 1 diabetic women.
Diabet Med 2003; 20(1): 46–50. Diabet Med 2009; 26(11): 1181–1182.
43. Mecacci F, Carignani L, Cioni R et al. Maternal metabolic con- 58. Sciacca L, Marotta V, Insalaco F et al. Use of insulin detemir dur-
trol and perinatal outcome in women with gestational diabetes ing pregnancy. Nutr Metab Cardiovasc Dis 2010; 20(4): e15–e16.
treated with regular or lispro insulin: Comparison with non- 59. Mathiesen ER, Hod M, Ivanisevic M et al. Maternal efficacy and
diabetic pregnant women. Eur J Obstet Gynecol Reprod Biol safety outcomes in a randomized, controlled trial comparing
2003; 111(1): 19–24. insulin detemir with NPH insulin in 310 pregnant women with
44. Pettitt DJ, Ospina P, Kolaczynski JW, Jovanovic L. Comparison of type 1 diabetes. Diabetes Care 2012; 35(10): 2012–2017.
an insulin analog, insulin aspart, and regular human insulin with 60. Hod M, Mathiesen ER, Jovanovic L et al. A randomized trial com-
no insulin in gestational diabetes mellitus. Diabetes Care 2003; paring perinatal outcomes using insulin detemir or neutral prot-
26(1): 183–186. amine Hagedorn in type 1 diabetes. J Matern Fetal Neonatal Med
45. Pettitt DJ, Ospina P, Howard C, Zisser H, Jovanovic L. Efficacy, 2014; 27(1): 7–13.
safety and lack of immunogenicity of insulin aspart compared 61. Pollex EK, Feig DS, Lubetsky A, Yip PM, Koren G. Insulin glargine
with regular human insulin for women with gestational diabetes safety in pregnancy: A transplacental transfer study. Diabetes
mellitus. Diabet Med 2007; 24(10): 1129–1135. Care 2010; 33(1): 29–33.
20 Gestational diabetes
mellitus: The consequences
of not treating
Oded Langer
In the current era of evidence-based medicine, it is sur-

GDM dilemma prising that the opposing positions are not the result of
It is self-evident that women with type 1 and type 2 poorly authors’ researched data but based upon opinions lack-
controlled diabetes have a significant increase in adverse ing evidence to support these positions. In order to deter-
perinatal outcome in comparison to nondiabetic patients. mine the research-driven answer to this dilemma, the most
The condition is especially detrimental when the targeted appropriate means would be to conduct randomized clini-
level of glycemia is not achieved. It is also realistic to assume cal trials (RCTs). However, in some cases, the RCT study
that similar adverse outcomes will result in gestational dia- design faces an ethical dilemma so an alternative design is
betes mellitus (GDM) since it is a looking-glass image of mandated.
type 2 diabetes. Since the late 1960s when O’Sullivan first GDM is defined as carbohydrate intolerance of variable
suggested the term “gestational diabetes,” an argumenta- severity with onset or first recognition during pregnancy.
tive environment has persistently surrounded this clinical Therefore, some of the women already have impaired glu-
entity and its association to adverse pregnancy outcome, i.e., cose tolerance (IGT) test and/or impaired fasting glucose.
macrosomia, birth trauma, and metabolic complications. The definition applies regardless of treatment modal-
However, opinions and anecdotes have surpassed research- ity and/or the persistence of the condition after pregnancy.
generated data. While it is likely that maternal carbohydrate intolerance
To date, there is no consensus regarding diagnostic crite- reflects a continuum of risk for adverse outcomes,10,11 it is
ria and the utility of universal screening for GDM. Lack of not known whether there is a benefit to identification and
accord on these issues have led to inconsistencies in results subsequent treatment of mild levels of severe carbohydrate
of published studies.1,2 For example, Jarrett concluded that intolerance during pregnancy.
GDM is “a nonentity” whose only clinical association is with Since GDM is considered the major medical complica-
an increased maternal risk of subsequent diabetes.3 Several tion in pregnancy, empirical attempts were made to address
studies and noted forums have stated that there is as yet no the controversy. However, the results of many studies are
consensus on the definition, management, or treatment of derived from administrative databases, i.e., billing systems.
GDM.4 Hunter and Milner stated “gestational diabetes is a The data should not be used for clinical studies since they
diagnosis still looking for a disease.” According to these phy- may lead to erroneous conclusions.12
sicians, GDM is not convincingly associated with increased The conclusions from other studies are the result of obser-
perinatal mortality or morbidity, and macrosomia per se, vational design. Most reported associations in observational
regardless of definition, is not a morbid condition.5 research are false, and the minority of associations that are
In contrast, Beard and colleagues6 concluded that GDM true are often exaggerated. This issue is problematic for weak
is a clinical entity associated with a significant incidence in associations, often defined as relative risks (RRs) or odds
diabetes in the later life of the mother and an increase in ratios (ORs). These relationships can usually be ascribed to
fetal and neonatal morbidity. In recent years, several stud- bias rather than to an underlying association. A prudent rule
ies demonstrated the association of GDM with perinatal to follow in statistical analysis is to assume that unless RRs
morbidity and mortality. The results of these studies dem- in cohort studies exceed 2–3 or ORs in case–control studies
onstrated that GDM is not only a clinical entity but a disease exceed 3 or 4, associations in observational study findings
with short- and long-term complications.7–9 should not be considered credible.13
157
158 Gestational diabetes mellitus
Change from physiology to 50% of subjects with normal glucose tolerance and previous
GDM. This was primarily due to decreased nonoxidative
pathophysiology glucose disposal. Ward et al., 31 using the Bergman intrave-
nous glucose minimal model technique, also found evidence
The first condition in order to establish a medical entity as a
of increased insulin resistance in women with previous
disease is to demonstrate change from normal to pathophysi-
GDM. Unfortunately, neither Catalano nor Ward controlled
ology. This will facilitate the development of interventions
for family history in the subjects, so it is unclear whether
aimed at prevention as well as treatment. Human pregnancy
family history or previous GDM was the marker for these
is characterized as an insulin-resistant condition. Although
early abnormalities in glucose metabolism.
there is a fourfold to fivefold range of insulin resistance in the
In summary, GDM is characterized by pathogenesis
general population, there is a relatively uniform 40%–50%
deviating from the normal physiology of pregnancy that
increase (from the pregravid condition) in insulin resistance
involves insulin resistance and decreased insulin secretion.
and increase in insulin secretion. A substantial number of
Furthermore, similarity exists between the pathogenesis of
studies have addressed this topic.14–19
GDM and type 2 diabetes, which are probably one disease at
Another factor which may contribute to increased insulin
different stages on the spectrum of glucose intolerance.
resistance during pregnancy is the rise in body fat or change
in body fat distribution that begins in early gestation. It is
well established that obesity per se causes insulin resis-
tance.20 It has been shown that total oxidative and nonoxi- Is there an associated increased
dative glucose metabolism is inversely related to increased adverse outcome in GDM?
visceral-to-subcutaneous-fat ratio in obese women and to
total fat content in lean women.21 Others have demonstrated Socrates once said that the beginning of wisdom is the defi-
decreased insulin sensitivity in subjects with a central pat- nition of terms. However, using different definitions for
tern of fat distribution. Whatever the cause for increased GDM diagnosis criteria or threshold for neonatal hypogly-
insulin resistance during pregnancy in women who main- cemia will result in varying rates of the problem. Not only
tain normal glucose tolerance, it is offset by a 3- to 3.5-fold different definitions but also diverse institutional policies
increase in insulin secretion.22 Therefore, the degree of insu- such as indication for neonatal intensive care unit (NICU)
lin resistance during late gestation appears to be dependent admission and for cesarean delivery will result in varying
primarily on pregravid maternal insulin resistance, which is rates reported in different studies.
quite variable, and secondarily on the 40%–50% increases Small sample sizes have resulted in the evolution of the
mediated through placental factors. concept of composite outcomes (Tables 20.1 through 20.3).
In 1981, Bergman et al.23 proposed that there is a predict- Composite outcomes are pooled individual outcomes
able relationship in the shape of a rectangular hyperbola (serious morbidities) in order to produce a single outcome
between the quantity of insulin produced by β-cells and (overall serious morbidity). As treatments and population
the sensitivity of tissues to the glucose-lowering effects of health improve, the number of individual adverse outcomes
insulin. Kahn et al.24 demonstrated that this relationship is has declined. Using composite outcomes can often over-
present across a wide range of insulin sensitivity in people come this reality by combining several outcomes, thereby
with normal glucose tolerance. A hyperbolic relationship increasing the efficiency of a clinical trial (power and size).
exists between insulin sensitivity and several measures of Outcome selection has clinical significance; however, when
β-cell insulin release in women with IGT.25 The amount of evaluating composite outcomes in various studies, it is

insulin released at a given level of insulin resistance is lower
in people with abnormal compared with normal glucose
tolerance.26,27 Although GDM is defined as carbohydrate Table 20.1 Sample sizes for various effects in the
intolerance of variable severity with onset or first recogni- primary outcome (composite)
tion during pregnancy,28 the definition applies whether the
condition persists after pregnancy but does not preclude the 30% reduction 40% reduction
possibility that the glucose intolerance may have predated
Outcome Outcome Outcome
the pregnancy. Women who develop GDM are, in fact, rate in rate in rate in
genetically predisposed to develop type 2 diabetes since they untreated treated Sample treated Sample
tend to have a strong family history of the disease. Some group (%) group (%) size group (%) size
stressors associated with pregnancy probably trigger them 20 14.0 611 12.0 325
to develop overt disease sooner than if they had not become 25 17.5 537 15.0 248
pregnant.
30 21.0 366 18.0 196
Similar to other groups genetically susceptible to type 2
35 24.5 295 21.0 159
diabetes, women with previous GDM and normal glucose
40 28.0 243 24.0 132
tolerance have been shown to have defects in glucose metab-
45 31.5 218 27.0 110
olism. Using the hyperinsulinemic–euglycemic clamp tech-
nique, Catalano et al.29,30 demonstrated insulin resistance in 50 35.0 169 30.0 93
Is there an associated increased adverse outcome in GDM? 159
Table 20.2 Prevalence of hyperbilirubinemia
Definition Authors Population n Outcome rate (%)

≥10 mg/dL Gabbe et al. GDM treated 261 15.7
>12 mg/dL Hod et al. GDM treated 731 5.2
>12 mg/dL Magee et al. GDM 96 15.6
Normal 521 10.6
≥12 mg/dL Langer et al. GDM intense tx 1145 7.6
GDM usual tx 1316 7.9
Not defined Garner et al. GDM treated 149 5.4
GDM untreated 150 6.6
Requiring treatment Persson et al. Mild GDM treated 233 4.5
>12 mg/dL Langer et al. GDM untreated 555 11.0
GDM treated 1110 9.0
No GDM 1110 4.0
>95th percentile Landon GDM untreated 473 12.9
Not available (n.a.) Crowter GDM untreated 524 n.a.
Table 20.3 Prevalence of neonatal hypoglycemia
Outcome rate
Definition Authors Population n (%)
<30 mg/dL if <2500 g Gabbe et al. GDM treated 261 6.5
<20 mg/dL if 2500 g
<30 mg/dL up to 72 hours; Hod et al. GDM treated 731 5.2
<40 mg/dL later
≤30 mg/dL × 2 <4 hours after birth Langer et al. GDM intense tx 1145 3.8
GDM conventional 1316 20.0
<31 mg/dL <2 hours after birth Jang et al. GDM treated 65 7.7
Symptomatic Persson et al. Mild GDM treated 233 0.9
GDM treated 222 2.3
<30 mg/dL <1 hour after birth Schader-Graf et al. GDM treated 154 40.3
40 mg/dL × 2 <4 hours after birth Langer et al. GDM treated with glyburide 201 9.0
GDM treated with insulin 203 5.9
<40 mg/dL Langer et al. GDM untreated 555 7.0
GDM treated 1110 14.0
No GDM 1110 2.0
Hypoglycemia requiring I.V. therapy Crowther et al. GDM untreated 506 5.0
<35 mg/dL Landon et al. Mild GDM untreated 473 15.0
prudent to verify that the variables included in each compos- should not be used as an outcome measure in GDM because
ite outcome are identical. the procedure is not directly related to disease morbidity.
In the compilation of an outcome variable, only those that Delivery through c/s has become a self-fulfilling prophecy
are a result of GDM pathophysiology should be included; in the treatment of GDM. Naylor et al.33 reported that the c/s
variables that reflect physician behavior, i.e., cesarean sec- rate was 34% in women with GDM compared with 20% in
tion (c/s) delivery, misrepresent the conclusion. Although women without GDM. The National Institute of Child Health
associated with maternal morbidity (e.g., infection, bleed- and Human Development (NICHD) study 9 resulted in 27%
ing), c/s is a physician-driven decision rather than a com- c/s rate in the treated versus 32% in the untreated group, while
plication of the disease. Rates of c/s have skyrocketed with the Australian Carbohydrate Intolerance Study in Pregnant
more repeat c/s’s,32 c/s by demand, and elective c/s for breech Women (ACHOIS)8 c/s data revealed 31% in the treated and
delivery. Although this procedure has outstripped historical 32% in the untreated group. What these numbers indicate
numbers for both diabetic and nondiabetic patients, c/s rates is that inherent in the study of a treatable entity is that this
new knowledge may alter the clinician’s behavior.33 On the drawbacks of this study included the retrospective design
other hand, there are care providers who have not inevita- and the lack of distinction between types of diabetes since
bly been influenced by the GDM diagnosis and have instead the results were derived from an administrative database.
evaluated the presenting conditions of the disease.34,35 In our However, it would be reasonable to assume that the majority
study,7 the c/s rate was similar for the untreated (24%) and of the diabetic patients had GDM that accounts for 90% of
the treated (23%) compared to 14% in nondiabetic subjects. all diabetic pregnancies.
Clinicians must, therefore, consider the merits of estab-
lishing a GDM diagnosis. GDM, if untreated or not rec-
Neonatal complications ognized, may be associated with an increased risk of
intrauterine fetal death and commonly reported morbidi-
The adverse outcomes most commonly associated with GDM ties such as macrosomia, birth trauma, neonatal hypoglyce-
include increased perinatal mortality, macrosomia, shoulder mia, hyperbilirubinemia, hypoglycemia, and polycythemia.
dystocia, birth trauma, preeclampsia, c/s, neonatal hypo- There is paucity of prospective data concerning some of
glycemia, hypocalcemia, hyperbilirubinemia, and polycy- these risks. However, it is generally agreed that women with
themia. In addition, there are long-term effects associated GDM with significantly elevated fasting blood glucose levels
with GDM pregnancies such as an increased maternal risk appear to have an increased risk of intrauterine fetal death.
of developing diabetes in the future and an increased risk of
obesity and glucose intolerance in the offspring.
Perinatal mortality is the most significant outcome, and
early, albeit flawed studies showed a fourfold increase in Macrosomia, shoulder dystocia,
perinatal mortality in women with GDM. These studies
and birth trauma
failed to control for variables affecting perinatal mortality
such as fetal malformations, maternal history of stillbirth, Being relatively common and easily documented, macrosomia
and advanced maternal age. Furthermore, these studies more is the adverse perinatal outcome most investigators refer to
than likely included women with unrecognized pregestational when addressing GDM. Macrosomia is the primary outcome
diabetes, thus contaminating the results. In addition, in most with relevant surrogate complications such as c/s, shoulder dys-
studies a labeling bias exists since a GDM diagnosis tends to tocia, and brachial plexus injury. The overall rate of macroso-
enhance surveillance and interventions that may have a major mia for the nondiabetic population is 7%–9%.28,41,42 In contrast,
impact on perinatal mortality. Some researchers have sug- the incidence reported for macrosomia in GDM is manage-
gested that GDM has no or negligible effect on mortality. This ment dependent. When good glycemic control is not achieved,
could be explained by two opposing views: GDM has no or the incidence of macrosomia can be as high as 20%–45%.35 The
negligible effect on mortality, or due to the overall decrease in macrosomic fetus is a result of diabetic fetopathy and is char-
perinatal mortality, excess fetal deaths due to unrecognized acterized by organomegaly.43,44 Complications, directly and
GDM could go unnoticed in smaller studies. indirectly associated to fetal macrosomia are neonatal hypo-
O’Sullivan and Mahan first reported an association glycemia, hypocalcemia, hyperbilirubinemia, and polycythe-
between GDM and perinatal death, documenting a 6.4% risk mia; in addition to birth trauma, these are all the consequence
only in women with GDM who were older than 25 years of of not treating or inadequate treatment of the disease.
age, and a RR of 4.3 over controls.36 Abell et al.37 reported Excessive fetal growth occurs in as many as 50% of preg-
similar results in women with GDM with a 3.9% overall peri- nancies complicated by GDM.45 It was shown that the accel-
natal mortality rate. However, an analysis of 1016 GDM preg- erated fetal growth is associated with the maternal glycemic
nancies from the same institution documented an increased profile.46,47 Although fetal growth can be measured by birth
perinatal mortality rate (3.2%) only among those meeting the weight, a more accurate way to characterize overgrowth is
stringent World Health Organization (WHO) criteria for the by estimation of body composition that includes lean body
diagnosis of GDM.38 Schmidt et al.39 evaluated the relation mass (LBM) and fat-free mass (FFM). LBM is a metaboli-
between the American Diabetes Association (ADA) and the cally active tissue and is relatively stable in utero. FFM is
WHO diagnostic criteria for GDM against pregnancy out- more variable and sensitive to factors that affect fetal growth.
come. Of the 4977 women in the study, 2.6% had GDM by the Therefore, to more accurately characterize the diabetic fetop-
ADA criteria and 7.2% by the WHO criteria. The perinatal athy, measurements that can identify even minimal devia-
death in the ADA group had an OR of 3.10, 95% C.I., 1.42– tions from the norm are needed. FFM and LBM may provide
6.47. Similarly, the perinatal mortality by the WHO criteria the means. Recent studies have shown conflicting results in
had an OR of 1.59, 95% C.I., 0.86–2.90 (nonsignificant). the evaluation of infant body composition.43,48,49 Catalano
Mondestin et al.40 reported the results of a retrospec- reported increased FFM in infants of women with GDM
tive cohort study of U.S. data (1995–1997). These included even when average weight for gestational age were compa-
10 million nondiabetic gravidas and 271,691 diabetic patients rable to infants of women with normal glucose tolerance.48
with fetal death rates of 4/1000 and 5.9/1000, respectively. Similarly, he demonstrated increased body fat in infants of
Fetal death rates increased when birth weight was >4250 g women with GDM requiring a cesarean delivery compared
for nondiabetic and 4000 g for diabetic patients with a two- with normal glucose tolerance despite similar birth weights.
fold increased rate in mortality in the diabetic group. The In contrast, we and Naeye43,44 found an increase in LBM at
Clinical studies 161
the time of autopsy in overgrown infants of women with dia- Motor Proficiency at ages 6, 8, and 9 years. They reported
betes. In a study evaluating body composition of macroso- that the children’s average scores on the Bruininks–
mic infants of diabetic women, we demonstrated increased Oseretsky test at ages 6–9 years correlated significantly with
body fat and decreased LBM in infants of women with GDM B-hydroxybutyrate in maternal second and third trimes-
when compared to women with normal glucose tolerance.50 ters. There was also a borderline association between chil-
Long-term effects of GDM. When addressing the issue dren’s scores on the psychomotor development index at age
of the long-term effects of GDM, one must differentiate 2 and B-hydroxybutyrate. Rizzo et al.59 correlated measures
between the long-term maternal effects and the prognosis of maternal glucose and lipid metabolism (fasting plasma
for the offspring. glucose levels, hemoglobin A1C levels, episodes of hypo-
glycemia and acetonuria, and plasma β-hydroxybutyrate
and free fatty acid levels) with two measures of intellectual
The mother development in the o ffspring using the Bayley Scales of
Infant Development for 2-year olds and the Stanford–Binet
The increased risk of developing type 2 diabetes later in life Intelligence Scale for 3- to 5-year olds expressed as an aver-
in women with GDM is well known with the magnitude of age of the three scores. The children’s mental development
the risk ranging from 20% to 50%; it is lower in Caucasians index scores at the age of two correlated inversely with the
and higher in Hispanic women, those of Mediterranean or mother’s third-trimester plasma β-hydroxybutyrate levels;
East Asian descent and the Canadian Aboriginal popula- the average Stanford–Binet Intelligence Scale scores corre-
tion. There is no evidence that the treatment of GDM will lated inversely with third-trimester plasma β-hydroxybutyrate
reduce this risk although recent data51 suggest that the rate and free fatty acid levels.
of progression to type 2 diabetes can be modified by lifestyle Maternal diabetes during pregnancy may affect behav-
changes, thus underscoring a possible benefit for increased ioral and intellectual development in the offspring. Petersen
surveillance of this high-risk population. et al. suggested that first trimester intrauterine growth delay
is associated with psychomotor deficit in the offspring at age
4–5 probably as a result of poor glycemic control.60,69 Sells
The neonate et al. reported that late entry into treatment programs in preg-
nancy in preexisting diabetic women resulted in lower scores
Since Barker’s primary epidemiologic studies in 1989,52,53 on language measures and intellectual development through
there is increasingly strong data to support the concept that age 2 in comparison to women who maintained good glu-
many fetal stresses may lead to fetal programming and the cose control during pregnancy.63 Stenninger et al. reported
alteration of the normal developmental gene expression pat- that children born to mothers with diabetes (probably
tern. Research indicates that the child of the diabetic mother GDM) who subsequently developed neonatal hypoglycemia
remains at increased risk for a variety of developmental dis- experienced long-term neurological dysfunction. Offspring
turbances: obesity,54–56 IGT or diabetes,57 and diminished had more difficulties in validated screening tests for mini-
neurobehavioral capacities.58–64 Therefore, it would be rea- mal brain dysfunction and were hyperactive, impulsive, and
sonable to speculate that the process whereby a stimulus or easily distracted. In addition, they had a lower developmen-
insult (glucose toxicity and other metabolic fuels) acting at a tal score in comparison to the offspring of normoglycemic
critical period of development in early and during intrauter- diabetic women and nondiabetic control patients.70 Olsson
ine life alters gene expression patterns for life.65,66 et al. concluded that impaired cognitive function may pre-
cede clinical onset of type 2 diabetes.71 In summary, the
existing evidence indicates that poorer glycemic control or
Cognitive development in children of other shared risk factors may influence both cognitive devel-
diabetic mothers opment and the risk for type 2 diabetes. The data suggest that
early detection of subclinical disease and treatment may be
There is paucity of data related to the cognitive development of value in protecting against cognitive deficits and adverse
of infants of diabetic mothers. A study of infants born to neurological outcomes in addition to the possibility of early
73 women with preexisting diabetes and 112 women with development of metabolic syndrome (hypertension, obesity,
GDM sought to determine the relationship between mater- and diabetes) when GDM is not treated or poorly managed.
nal fasting plasma glucose and hemoglobin A1C during
the second and third trimesters on neonatal performance
on the Brazelton Neonatal Behavioral Assessment Scale.58 Clinical studies
They found a significant correlation between poor glycemic
control in three out of the four newborn behavioral dimen- Care providers and their institutions repeatedly find them-
sions on the scale for both gestational and pregestational selves compared to one another as well as to national/
diabetics. In another study,67,68 they evaluated the offspring universal standards. Inherent in these comparisons is that
of 95 women with preexisting diabetes and 101 subjects providers with improved outcomes presumably provide
with GDM using the Bayley Scales of Infant Development higher-quality care. Nonetheless, there are three major
at 2 years of age and the Bruininks–Oseretsky Test of sources of distinction in outcomes: preexisting patient
characteristics, quality of care, and random variation. These

Table 20.4 The rate of large infants in untreated
factors, in addition to others, will enable the classification of
gestational diabetes
GDM as a clinical entity.
There has historically been two phases in which this
Macrosomia LGA (%)
issue has been addressed. In the past, there has been rela-
(%) (≥90th
tively scant data; shortcomings in the few existing studies (≥4000 g) percentile) Subjects
included small sample sizes and, therefore, insufficient power
(increased likelihood for types 1 and 2 errors), as well as lack O’Sullivan et al.36 15 30 n = 259
of controls for the confounding effects of ethnicity, maternal Coustan et al.91 41 47 n = 34
obesity, age, and parity. In evaluating the results of a study, Coetzee et al.92 32 55 n = 22
it is prudent to consider the sample size and the power des- Li et al.77 7 29 n = 73
ignated for the study. Power is the ability to detect a differ- Hod et al.93 22 34 n = 32
ence when one really exists; power increases as the sample Langer et al.7 17 29 n = 555
size increases. It also precludes a type 2 error (false-negative); Sermer et al.94 — 35 n = 150
power of 80% is usually acceptable. However, a very small dif- Adams et al.75 44 44 n = 16
ference can achieve statistical significance if the sample size is Ostlund et al.73 33 25 n = 211
large enough. On the other hand, a large effect may not attain Landon et al.9 14.3 14.5 n = 473
statistical significance if there is substantial variability aris- Crowther et al.8 21 22 n = 510
ing from a small sample size. Statistical significance does not
always equate with clinical significance. These unresolved
issues fanned the controversy over GDM as a clinical entity. significantly higher rate of LGA, 29% in the untreated group
Roberts et al.72 evaluated outcomes of women with (NDDG criteria) when compared to the nondiabetic controls.
untreated IGT (7.8 mmol/L or 140 mg/dL) during preg- Neither the women nor caregivers were blinded to the OGTT
nancy; there was an increased incidence of c/s with no dif- results, thus allowing the women to initiate dietary and other
ferences in neonatal outcome compared with women with lifestyle modifications that could potentially have affected
normal oral glucose tolerance test (OGTT). Ostlund et al.73 glycemic control while leaving the caregivers exposed to a
prospectively studied women with IGT (defined as fasting potential labeling bias. Increased morbidity in untreated
blood glucose level of <6.7 mmol/L [121 mg/dL] and 2-hour GDM was demonstrated in several studies (Table 20.4).
blood glucose level >9 and <11 mmol/L [162–198 mg/dL]) There is a wide range of research designs on a continuum
who were undiagnosed and untreated. The obstetrician from pilot studies to the venerated quasi-randomized,78
was not informed of the deviation in the glucose tolerance. randomized, and blinded RCT. Randomized and quasi-
They compared the untreated IGT to both the controls and randomized controlled trials are sometimes necessary if an
treated GDMs. The rate of macrosomia (>4000 g) was 33%, intervention cannot be blinded. In the past decade, studies
16%, and 30% and large for gestational age (LGA) 25%, 4%, utilizing either randomized or case–control design defini-
and 25%, respectively. The main problem in the Ostlund tively established GDM as a clinical entity and the efficacy of
study is that the authors used a nontraditional definition treatment. However, although the RCT is the gold standard,
for GDM, which was a modification of the Lind defini- it cannot always be the design of choice since it is contrain-
tion.74 Adams et al.75 identified 16 cases of clinically unrec- dicated from an ethical standpoint. For example, if women
ognized GDM diagnosed using the National Diabetes Data with even mild GDM are clinically treated, then not treating
Group (NDDG) criteria and compared them to 64 nondia- them for purposes of randomization in a study poses a poten-
betic controls. The study suggests that unrecognized GDM tial ethical dilemma. Then a realistic and ethical approach
increases the risks for neonatal complications. The major is the use of case–control design. As a rule, a well-executed
limitation of the study is the small sample size; the results retrospective study has fewer methodologic limitations than
could have been affected by both alpha and beta errors. a poorly performed prospective study.
The Toronto Tri-Hospital Gestational Diabetes Project76 Case–control studies are generally a survey study of one
explored pregnancy outcomes for the 3637 subjects with- point in time. Generally they are descriptive about preva-
out a diagnosis of GDM whose caregivers were blinded to lence or behavior, i.e., prevalence of smoking in pregnancy.
the OGTT results. There was a direct relationship between This type of study design is quick and relatively inexpensive.
OGTT results and a number of adverse pregnancy outcomes, The strengths of this format are the potential to study rare
i.e., cesarean delivery, neonatal macrosomia, and preeclamp- events for their antecedent associations as well as its use-
sia. Multivariate analysis was used to control for potential fulness to generate hypotheses for a prospective study. The
confounding effects; the OGTT results continued to have a weakness of this design is the necessity to match the control
significant independent impact. Li et al.77 randomly assigned to the case study within the same time frame (the common
209 women into three groups based on the OGTT results. denominator). If not, there is a significantly greater chance
The first group “mild GDM” (n = 75) based on NDDG crite- for bias in selecting the controls. Finally, a major drawback is
ria was untreated. The second group, GDM, diagnosed after the difficulty to show causality.
a 75 g OGTT by WHO criteria, was treated. The third group We,7 in the largest to date case–control study, described
was normal, nondiabetic controls. The results showed a perinatal outcomes in treated versus untreated GDM that
Clinical studies 163
Table 20.5 Intensified versus conventional management of

Conventional Intensified Untreated

Macrosomia (%) 13.6 7.01 16.8
LGA (%) 20.1 13.1 29.4
Metabolic complication (%) 13.3 3.1 29.0
Respiratory complication (%) 6.2 2.3 12.0
Shoulder dystocia (%) 1.4 0.4 2.5
Stillbirth (per 1000) 3.6 1.0 4.2
Cesarean section (%) 22.0 15.0 23.7
Subjects 1316 1145 555
Abbreviation: LGA, large for gestational age.
addressed many of the limitations posed by the aforemen- the GDM group treated by the conventional approach, the
tioned studies. A population of 555 women with untreated results were comparable. In contrast, the pregnancy outcome
GDM were matched with 1110 women treated for the dis- in GDMs treated with intensified therapy showed signifi-
ease as well as 1110 normal subjects. Patients in the untreated cantly lower rates of complications. This finding demon-
group were recruited to the study after 37 weeks gestation strates that outcome of pregnancy is directly dependent on
that left the fetus exposed to the glucose toxicity throughout treatment modality (Table 20.5).
pregnancy and precluded lifestyle modifications such as diet Fasting plasma glucose is accepted as the gold standard
that could have influenced pregnancy outcome. The diag- for severity of diabetes. This is true in individuals with type 2
nostic criteria used in the study are one of two accepted and diabetes and in women with GDM. In an attempt to control
recommended in the last decade since it was supported by for different GDM severity levels in the treated and untreated
two international workshops on GDM that represent inter- GDMs, we stratified the patients based on increases in fast-
national consensus,79,80 American College of Obstetricians ing plasma glucose (10 mg increments) for each severity cat-
and Gynecologists (ACOG),81 and the ADA recommenda- egory. In the treated GDMs, similar perinatal outcome exists
tions.82 Two nondiabetic controls were matched to each for all fasting severity categories reiterating the importance
untreated GDM on the basis of the following characteristics: of achieving targeted levels of glycemic control. In contrast,
ethnicity, parity, gestational age at delivery (within 1 week), in the untreated GDMs, significant morbidity was found
disease severity (based on fasting plasma glucose level), obe- in each fasting plasma category of severity. Logistic regres-
sity (BMI), and number of prenatal visits. The treated sub- sion revealed in the untreated group that fasting plasma
jects used self-monitoring blood glucose memory reflectance glucose (severity of disease) had a significant independent
meters seven times daily in order to receive accurate feed- impact when each 10 mg increment increased the likelihood
back regarding glycemic profile. of adverse outcome (composite) by 15%; for each pound
A composite adverse outcome comprised of LGA, neona- increase in obese patients, the likelihood of adverse outcome
tal respiratory disease, hypoglycemia, hyperbilirubinemia, increased by 3%. For the treated GDMs, parity was found to
shoulder dystocia, and stillbirth enabled the evaluation of have a 6% increment for every child and obesity and weight
neonatal disease (morbidity) in addition to specific morbidity gain had a negligible effect although both were found to be
components. A composite adverse outcome of the aforemen- statistically significant.
tioned variables was present in 59% of the untreated compared Even in patients with lesser degrees of glucose intoler-
with 18% of the treated and 11% for the nondiabetic popula- ance (fasting plasma glucose <96 mg/dL), there were signifi-
tion. The rates of LGA infants were 29% in untreated versus cantly higher rates of morbidity in untreated women. For
11% with treatment that was similar to the rate in the control the nondiabetic group, the analysis revealed that gestational
population. There was a twofold to fourfold increased risk for age at delivery, previous macrosomia, and glucose screening
large infants and shoulder dystocia, a twofold to sevenfold results had significant independent effects on the composite
increased risk for metabolic and respiratory complications, a outcome. Moreover, for the mild GDM group (fasting glu-
fourfold increased risk for NICU admissions, and a twofold cose level <96 mg/dL), treatment was associated with 10%
increased risk for c/s. Multiple logistic regression analysis LGA rate in comparison to 20% without the benefit of treat-
confirmed fasting glucose as the most predictive cause for ment, composite outcome of 31% compared to 17%, respec-
the composite perinatal outcome in untreated GDM. There tively. Glucose intolerance discovered during pregnancy
was a twofold to threefold decrease in adverse outcome with can be ameliorated with medical intervention in all levels of
treatment after controlling for maternal weight, parity, and GDM severity.
disease severity according to fasting glucose level. Moreover, Nearly 40 years have passed from the original O’Sullivan
when the outcome in the untreated group was compared to randomized study36 that demonstrated that women
with GDM are at higher risk for macrosomia (13.1% in

Table 20.6 Association between mild
untreated vs. 4.3% in treated subjects). The ACHOIS ran-
hyperglycemia (OGTT) and fetal macrosomia
domized study 8 published in 2005 demonstrated the ben-
efits of treatment of GDM patients. The authors used 75 g
Year Definition %LGA Criteria
OGTT result of 2-hour glucose level between 7.8 and 11
mmol/L (140–198 mg/dL) and fasting plasma <7.8 mmol/L Tallarigo et al. 1986 Normal 28 NDDG
(<140 mg/dL). Blood glucose was monitored 4×/daily with Oats et al. 1986 Impaired 22 WHO
a target fasting value of 3.5–5.3 mmol/L (63–99 mg/dL). Langer et al. 1987 1 abnormal 34 NDDG
Twenty percent of the subjects required insulin therapy to Langer et al. 1989 1 abnormal 24 NDDG
achieve targeted levels of glycemic control. However, the Li et al. 1987 Impaired 28 WHO
authors failed to report glycemic data and the relationship Herman et al. 1988 Normal 24 NDDG
between glucose levels and pregnancy outcome. The treated Lindsay et al. 1989 1 abnormal 18 C&C
group was associated with a significant reduction in the rate Neiger et al. 1991 2 values 21 C&C
of the composite outcome (perinatal death, shoulder dysto- Jensen et al. 2001 Normal 21 DPSG-U
cia, and birth traumas including fractures or nerve palsy)
Landon et al. 2009 Normal 15 C&C
resulting in an adjusted RR, 0.33; C.I., 0.14–0.75. Secondary
neonatal outcomes demonstrated a significantly higher rate
in LGA infants in the untreated group compared to the
treated group (22% vs. 13%). There were significant increases Many cases in the nondiabetic population may remain
in NICU rate (71% treated vs. 61% untreated) but did not unrecognized and, therefore, resulted in untreated GDM
demonstrate significant differences in the rate of hypogly- with glucose toxicity. They are reinstated into the “normal”
cemia (defined as requiring intravenous therapy), jaundice population in most maternal clinics; we deliver them every
(phototherapy), and respiratory complications (supplemen- day. This group of patients may unwittingly “contaminate”
tal oxygen). In addition, the rate of preeclampsia was 18% in the rates reported for the normal population by increasing
the untreated versus 12% in the treated group. the adverse pregnancy outcome (Table 20.6).
Both the NICHD and AICHOS studies made significant The scientific rationale for the use of two or more abnor-
contributions to the study of GDM. However, the rate of mal values is not based on evidence but rather on opinions.
adverse outcome in each of these studies are not compara- The explanation for those who support two abnormal val-
ble. There were different inclusion criteria; the NICHD used ues range from “just because” to “better reproducibility
fasting plasma glucose <95 mg/dL, while the AICHOS study of the test.” However, data supporting these positions are
recruited patients at all levels of the glucose spectrum up to lacking. If at all, the existing data suggest that one abnor-
<140 mg/dL (from mild to severe GDM and type 2 diabetes). mal value has the same characteristics and predictive value
Additionally, the primary outcome (composite) was different as two or more values. The use of one abnormal value for
in each study because each contained different variables. In GDM diagnosis is further supported by the fact that many
contrast, the previously described case–control study7 used obstetricians will use screening values of 180 mg/dL or
the same inclusion criteria as the AICHOS study of fasting greater as a single diagnosis for GDM and will treat based
plasma <140 mg/dL resulting in similar rates of univariant on this single result.
adverse outcomes. The components in each of the composite We, in 1987, suggested, in a case–control study, that
outcomes in each of the three studies were dissimilar, and women with one abnormal value on the OGTT results have
therefore, the rates of composite outcome were not compa- a significantly increased risk for adverse pregnancy outcome
rable. However, the similarity in all three studies in LGA when compared to nondiabetic women and women with
and macrosomia rates demonstrates that case–control stud- treated GDM (two or more abnormal values on the OGTT).83
ies can result in comparable results to those in randomized In a follow-up study, women with one abnormal value were
studies when designed appropriately. randomized into treatment and nontreatment and compared
to nondiabetic subjects. The incidence of large infants was
significantly higher in the untreated group. When patients
Mild untreated hyperglycemia were stratified into obese and nonobese for each study group
(treated, untreated, and control), there was a significantly
Another component of the GDM debate surrounds the higher rate of large infants and metabolic complications in
parameters of the lower threshold of the disease. This element the untreated group. There was no significant difference in
affects the controversy surrounding the diagnostic criteria the rate of LGA between obese and nonobese patients.84 In
that has evolved in the past few years from the International a third study, we compared the incidence of LGA infants in
Diabetes Study Group based on the HAPO study to the relation to the number of abnormal values on the OGTT. We
ADA, ACOG, and WHO criteria. The association between found a similar rate of LGA infants when one, two, or three
mild hyperglycemia and adverse neonatal outcome has values were abnormal. This was especially true in patients
been a major concern for the past two decades especially for with poor glycemic control.85 Similar findings by Lindsay
patients who could not reach the “gold standard” currently et al.86 showed 18% LGA in his study population when one
used for screening and diagnosis (WHO, NDDG, ADA). abnormal value was used. Neiger et al.87 showed that women
Mild untreated hyperglycemia 165
with one abnormal value retested after 4 weeks showed that outcome variables were macrosomia (≥4000 g), LGA, and
about 33% had at least two abnormal values on the OGTT. neonatal FFM. Maternal outcomes included preeclampsia,
Gruendhammer et al.88 studied 152 women with one abnor- induction of labor, and cesarean delivery. The rate of com-
mal glucose value match controlled to 304 nondiabetic posite outcome in the treated and untreated groups was
women with normal OGTT values. They found that women 32.4% versus 37.0% (nonsignificant), respectively. None of
with only one abnormal OGTT value had increased risk in the individual frequency components of the composite out-
comparison to the control subjects. Similar results were found come were statistically different between the two groups.
in other studies.89,90 Recently, the use of one abnormal value In the secondary outcomes, there were significantly lower
as part of GDM diagnosis was endorsed by the International rates of LGA, macrosomia, and shoulder dystocia (1.5%
Diabetes in Pregnancy Group and the American Diabetes treated vs. 4.0% untreated). Preeclampsia results showed
Association.11,82 This endorsement documented that mild 8.6% for the treated versus 13.6% for the untreated subjects.
to severe hyperglycemia when unrecognized and untreated There are several major dissimilarities in the research
caused higher rates of adverse pregnancy outcome. design of each study: the NICHD, AICHOS, and Langer.
The NICHD Maternal-Fetal Medical Units (MFMU)9 These variances resulted in different outcome rates. What
trial sought to identify the lower end of the glucose is more scientifically sound is to indicate and emphasize
spectrum (“mild” GDM, fasting plasma glucose below the direction, i.e., reduced or increased rather than fre-
95 mg/dL). The rationale for addressing this end of the quency of rates when describing outcomes (Table 20.7).
spectrum was also related to the ethical considerations of The ACHOIS study 8 was based on an ethnically homoge-
participation in a randomized study of GDM treatment. neous population, while the NICHD and our study were
By following this reasoning, they recognized that patients more ethnically diverse (Langer study more Hispanic
with abnormal OGTT results (fasting plasma glucose >95 women included). The ethnic diversity affects GDM
mg/dL) are GDM and, therefore, should be treated. Four prevalence. In contrast, successful treatment resulted in
hundred and eighty-five women who met the entry criteria controlling abnormal glucose levels, which in turn led
were randomized to treatment, and 473 received the stan- to lowering the adverse pregnancy outcome. Therefore,
dard prenatal care. The treatment arm performed daily self- treatment will not be affected by ethnicity but rather by
monitored blood glucose testing with the glycemic target achieving targeted levels of glycemic control. As noted
achieved in the vast majority of cases. The primary peri- earlier, each study used different definitions for compos-
natal outcome was composite outcome that included peri- ite outcome and metabolic complications (hypoglycemia,
natal death, neonatal hypoglycemia, hyperbilirubinemia, hyperbilirubinemia), therefore making study results dif-
elevated cord C-peptide level, and birth trauma. Secondary ficult to compare (Table 20.8).
Table 20.7 Perinatal outcome: In untreated gestational diabetes mellitus
Langer et al.7 Crowther et al.8 Landon et al.9

Inclusion criteria (fasting) <140 mg/dL <140 mg/dL <95 mg/dL
C and C WHO C and C
Composite not comparable Yes Yes Yes
Race or ethnic group (%)
White 11.0 78.0 25.2
Black 3.0 — 11.4
Hispanic 86.0 — 56.0
Asian and others — 22.0 7.4
Composite outcome (%) 59.0 4.0 37.0
Macrosomia (%) 16.8 21.0 14.3
LGA (%) 29.4 22.0 14.5
Ponderal index % (>2.85) 22.0 — —
Metabolic complication (%) 29.0 14.0 28.3
Respiratory complication (%) 12.0 4.0 2.9
Shoulder dystocia (%) 2.5 3.0 4.0
NICU admission (%) 24.0 61.0 11.6
Stillbirth (1000) 5.4 5.7 0.0
Cesarean section (%) 24.0 32.0 33.8
n = 555 n = 524 n = 473
Abbreviation: C and C, Carpenter and Coustan.

Table 20.8 Comparison of treatment effect for different outcome definitions in the
case–control study, ACHOIS study, and NICHD MFMU trial
Langer et al.7 Crowther et al.8 Landon et al.9

Inclusion criteria (fasting) <140 mg/dL <140 mg/dL WHO <95 mg/dL
C and C C and C
Composite not comparable Yes Yes Yes
Composite outcome Reduced Reduced No difference
Macrosomia Reduced Reduced Reduced
LGA (%) Reduced Reduced Reduced
Metabolic complication Reduced No difference No difference
Neonatal hypoglycemia Reduced No difference No difference
Neonatal hyperbilirubinemia Reduced No difference No difference
Respiratory complication (%) Reduced No difference No difference
Shoulder dystocia (%) Reduced No difference Reduced
NICU admission (%) Reduced Increased Reduced
Stillbirth (1000) No difference No difference 0.0
Preeclampsia No difference Reduced Reduced
Induction of labor Reduced Increased No difference
Cesarean section (%) No difference No difference Reduced
n = 555 n = 524 n = 473
Abbreviation: C and C, Carpenter and Coustan.
Summary individuals with GDM that could have a significant impact

on the provision of obstetrical care. GDM, when not
Considering the metabolic heterogeneity of women with treated, is associated with increased adverse outcome in
GDM and thus their likely broad range of perinatal risk, pregnancy. Treatment and achievement of targeted levels of
it is not surprising that descriptions of GDM have ranged glycemic control will result in perinatal outcome compa-
from “a major health problem” to “a diagnosis still look- rable to that in the general population. Perinatal outcome
ing for a disease.” We and others addressed the question and long-term complications for millions of mothers and
of GDM as a clinical entity and the benefits of treating their offspring are at stake.
REFERENCES
1. Langer O, Umans J, Miodovnik M. Perspectives on the proposed 10. HAPO Study Cooperative Research Group; Metzger B, Lowe L,
gestational diabetes mellitus diagnostic criteria. Obstet Gynecol Dyer A et al. Hyperglycemia and adverse pregnancy outcomes.
2013; 121(1): 177–182. N Engl J Med 2008; 358: 1991–2002.
2. Langer O, Umans J, Miodovnik M. The proposed GDM diagnos- 11. International Association of Diabetes and Pregnancy Study
tic criteria: A difference to be a difference, must make a differ- Groups consensus Panel; Metzger B, Gabbe S, Persson B et al.
ence. J Matern Fetal Neonatal Med 2013; 26(2): 111–115. International association of diabetes and pregnancy study groups
3. Jarrett, R. Gestational diabetes: A non-entity? Br Med J 1993; 306: recommendations on the diagnosis and classification of hypergly-
37–38. cemia in pregnancy. Diabetes Care 2010; 33: 676–682.
4. Sermer M, Naylor C, Farine D et al. The Toronto Tri-Hospital 12. Grimes D. Epidemiologic research using administrative data-
Gestational Diabetes Project. A preliminary review. Diabetes bases: Garbage in, garbage out. Obstet Gynecol 2010; 116(5):
Care 1998; 21: B33–B42. 1018–1026.
5. Hunter D, Milner R. Gestational diabetes and birth trauma [let- 13. Grimes D, Schultz K. False alarms and pseudo-epidemics: The
ter]. Am J Obstet Gynecol 1985; 152: 918–919. limitations of observational epidemiology. Obstet Gynecol 2012;
6. Beard R, Hoet J. Is gestational diabetes a clinical entity? 120(4): 920–927.
Diabetologia 1982; 23(4): 307–312. 14. Ryan E, Enns L. Role of gestational hormones in the induction of
7. Langer O, Yogev Y, Most O. Gestational diabetes: The con- insulin resistance. J Clin Endocrinol Metab 1988; 67: 431–437.
sequences of not treating. Am J Obstet Gynecol 2005; 192: 15. Pittas A, Joseph N, Greenberg A. Adipocytokines and insulin
989–997. resistance. J Clin Endocrinol Metab 2004; 89: 447–452.
8. Crowther C, Hiller J, Moss J et al. Effect of treatment of gestational 16. Matsubara M, Maruoka S, Katayose S. Decreased plasma adi-
diabetes mellitus on pregnancy outcomes. N Engl J Med 2005; ponectins concentrations in women with dyslipidemia. J Clin
352(24): 2477–2486. Endocrinol Metab 2002; 87: 2764–2769.
9. Landon M, Spong C, Thom E et al. A multicenter, randomized 17. Yamauchi T, Kamon J, Waki H et al. The fat-derived hormone
trial of treatment for mild gestational diabetes. N Engl J Med adiponectins reverses insulin resistance associated with both
2009; 361(14): 1339–1348. lipoatrophy and obesity. Nat Med 2001; 7: 941–946.
References 167
18. Retnakaran R, Hanley A, Raif N et al. C-reactive protein and 41. ACOG Practice Bulletin Clinical Management Guidelines. Fetal
gestational diabetes: The central role of maternal obesity. J Clin macrosomia. Obstet Gynecol 2000; 96(5).
Endocrinol Metab 2003; 88: 3507–3512. 42. Rouse D, Owen J, Goldenberg R et al. The effectiveness and costs
19. Catalano P. Editorial: Obesity and pregnancy—The propaga- of elective cesarean delivery for fetal macrosomia diagnosed by
tion of a viscous cycle? J Clin Endocrinol Metab 2003; 88(8): ultrasound. J Am Med Assoc 1996; 276(18): 1480–1486.
3505–3506. 43. Naeye R. Infants of diabetic mothers: A quantitative, morpho-
20. DeFronzo R, Ferrannini E. Insulin resistance: A multifaceted syn- logic study. Pediatrics 1965; 35(6): 980–988.
drome responsible for NIDDM, obesity, hypertension, dyslipid- 44. Langer O, Kagan-Hallet K. Diabetic vs. non-diabetic infants: A
emia, and atherosclerotic heart disease. Diabetes Care 1991; 14: quantitative morphological study. In: Proceedings of the 38th
173–194. Annual Meeting of the Society for Gynecologic Investigation, San
21. Bonora E, Del Prato S, Bonadonna R et al. Total body fat content Antonio, TX, 1992.
and fat topography are associated differently with in vivo glu- 45. Chengi Y, Chung J, Block-Kurbisch I et al. Treatment of gesta-
cose metabolism in nonobese and obese nondiabetic women. tional diabetes mellitus: Glyburide compared to subcutaneous
Diabetes 1992; 41: 1151–1159. insulin therapy and associated perinatal outcomes. J Matern Fetal
22. Catalano P, Tyzbir E, Roman N et al. Longitudinal changes in insu- Neonatal Medicine 2012; 25(4): 379–384.
lin release and insulin resistance in non-obese pregnant women. 46. Langer O, Levy J, Brustman L et al. Glycemic control in g estational
Am J Obstet Gynecol 1991; 165: 1667–1672. diabetes mellitus—How tight is tight enough: Small for gesta-
23. Bergman R, Phillips L, Cobelli C. Physiologic evaluation of factors tional age versus large for gestational age? Am J Obstet Gynecol
controlling glucose sensitivity from the response to intravenous 1989; 161: 646–653.
glucose. J Clin Invest 1981; 68: 1456–1467. 47. Langer O, Mazze R. The relationship between large-for-
24. Kahn S, Prigeon R, McCullough D. Quantification of the rela- gestational-age infants and glycemic control in women with
tionship between insulin sensitivity and β-cell function in human GDM. Am J Obstet Gynecol 1988; 159(6): 1478–1483.
subjects: Evidence for a hyperbolic function. Diabetes 1993; 42: 48. Catalano P, Thomas A, Huston-Presley L et al. Increased fetal
1663–1672. adiposity: A very sensitive marker of abnormal in-utero develop-
25. Buchanan T, Xiang A, Peters R. Response of pancreatic β-cells ment. Am J Obstet Gynecol 2003; 189(6): 1698–1704.
to improved insulin sensitivity in women at high risk for type 2 49. Bernstein I, Catalano P. Increased neonatal fat is associated with
diabetes. Diabetes 2000; 49: 782–788. risk of cesarean section in infants of mothers with gestational
26. Xiang H, Peters K, Trigo E et al. Multiple metabolic defects during diabetes. Society of Perinatal Obstetricians, San Francisco,
late pregnancy in women at high risk for type 2 diabetes mellitus. CA,1993.
Diabetes 1999; 48: 848–854. 50. McFarland M, Trylovich C, Langer O. Anthropometric differ-
27. Buchanan T. Commentary: Pancreatic β-cell defects in gesta- ences in macrosomic infants of diabetic and non-diabetic moth-
tional diabetes: Implications for the pathogenesis and preven- ers. J Matern Fetal Med 1998; 7(6): 292–295.
tion of type 2 diabetes. J Clin Endocrinol Metab 2001; 86(3): 51. Dyck R, Klomp H, Tan L et al. A comparison of rates, risk factors,
989–993. and outcomes of gestational diabetes between aboriginal and
28. American College of Obstetricians and Gynecologists. Screening non-aboriginal women in the Saskatoon health district. Diabetes
and diagnosis of gestational diabetes mellitus. Committee Care 2002; 25(3): 487–493.
Opinion No. 504. Obstet Gynecol 2011; 118: 751–753. 52. Barker D, Osmond C, Golding J et al. Growth in utero, blood
29. Catalano P, Bernstein I, Wolfe R et al. Subclinical abnormalities pressure in childhood and adult life, and mortality from cardio-
of glucose metabolism in subjects with previous gestational dia- vascular disease. Br Med J 1989; 298: 564–567.
betes. Am J Obstet Gynecol 1986; 155: 1255–1262. 53. Barker D. In utero programming of cardiovascular disease.
30. Catalano P, Vargo K, Bernstein I et al. Incidence and risk fac- Theriogenology 2000; 53: 555–574.
tors associated with abnormal postpartum glucose tolerance in 54. Silverman B, Rizzo T, Cho N et al. Long-term effects of the intra-
women with gestational diabetes. Am J Obstet Gynecol 1991; uterine environment. The Northwestern University Diabetes in
165: 914–919. Pregnancy Center. Diabetes Care 1998; 21(Suppl. 2): B142–B149.
31. Ward W, Johnston C, Beard J et al. Insulin resistance and impaired 55. Pettitt D, Nelson R, Saad M et al. Diabetes and obesity in the
insulin secretion in subjects with histories of gestational diabetes offspring of Pima Indian women with diabetes during pregnancy.
mellitus. Diabetes 1985; 34: 861–868. Diabetes Care 1993; 16(1): 310–314.
32. Landon M, Hauth J, Leveno K et al. Maternal and perinatal out- 56. Dabelea D, Pettitt D. Intrauterine diabetic environment confers
comes associated with a trial of labor after prior cesarean deliv- risks for type 2 diabetes mellitus and obesity in the offspring,
ery. N Engl J Med 2004; 351: 2581–2589. in addition to genetic susceptibility. J Pediatr Endocrinol Metab
33. Naylor C, Sermer M, Chen E et al. Cesarean delivery in relation to 2001; 14(8): 1085–1091.
birth weight and gestational glucose tolerance: Pathophysiology 57. Silverman B, Metzger B, Cho N et al. Fetal hyperinsulinism and
or practice style? Toronto Tri-Hospital Gestational Diabetes impaired glucose tolerance in adolescent offspring of diabetic
Investigators. J Am Med Assoc 1996; 275: 1165–1179. mothers. Diabetes Care 1995; 18: 611–617.
34. Conway D, Langer O. Elective delivery of infants with macro- 58. Rizzo T, Freinkel N, Metzger B et al. Correlations between
somia in diabetic women: Reduced should dystocia versus antepartum maternal metabolism and newborn behavior. Am J
increased cesarean deliveries. Am J Obstet Gynecol 1998; 178(5): Obstet Gynecol 1990; 163: 1458–1464.
922–925. 59. Rizzo T, Metzger B, Burns W et al. Correlations between antepar-
35. Langer O, Rodriguez D, Xenakis E et al. Intensified versus con- tum maternal metabolism and intelligence of offspring. N Engl J
ventional management of gestational diabetes. Am J Obstet Med 1991; 325: 911–916.
Gynecol 1994; 170(4): 1036–1047. 60. Petersen M, Pedersen S, Greisen G et al. Early growth delay in
36. O’Sullivan J, Mahan C, Charles D. Screening criteria for high- diabetic pregnancy: Relation to psychomotor development at age
risk gestational diabetic patients. Am J Obstet Gynecol 1973; 116: 4. Br Med J 1988; 296: 598–600.
895. 61. Visser G, Bekedam D, Mulder E et al. Delayed emergence of fetal
37. Abell D, Beischer N, Wood C. Routine testing for gestational behavior in type 1 diabetic women. Early Hum Dev 1985; 12:
diabetes, pregnancy, hypoglycemia, and fetal growth retardation 167–172.
and results of treatment. J Perinat Med 1976; 4: 197–212. 62. Mulder E, Visser G, Bekedam D et al. Emergence of behavioral
38. Oats J, Beischer N. Gestational diabetes. Aust N Z J Obstet states in fetuses of type 1 diabetic women. Early Hum Dev 1987;
Gynaecol 1986; 26: 2–9. 15: 231–251.
39. Schmidt M, Duncan B, Reichelt A et al. Gestational diabetes 63. Sells C, Robinson N, Brown Z et al. Long-term developmental fol-
mellitus diagnosed with a 2-h 75-g oral glucose tolerance test low-up infants of diabetic mothers. J Pediatr 1994; 125: S9–S17.
and adverse pregnancy outcomes. Diabetes Care 2001; 24(7): 64. Lincoln N, Faleiro R, Kelly C. Effect of Long-term glycemic con-
1151–1153. trol on cognitive function. Diabetes Care 1996; 19(6): 656–658.
40. Mondestin M, Ananth C, Smulian J et al. Birth weight and fetal 65. Vohr B, McGarvey S, Coll C. Effects of maternal gestational dia-
death in the United States: The effect of maternal diabetes during betes and adiposity on neonatal adiposity and blood pressure.
pregnancy. Am J Obstet Gynecol 2002; 187: 922–926. Diabetes Care 1995; 18(4): 467–475.
66. Vohr B, McGarvey S. Growth patterns of large-for-gestational age 80. Metzger BE. Summary and Recommendations of the Third
and appropriate-for-gestational age infants of gestational diabetic International Workshop Conference on Gestational Diabetes
mothers and control mothers at age 1 year. Diabetes Care 1997; Mellitus. Diabetes 1991; 40(Suppl. 2): 197–201.
20(7): 1066–1072. 81. American College Obstetricians and Gynecologists. Practice
67. Rizzo T, Dooley S, Metzger B et al. Prenatal and perinatal influ- Bulletin: Gestational diabetes mellitus. Obstet Gynecol 2013;
ences on long-term psychomotor development in offspring of 122(2), Part 1: 406–416.
diabetic mothers. Am J Obstet Gynecol 1995; 173: 1753–1758. 82. ADA Standards of Care 2014. Clinical Practice Recommendations.
68. Rizzo T, Ogata E, Dooley S et al. Perinatal complications and Diabetes Care 37(Suppl. #1).
cognitive development in 2–5 year-old children of diabetic moth- 83. Langer O, Brustman L, Anyaegbunam A. The significance of one
ers. Am J Obstet Gynecol 1994; 171: 706–713. abnormal glucose tolerance test value on adverse outcome in
69. Petersen M. Status at 4–5 years in 90 children and insulin-depen- pregnancy. Am J Obstet Gynecol 1987; 157: 758–763.
dent diabetic mothers. In: Sutherland H, Stowers J, Pearson D, 84. Langer O, Anyaegbunam A, Brustman L et al. Management of
eds., Carbohydrate Metabolism in Pregnancy and the Newborn women with one abnormal oral glucose tolerance test value
IV, Springer-Verlag: London, U.K., 1989, pp. 353–361. reduces adverse outcome in pregnancy. Am J Obstet Gynecol
70. Stenninger E, Flink R, Eriksson B et al. Long-term neurological 1989; 161: 593–599.
dysfunction and neonatal hypoglycemia after diabetic preg- 85. Berkus M, Langer O. Glucose tolerance test: Degree of glucose
nancy. Arch Dis Child Fetal Neonatal Ed 1998; 79(3): F174–F179. abnormality correlates with neonatal outcome. Obstet Gynecol
71. Olsson G, Hulting A, Montgomery S. Cognitive function in chil- 1993; 81(3): 344–348.
dren and subsequent type 2 diabetes. Diabetes Care 2008; 31: 86. Lindsay M, Graves W, Klein L. The relationship of one abnor-
514–516. mal glucose tolerance test value and pregnancy complications.
72. Roberts R, Moohan J, Foo R et al. Fetal outcome in mothers with Obstet Gynecol 1989; 73: 103–106.
impaired glucose tolerance in pregnancy. Diabet Med 1993; 10: 87. Neiger R, Coustan D. The role of repeat glucose tolerance tests in
438–443. the diagnosis of gestational diabetes. Am J Obstet Gynecol 1991;
73. Ostlund I, Hanson U, Bjorklund A et al. Maternal and fetal out- 165: 787–790.
comes if gestational impaired glucose tolerance is not treated. 88. Gruendhammer M, Brezinka C, Lechleitner M. The number of
Diabetes Care 2003; 26(7): 2107–2111. abnormal plasma glucose values in the oral glucose tolerance
74. Lind T, Phillips P. Influence of pregnancy on the 75-g OGTT: A test and the feto-maternal outcome of pregnancy. Eur J Obstet
prospective multicenter study. The Diabetic Pregnancy Study Gynecol Reprod Biol 2003; 108: 131–136.
Group of the European Association for the Study of Diabetes. 89. Forest J, Masse J, Garrido-Russo M. glucose tolerance test dur-
Diabetes 1991; 40(Suppl. 2): 8–13. ing pregnancy: The significance of one abnormal value. Clin
75. Adams K, Li H, Nelson R et al. Sequelae of unrecognized gesta- Biochem 1994; 27: 200–204.
tional diabetes. Am J Obstet Gynecol 1998; 178(6): 1321–1332. 90. Bo S, Menato G, Gallo M et al. Mild gestational hyperglycemia,
76. Sermer M, Naylor C, Gare D et al. Toronto Tri-Hospital the metabolic syndrome and adverse neonatal outcomes. Acta
Gestational Diabetes Investigators: Impact of increasing carbo- Obstet Gynecol Scand 2004; 83: 335–340.
hydrate intolerance on maternal-fetal outcomes in 3,637 women 91. Coustan DR, Lewis SB. Insulin therapy in gestational diabetes.
without gestational diabetes. Am J Obstet Gynecol 1995; 173: Ob. Gyn. 1978; 51: 306.
146–156. 92. Coetzee EJ, Jackson WPU. The management of non-insulin-
77. Li D, Wong V, O’Hoy K et al. Is treatment needed for mild impair- dependent diabetes during pregnancy. Diabetes Res Clin Practice
ment of glucose tolerance in pregnancy? A randomized con- 1986; 1: 281–287.
trolled trial. Br J Obstet Gynaecol 1987; 94(9): 851–854. 93. Hod M, Robinerson D, Peled Y. Gestational diabetes: Is it a clini-
78. Alwan N, Tuffnell D, West J. Treatment for gestational diabetes. cal entity? Diabetes Reviews 1995; 3: 602–613.
Cochrane Pregnancy and Childbirth Group. 2009. 94. Sermer M, Naylor CD, Gare DJ et al. Impact of increasing car-
79. Metzger B, Coustan D. Summary and Recommendations of bohydrate intolerance on maternal-fetal outcomes in 3637
the Fourth International Workshop-Conference on Gestational women without gestational diabetes. The Toronto Tri-Hospital
Diabetes Mellitus. The Organizing Committee. Diabetes Care Gestational Diabetes Project. Am J Obstet Gynecol 1995; 173(1):
1998; 21(Suppl. 2): B161–B167. 146–156.
in multiple pregnancies
Matteo Andrea Bonomo and Angela Napoli
Outside the United States, data extracted from the New

Epidemiology of multiple pregnancies Zealand medical records of 4939 deliveries, with discharges
Multiple gestations: An epidemic between 1994 and 1995, reported 1.1% of twin pregnancies,
A true “epidemic” of multiple gestations1,2 has been recorded with no differences on an ethnic basis (Europeans 1.2%,
over the last two decades: the 2014 Practice Bulletin of the Maori 0.8%, Pacific Islanders 1.4%, others 0.4%). The num-
American College of Obstetricians and Gynecologists (ACOG)3 bers were too small to look at the risk of gestational diabetes
reports that in the United States, the rate of twin births mellitus (GDM) within ethnic groups, but the groups were
increased 76% between 1980 and 2009, from 18.9 to 33.3 per well matched for this.6
1000 births (Figure 21.1). These impressive rises c oncerned In Europe, the incidence of twinning rates shows a grad-
not only simple twin pregnancies. Mainly as a consequence ual decrease from north to south. The frequency of multiple
of the spread of assisted reproductive technology (ART) pro- pregnancies is high among “Anglo-Saxons,” and even higher
cedures, the rate of triplet and higher-order multifetal ges- in Scandinavia7,8; in contrast, the rate is medium among Slav
tations rose more than 400% during the 1980s and 1990s, populations and low for Latins (although data are scanty for
peaking at 193.5 per 100,000 births in 1998, followed by a Mediterranean areas).9,10 In Italy, the rate of spontaneous
modest decline to 153.4 per 100,000 births by 2009.3 multiple deliveries in 2010 was 1.6%, ranging from 1.0% to
Despite this partial reversal, reflecting the reduction in 3.1% (Figure 21.2), with no significant differences between
the number of embryos transferred with in vitro fertiliza- the north, center, and south.11
tion (IVF) and the larger numbers of multifetal pregnancy In East Asian countries, twin maternities are rare; for
reduction procedures, the most notable feature remains the instance, in Japan the twinning rate has varied between 3
increased incidence in twin gestations. This has been attrib- and 9 per 1000 from the 1920s to the 1990s. The patterns of
uted largely to two concomitant reasons: an older parturient twinning rates in Singapore and Hong Kong are comparable
population resulting from delayed childbearing and the rise to Japan, though slightly higher.10
in use of infertility therapies.4,5 Among women attempting
conception, the rate of infertility treatment is now estimated
at 11%, equally distributed between 5.4% ART procedures Risk factors
and 5.5% ovulation.2 This general trend is seen, though with Maternal age
obvious differences, in almost all developed countries. Older women with lower parity are much more likely to give
birth to twins or triplets, so the rate of multifetal pregnan-
cies rises with maternal age. As reported in the ACOG bul-
Twinning rates by ethnics and geographical area letin, the multiple-birth rate in the United States rises from
Ethnicity seems to play a role in the frequency of multife- 16.3 per 1000 live births for women younger than 20 years to
tal gestations. As reported in the National Vital Statistics 71.1 per 1000 for women 40 years and older.3 A similar relation
Report 2013, among the three largest racial/origin groups in with age has been reported in Europe. Eriksson et al. found a
the United States, the multiple-birth rates in 2011 were as steady increase in multiple pregnancies starting around 1980,
follows: non-Hispanic white 38.3‰ (36.6 twins + 1.71 trip- with a marked rise in the age-specific rate in the group aged
lets or higher order), non-Hispanic black 38.3‰ (37.2 + 1.1), 40+ years in England and Wales in 2007,7 and in Sweden a
and Hispanic 23.9‰ (23.1 + 0.8). This also helps explain the few years before.8 The same was reported by Pison in France
differences between different U.S. regions. Multiple-birth and in other western developed countries.9 There are prob-
rates ranged widely by state, with the highest in New Jersey ably numerous reasons for these increases. Hypothetically,
and the lowest in New Mexico.5 premenopausal mothers might have hyperstimulation by
169
170 Gestational diabetes mellitus in multiple pregnancies
36 endogenous follicle-stimulating hormones (FSH) caused by

Twinning rate in the United States (per 1000 births) 75 neuroendocrine, hypothalamic, or pituitary mechanisms.2
Assisted reproductive therapy
Relative change in twinning rate since 1980 (%)

65
31
Infertility therapies influencing the rate of multiple preg-
55 nancies include ART procedures (mainly IVF and related
procedures) and controlled ovarian hyperstimulation with
26
gonadotropins.12,13 Fertility treatments are complex, and each
45
ART cycle consists of several steps that include hormonal
treatments to trigger and stimulate the growth of multiple
35 ovarian follicles, medications to suppress the natural men-
21 strual cycle and downregulate the pituitary gland, drugs to
25 bring about final maturation of the eggs (ovulation trigger-
ing), and endometrial preparation for placing the embryos in
the uterus. Finally, the luteal phase support involves several
15
16 options including progesterone, estrogen (E2), and human
chorionic gonadotropin.
Twinning rate per 1000 5 Multiple stimulation regimens are therefore possible,
Relative change since 1980 including clomiphene alone, human menopausal gonado-
11 –5 tropin (HMG), clomiphene plus HMG, FSH, pulsatile
1980 1984 1988 1992 1996 2000 2004 2008 gonadotropin-releasing hormone (GnRH), and FSH plus
HMG. The use of GnRH agonists and antagonists has
Figure 21.1 Temporal trends and changes in twinning
rates in the United States, 1980–2009. (From Ananth, C.V. improved pregnancy rates by preventing natural ovulation.
and Chauhan, S.P., Semin. Perinatol., 36, 156, 2012. With Although these approaches influence the maternal endo-
permission.) crine milieu differently, published reports do not usually
specify the type of protocol adopted.
According to the most recent report from the National
Center for Chronic Disease Prevention and Health Pro
motion,14 among 4,046,553 infants born in the United States
in 2010, 59,119 (1.5%) were conceived with ART procedures
1.5 done in 2009 and 2010. Although 96.5% of total infants
1.2
1.7 born in 2010 were singletons, only 53.6% of all ART infants
1.6
1.7
were. Nationwide, 46.4% of ART infants were multiples,
1.0 compared with only 3.4% of all infants; looking at high-
order multiple pregnancies (triplets or higher), the frequency
1.6 after ART procedures was 3.0%, compared with only 0.1% of
2.0 all infants. Overall, in 2010, 19.2% of all twin and 32.5% of
Total:1.6
triplet or higher-order births in the United States were con-
1.6
1.4 ceived with ART.
1.3 These treatments in women who on average are older and
weigh more may further influence the maternal hormone
1.2
1.9 concentrations in pregnancies that are often multiple.15
1.4
1.7 1.5
1.2 3.1 Physiology of multiple pregnancies
Weight and caloric intake
1.3 Given the increased metabolic rate and higher caloric
needs in twin pregnancies, the U.S. Institute of Medicine
Guidelines established specific nutritional and weight guide-
lines.16 Women with a twin gestation and normal body mass
1.4 index (BMI) are recommended to gain 37–54 lb (17–24 kg),
overweight women 31–50 lb (14–22 kg), and obese women
25–42 lb (11–19 kg). The daily recommended caloric intake
for women with normal BMI who are pregnant with twins
Figure 21.2 Twinning rates (%) in Italy, 2010: regional differ-
ences. (Data from Basili, F. et al., Analisi dell’evento nascita – is 40–45 kcal/kg during the first trimester, then adjusted as
Anno, 2013, 1, 2010.) necessary for the weight gain goal.17
Glycemic homeostasis in pregnancy 171
Table 21.1 Twin pregnancy nutritional recommendations: Micronutrients
Micronutrient supplement (daily total intake) First trimester Second trimester Third trimester
MVI with iron (30 mg elemental tablet) 1 2 2
Calcium (mg) 1500 2500 2500
Vitamin D (international units) 1000 1000 1000
Magnesium (mg) 400 800 800
Zinc (mg) 15 30 30
DHA/EPA (mg) 300–500 300–500 300–500
Folic acid (mg) 1 1 1
Vitamin C (mg) 500–1000 500–1000 500–1000
Vitamin E (international units) 400 400 400
Source: Modified from Goodnight, W. et al., Obstet. Gynecol., 114, 1121, 2009. With permission.
Abbreviations: DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; MVI, multivitamin.
Iron and other micronutrients glucose, lipids, and amino acids are available for the fetus.
Women with twin gestations require nearly double the iron Even with hepatic glucose production increased by approxi-
replacement (about 869 mg elemental iron) of a singleton mately 30%, fasting glucose concentrations fall, probably
pregnancy, on account of the larger amount being trans- secondary to the fetoplacental utilization. Placental and fetal
ported to the fetuses. When iron-deficiency anemia is man- demand for glucose is considerable and may even approach
ifest, 60–120 mg of elemental iron daily for 4 weeks raises the equivalent of 150 g/day of glucose in the third trimester.
the hemoglobin concentration by 1 g/dL. Considering the Glucose transport to the fetus is in direct proportion
increased incidence of hypochromic anemia already pres- to the maternal levels and is boosted by a fivefold increase
ent in twin pregnancies, appropriate iron supplementation in placental glucose flux even in the absence of maternal
is strongly indicated in these women, as both in singleton hyperglycemia. If fetal glucose requirements cannot be met
and multiple pregnancies it can reduce the 3–17 times higher because of maternal hypoglycemia or placental insufficiency,
risk of requiring blood transfusion during cesarean section the fetus can draw on alternate substrates, such as ketone
because of mild-to-severe preoperative anemia as well as bodies derived from beta-oxidation of fatty acids.
the risk of premature labor and the severity of postpartum Insulin does not cross the placenta, and fetal insulin pro-
hemorrhage.18,19 duction starts early.
Recommendations for micronutrient supplementation in
twin pregnancies, summarized in Table 21.1, are based on
expert opinion, and the long-term implications for a woman’s Multiple pregnancy: A potentially diabetogenic
health due to nutrient deficiency are not known. However, condition
calcium, vitamin D, zinc, magnesium, folate, vitamin C, and Women with multifetal gestations undergo significant physio-
omega-3 fatty acids are all important in fetal development.18 logical adaptations beyond the metabolic, cardiovascular, and
hematologic changes expected for a singleton pregnancy. The
levels of pregnancy-related diabetogenic hormones, including
Glycemic homeostasis in pregnancy human placental lactogen, estrogens, progesterone, and corti-
sol,23 are higher in multiple than singleton pregnancies because
Singleton pregnancy: A potentially diabetogenic of the growing placental mass. At term, the median weight of
condition twin placentas is about 1119 g, almost double the median pla-
Pregnancy itself is to some extent a “diabetogenic” state on cental weight for singleton boys (679 g) and girls (668 g).24
account of the continuous changes in maternal hormone Although the larger placental mass in twin gestation
concentrations. Over the course of gestation, the insulin results in an increase in placental-derived steroid hormones
response to nutrients progressively rises, while glucose tol- that may predispose to glucose intolerance by increasing
erance changes only slightly. The balance is maintained insulin resistance, the risk is likely to be offset by the increase
by progressive increases in basal and postprandial insulin in caloric expenditure with the higher metabolic rate neces-
concentrations to counteract the rising insulin resistance.20 sary to maintain a twin/multiple gestation.25 Twin gestations
At the end of pregnancy, the basal and 24 hours mean insu- have higher metabolic rates than a singleton pregnancy, with
lin concentrations may be double those before pregnancy, maternal resting energy expenditure 10% greater, resulting
with the first and second phases of insulin secretion three in a 40% increase in caloric requirements.17 In addition, BMI
times greater.20–22 and weight gain during pregnancy, which are important
The increases in insulin resistance mean slightly higher contributors to the insulin resistance, are also higher in twin
postprandial concentrations of metabolic fuels including than singleton pregnancies.26–29
Gestational diabetes pregnancies were more prone to GDM.6 Therefore these

women should be screened again in late pregnancy too.
Glycemic goals Yogev compared the performance of the 50 g GCT in 529
Glucose control is essential during pregnancy for moth- twin and 14,268 singleton pregnancies followed at a single
ers with diabetes and their infants, as the risks for adverse center in Israel, through the Clalit Health Services data-
health outcomes and quality of life correlate with mater- base.33 This study comprises one of the largest cohorts in a
nal glucose levels. There are no specific recommendations single center addressing the issue of the GCT in twin preg-
on glycemic goals in twin diabetic pregnancies, but recent nancies. The authors found that women with twin pregnan-
research suggests weekly self-monitoring blood glucose cies were approximately twice as likely to have an abnormal
(SMBG) + HbA1C provides more complete data for glucose GCT and were also more likely to have a normal OGTT than
control during pregnancy, as SMBG alone can miss certain women with singleton pregnancies (Tables 21.2 and 21.3).
high glucose values.30 Women in the twins group had higher mean GCT results
(104.7 ± 28 vs. 98.5 ± 25 mg/dL, p < 0.001), more GCT results
higher than 130 mg/dL (20.2% vs. 11.8%, p < 0.001) (odds
ratio [OR] 2.2, 95% confidence interval [CI] 1.7–2.7) or
Multiple pregnancies and GDM GCT >140 mg/dL (13.8% vs. 9.6%, p = 0.001) (OR 1.9, 95% CI
Diagnosis 1.5–2.5) even after adjustment for maternal age, parity, and
Considering the various differences between twin and sin- fetal sex. The PPV for a GCT >140 mg/dL was significantly
gleton pregnancies, one might assume that the accuracy lower in the twins group for either one or more or two or
and characteristics of the diagnostic tests would differ too. more abnormal values in the OGTT (Table 21.3). The overall
Surprisingly, however, data in support of this hypothesis are rates of GDM, when diagnosed with CGT >200, were similar
limited and not concordant.31 in the twin and singleton groups (3.0% vs. 3.6%).
In a retrospective study of 2554 Italian women screened Although these were all retrospective studies, with no
for gestational diabetes, Corrado et al. found that twin ges- information on potential confounders, the 50 g GCT appears
tation was independently associated with an increased risk to give a higher false-positive rate and a lower PPV in twins
for an abnormal 50 g glucose challenge test (GCT) result. than singleton pregnancies, since a mild form of glucose
However, the positive predictive value (PPV) of an abnormal intolerance does not translate into a difference in the rate
GCT for gestational diabetes was lower in twin gestations, of GDM. Some authors suggest that the diabetogenic effect
as the overall rate of GDM was similar in singleton and twin resulting from the rising levels of placental hormones is par-
pregnancies.32 tially counteracted by the increased glucose demand due to
Similarly, in a multiethnic population, Simmons et al. the presence of multiple fetuses and the higher basal meta-
reported that women with twin pregnancy had higher 1-hour bolic rate.25
GCT results, but not significantly higher rates of positive
GCT. In the small numbers of women with an oral glucose Frequency of gestational diabetes in twin pregnancies
tolerance test (OGTT), the incidence of GDM was not signif- Few studies have examined the risk of GDM in women
icantly higher than in singleton pregnancies. However, as the carrying multiple pregnancies, and their findings are con-
significantly wider screening for GDM among women with flicting. This could be partially due to differences in the cri-
twin pregnancies (77.8% vs. 50.6%) should have reduced the teria for diagnosing GDM (World Health Organization,34
difference in GDM risk, these authors concluded that twin Australasian Diabetes in Pregnancy Society ADIPS,35
Table 21.2 Common screening tests for gestational diabetes mellitus in twin pregnancies
50 g GCT Abnormal GCT OGTT+ after GCT +

Twin versus Twin versus (≥140 mg/dL)
Authorref (country) Study Year singleton (no.) singleton (%) Twin versus singleton (%)
Corrado 32 Retrospective 2003 70 versus 2,484 31.4 versus 20.0 18.1 versus 16.1, NS
(Italy) Single center
Simmons6 Retrospective 2002 54 versus 4,885 17.9 versus 12.6 50 versus 28.3, NS
(New Zealand) Multicenter
Multiethnic
Yogev33 Retrospective 2014 529 versus 14,268 13.8 versus 9.6 OGTT
(Israel) Single center ≥1 abnormal result
21.1% versus 33.8% p = 0.03
≥2 abnormal results
12.7% versus 23% p = 0.04
Abbreviation: GCT, glucose challenge test.

Multiple pregnancies and GDM 173
Table 21.3 OGTT results in twin and singleton pregnancies with an abnormal
glucose challenge test result (>140 mg/dL)
Twins Singletons
Outcome (71) (1222) p value
Mean gestational week at testing 27.2 ± 1.7 27.5 ± 2.6 0.3
Result 1 (fasting)
Mean (mg/dL) 75.0 ± 11.7 75.2 ± 14.5 0.9
>95 mg/dL 2 (2.8) 74 (6.1) 0.3
Result 2 (60 min)
Mean (mg/dL) 160.0 ± 33.9 162.0 ± 31.8 0.6
>180 mg/dL 13 (18.3) 338 (27.7) 0.09
Result 3 (120 min)
Mean (mg/dL) 129.3 ± 30.7 128.7 ± 29.5 0.9
>155 mg/dL 11 (15.5) 229 (18.7) 0.5
Result 4 (180 min)
Mean (mg/dL) 93.6 ± 15.8 97.3 ± 21.2 0.1
>140 mg/dL 3 (4.2) 47 (3.4) 0.7
No abnormal results 56 (78.9) 809 (66.2) 0.03
(false-positive GCT)
One or more abnormal results 15 (21.1) 413 (33.8) 0.03
Two or more abnormal results 9 (12.7) 281 (23.0) 0.04
Source: Modified from Yogev, Y. et al., J. Matern. Fetal. Neonatal Med., 27, 57, 2014. With permission.
Note: Results are presented as mean ± SD, or no. (%).
Abbreviation: GCT, glucose challenge test.
National Diabetes Data Group,36 and Carpenter and was no information on the clinical subtypes of diabetes in
Coustan,37 as well as to the small populations studied and pregnancy. However, during the study reference period, uni-
differences in study design). versal screening for GDM was in place. The 3-hour 100 g
The oldest studies, in the 1980s, used an i.v. glucose toler- OGTT was routinely used for the diagnosis of GDM, accord-
ance test to evaluate the carbohydrate metabolism, but the ing to the diagnostic criteria recommended by the American
results in twin and singleton pregnancies were not univo- Diabetes Association or the National Diabetes Data Group.
cal.23,38,39 Subsequently, however, with the OGTT, some stud- The actual role of multiple pregnancies in the frequency
ies have indicated an increase in the incidence of GDM in of GDM is also hard to establish because of the confound-
twin pregnancies,31,40–42 while others find no difference from ing effect of ART and other infertility treatments that, as we
singleton pregnancy rates43–45 (Table 21.4). In the United have seen earlier, often give rise to multifetal pregnancies.
States, a retrospective cohort-based analysis of 281,505 twin Many recent studies report higher rates of GDM in pregnant
births between 1995 and 2000 confirmed a 3.5% rate of dia- women treated with various types of ART or ovarian hor-
betes46; mothers were more likely to be black (6.9% vs. 3.8%) monal stimulation; this finding, however, was seen both in
and older than 35 (28.8% vs. 18.5%). Unfortunately, there singleton and multiple pregnancies, suggesting a direct effect
Table 21.4 Prevalence of gestational diabetes mellitus in singleton and multiple pregnancies, based on oral
glucose tolerance test
No. pregnancies % GDM

Authorref Study Year Multiple/singleton Diagnostic criteria Multiple/singleton p value
Naicker45 Case control 1983 26/26 Response to GTT 50 g Glucose similar response NS
Wein41 Retrospective 1992 798/61,914 WHO 7.4/5.6 0.012
Henderson44 Case control 1995 130/130 NDDG 5.8/5.4 NS
Roach40 Retrospective 1998 71/824 ADIPS AU 15.5/9.1 <0.001
Schwartz31 Retrospective 1999 429/29,215 NDDG 7.7/4.1 <0.05
Simmons6 Retrospective 2002 54/4,885 ADIPS NZ 11.9/5.1 0.05
Buhling43 Case/control 2003 89/178 C&C 3.4/3.4 NS
Rauh-Hain42 Retrospective 2008 553/22,503 NDDG 4.0/2.3 0.01
Risk ratio between 2007 and 2009 in the Australian National Perinatal
M-H, fixed, 95% Cl Data Collection.57 Comparing ART and non-ART pregnan-
cies, the overall frequency of GDM was significantly higher
in the ART group (7.6% vs. 5.0%). Looking then at multiple
pregnancies, even though the prevalence of GDM was higher
in ART than non-ART cases, multivariate analysis did not
show a significant difference (Table 21.5).
These data suggest that the higher risk of GDM after ART
is not due to an effect on multifetality and that other factors
must be considered—for example, the cause of infertility.
GDM metabolic management in multiple pregnancies

Considering the high prevalence of twin pregnancies, and
the frequency of GDM in these mothers, it is surprising and
disappointing that there is so little data on the best metabolic
approach to multiple pregnancies complicated by glucose
intolerance. Only one study, by Schwartz et al., published
in 1999 in American Journal of Obstetrics and Gynecology
(AJOG),31 addressed this question in detail. They examined
0.01 0.1 1 10 100
the obstetric database of the Detroit Medical Center–Sinai
Spontaneous conception IVF/ICSI Hospital between January 1, 1990, and June 30, 1998. In about
Figure 21.3 Relative risk of gestational diabetes in assisted
30,000 deliveries, GDM was found in 1222 (4.1%) of single-
reproductive technology compared with spontaneous singleton and 33 (7.7%) of twin pregnancies. Both the proportion
ton conceptions. (From Pandey, S. et al., Hum. Reprod. of women requiring insulin and the average daily insulin
Update, 18, 485, 2012. With permission.) doses were only minimally higher in twin than singleton
GDM pregnancies (30.3% vs. 24.4% and 67.8 U [units] vs.
of these therapies on glucose tolerance in pregnancy, sepa- 56.2 U, respectively, NS [not significant]). Fasting and 2-hour
rate from their part in multifetal gestations. A meta-analysis postprandial SMBG were also similar in the two groups, well
of obstetric and perinatal outcomes in singleton pregnan- below the therapeutic target values, set at 100 mg/dL in the
cies resulting from IVF/ICSI, by Pandey and colleagues, was fasting state and 120 mg/dL 2 hours after meals.
published in 201247; in six studies reporting GDM, regarding Other studies also report no differences in the therapeutic
13,399 pregnancies,48–53 the relative risk of GDM after ART approach to GDM in singleton and multiple pregnancies,58 at
procedures was 1.48 compared with spontaneous conception least for twins; higher-order multiple pregnancies might pres-
(Figure 21.3). It is worth noting that the higher risk persisted ent different, more serious problems, due to the larger placen-
in subgroups of matched cohort studies and when sensitivity tal mass resulting in higher production of placental-derived
analysis was done only in good-quality studies. hormones that may lead to greater insulin resistance and,
Other studies in recent years54–56 mostly confirm the consequently, metabolic alterations. No studies are available,
higher risk of GDM in singleton pregnancies with ART. Of however, on metabolic control in GDM complicating triplets
particular interest is the article by Wang et al. on a retro- and up, probably because of the small numbers—although
spective cohort of more than 400,000 pregnancies registered they are not so rare any more in most developed countries.59
Table 21.5 Prevalence of gestational diabetes mellitus in pregnancies conceived with assisted
reproductive technology procedures, compared with spontaneous conception (nonassisted
reproductive technology)
Non-ART ART
Total GDM Total OR AORa
(no.) (%) (no.) GDM (%) (95% CI) (95% CI)
All mothers 386.660 5.0 13.732 7.5 1.53 (1.43–1.64) 1.26 (1.18–1.36)
Singleton 381.402 7.3 12.105 8.8 1.22 (0.99–1.49) 1.18 (0.94–1.48)
Twin 5.208 5.0 1.571 7.6 1.55 (1.45–1.65) 1.28 (1.20–1.37)
Source: Modified from Wang, Y.A. et al., Hum. Reprod., 28(9), 2554, 2013. With permission.
Note: Data from the Australian National Perinatal Data Collection (NDPC), 2007–2009.
Abbreviations: AOR, adjusted odds ratio; CI, confidence interval; OR, odds ratio.
a Adjusted for age, parity, body mass index, health insurance, smoking during pregnancy, and essential hypertension.
Multiple pregnancies and GDM 175
Coming back to simple twin GDM pregnancies, prob- present high rates of hypertensive disorders and preeclamp-
ably the period of metabolic derangement after diagnosis sia,66 and the most frequent perinatal complications are
(few weeks) is too short to make it necessary to introduce excessive fetal growth and macrosomia,67 with consequent
significant changes in the therapeutic approach usually increased frequency of birth trauma, shoulder dystocia and
adopted in singleton pregnancies complicated by the same maternal morbidity due to operative delivery, and preterm
problem. This is not true, however, for twin pregnancies in deliveries. Perinatal mortality rates (stillbirths and neonatal
mothers who already have type 1 diabetes. death) may also be higher than normal.
In a retrospective study of 15 twin pregnancies with type 1 On the other hand, as reported in Table 21.6, a twin or
diabetes followed at the same center in Copenhagen, Callesen multiple pregnancy is itself a risk factor for several perina-
and colleagues60 investigated the insulin requirement during tal negative outcomes,18,68,69 such as low birth weight (LBW)
gestation, compared with a group of 108 singleton pregnan- and prematurity,70 stillbirth and neonatal death,71–73 mater-
cies with type 1 diabetes. With a HbA1c target ≤5.6% in the nal hypertension and preeclampsia, and hemolysis, elevated
second part of pregnancy, they reported small increases in liver enzymes, and low platelets syndrome.74,75 It is therefore
insulin doses from prepregnancy until 8 weeks, a decrease interesting to see how the combination of these two poten-
from 8 to 14 weeks, a substantial increase from 14 to 27 weeks, tially dangerous conditions—GDM and twin pregnancy—
and a stable insulin requirement from 27 to 33 weeks. Thus, influence pregnancy outcomes. The few studies reported in
the insulin requirement before pregnancy until 14 weeks was the last decade have produced controversial results.
comparable in twin and singleton pregnancies; the total insu- A small group of 28 twin pregnancies with GDM was
lin requirement increase doubled weekly between 14 and 27 examined in 2003 by Moses et al. in Australia76 in compari-
gestational weeks in twin pregnancies; this was explained by son with 29 control twin normoglycemic pregnancies; no
the growth of the total fetal and placental mass. From 27 to differences were found in pregnancy outcome (birth weight,
33 weeks, the insulin requirement remained fairly stable com- Apgar score, gestational week of delivery), with the excep-
pared with the continuous steady rise usually seen in single- tion of a higher rate of elective cesarean sections for twin
ton diabetic pregnancies; this was very likely due to the more pregnancies.
limited placental and fetal growth often seen toward the end In 2006, a Korean group77 examined 99 twin pregnancies
of twin pregnancies, possibly associated with some stagnation delivered at Cheil General Hospital in Seoul in 1998–2002,
in hormonal production, although the final placental weight in a case–control study comparing 33 women with GDM
was 74% higher in twins than singleton pregnancies with and 66 nondiabetic mothers. Again, no significant differ-
type 1 diabetes (1220 vs. 700 g). ences were found in birth weight, Apgar score, respiratory
As placental growth can be expected to follow simi- distress syndrome, and other possible complications such
lar patterns in twin pregnancies with GDM, the smaller as meconium aspiration pneumonia, transient tachypnea of
increase in insulin needs in these women might conceivably newborn, hyperbilirubinemia, hypoglycemia, hypocalcemia,
be attributable to a lower degree of glycometabolic alteration, and congenital anomalies. Intrauterine fetal death occurred
and to its shorter duration, insufficient to cause a substantial in two pregnancies in each group (NS). Women with GDM
difference from singleton pregnancies. were followed with a glycemic target of 60–95 mg/dL fasting,
and 120 mg/dL 2 hours postprandial; insulin (added when at
least 20% of measurements were above the target) was nec-
Glucose control and pregnancy outcome essary in 6 out of 33 (18%) cases, achieving optimal glucose
GDM generally raises the risks of adverse complications for control in all cases.
both mother and child,61,62 but these risks can be reduced Three other studies have since been published on the
with appropriate diagnosis and treatment.63–65 Pregnancies same subject: in 2011 in Portugal by Simões,78 in 2012 by
complicated by untreated or suboptimally controlled GDM González González79 in nine tertiary university centers in
Table 21.6 Obstetric and perinatal complications in multiple pregnancies
Outcome Singleton Twins Triplets Authorref

Birth weight <1500 g (%) 1.1 9.8 35.3 Martin5
Birth weight <2500 g (%) 6.3 56.3 94.4 Martin5
<32 weeks gestation (%) 1.6 11.3 36.5 Martin5
<37 weeks gestation (%) 10.0 57.3 93.4 Martin5
Cerebral palsy (‰ live births) 1.6 7.0 28.0 Petterson71
Infant mortality (‰ live births) 5.4 23.6 52.5 Luke68
Maternal hypertensive disorders (%) 6.5 12.7 20.0 Day75
HELLP syndrome 0.2 0.9 2.1 Day75
Abbreviation: HELLP, hemolysis, elevated liver enzymes, and low platelets syndrome.
Spain, and in 2013 by Luo et al.,46 using 1995–2000 data of complications and patient or provider preference, there were
the U.S. National Center for Health Statistics (NCHS). 2.5 times the number of cesarean sections compared with
Simões et al.78 reported a retrospective case–control study singleton deliveries.
on 105 twin pregnancies with GDM matched, with a 3:1
ratio, with 315 non-GDM controls. The only significant dif-
ference was in the prevalence of respiratory distress at birth Discussion and conclusions
and jaundice, which were more frequent in GDM pregnan-
cies; a higher rate of perinatal mortality in the same group The substantial increase in the frequency of multiple preg-
did not reach statistical significance. No information was nancies recorded in the last few decades, mainly as a result
available on the glucose control in diabetic mothers. of the growth in the use of ART techniques and therapies
The Spanish group conducted a retrospective observa- linked to ovarian stimulation, led us to expect some negative
tional study on 534 twin pregnancies, 257 with GDM and consequences related to the potential “diabetogenic” effect of
277 with normal glucose tolerance.79 GDM was associated the growing placental mass. There was worry not only about
with a higher risk of hypertensive complications, prematu- an increase in the prevalence of GDM and other alterations
rity, and macrosomia (not large for gestational age [LGA]), in the carbohydrate metabolism in these cases, but also
but a significantly lower risk of small-for-gestational-age about unfavorable outcomes in women with both these inde-
(SGA) infants. Prematurity was related not to GDM but to pendent risk factors—GDM and multiple pregnancy.
other pregnancy complications. Apgar scores, rate of admis- In actual fact, the figures on the prevalence of GDM in
sion to neonatal intensive care units, and perinatal mortality multiple pregnancies are discordant. There does appear to be a
were similar in mothers with GDM and controls. greater frequency of minor glucose tolerance disorders, partic-
The study by Luo et al.46 is important because of the huge ularly with frequently positive 50 g OGTT, but there seems to be
number of pregnancies examined. From the NCHS matched no rise in the diagnoses of GDM. There may in fact, therefore,
multiple-birth data for 1995–2000, the authors obtained be some sort of compensatory effect, resulting from the greater
information on about 20,000 twin pregnancies complicated calorie consumption and higher metabolic rate, which would
by diabetes (the type of diabetes was not specified, but prob- reduce the consequences on the material carbohydrate metabo-
ably mainly GDM), and more than 540,000 nondiabetic lism of the altered endocrine milieu due to the high concentra-
twin pregnancies. A previous article by the same group80 tions of placental hormones with “anti-insulin” action.
had reported a potential protective effect of diabetes against Similarly, in women with GDM complicating a multiple
neonatal death in twin pregnancies. Therefore, to exclude pregnancy, the obstetric and perinatal outcomes seem no
the hypothesis that this apparently positive result could different from single pregnancies. This obviously depends on
“mask” serious problems preceding delivery causing an how metabolic control is maintained after diagnosis, but it
increase in fetal deaths, this study was designed to assess does not seem any harder to manage diabetes during a mul-
the consequences of diabetes complicating twin pregnancies tiple pregnancy. The proportions of women requiring insulin
on perinatal mortality, considering both fetal and neonatal are in fact close to those for single pregnancy, and the insu-
deaths. The results, however, were not univocal: the risks of lin doses needed in such cases do not seem any higher; the
stillbirths and neonatal death were lower in very preterm or situation is different, however, in multiple pregnancies with
very-low-birth-weight twin diabetic pregnancies, but not pregestational diabetes.
among neonates with normal birth weight, where these rates There even seems to be a protective effect against peri-
were higher in diabetic women. Moreover, the lower risk of natal mortality, at least for preterm births and very-low-
perinatal death in diabetic twin pregnancies was observed birth-weight neonates, as if slight hyperglycemia in early
for vaginal but not cesarean section births. Overall, however, pregnancy gave some sort of “survival advantage.” Luo’s
perinatal death was less frequent in twin diabetic than non- findings, however, are surprising and certainly call for fur-
diabetic pregnancies (2.1% vs. 3.3%), confirming the trend ther confirmation, with greater details of the type of meta-
evidenced in 2011. bolic alteration present.
Looking at secondary outcomes, in diabetic pregnancies, In conclusion, the question of how GDM and multiple
mothers were more likely to have reported any other mater- fetuses are related is definitely important, especially in view
nal major illness (10.6% vs. 4.6%) or to have a cesarean sec- of the current “epidemic” of multiple pregnancies. However,
tion delivery (60.8% vs. 51.9%). Preterm (61.4% vs. 56.1%) or findings so far on the interaction between the two condi-
LGA (13.3% vs. 9.3%) births and congenital anomalies (2.8% tions seem reassuring. Once more, the main factor is the
vs. 2.1%) were also more frequent in diabetic pregnancies. importance of metabolic control on the outcome of a preg-
Extremely preterm births were slightly less frequent (10.5% nancy complicated by GDM, so it is essential to achieve and
vs. 12.1%) and slightly preterm births more frequent (50.9% maintain optimal glycemic levels, regardless of the number
vs. 44.0%) in diabetic pregnancies. LBW (52.5% vs. 54.6%) of fetuses. Diagnostic and therapeutic strategies regard-
and SGA (9.9% vs. 10.8%) infants were slightly less fre- ing the carbohydrate metabolism in women with multiple
quent in diabetic pregnancies. Among the extreme preterm pregnancies should therefore be no different from the uni-
births in diabetic and nondiabetic pregnancies, SGA was versally accepted approach for single pregnancies, which
slightly less frequent (10.0% vs. 11.7%), while LGA was simi- have achieved progressive improvements in maternofetal
lar (10.1% vs. 9.7%). Given the higher rates of twin pregnancy outcomes in recent years.
References 177
REFERENCES
1. Avraham S, Seidman DS. The multiple birth epidemic: Revisited. 27. Chu SY, D’Angelo DV. Gestational weight gain among US women
J Obstet Gynecol India 2012; 62: 386–390. who deliver twins, 2001–2006. Am J Obstet Gynecol 2009; 200:
2. Ananth CV, Chauhan SP. Epidemiology of twinning in developed 390.e1–390.e6.
countries. Semin Perinatol 2012; 36: 156–161. 28. Chu SY, Callaghan WM, Bish CL et al. Gestational weight gain
3. American College of Obstetricians and Gynecologists. ACOG by body mass index among US women delivering live births,
Practice Bulletin No 144: Multifetal gestations: Twin, triplet, and 2004–2005: Fueling future obesity. Am J Obstet Gynecol 2009;
higher-order multifetal pregnancies. Obstet Gynecol 2014; 123: 200: e271–e277.
1118–1132. 29. Fox NS, Rebarber A, Roman AS et al. Weight gain in twin preg-
4. Martin JA, Hamilton BE, Osterman MJ. Three decades of twin nancies and adverse outcomes: Examining the 2009 Institute of
births in the United States, 1980–2009. NCHS Data Brief 2012; Medicine guidelines. Obstet Gynecol 2010; 116: 100–106.
80: 1–8. 30. Castorino K, Jovanovic L. Pregnancy and diabetes management:
5. Martin JA, Hamilton BE, Ventura SJ et al. Births: Final data for Advances and controversies. Clin Chem 2011; 57: 221–230.
2011. Natl Vital Stat Rep 2013; 62: 1–17. 31. Schwartz DB, Daoud Y, Zazula P et al. Gestational diabetes mel-
6. Simmons D, Yapa M. Association between twin pregnancy and litus: Metabolic and blood glucose parameters in singleton versus
hyperglycemia in a multiethnic community in New Zealand. twin pregnancies. Am J Obstet Gynecol 1999; 181: 912–914.
Diabetes Care 2002; 25: 934–935. 32. Corrado F, Caputo F, Facciola F et al. Gestational glucose intoler-
7. Eriksson AW, Fellman J. Temporal trends in the rates of multiple ance in multiple pregnancy. Diabetes Care 2003; 26: 1646.
maternities in England and Wales. Twin Res Hum Genet 2007; 33. Yogev Y, Eisner M, Hiersch L et al. The performance of the
10: 626–632. screening test for gestational diabetes in twin versus singleton
8. Eriksson AW, Fellman J. Demographic analysis of the variation in pregnancies. J Matern Fetal Neonatal Med 2014; 27: 57–61.
the rates of multiple maternities in Sweden since 1751. Hum Biol 34. World Health Organization. WHO Expert Committee on
2004; 76: 343–359. Diabetes Mellitus: Second Report. Technical Report Series, no.
9. Pison G, D’Addato AV. Frequency of twin births in developed 646. World Health Organization: Geneva, Switzerland, 1980.
countries. Twin Res Hum Genet 2006; 9: 250–259. 35. Martin FI. The diagnosis of gestational diabetes. Ad Hoc Working
10. D’Addato A. Secular trends in twinning rates. J Biosoc Sci 2007; Party. Med J Aust 1991; 155: 112.
39: 147–151. 36. National Diabetes Data Group. Classification and diagnosis of
11. Basili F, Cocchi M, Di Rosa A et al. Certificato di assistenza al diabetes mellitus and other categories of glucose intolerance.
parto (CeDAP). Analisi dell’evento nascita - Anno 2010; 2013: Diabetes 1979; 28: 1039–1057.
1–79. 37. Carpenter MW, Coustan DR. Criteria for screening tests for gesta-
12. Beall SA, DeCherney A. History and challenges surrounding tional diabetes. Am J Obstet Gynecol 1982; 144: 768–773.
ovarian stimulation in the treatment of infertility. Fertil Steril 2012; 38. Naidoo L, Jailal I, Moodley J et al. Intravenous glucose tolerance
97: 795–801. tests in women with twin pregnancy. Obstet Gynecol 1985; 66:
13. Kulkarni AD, Jamieson DJ, Jones HW et al. Fertility treatments 500–502.
and multiple births in the United States. New Engl J Med 2013; 39. Dwyer PL, Oats JN, Walstab JE et al. Glucose intolerance in twin
369: 2218–2225. pregnancy. Aust N Z J Obstet Gynaecol 1982; 22: 131–135.
14. Sunderam S, Kissin DM, Crawford S et al. Assisted reproduc- 40. Roach VJ, Lau TK, Wilson D et al. The incidence of gestational
tive technology surveillance—United States, 2010. Morb Mortal diabetes in multiple pregnancy. Aust N Z J Obstet Gynaecol
Wkly Rep Surveill Summ 2013; 62: 1–24. 1998; 38: 56–57.
15. Farquhar C, Rishworth JR, Brown J et al. Assisted reproduc- 41. Wein P, Warwick MM, Beischer NA. Gestational diabetes in twin
tive technology: An overview of Cochrane Reviews. Cochrane pregnancy: Prevalence and long term complications. Aust N Z J
Database Syst Rev 2013; 8: CD010537. Obstet Gynaecol 1992; 32: 325–327.
16. Institute of Medicine. Weight Gain during Pregnancy: Reexamining 42. Rauh-Hain JA, Rana S, Tamez H et al. Risk for developing gesta-
the Guidelines, National Academies Press: Washington, DC, tional diabetes in women with twin pregnancies. J Matern-Fetal
2009. Neonatal Med 2009; 22: 293–299.
17. Goodnight W, Newman R, Society of Maternal-Fetal Medicine. 43. Buhling KJ, Henrich W, Starr E et al. Risk for gestational diabetes
Optimal nutrition for improved twin pregnancy outcome. Obstet and hypertension for women with twin pregnancy compared to
Gynecol 2009; 114: 1121–1134. singleton pregnancy. Arch Gynecol Obstet 2003; 269: 33–36.
18. Young BC, Wylie BJ. Effects of twin gestation on maternal mor- 44. Henderson CE, Scarpelli S, LaRosa D et al. Assessing the risk of
bidity. Semin Perinatol 2012; 36: 162–168. gestational diabetes in twin gestation. J Natl Med Assoc 1995; 87:
19. Jauniaux E, Ben-Ami I, Maymon R. Do assisted-reproduction 757–758.
twin pregnancies require additional antenatal care? Reprod 45. Naicker RS, Subrayen KT, Jialal I. Carbohydrate metabolism in
BioMed Online 2013; 26: 107–119. twin pregnancy. S Afr Med J 1983; 63: 538–544.
20. Catalano PM, Huston L, Amini SB et al. Longitudinal changes 46. Luo ZC, Zhao YJ, Ouyang F et al. Diabetes and perinatal mortal-
in glucose metabolism during pregnancy in obese women with ity in twin pregnancies. PLOS ONE 2013; 8(9): e75354.
normal glucose tolerance and gestational diabetes mellitus. Am J 47. Pandey S, Shetty A, Hamilton M et al. Obstetric and perinatal
Obstet Gynecol 1999; 1809: 903–916. outcomes in singleton pregnancies resulting from IVF/ICSI: A sys-
21. Feldt-Rasmussen U, Mathiesen ER. Endocrine disorders in preg- tematic review and meta-analysis. Hum Reprod Update 2012; 18:
nancy: Physiological and hormonal aspects of pregnancy. Best 485–503.
Pract Res Clin Endocrinol Metabol 2011; 25: 875–884. 48. Delgadillo JCB, Mendez LMA, Chavez VG et al. Perinatal results
22. Lain KY, Catalano PM. Factors that affects maternal insulin resis- in pregnancies obtained with embryo transfer in vitro fertil-
tance and modify fetal growth and body composition. Metab ization: A case–control study. Ginecol Obstet Mex 2006; 74:
Syndr Relat Disord 2006; 4: 91–100. 626–639.
23. Spellacy WN, Buhl WC, Birk SA. Carbohydrate metabolism 49. Ochsenkuhn R, Strowitzki T, Gurtner M et al. Pregnancy compli-
in women with twin pregnancy. Obstet Gynecol 1980; 55: cations, obstetric risks, and neonatal outcome in singleton and
687–693. twin pregnancies after GIFT and IVF. Arch Gynecol Obstet 2003;
24. Allmog B, Shehata F, Aljabri S et al. Placenta weight percentile 268: 256–261.
curves for singleton and twins deliveries. Placenta 2011; 32(1): 50. Reubinoff BE, Samueloff A, Ben-Haim M et al. Is the obstet-
58–62. ric outcome of in vitro fertilized singleton gestations different
25. Shinagawa S, Suzuki S, Chihara H et al. Maternal basal meta- from natural ones? A controlled study. Fertil Steril 1997; 67:
bolic rate in twin pregnancy. Gynecol Obstet Invest 2005; 60: 1077–1083.
145–148. 51. Schieve LA, Cohen B, Nannini A et al. A population-based study
26. Lantz ME, Chez RA, Rodriguez A et al. Maternal weight gain pat- of maternal and perinatal outcomes associated with assisted
terns and birth weight outcome in twin gestation. Obstet Gynecol reproductive technology in Massachusetts. Matern Child Health
1996; 87: 551–556. J 2007; 11: 517–525.
52. De Geyter C, De Geyter M, Steimann S et al. Comparative birth 66. Colatrella A, Loguercio V, Mattei L et al. Hypertension in diabetic
weights of singletons born after assisted reproduction and natural pregnancy: Impact and long-term outlook. Best Pract Res Clin
conception in previously infertile women. Hum Reprod 2006; Endocrinol Metab 2010; 24: 635–651.
21: 705–712. 67. Lapolla A, Dalfrà MG, Bonomo M et al. Gestational diabetes mel-
53. Sazonova A, Kllen K, Thurin-Kjellberg A et al. Factors affect- litus in Italy: A multicenter study. Eur J Obstet Gynecol Reprod
ing obstetric outcome of singletons born after IVF. Hum Reprod Biol 2009; 145: 149–153.
2011; 26: 2878–2886. 68. Luke B, Brown MB. Contemporary risks of maternal morbidity
54. Silberstein T, Levy A, Harlev A et al. Perinatal outcome of preg- and adverse outcomes with increasing maternal age and plurality.
nancies following in vitro fertilization and ovulation induction. J Fertil Steril 2007; 88: 283–293.
Matern Fetal Neonatal Med 2014; 27(13): 1316–1319. 69. Conde-Agudelo A, Belizan JM, Lindmark G. Maternal morbid-
55. Farhi A, Reichman B, Boyko V et al. Maternal and neonatal health ity and mortality associated with multiple gestations. Obstet
outcomes following assisted reproduction. Reprod Biomed Gynecol 2000; 95: 899–904.
Online 2013; 26: 454–461. 70. Institute of Medicine. Preterm Birth: Causes, Consequences, and
56. Stojnic J, Radunovic N, Jeremic K et al. Perinatal outcome of sin- Prevention, National Academies Press: Washington, DC, 2007.
gleton pregnancies following in vitro fertilization. Clin Exp Obstet 71. Petterson B, Nelson KB, Watson L et al. Twins, triplets and cere-
Gynecol 2013; 40: 277–283. bral palsy in births in Western Australia in the 1980s. Br Med
57. Wang YA, Nikravan R, Smith HC et al. Higher prevalence of ges- J 1993; 307: 1239–1243.
tational diabetes mellitus following assisted reproduction tech- 72. Scher AI, Petterson B, Blair E et al. The risk of mortality or cere-
nology treatment. Hum Reprod 2013; 28(9): 2554–2561. bral palsy in twins: A collaborative population-based study.
58. Norwitz ER, Edus V, Shin Park J. Maternal physiology and com- Pediatr Res 2002; 52: 671–681.
plications of multiple pregnancy. Semin Perinatol 2005; 29: 73. Glinianaia SV, Obeysekera MA, Sturgiss S et al. Stillbirth and
338–348. neonatal mortality in monochorionic and dichorionic twins:
59. Mordel N, Ezra Y, Dorembus D et al. Triplets are not so rare any A population-based study. Hum Reprod 2011; 26: 2549–2557.
more. J Perinat Med 1992; 20: 117–122. 74. Sibai BM, Hauth J, Caritis S et al. Hypertensive disorders in twin
60. Callesen N, Ringholm L, Stage E et al. Insulin requirements in versus singleton gestations. National Institute of Child Health
Type 1 diabetic pregnancy. Do twin pregnant women require and Human Development Network of Maternal-Fetal Medicine
twice as much insulin as singleton pregnant women? Diabetes Units. Am J Obstet Gynecol 2000; 182: 938–942.
Care 2012; 35: 1246–1248. 75. Day MC, Barton JR, O’Brien JM et al. The effect of fetal number
61. Langer O, Yogev Y, Most O et al. Gestational diabetes: The on the development of hypertensive conditions of pregnancy.
consequences of not treating. Am J Obstet Gynecol 2005; 192: Obstet Gynecol 2005; 106: 927–931.
989–997. 76. Moses RG, Webb AJ, Lucas EM et al. Twin pregnancy outcomes
62. HAPO Study Cooperative Research Group. Hyperglycemia for women with gestational diabetes mellitus compared with
and adverse pregnancy outcomes. N Engl J Med 2008; 358: glucose tolerant women. Aust N Z J Obstet Gynaecol 2003; 43:
1991–2002. 38–40.
63. Bonomo M, Cetin I, Pisoni MP et al. Flexible treatment of ges- 77. Cho HJ, Shin JS, Yang JH et al. Perinatal outcome in twin preg-
tational diabetes modulated on ultrasound evaluation of intra- nancies complicated by gestational diabetes mellitus: A compar-
uterine growth: A controlled randomized clinical trial. Diabetes ative study. J Korean Med Sci 2006; 21: 457–459.
Metab 2004; 30: 237–244. 78. Simões T, Queirós A, Correia L et al. Gestational diabetes mel-
64. Crowther CA, Hiller JE, Moss JR et al., for the Australian litus complicating twin pregnancies. J Perinat Med 2011; 39:
Carbohydrate Intolerance Study in Pregnant Women (ACHOIS) 437–440.
Trial Group. Effect of treatment of gestational diabetes mellitus 79. González González NL, Goya M, Bellart J et al. Obstetric
on pregnancy outcomes. N Engl J Med 2005; 352: 2477–2486. and perinatal outcome in women with twin pregnancy and
65. Landon MB, Spong CY, Thom E et al., Eunice Kennedy Shriver gestational diabetes. J Matern Fetal Neonatal Med 2012; 25:
National Institute of Child Health and Human Development 1084–1089.
Maternal-Fetal Medicine Units Network. A multicenter, random- 80. Luo ZC, Simonet F, Wei SQ et al. Diabetes in pregnancy may
ized trial of treatment for mild gestational diabetes. N Engl J Med differentially affect neonatal outcomes for twins and singletons.
2009; 361: 1339–1348. Diabet Med 2011; 28: 1068–1073.
22 Glycemic goals in diabetic
pregnancy and defining
“good control”: Maternal
and fetal perspective
Liran Hiersch and Yariv Yogev
Introduction Glucose thresholds and perinatal

The association between diabetes during pregnancy and complications
the risk of adverse maternal and fetal outcome is well
Several authoritative bodies have recommended varying
established. Maternal complications include spontaneous
levels of glycemia that need to be targeted in pregnant

abortions, preterm deliveries, preeclampsia, nephropa-
patients with diabetes. The sources of these recommenda-
thy, cesarean section, and among others birth trauma.
tions utilized the concept of isolated normality based on
The main fetal and neonatal complications are congeni-
nondiabetic glucose profiles and are usually well-defined
tal anomalies, deviant fetal growth, birth-related trauma,
cutoff levels.10–12 Yet, it appears that one cannot address
metabolic abnormalities, and stillbirth. Moreover, evolving
glucose as a dichotomous variable but as a continuous one,
evidence indicate that diabetes, among other components
with type of complications and rate depending upon the
of the metabolic syndrome, is a significant risk factor for
level of glycemic control. The association between glycemic
childhood and adult obesity and cardiovascular d isease,
control and several main perinatal complications will be
due to in utero programming mechanisms including
further discussed in the following paragraphs.
altered organ development, cellular signaling responses,
and epigenetic modifications of gene expression.1–4
Not only that improving glycemic control reduces the Perinatal mortality
risk for adverse outcome in pregestational diabetes, 5,6 but The most devastating adverse fetal outcome is perinatal mor-
also in gestational diabetes mellitus (GDM), treatment tality. However, since congenital anomalies, which are also
and achieving desired level of glycemic status effectively important complications of maternal diabetes, are a major
diminishes the risk for complications.7–9 Yet several issues contributing factor in the overall perinatal mortality rate,
should be addressed when discussing the relation between it is important to adjust for the effect of congenital malfor-
glycemic control and adverse pregnancy outcome. First, mations when calculating fetal and neonatal death rates.
how is good glycemic control defined? What is the glu- Karlsson and Kjellmer evaluated the effect of the degree of
cose threshold that should be reached in order to dimin- maternal glycemia (expressed as the mean daily blood glu-
ish the risk for adverse outcome? Second, is maintaining cose value) on perinatal mortality in 179 women with pre-
glycemic control according to the recommendations and existing diabetes.13 Patients were divided into three groups
guidelines can truly abolish the increased risk for com- with the following mean blood glucose: <100, 100–150, and
plications as compared to the general population? Finally, >150 mg/dL. Perinatal mortality was 3.8%, 16%, and 24%,
how glucose status should be monitored during pregnancy respectively, although no correction was made for congenital
in regard to the methods used, diurnal occasions, and anomalies. Yet these data suggest that strict glycemic con-
frequencies? trol (<100 mg/dL) is necessary for diminishing the risk for
179
180 Glycemic goals in diabetic pregnancy and defining “good control”
perinatal mortality in this population. However, perinatal infants based on its mathematical definition (>90th percen-
mortality rate is increased not only in women with preexist- tile) is 10%. Similarly, the rate of small-for-gestational-age
ing diabetes but also in women with GDM.14–16 In GDM, the (SGA) infants (<10th percentile) should be 10%. The rates
coeffect of other factors, such as the relatively older pregnant found in those of pregnant diabetic women should, there-
mother in comparison to the gravid nondiabetic population fore, be compared to the aforementioned baselines. Previous
and the presence of comorbidities (e.g., hypertension, obe- studies have reported rates of macrosomia and LGA of
sity), may also attribute to the increased risk of perinatal 10%–20% and 15%–30%, respectively.26–29 However, when
death. In a large cohort study of 4,757 women with GDM and exploring the association between the level of glycemic
10,804 nondiabetic gravidas, the rate of perinatal mortality control and fetal growth pattern, a control should be made
was evaluated.17 In this cohort, 79% of the study population for important confounding variables that were proven to be
achieved targeted levels of glycemic control (mean blood glu- associated with neonatal over-/undergrowth such as mater-
cose < 105 mg/dL). The incidence of stillbirth was 4.8/1000 nal obesity, gestational weight gain, and preeclampsia.30–32
for the subjects with GDM and 4.2/1000 for the nondiabetic Numerous studies have shown that by achieving the
subjects with neonatal death rates of 5.2/1000 and 5.3/1000, desired glycemic control, the rate of macrosomia and LGA
respectively. In seems that achievement of targeted levels of of neonates among diabetic gravidas can be reduced to
glycemic control in the GDM population can reduce the peri- closer, if not similar, to the general population.18,33–35 Landon
natal mortality to rates comparable to the general popula- et al. reported an LGA rate of 9.3% with a mean blood glu-
tion. Similar rates of perinatal mortality between nondiabetic cose <110 mg/dL in 43 patients with well-controlled type 1
gravidas and patients with intensified managed GDM were diabetes.35 With mean blood glucose levels >126 mg/dL in
reported.18 This study demonstrates again the relative protec- 32 patients, the LGA rate was 34%. Langer and associates
tive effect of controlling the abnormal levels of glycemia. found a rate of LGA in women with preexisting diabetes
The data suggest that both gestational and preexisting comparable to the rate in the general population (10%) when
diabetes are associated with increased perinatal mortality the mean self-monitoring blood glucose (SMBG) was within
when established level of glycemia is not achieved. Although 90–95 mg/dL.18 In a secondary analysis of the Metformin
the increased rate for perinatal mortality may be multifac- in Gestational Diabetes (MiG) trial evaluating 724 women
torial, it appears that a threshold of mean blood glucose with GDM who failed to achieve desired glycemic control
<100–110 mg/dL will be sufficient for the prevention of this under exclusive diet therapy, the lowest risk for LGA was
complication. when mean fasting glucose value was <90 mg/dL and when
2-hours postprandial values were <105–115 mg/dL unrelated
to maternal body mass index or treatment modality.29
Congenital anomalies However, it appears that tighter is not always better.
Since GDM is expressed in the late second or early third tri- Langer et al.36 have found that in GDM a threshold of mean
mester, which is beyond the period of fetal organogenesis, blood glucose <87 mg/dL was associated with an increased
congenital anomalies are a complication of mainly preexist- risk for SGA neonates. Similar relationship was found
ing diabetes. In studies reporting the association between between lower HbA1C levels before delivery (<5.7%) and
glycemic control and congenital anomalies, the method increased risk for SGA infants in women with type 1 diabe-
that was used for glycemic control evaluation was hemoglo- tes.28 Thus, in respect to deviant fetal growth patterns, both
bin A1C (HbA1C), mean blood glucose, mean fasting glu- upper and lower threshold should be used to define desired
cose, or mean postprandial glucose levels depending upon glycemic control.
the study.19–22 Temple et al., in a study measuring HbA1C,
found an increased rate of congenital anomalies already in
threshold of 7.5% (1% vs. 8%, p = 0.03), which translates to Metabolic abnormalities and respiratory complications
mean blood glucose of approximately 170 mg/dL.19 Others, Metabolic complications include fetal hypoglycemia, poly-
who used glucose profile to define the threshold for anoma- cythemia, hyperbilirubinemia, and hypocalcemia. The main
lies, suggested fasting plasma glucose of <120 mg/dL, post- contributor to metabolic complications is fetal hyperin-
prandial <140 mg/dL, preprandial <120 mg/dL, and overall sulinemia, which is also responsible for fetal macrosomia.
mean <110 mg/dL.22–24 In these studies, the preconception Thus, it is reasonable that the recommended thresholds for
rate of anomalies was 1%–1.5%, and in patients above these macrosomia prevention will likewise be advisable for meta-
thresholds, the rate of anomalies ranged from 6% to 12%. bolic complications. Studies in women with type 1 diabetes
With regard to congenital anomalies, the critical period for reported that a mean blood glucose <110 mg/dL was associ-
intervention and risk reduction is before or early in gestation ated with a significant reduction in rates of neonatal hypo-
making preconceptional care mandatory.25 glycemia and respiratory complications compared to poorly
controlled gravidas13 and that the rates were comparable to
the nondiabetes population.35 In GDM, good glycemic con-
Deviant fetal growth: Macrosomic and trol (mean fasting glucose < 95 mg/dL and either 1 hour post-
growth-restricted fetuses prandial <140 mg/dL or 2 hour postprandial <120 mg/dL)
In the general population, the rate of macrosomia (≥4000 g) was associated with decreased rates of neonatal hypoglyce-
is 8%–10%, and the rate of large-for-gestational-age (LGA) mia (7.1% vs. 9.3%, p = 0.031) compared to those who did
Recommended glucose thresholds and normal glucose values during pregnancy: Are they the same? 181
not reach good glycemic control.37 Langer et al. found that Preeclampsia
mean blood glucose <100 mg/dL was associated with simi- Although many studies have reported an increased risk
lar rates of metabolic complications in gravidas with GDM for hypertension and preeclampsia among women with
compared to nondiabetic controls, and that higher levels diabetes,42,51–54 some have disputed this relationship. 55,56
were associated with up to eightfold increased risk.38 As In women with pregestational diabetes, Cohen et al.
regards to respiratory complications, mean blood glucose found that first-trimester HbA1C levels were higher in
>106 mg/dL was associated with delayed appearance of bio- those who subsequently developed preeclampsia com-
chemical markers of pulmonary maturity, though without pared to those who did not (7.7% vs. 6.7%, p = 000.1). 57
correlation with clinical neonatal respiratory morbidity.39 In The association between high HbA1C levels (>8%) during
contrast, others failed to demonstrate the relation between early pregnancy and the rate of subsequent preeclampsia
glycemic control and appearance of biochemical markers of was further reported by others.6 These findings further
pulmonary maturity,40,41 except for uncontrolled diabetes emphasize the importance of prepregnancy optimization
(mean glucose >120 mg/dL) at preterm.40 of glycemic profile in women with type 1 and 2 diabetes.
The association between glycemic control and the rate
of preeclampsia was also reported in women with GDM.
Preterm delivery Yogev et al. found in 1183 women with GDM that those
Preterm delivery is considered the most important etiology with mean blood glucose <95 mg/dL had similar rates of
for neonatal morbidity and mortality. There is a comparable preeclampsia regardless of the severity of GDM (defined
incidence of preterm delivery in women with pregestational by fasting blood glucose levels on the oral glucose toler-
diabetes between types 1 and 2 diabetes (33.6% in women ance test). 54 However, in women with mean blood glucose
with type 1 vs. 32% in women with type 2).42 According to a >95 mg/dL, the rate of preeclampsia was higher in those
multicenter survey (435 pregnancies in women with preges- with fasting glucose >115 mg/dL compared to <115 mg/dL
tational diabetes), diabetes was directly implied in preterm (18.0% vs. 9.8%, OR 2.56, 95% CI 1.5–4.3). In addition, the
delivery risk particularly when first trimester HbA1C >8% severity of GDM (OR 1.7, 95% CI 1.21–2.38) was indepen
occurred and in cases of preexisting nephropathy.43 Others dently and significantly associated with an increased risk
reported that the risk of delivering preterm was more than of preeclampsia.54 Others found that fasting glucose levels
40% when HbA1C was above 7.7% in week 8.44 In a study of >90 mg/dL and/or 2-hour postprandial levels >120 mg/dL
253 gravidas with pregestational diabetes, those with sponta- were associated with increased risk for preeclampsia (OR
neous preterm delivery had at delivery higher mean HbA1C 8.40, 95% CI 4.57–15.42, p < 0.001). 58
levels (8.1% ± 1.4% vs. 7.4% ± 1.2%, p = 0.002) and blood glu- In summary, the level of glycemic control in women
cose (148 ± 30 vs. 109 ± 24 mg/dL, p = 0.043) compared to with pregestational diabetes and GDM is associated with
those delivered at term.45 However, there are differing opin- increased risk for maternal and neonatal complications.
ions if GDM as a single indicator increases the risk of pre- However, glucose values are best described as a continuous
term delivery. Several authors have suggested that there was variable, and the risk to the fetus increases in direct relation
a higher incidence of preterm delivery in association with to the increased level of maternal glycemia. In addition, the
different levels of glucose intolerance,46–48 while others found risk for each outcome measure was shown to correlate with
no difference in the rate of preterm delivery in comparison different glucose threshold. Thus, even if optimal glycemic
to the nondiabetic population.49,50 Scarce data, however, goals cannot be reached, it is important to keep on strug-
exist regarding the effect of maternal glycemic control and gling as it would still minimize the risk for complications
the risk of preterm delivery. Bar-Hava et al.49 demonstrated even if not abolish it.
in a small study similar overall rates of preterm delivery in
both subjects with GDM (n = 34 women) and without GDM
(n = 68 women). Moreover, women with GDM who delivered Recommended glucose thresholds
at term or preterm had similar glycemic profiles for both the
entire treatment period and the week preceding delivery. In and normal glucose values during
contrast, in a large study (n = 1526), Yogev et al.48 reported pregnancy: Are they the same?
that the mean blood glucose was independently associated
with the risk for spontaneous preterm delivery in women Various glycemic thresholds were recommended in patients
with GDM (odds ratio [OR] of 1.94, confidence interval [CI] with diabetes complicating pregnancy (Table 22.1).10,59–61
1.25–3.02) and that only 35% of women with spontaneous Although not identical, the cutoff levels proposed by the
preterm delivery have reached the desired glycemic control different authoritative bodies for each diurnal occasion
(defined as mean blood glucose <105 mg/dL) compared to are within a 10–15 mg/dL range. For example, while the
54% in those delivered at term (p = 0.004). It seems that in American College of Obstetricians and Gynecologists
pregestational diabetes, HbA1C level <7.7%–8% in early have recommended fasting levels of 60–90 mg/dL,10 the
gestation is desired for the prevention of preterm delivery, Fifth International Workshop-Conference on GDM recom-
whereas mean blood glucose of <105–110 mg/dL could alter mended fasting glucose level <95 mg/dL.60 However, the
the risk of preterm delivery in both gravidas with GDM and sources of these recommendations utilized the concept of
with pregestational diabetic. “mimicking” normality as they were based on nondiabetic
Table 22.1 Comparison between recommended glycemic thresholds in pregnancy and glycemic profile of
nondiabetic gravidas
Glycemic profile
Recommended glycemic thresholds in nondiabetics
Fifth International
American College Workshop- Canadian
of Obstetrics and American Diabetes Conference on Diabetes
Gynecologists10 Association59 GDM60 Association61 Yogev et al.62
Fasting (mg/dL) 60–90 <105 <96 <95 75 ± 12
Premeal (mg/dL) 60–105 GDM <95/type I or — — 78 ± 11
II DM—60–99
Postmeal (mg/dL)
1 hour <130–140 <155 <140 <140 105 ± 13
2 hours <120 <130 <120 <120 97 ± 11
Mean (mg/dL) 100 — — — 84 ± 18
Abbreviations: DM, diabetes mellitus; GDM, gestational diabetes mellitus.
glucose profiles. One of only few studies who reported data research, advances in glucose monitoring, and treatment
concerning glycemic profile in nondiabetic gravidas was modalities, women with diabetes complicating pregnancy
conducted by Yogev et al., who measured interstitial glu- still experience higher rate of adverse outcome compared to
cose levels in subcutaneous tissue every 5 minutes using the general population. Several studies have shown that even
continuous glucose monitoring (CGM) for 72 consecu- meticulous glucose control aiming and achieving strict gly-
tive hours in nondiabetic gravidas.62 They found that for cemic targets, as advised by present guidelines and recom-
nonobese women, the fasting blood glucose level was 75 ± mendations, did not eliminate adverse perinatal outcome,
12 mg/dL, the mean blood glucose level was 83.7 ± 18 mg/ especially macrosomia, compared to the general population.
dL, the postprandial peak glucose value level was 110 ± In a small study, the rate of macrosomic fetuses of 14 women
16 mg/dL, and the time interval that was needed to reach with type 1 diabetes who were treated with continuous sub-
peak postprandial glucose level was 70 ± 13 minutes. A cutaneous insulin infusion and reached a desired glycemic
similar postprandial glycemic profile was obtained for control was assessed.65 Mean HbA1C level by trimesters was
breakfast, lunch, and dinner. Obese women were charac- 6.5% ± 0.9%, 5.9% ± 0.7%, and 5.8% ± 0.6%. Yet average birth
terized by a significantly higher postprandial glucose peak weight was 3312.1 ± 750 g with macrosomia (>4000 g) rate of
value, increased 1- and 2-hour postprandial glucose levels, 35%. Evers et al. reported similar findings in a cohort of 289
increased time interval for glucose peak, and significantly gravidas with type 1 diabetes, of whom 25.3% had a birth
lower mean blood glucose during the night. No difference weight greater than 4000 g.66 The mean HbA1C (%) levels
was found in fasting and mean blood glucose between achieved were 6.5 + 1.0, 6.0 + 0.9, and 6.2 + 1.1 during the
obese and nonobese subjects. Almost similar results in first, second, and third trimester, respectively. Evidence from
nonobese nondiabetic gravidas were obtained by Parretti studies by Mello et al.67 and Howorka et al.,68 who showed
et al.63 The thresholds endorsed by various authoritative that mean glucose levels required for normal birth weight
bodies represent merely the upper 1 or 2 standard devia- was <100 mg/dL with standard deviation (SD) of 12 mg/dL
tions from mean glucose value of nondiabetic gravidas, (noting a high rate of hypoglycemia) also supports the argu-
while other factors such as maternal BMI or the specific ment that macrosomia prevention requires more strict gly-
outcome variable, which was meant to be prevented, was cemic control than usually recommended during pregnancy
not encountered. In addition, even if nondiabetic glucose complicated by diabetes.
profile is achieved by a diabetic gravida, is it sufficient for
preventing diabetes-associated complications?
Glucose monitoring: The effect on
Does achieving good glycemic control glycemic control and pregnancy
eliminate complications associated outcome
with diabetes? The ability to monitor glucose values is of historical signifi-
cance, since throughout recorded history, physicians have
The management of diabetes in pregnancy, as put forward by been familiar with diabetes and have “finger dipped” in
the “St. Vincent” declaration, was aimed at achieving near- order to detect the “sweetness” of patients’ urine. This imper-
normal pregnancy outcomes.64 However, despite aggressive fect technique was used until the development of urine sticks
Glucose monitoring: The effect on glycemic control and pregnancy outcome 183
in the early twentieth century that were sensitive enough to In patients with type 1 diabetes, Kerssen et al.74 reported
detect glycosuria. Research efforts have continuously been that the detection of hyper- and hypoglycemia was sig-
directed toward the development of a process for testing nificantly higher in women with 10 or more SMBG deter-
blood glucose using either visual or electronic interpreta- minations daily than in patients with fewer. However, no
tion with a reflectance meter. The original meters used a wet correlation with pregnancy outcome was reported. In con-
method that often required as many as four steps (approxi- trast, Homko et al.75 randomly assigned 58 women with diet-
mately 10 minutes/step) to obtain one test result. Today’s controlled GDM and a fasting blood glucose level <95 mg/dL
reflectance meters are more user-friendly and include a to two groups. The experimental group measured their blood
memory chip for data storage and one-step testing. Recently, glucose levels four times daily using a reflectance meter with
a CGM technique was developed that facilitates the collec- a memory chip. Metabolic status was assessed in the con-
tion of even more accurate glucose data. However, is it the trol group by periodic monitoring at prenatal visits. They
most sophisticated method of glucose monitoring associated reported no significant differences with regard to dietary
with the best clinical results? compliance, birth weight, gestational age at delivery, rates of
macrosomia, delivery by cesarean section, Apgar scores, and
neonatal complications. It seems that the optimal frequency
Glycosylated hemoglobin: Can a single measure of for blood glucose monitoring should probably be different
glycemic level be valuable? among the various diabetes groups with diet-controlled
Traditionally, in nonpregnant diabetic patients, glycosyl- GDM needing the less frequent monitoring.
ated hemoglobin (HbA1C) became the indicator of long-
term glycemia. It is a modification of hemoglobin caused
by the attachment of glucose to the N-terminus of the beta The role of continuous glucose monitoring during
chain. The rate of attachment is determined by the glucose pregnancy
concentration in the blood. Based on the lifespan of the red The main advantage of CGM is by recognizing period of
blood cells that averages 120 days, different reports have hypo- or hyperglycemia that could not be detected using
suggested that the predictability of HbA1C ranges from 4 to routine determinations using SMBG. In pregnancies com-
10 weeks.69,70 Thus, it is virtually impossible to alter treat- plicated with type 1 diabetes, Yogev et al.76 demonstrated a
ment modality based on retrospective data, especially in mean total time (192 ± 28 minutes/day) of undetected hyper-
GDM since the window of opportunity is so small. Yet lately, glycemia (glucose levels > 140 mg/dL) identified by CGM,
Jovanovic et al. found that in 24 women with GDM and ini- which would not be recognized if SMBG was used alone.
tial HbA1C ≥7.0%, a decline of 0.47% per week in HbA1C Moreover, when GDM patients were evaluated, the mean
levels was observed in response to carbohydrate-restricted total time of hyperglycemia was 132 ± 31 minutes/day for
diet or insulin treatment.71 This finding, if proven by larger- those treated with insulin and 94 ± 23 minutes/day for those
scale studies, could reflect a possible utility of HbA1C mea- treated with diet only.77
surement during therapy. Until then, HbA1C measurement The efficacy and effectiveness of CGM during pregnancy
should be reserved for preexisting diabetes especially at was assessed in a recently published systematic review by
the first office visit for counseling for the risk of congenital Voormolen et al.78 They retrieved 5032 articles, 11 of which
anomalies and macro and micro complications throughout remained as relevant after selection according to predefined
pregnancy. criteria. Of those, only two studies were randomized con-
trolled trials (RCTs) evaluating the effect on pregnancy
outcome with conflicting results. Murphy et al.79 ran-
Is there an optimal frequency for blood glucose domly allocated pregnant women with type 1 or 2 DM
determinations? to either standard antenatal care with SMBG (n = 33) or to
Nowadays, the role of SMBG in nonpregnant and pregnant standard antenatal care with additional CGM use (n = 38).
women has become the standard of care for achieving tar- Retrospective CGM was planned every 4–6 weeks for a
geted levels of glycemic control.72 However, the optimal fre- period of 5–7 consecutive days and was executed on aver-
quency for blood glucose determinations is unclear. Langer age 4.2 times per pregnancy. Lower HbA1C values were seen
et al.18 demonstrated in a prospective study that a mean throughout pregnancy in the CGM group, but this difference
blood glucose obtained from SMBG (seven times/day, inten- did not reach statistical significance until at 32–36 weeks
sified treatment group), using memory reflectance meters, of gestation (5.8% ± 0.6% vs. 6.4% ± 0.7%, p = 0.007). The
identified more fetal macrosomia and other neonatal mor- incidence of LGA was significantly lower in the CGM group
bidities in comparison to weekly fasting and 2-hour labora- with OR for reducing LGA rate of 0.36 (95% CI, 0.13–0.98;
tory glucose determinations supplemented by four times/ p = 0.05). Comparison of other outcomes such as mode of
day (unverified) self-monitoring with only test strips and no delivery and neonatal morbidity was not significantly differ-
meters. Of note, the compliance in the intensified treatment ent. It must be noted that the intervention group included
group was not ideal since the mean test per day was only three twin pregnancies and four children with birth weights
5.2 instead of 7 as instructed. Others have reported similar under the 10th percentile, whereas there were none in the
results with four/day glucose determinations as compared to control group (p = 0.1). Logically, this affected the mean
less intensified treatment.73 birth weight of the intervention group for the better, though a
birth weight below the 10th percentile must be considered an Summary

unwanted result. In contrast, a high-quality RCT by Secher
et al.80 randomized 154 women with either type 1 or 2 DM Maternal hyperglycemia is associated with increased risk
and a singleton pregnancy for the additional use of real-time for adverse perinatal outcome with no single threshold that
CGM (n = 79) or routine antenatal care with SMBG seven eliminates the risk for all adverse outcomes. Yet maintaining
times daily (n = 75). The CGM was intermittently planned glucose level as close to recommended values may minimize
at 8, 12, 21, 27, and 33 weeks of gestation and was executed the risk for most complications. In addition, it is obligatory
at least three times by 75% of the women. Women assigned that women with type 1 or 2 DM receive proper prepregnancy
to real-time CGM had baseline HbA1C similar to that of counseling emphasizing the need for reaching euglycemia
women in the control arm (median 6.6% vs. 6.8%, p = 0.67). prior to conception and close surveillance during gestation.
At 33 gestational weeks, HbA1C levels remained comparable The evolving research regarding the use of continues glucose
regardless of the real-time CGM use. The prevalence of LGA monitoring and its effect on glycemic control and pregnancy
infants (45% vs. 34%; p = 0.19) and other perinatal outcomes outcome combined with improvement in pharmacotherapy
(e.g., preeclampsia, cesarean delivery, preterm delivery, and options will hopefully lead to safer pregnancy for both the
neonatal hypoglycemia) were comparable between the arms. mother and fetus.
REFERENCES
1. Boney CM, Verma A, Tucker R, Vohr BR. Metabolic syndrome in 16. Pettitt DJ, Knowler WC, Baird HR, Bennett PH. Gestational dia-
childhood: Association with birth weight, maternal obesity, and betes: Infant and maternal complications of pregnancy in relation
gestational diabetes mellitus. Pediatrics 2005; 115: e290–e296. to third-trimester glucose tolerance in the Pima Indians. Diabetes
2. Lau C, Rogers JM, Desai M, Ross MG. Fetal programming of adult Care 1980; 3: 458–464.
disease: Implications for prenatal care. Obstet Gynecol 2011; 117: 17. Langer O. Glycemic targets for the optimal treatment of GDM.
978–985. Clin Obstet Gynecol 2013; 56: 788–802.
3. Ornoy A. Prenatal origin of obesity and their complications: 18. Langer O, Rodriguez DA, Xenakis EM, McFarland MB, Berkus
Gestational diabetes, maternal overweight and the paradoxi- MD, Arrendondo F. Intensified versus conventional management
cal effects of fetal growth restriction and macrosomia. Reprod of gestational diabetes. Am J Obstet Gynecol 1994; 170: 1036–
Toxicol 2011; 32: 205–212. 1046; discussion 1046–1047.
4. Hadar E, Ashwal E, Hod M. The preconceptional period as an 19. Temple R, Aldridge V, Greenwood R, Heyburn P, Sampson M,
opportunity for prediction and prevention of noncommunica- Stanley K. Association between outcome of pregnancy and gly-
ble disease. Best Pract Res Clin Obstet Gynaecol 2015; 29(1): caemic control in early pregnancy in type 1 diabetes: Population
54–62. based study. Br Med J 2002; 325: 1275–1276.
5. Petrovski G, Dimitrovski C, Bogoev M, Milenkovic T, Ahmeti I, 20. Greene MF, Hare JW, Cloherty JP, Benacerraf BR, Soeldner JS.
Bitovska I. Is there a difference in pregnancy and glycemic out- First-trimester hemoglobin A1 and risk for major malformation
come in patients with type 1 diabetes on insulin pump with con- and spontaneous abortion in diabetic pregnancy. Teratology
stant or intermittent glucose monitoring? A pilot study. Diabetes 1989; 39: 225–231.
Technol Ther 2011; 13: 1109–1113. 21. Rosenn B, Miodovnik M, Combs CA, Khoury J, Siddiqi TA.
6. Holmes VA, Young IS, Patterson CC et al. Optimal glycemic con- Glycemic thresholds for spontaneous abortion and congenital
trol, pre-eclampsia, and gestational hypertension in women with malformations in insulin-dependent diabetes mellitus. Obstet
type 1 diabetes in the diabetes and pre-eclampsia intervention Gynecol 1994; 84: 515–520.
trial. Diabetes Care 2011; 34: 1683–1688. 22. Kitzmiller JL, Gavin LA, Gin GD, Jovanovic-Peterson L, Main EK,
7. Langer O, Yogev Y, Most O, Xenakis EM. Gestational diabetes: Zigrang WD. Preconception care of diabetes. Glycemic con-
The consequences of not treating. Am J Obstet Gynecol 2005; trol prevents congenital anomalies. J Am Med Assoc 1991; 265:
192: 989–997. 731–736.
8. Crowther CA, Hiller JE, Moss JR, McPhee AJ, Jeffries WS, 23. Fuhrmann K, Reiher H, Semmler K, Fischer F, Fischer M, Glockner
Robinson JS. Effect of treatment of gestational diabetes mellitus E. Prevention of congenital malformations in infants of insulin-
on pregnancy outcomes. N Engl J Med 2005; 352: 2477–2486. dependent diabetic mothers. Diabetes Care 1983; 6: 219–223.
9. Landon MB, Spong CY, Thom E et al. A multicenter, random- 24. Schaefer-Graf UM, Buchanan TA, Xiang A, Songster G, Montoro
ized trial of treatment for mild gestational diabetes. N Engl J Med M, Kjos SL. Patterns of congenital anomalies and relationship to
2009; 361: 1339–1348. initial maternal fasting glucose levels in pregnancies complicated
10. Committee on Practice Bulletins—Obstetrics. Practice Bulletin by type 2 and gestational diabetes. Am J Obstet Gynecol 2000;
No. 137: Gestational diabetes mellitus. Obstet Gynecol 2013; 182: 313–320.
122: 406–416. 25. McElvy SS, Miodovnik M, Rosenn B et al. A focused preconcep-
11. Landon MB, Mele L, Spong CY et al. The relationship between tional and early pregnancy program in women with type 1 diabe-
maternal glycemia and perinatal outcome. Obstet Gynecol 2011; tes reduces perinatal mortality and malformation rates to general
117: 218–224. population levels. J Matern Fetal Med 2000; 9: 14–20.
12. Mazze R, Yogev Y, Langer O. Measuring glucose exposure and 26. Lurie S, Matzkel A, Weissman A, Gotlibe Z, Friedman A.
variability using continuous glucose monitoring in normal and Outcome of pregnancy in class A1 and A2 gestational diabetic
abnormal glucose metabolism in pregnancy. J Matern Fetal patients delivered beyond 40 weeks’ gestation. Am J Perinatol
Neonatal Med 2012; 25: 1171–1175. 1992; 9: 484–488.
13. Karlsson K, Kjellmer I. The outcome of diabetic pregnancies in 27. Kjos SL, Henry OA, Montoro M, Buchanan TA, Mestman JH.
relation to the mother’s blood sugar level. Am J Obstet Gynecol Insulin-requiring diabetes in pregnancy: A randomized trial
1972; 112: 213–220. of active induction of labor and expectant management. Am J
14. O’Sullivan JB, Charles D, Mahan CM, Dandrow RV. Gestational Obstet Gynecol 1993; 169: 611–615.
diabetes and perinatal mortality rate. Am J Obstet Gynecol 1973; 28. Kanda E, Matsuda Y, Makino Y, Matsui H. Risk factors associated
116: 901–904. with altered fetal growth in patients with pregestational diabetes
15. Beischer NA, Wein P, Sheedy MT, Steffen B. Identification and mellitus. J Matern Fetal Neonatal Med 2012; 25: 1390–1394.
treatment of women with hyperglycaemia diagnosed during 29. Rowan JA, Gao W, Hague WM, McIntyre HD. Glycemia and its
pregnancy can significantly reduce perinatal mortality rates. Aust relationship to outcomes in the metformin in gestational diabetes
N Z J Obstet Gynaecol 1996; 36: 239–247. trial. Diabetes Care 2010; 33: 9–16.
References 185
30. Most O, Langer O. Gestational diabetes: Maternal weight gain in 51. Suhonen L, Teramo K. Hypertension and pre-eclampsia in women
relation to fetal growth, treatment modality, BMI and glycemic with gestational glucose intolerance. Acta Obstet Gynecol Scand
control. J Matern Fetal Neonatal Med 2012; 25: 2458–2463. 1993; 72: 269–272.
31. Yu Z, Han S, Zhu J, Sun X, Ji C, Guo X. Pre-pregnancy body mass 52. Garner PR, D’Alton ME, Dudley DK, Huard P, Hardie M.
index in relation to infant birth weight and offspring overweight/ Preeclampsia in diabetic pregnancies. Am J Obstet Gynecol
obesity: A systematic review and meta-analysis. PLOS ONE 2013; 1990; 163: 505–508.
8: e61627. 53. Ros HS, Cnattingius S, Lipworth L. Comparison of risk factors for
32. Olmos PR, Araya-Del-Pino AP, Gonzalez-Carvello CA et al. preeclampsia and gestational hypertension in a population-based
Near-optimal glycemic control in Chilean women with pregesta- cohort study. Am J Epidemiol 1998; 147: 1062–1070.
tional type-2 diabetes: Persistent macrosomia relates to maternal 54. Yogev Y, Xenakis EM, Langer O. The association between
pre-pregnancy overweight. Diabetes Res Clin Pract 2009; 85: preeclampsia and the severity of gestational diabetes: The

53–60. impact of glycemic control. Am J Obstet Gynecol 2004; 191:
33. Roversi GD, Gargiulo M, Nicolini U et al. Maximal tolerated 1655–1660.
insulin therapy in gestational diabetes. Diabetes Care 1980; 3: 55. Schaffir JA, Lockwood CJ, Lapinski R, Yoon L, Alvarez M.
489–494. Incidence of pregnancy-induced hypertension among gestational
34. Jovanovic-Peterson L, Durak EP, Peterson CM. Randomized trial diabetics. Am J Perinatol 1995; 12: 252–254.
of diet versus diet plus cardiovascular conditioning on glucose 56. van Hoorn J, Dekker G, Jeffries B. Gestational diabetes versus
levels in gestational diabetes. Am J Obstet Gynecol 1989; 161: obesity as risk factors for pregnancy-induced hypertensive disor-
415–419. ders and fetal macrosomia. Aust N Z J Obstet Gynaecol 2002; 42:
35. Landon MB, Gabbe SG, Piana R, Mennuti MT, Main EK. 29–34.
Neonatal morbidity in pregnancy complicated by diabetes mel- 57. Cohen AL, Wenger JB, James-Todd T et al. The association of
litus: Predictive value of maternal glycemic profiles. Am J Obstet circulating angiogenic factors and HbA1c with the risk of pre-
Gynecol 1987; 156: 1089–1095. eclampsia in women with preexisting diabetes. Hypertens
36. Langer O, Levy J, Brustman L, Anyaegbunam A, Merkatz R, Pregnancy 2014; 33: 81–92.
Divon M. Glycemic control in gestational diabetes mellitus— 58. Phaloprakarn C, Tangjitgamol S. Risk assessment for preeclamp-
How tight is tight enough: Small for gestational age versus large sia in women with gestational diabetes mellitus. J Perinat Med
for gestational age? Am J Obstet Gynecol 1989; 161: 646–653. 2009; 37: 617–621.
37. Gonzalez-Quintero VH, Istwan NB, Rhea DJ et al. The impact 59. Executive summary: Standards of medical care in diabetes—2014.
of glycemic control on neonatal outcome in singleton pregnan- Diabetes Care 2014; 37(Suppl. 1): S5–S13.
cies complicated by gestational diabetes. Diabetes Care 2007; 60. Metzger BE, Buchanan TA, Coustan DR et al. Summary and rec-
30: 467–470. ommendations of the Fifth International Workshop-Conference on
38. Langer O, Conway DL. Level of glycemia and perinatal outcome Gestational Diabetes Mellitus. Diabetes Care 2007; 30(Suppl. 2):
in pregestational diabetes. J Matern Fetal Med 2000; 9: 35–41. S251–S260.
39. Piper JM, Xenakis EM, Langer O. Delayed appearance of pulmo- 61. Canadian Diabetes Association Clinical Practice Guidelines
nary maturation markers is associated with poor glucose control Expert Committee. Canadian Diabetes Association 2013 Clinical
in diabetic pregnancies. J Matern Fetal Med 1998; 7: 148–153. Practice Guidelines for the Prevention and Management of
40. De Luca AK, Nakazawa CY, Azevedo BC, Rudge MV, De Araujo Diabetes in Canada. Can J Diabetes 2013; 37: S1–S212.
Costa RA, Calderon IM. Influence of glycemic control on fetal 62. Yogev Y, Ben-Haroush A, Chen R, Rosenn B, Hod M, Langer O.
lung maturity in gestations affected by diabetes or mild hypergly- Diurnal glycemic profile in obese and normal weight nondiabetic
cemia. Acta Obstet Gynecol Scand 2009; 88: 1036–1040. pregnant women. Am J Obstet Gynecol 2004; 191: 949–953.
41. Berkowitz K, Reyes C, Saadat P, Kjos SL. Fetal lung maturation. 63. Parretti E, Mecacci F, Papini M et al. Third-trimester maternal
Comparison of biochemical indices in gestational diabetic and glucose levels from diurnal profiles in nondiabetic pregnan-
nondiabetic pregnancies. J Reprod Med 1997; 42: 793–800. cies: Correlation with sonographic parameters of fetal growth.
42. Gizzo S, Patrelli TS, Rossanese M et al. An update on diabetic Diabetes Care 2001; 24: 1319–1323.
women obstetrical outcomes linked to preconception and preg- 64. Diabetes care and research in Europe: The Saint Vincent declara-
nancy glycemic profile: A systematic literature review. Scientific tion. Diabet Med 1990; 7: 360.
World J 2013; 2013: 254901. 65. Cohen O, Keidar N, Simchen M, Weisz B, Dolitsky M, Sivan E.
43. Boulot P, Chabbert-Buffet N, d’Ercole C et al. French multicentric Macrosomia in well controlled CSII treated Type I diabetic preg-
survey of outcome of pregnancy in women with pregestational nancy. Gynecol Endocrinol 2008; 24: 611–613.
diabetes. Diabetes Care 2003; 26: 2990–2993. 66. Evers IM, de Valk HW, Mol BW, ter Braak EW, Visser GH.
44. Lauszus FF, Fuglsang J, Flyvbjerg A, Klebe JG. Preterm delivery Macrosomia despite good glycaemic control in Type I diabetic
in normoalbuminuric, diabetic women without preeclampsia: pregnancy; results of a nationwide study in The Netherlands.
The role of metabolic control. Eur J Obstet Gynecol Reprod Biol Diabetologia 2002; 45: 1484–1489.
2006; 124: 144–149. 67. Mello G, Parretti E, Mecacci F et al. What degree of maternal
45. Kovilam O, Khoury J, Miodovnik M, Chames M, Spinnoto J, Sibai metabolic control in women with type 1 diabetes is associated
B. Spontaneous preterm delivery in the type 1 diabetic preg- with normal body size and proportions in full-term infants?
nancy: The role of glycemic control. J Matern Fetal Neonatal Med Diabetes Care 2000; 23: 1494–1498.
2002; 11: 245–248. 68. Howorka K, Pumprla J, Gabriel M et al. Normalization of preg-
46. Rosenberg TJ, Garbers S, Lipkind H, Chiasson MA. Maternal obe- nancy outcome in pregestational diabetes through functional
sity and diabetes as risk factors for adverse pregnancy outcomes: insulin treatment and modular out-patient education adapted for
Differences among 4 racial/ethnic groups. Am J Public Health pregnancy. Diabet Med 2001; 18: 965–972.
2005; 95: 1545–1551. 69. Widness JA, Schwartz HC, Kahn CB, Oh W, Schwartz R.
47. Sendag F, Terek MC, Itil IM, Oztekin K, Bilgin O. Maternal and Glycohemoglobin in diabetic pregnancy: A sequential study. Am
perinatal outcomes in women with gestational diabetes mellitus J Obstet Gynecol 1980; 136: 1024–1029.
as compared to nondiabetic controls. J Reprod Med 2001; 46: 70. Hall PM, Cook JG, Sheldon J, Rutherford SM, Gould BJ.
1057–1062. Glycosylated hemoglobins and glycosylated plasma proteins in
48. Yogev Y, Langer O. Spontaneous preterm delivery and gesta- the diagnosis of diabetes mellitus and impaired glucose toler-
tional diabetes: The impact of glycemic control. Arch Gynecol ance. Diabetes Care 1984; 7: 147–150.
Obstet 2007; 276: 361–365. 71. Jovanovic L, Savas H, Mehta M, Trujillo A, Pettitt DJ. Frequent
49. Bar-Hava I, Barnhard Y, Scarpelli SA, Orvieto R, Ben R, Divon monitoring of A1C during pregnancy as a treatment tool to guide
MY. Gestational diabetes and preterm labour: Is glycaemic con- therapy. Diabetes Care 2011; 34: 53–54.
trol a contributing factor? Eur J Obstet Gynecol Reprod Biol 1997; 72. Blumer I, Hadar E, Hadden DR et al. Diabetes and pregnancy:
73: 111–114. An endocrine society clinical practice guideline. J Clin Endocrinol
50. Nordin NM, Wei JW, Naing NN, Symonds EM. Comparison Metab 2013; 98: 4227–4249.
of maternal-fetal outcomes in gestational diabetes and lesser 73. Goldberg JD, Franklin B, Lasser D et al. Gestational diabetes:
degrees of glucose intolerance. J Obstet Gynaecol Res 2006; 32: Impact of home glucose monitoring on neonatal birth weight.
107–114. Am J Obstet Gynecol 1986; 154: 546–550.
74. Kerssen A, de Valk HW, Visser GH. Do HbA1c levels and the 78. Voormolen DN, DeVries JH, Evers IM, Mol BW, Franx A. The effi-
self-monitoring of blood glucose levels adequately reflect gly- cacy and effectiveness of continuous glucose monitoring during
caemic control during pregnancy in women with type 1 diabetes pregnancy: A systematic review. Obstet Gynecol Surv 2013; 68:
mellitus? Diabetologia 2006; 49: 25–28. 753–763.
75. Homko CJ, Sivan E, Reece EA. The impact of self-monitoring 79. Murphy HR, Rayman G, Lewis K et al. Effectiveness of con-
of blood glucose on self-efficacy and pregnancy outcomes in tinuous glucose monitoring in pregnant women with diabetes:
women with diet-controlled gestational diabetes. Diabetes Educ Randomised clinical trial. Br Med J 2008; 337: a1680.
2002; 28: 435–443. 80. Secher AL, Ringholm L, Andersen HU, Damm P, Mathiesen
76. Yogev Y, Chen R, Ben-Haroush A, Phillip M, Jovanovic L, Hod M. ER. The effect of real-time continuous glucose monitoring in
Continuous glucose monitoring for the evaluation of gravid women pregnant women with diabetes: A randomized controlled trial.
with type 1 diabetes mellitus. Obstet Gynecol 2003; 101: 633–638. Diabetes Care 2013; 36: 1877–1883.
77. Chen R, Yogev Y, Ben-Haroush A, Jovanovic L, Hod M, Phillip M.
Continuous glucose monitoring for the evaluation and improved
control of gestationals diabetes mellitus. J Matern Fetal Neonatal
Med 2003; 14: 256–260.
23 Insulin therapy in pregnancy
Lois Jovanovic and John L. Kitzmiller
complete by 7 weeks postconception. Thus, a woman may not

Introduction even know she is pregnant at this time. It is for this reason that
Before the advent of insulin, few diabetic women lived to prepregnancy counseling and planning is essential in women
childbearing age. Before 1922, fewer than 100 pregnancies in of childbearing age who have diabetes. Because organogenesis
diabetic women were reported, and most likely these women is complete so early on, if a woman presents to her health-
had type 1 and not type 2 diabetes.1 Even with this assump- care team and announces that she has missed her period by
tion, these cases of diabetes and pregnancy were associated only a few days, if the blood glucose levels are immediately
with a >90% infant mortality rate and a 30% maternal mor- normalized, then there is still is a chance to prevent cardiac
tality rate. As late as 1980, physicians were still counseling anomalies, although neural tube defects are already “set in
diabetic women to avoid pregnancy. This philosophy was stone” by the time the first period is missed. These findings
justified because of the poor obstetric history in 30%–50% emphasize the importance of glycemic control at the earliest
of diabetic women. Improved infant mortality rates finally stages of conception.2–4 Ideally, if a diabetic woman plans her
occurred after 1980, when treatment strategies stressed better pregnancy, then there is time to create algorithms of care that
control of maternal plasma glucose levels, and once self-mon- are individualized and a woman can be given choices. When
itoring blood glucose (SMBG) and glycosylated hemoglobin a diabetic woman presents in her first few weeks of pregnancy,
(A1C) became available. As the pathophysiology of pregnancy there is no time for individualization, but rather rigid pro-
complicated by diabetes has been elucidated, and as manage- tocols must be urgently instituted to provide optimal con-
ment programs have achieved and maintained near-normal trol within 24–48 hours. After the period of organogenesis,
glycemia throughout pregnancy complicated by type 1 dia- maternal hyperglycemia interferes with normal fetal growth
betes, perinatal mortality rates have decreased to levels seen and development during the second and third trimesters.5 The
in the general population. This review is intended to help maternal postprandial glucose level has been shown to be the
the clinician understand the increasing insulin requirements most important variable to impact on the subsequent risk of
of pregnancy and to design treatment protocols to achieve neonatal macrosomia.6–8 The fetus thus is “overnourished”
and maintain normoglycemia throughout pregnancy. by the peak postprandial glucose level.9 This peak response
occurs in >90% of woman at 1 hour after beginning a meal.
Therefore, 1 hour after beginning a meal, the glucose level
needs to be measured and treatment designed to maintain this
Glucose toxicity and the role of blood glucose in the normal range. Studies have shown than if
postprandial hyperglycemia the postprandial glucose levels are maintained from the sec-
ond trimester onward to <120 mg/dL 1 hour after beginning a
If the mother has hyperglycemia, the fetus will be exposed meal, then the risk of macrosomia is minimized.8
to either sustained hyperglycemia or intermittent pulses of
hyperglycemia; both situations prematurely stimulate fetal
insulin secretion. Fetal hyperinsulinemia may cause increased Diabetogenic forces of normal
fetal body fat (macrosomia) and, therefore, a difficult deliv-
ery, or cause inhibition of pulmonary maturation of surfac- pregnancy increase insulin
tant and, therefore, respiratory distress of the neonate. The requirements10
fetus may also have decreased serum potassium levels caused
by the elevated insulin and glucose levels and may therefore The fetal demise associated with pregnancy complicated by
have cardiac arrhythmias. Neonatal hypoglycemia may cause type 1 diabetes seems to arise from glucose extremes. Elevated
permanent neurological damage. There is also an increased maternal plasma glucose levels should always be avoided,
prevalence of congenital anomalies and spontaneous abor- because of the association of maternal hyperglycemia with
tions in diabetic women who are in poor glycemic control subsequent congenital malformation and spontaneous abor-
during the period of fetal organogenesis, which is nearly tions.2,5 To achieve normoglycemia, a clear understanding
187
of “normal” carbohydrate metabolism in pregnancy is because some reports claim that neonatal complications
paramount. Thus, the amount of insulin required to treat occur in spite of excellent metabolic control, although they
women with type 1 diabetes throughout pregnancy needs to fail to measure postprandial glucose levels.11,12 Postprandial
be sufficient to compensate for (1) increasing caloric needs, glucose control has been suggested as a key to neonatal out-
(2) increasing adiposity, (3) decreasing exercise, and (4) come for the pregnant woman with either type 1 diabetes
increasing anti-insulin or diabetogenic hormones of preg- or GDM.6–8 Alternatively, some have suggested that neo-
nancy. The major diabetogenic hormones of the placenta are natal morbidity is secondary to the variability of maternal
human chorionic somatomammotropin (hCS), previously serum glucose and the presence of antibodies to insulin.13
referred to as human placental lactogen, estrogen, and pro- Placental transfer of insulin complexed with immunoglob-
gesterone. Also, serum maternal cortisol levels (both bound ulin (Ig) G has also been associated with fetal macrosomia
and free) are increased. In addition, at the elevated levels in mothers with near-normal glycemia during gestation.
seen during gestation, prolactin has a diabetogenic effect.10 Menon et al.13 reported that antibody-bound insulin trans-
The strongest insulin antagonist of pregnancy is hCS. This ferred to the fetus was proportional to the concentration of
placental hormone appears in increasing concentrations antibody-bound insulin measured in the mother. Also, the
beginning at 10 weeks of gestation. By 20 weeks of gesta- amount of antibody-bound insulin transferred to the fetus
tion, plasma hCS levels are increased 300-fold, and by term, correlated directly with macrosomia in the infant and was
the turnover rate is 1000 mg/dL. The mechanism of action independent of maternal blood glucose levels. In contrast,
whereby hCS raises plasma glucose levels is unclear, but prob- Jovanovic et al.14 discovered only improved glucose control,
ably originates from its growth hormone-like properties. as evidenced by lower postprandial glucose excursions, but
hCS also promotes free fatty acid production by stimulating not lower insulin antibody levels, correlated with lower fetal
lipolysis, which promotes peripheral resistance to insulin. weights. They showed that insulin antibodies to exogenous
Placental progesterone rises 10-fold above nonpregnant levels insulin do not influence infant birth weight. It has been
and is associated with an insulin increase in normal healthy reported that both insulin lispro and insulin aspart (analogs
pregnant women by two- to fourfold. Most of the marked rise of human insulin with a peak insulin action achieved within
of serum cortisol during pregnancy can be attributed to the 1 hour after injection) significantly improve the postprandial
increase of cortisol-binding globulin induced by estrogen. glucose levels in nonpregnant diabetic patients.15–21 Because
However, free cortisol levels are also increased. This increase normoglycemia is important in the treatment of pregnant
potentiates the diurnal fluctuations of cortisol with the high- diabetic women, the use of insulin analogs would appear
est levels occurring in the early morning hours. The rising beneficial in the care of these women if the safety profile can
estrogen levels also trigger the rise in pituitary prolactin early be documented. This review presents the reports that studied
in pregnancy. Prolactin’s structure is similar to a growth hor- the safety and efficacy of insulin analogs in pregnancy and
mone, and at concentrations reached by the second trimes- offers an opinion as to the utility of insulin analogs for the
ter (>200 ng/mL), prolactin can affect glucose metabolism. treatment of the diabetes during pregnancy.
Although there are no studies that have examined prolactin
alone as an insulin antagonist, there is indirect evidence that
suppressing prolactin in women with gestational diabetes
with large doses of pyridoxine improves the glucose toler- Concern about anti-insulin antibody
ance. In addition to the increasing anti-insulin hormones formation during pregnancy
of pregnancy, there is also increased degradation of insulin
in pregnancy caused by placental enzymes comparable to Anti-insulin antibodies that cross the placenta may contri
liver insulinases. The placenta also has membrane-associated bute to hyperinsulinemia in utero and thus potentiate the
insulin-degrading activity. Concomitant with the hormon- metabolic aberrations in the fetus. Although insulin does not
ally induced insulin resistance and increased insulin deg- cross the placenta, antibodies to insulin do cross and may
radation, the rate of disposal of insulin slows. The normal bind fetal insulin; this necessitates the increased produc-
pancreas can adapt to these factors by increasing the insulin tion of free insulin to reestablish normoglycemia. Thus, the
secretory capacity. If the pancreas fails to respond adequately anti-insulin antibodies may potentiate the effect of maternal
to these alterations, then gestational diabetes mellitus (GDM) hyperglycemia to produce fetal hyperinsulinemia.
results. In a woman with type 1 diabetes, her insulin require- Human and highly purified insulins are significantly
ment will rise progressively. Failure to increase her insulin less immunogenic than mixed beef–pork insulins.16 Human
doses appropriately will result in increasing hyperglycemia.10 insulin treatment has been reported to achieve improved
pregnancy and infant outcome compared to using highly
purified animal insulins.14 Insulin lispro (which has the
Rationale for the use of human insulin amino acid sequence in the beta chain reversed at positions
during pregnancy B28 and B29) has been reported to be more efficacious than
human regular insulin to normalize the blood glucose levels
Although controversial, the rate of complications in preg- in women with GDM. This insulin rapidly lowered the post-
nancies of diabetic women has been tied to the metabolic prandial glucose levels, thereby decreasing the A1C levels,
control of maternal glucose.1–5 Perhaps the debate remains with fewer hypoglycemic episodes and without increasing
Use of rapid-acting insulin analogs in pregnancies complicated by diabetes 189
the anti-insulin antibody levels.17 In addition, the safety and to result in little, if any, placental transfer of insulin lispro
efficacy of insulin lispro in the treatment of type 1 diabetes in to the neonate, as was demonstrated in the present authors’
women throughout pregnancy has recently been reported.22 study. Thus, the overall decrease in circulating insulin lis-
Also, insulin aspart has recently been reported to be safe and pro, plus the lower immunogenic response of lispro, leads to
efficacious in type 1 diabetic women.23 However, there are less maternal antibody formation and, therefore, less insulin
only case reports and a small series of the use of the long- transfer to the fetus with a reduction in the risk for physical
acting insulin analogs24–27 during pregnancy. They appear malformations.17,21 Menon et al.13 attempted to link mater-
to be associated with increased macrosomia, however.27 The nal antibody formation to negative fetal outcomes. Careful
following discussion helps the clinician decide if the newer review of the paper reveals, instead, better overall control of
insulin’s benefit outweighs any risks. maternal hyperglycemia with attendant reductions in fetal
macrosomia. This may have ultimately diminished the risk
of neonatal complications, including macrosomia. Previous
investigations have demonstrated that birth weight could be
Use of rapid-acting insulin analogs in normalized with regular human insulin.22–24 This aggressive
pregnancies complicated by diabetes therapeutic intervention may explain the apparent lack of
macrosomia in both patient groups. There are no differences
Postprandial glucose control in a patient with GDM is impor- in fetal parameters, as would be expected in the clinical set-
tant to neonatal outcome.6–8 The Diabetes in Early Pregnancy ting where euglycemia is a goal of therapy, which reduce the
(DIEP) study identified 28.5% of infants from diabetic moth- risk to the fetus. Although the present authors were inter-
ers who were >90th percentile in infant birth weights.8 The ested in the metabolic effects of insulin lispro in women with
birth weight in this 28.5% correlated positively with fasting GDM, the primary concern was safety, specifically the risk
blood glucose and A1C. When adjusted for fasting blood glu- of hypoglycemia, hyperglycemia, and antibody production
cose and A1C, the nonfasting blood glucose concentration that might cause the insulin to cross the placenta to the fetal
in the third trimester was even a stronger predictor of infant side. In the study, 42 women with GDM were randomized to
birth weight and fetal macrosomia. Combs et al.7 confirmed receive regular human insulin or insulin lispro prior to con-
these findings, as they associated macrosomia with higher suming a test meal.17 Throughout the remainder of gestation,
postprandial glucose concentrations obtained between subjects received premeal insulin lispro or regular human
weeks 29 and 32 of gestation. In addition, they described a insulin (with and without basal insulin) and performed
higher risk of small-for-gestational-age (SGA) infants in blood glucose monitoring before and after each meal. During
those with lower (<130 mg/dL [7.2 mmol/L]) 1-hour post- the test meal, the areas under the glucose curve, and those for
prandial glucose levels. DeVeciana et al.6 described improved insulin and C-peptide levels, were significantly lower in the
fetal outcome and less risk of neonatal hypoglycemia, macro- insulin lispro group. The incidence of postprandial hyper-
somia, and cesarean delivery in patients who managed GDM glycemia (>120 mg/dL) was significantly lower in the lispro
by controlling 1-hour postprandial glucose concentrations group. Overall metabolic control also improved significantly
than in those who managed by only the preprandial glucose in the insulin lispro group, which showed the greatest abso-
concentrations. Therefore, rapid-acting insulin analogs that lute decrease in A1C levels as compared to the regular human
possess unique properties may make it a valuable therapeu- insulin group. The reduction from baseline A1C concentra-
tic option in the treatment of GDM and the prevention of tions at 6 weeks was statistically significant for the insulin
neonatal complications. First, the rapid absorption of insulin lispro group but not for the regular human insulin group.17
lispro from the subcutaneous site allows for a faster insulin To determine the immunologic effects of insulin lispro com-
peak concentration versus regular human insulin.17,18 This pared with regular human insulin, three different types
effect more closely mimics the physiologic first-phase insulin of antibodies were studied: (1) lispro-specific antibodies,
release and results in lower postprandial glucose concentra- (2) regular specific antibodies, and (3) cross-reactive anti-
tions, and may lead to improved postprandial coverage.15 In bodies. Levels of all three antibodies were evaluated at the
addition, insulin lispro is known to upregulate insulin recep- time of enrollment, at 6 weeks after enrollment, at delivery
tors.19,20 In the present authors’ study,17 the postprandial glu- (both in the maternal serum and the umbilical cord blood),
cose response to the test meal was more frequently within and at the postpartum follow-up visit in maternal serum.
the normal glucose range after a standardized dose of insu- No statistically significant differences were seen between the
lin lispro as compared with regular human insulin. Second, insulin lispro and regular human insulin groups. Now there
the elimination of insulin from the venous space is the same are several reports of the safety and efficacy of both insulin
as with regular human insulin, but the faster absorption of lispro and insulin aspart in pregnancy to confirm the results
insulin lispro allows both the glucose-lowering effect and the of the original report.17–25 Insulin aspart has also been suited
patient’s exposure to insulin to be less, which may result in in a similar fashion as insulin lispro. Wyatt et al.22 have
a diminished antibody response. Certainly, in clinical trials, reported that insulin aspart is not immunogenic, improved
there has not been any increase in antibody response associ- the area under the curve for the postprandial glucose excur-
ated with insulin lispro use.15,18,19,22 Since placental transfer of sion, and facilitates term delivery of a healthy infant. In
insulin occurs when it is complexed with Ig, the lack of insu- addition, the patient satisfaction was improved using insu-
lin lispro–induced antibody formation could be expected lin aspart compared to regular human insulin. Conclusions
from these studies are that those women with GDM who are malformations were due to a medication, it is perhaps unfair
not optimally managed with diet and exercise need insulin to single out insulin lispro. In spite of the medication used,
therapy. Insulin lispro and insulin aspart cause fewer hypo- the malformations reported are more indicative of poor
glycemic events than human regular insulin, and it attenu- glycemic control: situs inversus, one of the abnormalities
ates a greater postprandial response than regular human in the first infant, occurs almost exclusively in children of
insulin. Furthermore, the antibody levels are not increased diabetic mothers.30,31 During the initial clinical trials testing
over those seen with regular human insulin, and therefore, insulin lispro, pregnant women were excluded. However,
insulin lispro and insulin aspart may be considered a treat- some participants became pregnant unexpectedly during
ment option in patients with GDM. the trials, and 19 infants were born by these mothers who
were using insulin lispro. Of these births, 1 child had a
right dysplastic kidney but the other 18 were healthy.32 Now
Theoretical risks of the use of insulin there is a report that shows that insulin lispro is not tera-
togenic. Wyatt et al.22 report has shown that insulin lispro
analogs during pregnancy complicated is not associated with increased malformations. This report
by pregestational diabetes clearly showed that the 27 malformations that occurred in
this data set of 500 exposed pregnancies only occurred in
Diamond and Kormas28 first questioned the safety of using women whose A1C levels were greater than 2 standard devi-
insulin lispro during pregnancy in 1997. They reported on ations (SD) above the mean of the normal. There were no
two patients who used insulin lispro during pregnancies malformations in the sample of women whose A1C levels
and deliveries. One of these pregnancies was terminated at were within two SD of the normal. Thus the null hypothesis
20 weeks gestation and the second pregnancy resulted in a has been proven: insulin lispro does not by itself cause mal-
seemingly healthy infant after an elective cesarean delivery, formations. Malformations are only associated with hyper-
but who subsequently died unexpectedly 3 weeks later. Both glycemia in the first trimester. Mathiesen et al.23 have now
infants were discovered to have congenital abnormalities, reported that insulin aspart is safe and efficacious in preg-
which led the authors to question whether insulin lispro nancy with no increase in malformation rate.
might have teratogenic effects on the fetus, in which case
it should not be used during pregnancy. The report cites
concerns about insulin lispro use during pregnancy, yet it Possible effects of rapid-acting insulin
does not provide conclusive evidence that insulin lispro was analogs on the mother
responsible for the malformations of the infants mentioned.
In fact, there is sufficient reason to doubt that insulin lispro There are three situations in life in which rapid normalization
was to blame, since these isolated case reports were not part of blood glucose levels increase the risk for deterioration of
of a study and there was no control group. Therefore, the find- diabetic retinopathy: puberty, pregnancy, and rapid normal-
ings should stimulate initiation of clinical trials testing the ization of blood glucose levels. If two of these events occur
safety of insulin lispro during pregnancy and not be taken in the same patient, the risk for retinopathy progression
as evidence that it is unsafe. Despite the opinion of Diamond is potentiated.33,34 All three situations are associated with
and Kormas28 that poor glycemic control was not responsible increased serum concentrations of growth-promoting fac-
for the abnormalities of the infants in the cases described tors.33 It is hypothesized that when the blood glucose level
earlier, there is insufficient evidence to support this claim. is rapidly decreased, there is increased retinal extravasation
The letter reports that A1C levels were determined every of serum proteins. If there is a concomitant increase in the
3 months and that both women had values <7% at each test. concentration of serum growth-promoting factors, a predis-
However, an A1C of 7% may be associated with an increased posed retina may deteriorate. Pregnancy per se is the most
risk of fetal malformations. Since organogenesis is complete frequently reported situation in which rapid normalization
within the first 7 weeks of pregnancy29 and women tend to of blood glucose is associated with deterioration of retinal
improve their glycemic control as the pregnancy progresses, status. Normal pregnancy is associated with high concen-
an A1C measured at 3 months of pregnancy is a poor reflec- trations of many growth-promoting factors.35–40 Hill et al.37
tion of the mother’s blood glucose levels at conception and reported that a potent mitogen and angiogenic factor nor-
during the critical first organogenic weeks of pregnancy. mally absent from the adult circulation become detectable
The report also indicated that both women maintained a by 14 weeks of gestation and is maximal at 22–32 weeks of
mean blood glucose level of <108 mg/dL. A pregnant wom- gestation. A placental growth hormone variant had been
an’s target blood glucose should be <90 mg/dL fasting and found to increase throughout pregnancy, along with hCS
<120 mg/dL postprandially.29 If the women measured their and prolactin.38
fasting blood glucose only, the reported mean is obviously The production of maternal insulin-like growth factor
too high. If postprandial measurements were also taken into (IGF)-I has also been shown to increase significantly above
account, the mean is still too high, although less so. These nonpregnant levels.34 It is well known that diabetes mellitus
women would be categorized as being at high risk for bear- is associated with perturbations of growth hormone IFG-I in
ing infants with malformations. Throughout pregnancy, the cases of poor metabolic control.37 Kitzmiller et al.41 have sug-
second mother was being treated for hypertension, and if the gested that treatment with insulin lispro during pregnancies
Possible effects of rapid-acting insulin analogs on the mother 191
complicated by diabetes may be associated with accelera- retinopathy, 1 did develop moderate hemorrhages, exudates,
tion of diabetic retinopathy. If treatment with lispro insulin and intraretinal microaneurysms. Of their 20 patients with
did play a role in the rapid deterioration of retinopathy in initial background retinopathy, 2 progressed to proliferative
the case reports, it most likely was not mediated by IFG-I retinopathy. Laatikainen et al.45 confirmed that the decrease
activity of lispro. Human insulin binds to the IFG-I receptor in A1C levels was most rapid in the two patients with the
with an affinity of 0.1%–0.2% that of IFG-I. A comparison of worst progression. They concluded that a rapid near normal-
insulin lispro and human insulin was made to determine the ization of glycemic control during pregnancy could accelerate
relative IFG-I receptor binding affinity in human placenta the progression of retinopathy in patients with poorly con-
membranes, skeletal muscle, smooth muscle cells, and mam- trolled diabetes. The DIEP study34 reported that 10.3% of
mary epithelial cells. Insulin lispro had a slightly higher diabetic women who progressed, despite no retinopathy at
affinity for the human placenta membranes when com- baseline, had an initial A1C elevation of four SD above the
pared with human insulin (1.3 times greater than human mean of a normal population (risk progression 40%, odds
insulin). No other differences were observed in any other ratio [OR] 2.4). Independent of retinal status, the DIEP study
cell lines. Despite the suggested increased affinity, it should also reported that the duration of diabetes increased the risk
be noted that the absolute affinity for the IFG-I receptor is of progression such that after 6 years of duration of diabe-
extremely low for both insulin lispro and human insulin. tes, the OR was 3.0; by 11–15 years, it was 9.7; and >16 years
Concentrations >1000 times above the normal physiologic it was 15.0, but hyperglycemia was a stronger risk factor.
range are needed to reach 50% receptor binding. IFG-I is a Additional evidence has been reported by the Diabetes
much larger protein chain than insulin, and there is a 49% Control and Complications Trial (DCCT).46 In the con-
homology between human insulin and IFG-I. The reversal ventional care group who became pregnant (n = 135), and
of the B28 and B29 amino acids in insulin lispro increases thus had immediate intensification of glucose control, 47%
this homology to 51%, because of the analogous position worsened their retinal status, and the OR for progression by
in the IGF molecule. It has been shown that insulin lispro the second trimester was 2.6 compared to diabetic women
has the same affinity for the IFG-I receptor as does human in the conventional group who did not become pregnant. In
insulin; also, the dissociation kinetics of insulin lispro on order to compare the A1C levels reported by the DCCT43 to
the insulin receptor are identical to those of insulin, indi- other published reports, the approximate equivalent baseline
cating that insulin lispro should have no excess mitogenic A1C levels, using the DCCT normal range, is 7.1%, 9.8%, and
effect via either the IFG-I or the insulin receptor.42,43 Patients 9.5%, respectively. In addition, the rate of fall of the A1C level
in the Kitzmiller reports41 all had elevated levels of IFG-I in the three case reports was faster than the reported rate of
due to poor control of their diabetes and due to pregnancy fall associated with deterioration of retinal status.
per se, independent of the possible IFG-I activity of lispro. There is one case report in the literature that clearly
However, anecdotal cases should never be used to infer a shows that the combination of pregnancy and rapid normal-
cause–effect relationship. In controlled clinical trials of ization of severe hyperglycemia is sufficient to “explode” a
>2000 patients with insulin lispro, no significant differences previously normal retina. Hagay et al.47 reported a case of
in retinopathy were observed, but there were no pregnant a woman with no previously documented hyperglycemia
women in this trial.15 The factors that emerge as the indepen- who presented at 8 weeks of gestation with an HbA1C level
dent risk factors for retinopathy progression include elevated of 16%, and her ophthalmic examination was reported to be
A1C at baseline, duration of diabetes, significant proteinuria “completely normal.” She was treated with intensive insulin
(>300 mg/24 hours), pregnancy, and rapid normalization of therapy and at 12 weeks her A1C level was 5.9%. By the sec-
blood glucose (in <14 weeks). In fact, the strongest risk factor ond trimester, she had severe bilateral proliferative diabetic
for retinopathy progression, independent of baseline retinal retinopathy needing photocoagulation. In a report by Omori
status, is baseline elevation of A1C associated with a rapid et al.48 studying Japanese pregnant diabetic women, the
decline to normal. Of 14 patients who were treated with prevalence of retinopathy was 34.4% in their type 2 diabetic
insulin lispro during pregnancy, described by Kitzmiller population. In addition, the prevalence of proliferative reti-
et al.,41 11 had risk factors, including the evidence of baseline nopathy was as high in the women with type 2 diabetes as in
retinopathy, which have been associated with progression to the pregnant women with type 1 diabetes, despite the shorter
proliferative retinopathy during pregnancy. Therefore, we duration of documented diabetes in their patients with type
can learn the fact that if there is no retinopathy during the 2 diabetes. The need for photocoagulation occurred in 50%
first trimester of a pregnancy complicated by diabetes, proof their pregnant diabetic patients with type 2 diabetes with
gression to proliferative retinopathy needing laser therapy is greater than background retinopathy at the beginning of
rare; however, many of the cited references earlier empha- pregnancy. If insulin lispro did play a role in the progression
size that baseline elevation of glucose associated with rapid of retinopathy, it is more likely that the insulin lispro facili-
normalization can accelerate retinopathy. Phelps et al.44 tated the rapid normalization of the blood glucose levels. In
clearly showed that deterioration of retinopathy correlated pregnant diabetic patients, it has been shown that insulin
significantly with the levels of plasma glucose at entry and lispro improves glucose control and thus significantly lowers
with the magnitude of improvement in glycemia during the the A1C level compared to patients who are administered by
first 6–14 weeks after entry. Although the 13 patients with human regular insulin.17,22 There is danger in normalizing
no retinopathy at baseline did not progress to proliferative blood glucose quickly, regardless of the type of insulin used,
in pregnant women with a long duration of diabetes and ele- blood glucose. Before 1985, impure animal insulin was used,
vated A1C levels in the first trimester, in those with protein- with a result that the IgG antibody levels rose the longer
uria and perhaps those with type 1 diabetes. To date, there the women were treated. After 20 years of diabetes, women
have not been extensive clinical trials on the retinal status had antibody levels >10,000. Purified human insulin has
of women treated with insulin aspart (or any other insulin been available for >20 years; thus there is a new generation
analog) in pregnancy. Thus, we await clinical trials of the use of type 1 diabetes in women who have never been treated
of all of the insulin analogs in pregnancy to make a decision with animal insulins and therefore have negligible antibod-
about the role of analogs in the progression of retinopathy ies. Before giving these women insulin analogs, it must be
during pregnancy. Busy clinics may have a decreased abil- proved that (1) they do not cause an immunologic response,
ity to examine retinae thoroughly. Mild background reti- (2) they do not cross the placenta, (3) they do not increase the
nopathy may be missed, even in the best of settings. Any risk of congenital anomalies or spontaneous abortions, and
retinopathy increases the risk, especially if the blood glucose (4) they do not significantly increase the serum IFG-I levels
level is elevated. Rather than recommending angiography or accelerate diabetic retinopathy.
to all women before each pregnancy is planned, in the case
of no retinopathy seen on retinal examination, it is prudent
to improve the glucose control slowly. These case reports
reinforce the need to intensify preconceptional care pro- Definition of normoglycemia based
grams to allow the luxury of slowly normalizing the blood on infant outcome
glucose and to plan the pregnancy only after the blood glu-
cose levels have been stabilized in the normal range for at Previously, glucose control and targets for treatment were
least 6 months.46–48 If a patient presents pregnant, with high based on clinical judgment and concern for hypoglycemia.
A1C levels, regardless of the retinal status, as suggested by In fact, most clinicians preferred to maintain hyperglycemia
these cases, then a retinal specialist needs to be on the team, rather than increase the risk of a hypoglycemic reaction. As
be vigilant, and treats any developing angiopathy while the tools and techniques improved to achieve near-normal glu-
blood glucose is normalized. If an insulin analog was avail- cose levels during pregnancy, the emphasis has changed to
able that was not immunogenic and had the rapid action of strive for the degree of maternal metabolic control that is
insulin lispro but had less IFG-I activity than human insulin, associated with normal body size and proportions in full-
then even if there was no proof that the IFG-I activity of the term infants. Mello et al.50 published a study in which they
insulin plays a role in the acceleration of diabetic retinopa- investigated the anthropomorphic characteristics of 98 full-
thy when the blood glucose level is normalized quickly, such term singleton infants born to 98 women with type 1 diabe-
an insulin would become the treatment of choice. Insulin tes. They reported that those women who had a mean daily
aspart, an insulin analog that has been shown to produce a blood glucose level <95 mg/dL had normal infants whereas
peak blood level at 40 minutes and lowers the postprandial the women with mean blood glucose levels >95 mg/dL
glucose levels significantly better than human insulin, has delivered infants with an increased prevalence of being large
only 69% the IFG-I activity of human insulin. for gestational age, with a significantly greater ponderal
Long-acting insulin analogs have only recently been used index and thoracic circumference with respect to the con-
in clinical practice.49 The first clinically available long-acting trol group. Others have confirmed that overall mean glucose
insulin analog is insulin glargine. Insulin glargine has a gly- levels of <95 mg/dL can avoid alterations in fetal growth.5
cine substitution in the alpha chain at the 21 position and Jovanovic et al.51 studied 52 women with type 1 diabetes and
two glargines attached to the beta chain terminal at posi- found that when the mean blood glucose was maintained
tion 30. It is a soluble insulin and has been shown to pro- between 80 and 84 mg/dL, the outcome of pregnancy was
vide peakless, sustained, and predictable 24 hours action. Of normal. Langer et al.52 assessed the relationship between
note, insulin glargine has a sixfold increase in IFG-I activity optimal levels of glycemic control and perinatal outcome in
as compared to human insulin. There are no clinical trials a prospective study of 334 women with GDM and found that
using insulin glargine in pregnancy, nor are there reports when the mean glucose levels were <86 mg/dL, this group
of the retinal status in any of the 32 women in the previ- had a significantly higher prevalence of SGA infants. In con-
ously published case reports of the use of glargine in preg- trast, when the mean glucose levels were >105 mg/dL, there
nancy.24–27 However, a recent paper27 describing 118 women was a 20% prevalence of large-for-gestational-age infants.
treated with glargine during pregnancy showed that the Hellmuth et al.53 found that the frequency of hypoglycemia,
macrosomia rate was increased fourfold over that reported especially nocturnal hypoglycemia, was seen with a preva-
using human insulins. Of note, insulin glargine has a sixfold lence of 37% in the first trimester of pregnancies treated with
increase in IFG-I activity as compared to human insulin. intensified insulin therapy. Sacks et al.53 studied 48 women
Another insulin analog that is currently in clinical trials is with type 1 diabetes and 113 women with type 2 diabetes
insulin detemir. Here again there are no trials using this during pregnancy. They found that the mean glucose lev-
insulin in pregnancy, but the studies show that detemir has els were higher in the patients with type 1 diabetes and at
the same IFG-I activity as human insulin.27 Insulin detemir least one daily glucose level was <50 mg/dL during 19% of
is now category B. The clinician has to keep in mind that the observational days compared to only 2% of days in the type 2
most important concern is to safely normalize the maternal diabetes group. Rosenn et al.54 and Rosenn and Miodovnik55
Insulin requirements 193
published papers on the topic of the increased risk of hypo- decrease in insulin requirement. There are also reports of
glycemia during pregnancies complicated by diabetes. They rising insulin requirement in the first trimester. The pres-
found that at least 40% of mothers reported hypoglycemia ent authors have described the insulin requirements during
during pregnancy. Clinically significant hypoglycemia requir- pregnancy of a population of women with well-controlled
ing assistance from another person occurred in 71% of the type 1 diabetes, which possibly lends credence to the notion
84 women studied, with a peak incidence occurring between that first-trimester overinsulinization may be the cause of
10 and 15 weeks of gestation. They did not observe any the hypoglycemia seen by some workers in the first trimester.
increase in embryopathy. Jovanovic et al.56 then published In addition, together with the DIEP study group, the present
the first-trimester insulin requirements of women studied in authors have analyzed the insulin requirement and substrat-
the DIEP study. They showed that there was a drop in insulin ified based on the degree of glucose control in the first tri-
requirements during 8–11 weeks of gestation, which was seen mester.56 The weekly insulin requirement (as units/kg/day)
in all women with type 1 diabetes whose glucose concentra- were examined in the first trimester of diabetic women in the
tions were insensitively managed. The majority of decreases DIEP study with accurate gestational dating, regular glucose
in the insulin requirement were seen in the overnight monitoring, daily insulin dose recording, and monthly A1C
period. It was concluded that the insulin doses needed to be measurements. In pregnancies that resulted in live-born,
decreased for the overnight insulin requirement to prevent term, singleton infants, a significant increase in mean weekly
nocturnal hypoglycemia during the late first trimester. dosage was observed in weeks 3–7 (p < 0.001), followed by a
significant decline in weeks 8–15 (p < 0.001). The Friedman
nonparametric test localized significant changes to the
Insulin requirements interval between weeks 7–8 and weeks 11–12. To determine
if prior poor glucose control exaggerated these trends, the
Women with type 1 diabetes must increase their insulin women were divided based on their A1C values: <2 SD above
dosage to compensate for the diabetogenic forces of normal the mean of a normal population (group 1), 2–4 SD (group 2),
pregnancy. However, the exact pattern of insulin dosage and >4 SD (group 3) at baseline. Late first-trimester declines
increase is still controversial. Many observers have detected in dosage were statistically significant in group 2 (p = 0.002)
a decline in insulin requirement late in the first trimester of and in groups 2 and 3 together (p = 0.003). Similarly, women
diabetic pregnancies. Jorgen Pedersen, the father of the study with body mass index >27.0 had a greater initial insulin rise
of diabetes in pregnancy, was among the first to observe that and then a fall compared to leaner women. Observations
first-trimester hypoglycemia was a symptom of pregnancy in the DIEP study cohort disclosed a mid-first-trimester
and that it was a common knowledge among the physicians decline in insulin requirement in pregnant women with
of today. Pedersen wrote: insulin-dependent diabetes. Possible explanations include
overinsulinization of previously poorly controlled diabetes
Those physicians who manage diabetic women should be and a transient decline in progesterone during the late first-
particularly alert for hypoglycemia in women who have trimester luteoplacental shift in progesterone production.
recently become pregnant. About the 10th week of gesta- Clinicians should anticipate a reduction in insulin require-
tion there is an improvement in glucose tolerance mani- ment in the 4-week interval between weeks 8 and 12 of gesta-
festing itself as insulin coma, milder insulin reaction or tion. Based on these studies of women with well-controlled
an improvement in the degree of compensation. When a diabetes, an algorithm for care and an insulin requirement
reduction in insulin dosage is called for it amounts to an protocol has been created based on gestational week and
average of 34%. the woman’s current pregnant body weight. The total daily
dose of insulin in the first-trimester (weeks 5–12) insu-
Indeed, he even claimed that “Once in a while pregnancy lin requirement is 0.7 unit/kg/day; in the second trimester
may be diagnosed on account of inexplicable hypoglycemic (weeks 12–26), the insulin requirement is 0.8 unit/kg/day;
attacks.” In a total of 26 cases of insulin coma collected, all in the third trimester (weeks 26–36), the insulin require-
of the cases in his series occurred in months 1–4 of gesta- ment is 0.9 unit/kg/day; and at term (weeks 36–40), the insu-
tion, with the majority occurring at months 2–3. He also lin requirement is 1.0 unit/kg/day (Table 23.1). The insulin
noted that by late gestation, regardless of the metabolic con- needs to be divided throughout the day to provide the basal
trol and duration of diabetes, average daily insulin require- need (the dose of insulin that keeps a woman normal in the
ments increased by twofold from earlier in pregnancy. fasting state) and meal-related need.
Earlyfirst-trimester overinsulinization might explain a later When multiple insulin injections are used to provide the
first-trimester drop in insulin requirement. One example basal need, NPH insulin is preferred because it has a more
of this effect may be the significantly greater weight gain predictable absorption pattern than Lente or Ultralente
seen in the first trimester by diabetic women compared to insulin. Also, the recently developed long-acting insulin
healthy nondiabetic women. Perhaps the drive to increase analog, insulin determir, has just received the FDA category
caloric intake to prevent hypoglycemia in the first trimes- B determination. The preferred use of NPH is to give one-
ter may have been the cause of the first-trimester excessive sixth of the total daily dose of insulin (I) as morning, dinner,
weight gain in the diabetic women compared to the controls. and bedtime injections (i.e., NPH dose equals 50% of daily
On the other hand, others have not seen the first-trimester dose divided into three equal injections of NPH given every
Table 23.1 Insulin dosage regimen for diabetic pregnancy
Pregnancy NPH plus rapid-acting insulin schedule Patient weight in kg = Date and time:
“Big I” = total daily units of
insulin
Circle one: gestational weeks 13–28 29–34 35–40 Other
= 0–12
k = 0.7 0.8 0.9 1.0
Calculate desired units of insulin from above line.
“Big I” =___(k units ×weight kg)/24 hours
“Big I” = basal insulin requirement + bolus (meal related) insulin requirement
Basal = ½ “Big I,” bolus = ½ “Big I”
Basal: Divide so that ⅙ of “Big I” is NPH given before breakfast, ⅙ of “Big I” is NPH given before dinner, and ⅙ of “Big I” is
NPH given before bedtime.
Bolus: Divide so that ⅙ of “Big I” is rapid-acting insulin given before breakfast, ⅙ of “Big I” is rapid-acting insulin given
before lunch, and ⅙ of “Big I” is rapid-acting insulin given before dinner. The rapid-acting insulin is then titrated based on the
blood glucose.
0800 Prebreakfast: NPH = ⅙ “Big I” = ____
Check yesterday’s predinner BS:
If yesterday’s predinner BS < 60, then decrease today’s a.m. NPH by 2 units.
If yesterday’s predinner BS 61–90, no change in today’s a.m. NPH.
If yesterday’s predinner BS > 91, then increase today’s a.m. NPH by 2 units.
Do not feed the patient until Rapid-acting insulin = “Big I” = : to be adjusted according to the following scale:
the blood sugar is below
Prebreakfast BS <60 = ___ = (⅙ “Big I” dose) – 3% of the “Big I.”
120 mg/dL.
61–90 = ___ = ⅙ “Big I” dose.
91–120 = ___ = (⅙ “Big I” dose) + 3% of the “Big I.”
>121 = ___= (⅙ “Big I” dose) + 6% of the “Big I.”
If today’s BS 1 hour after breakfast is <110, then decrease tomorrow’s prebreakfast rapid-
acting insulin by 2 units.
If today’s BS 1 hour after breakfast is 111–120, no change in tomorrow’s prebreakfast rapid-
acting insulin.
If today’s BS 1 hour after breakfast is >121, then increase tomorrow’s prebreakfast rapid-acting
insulin by 2 units.
1200 Prelunch: Rapid-acting insulin is Vs “Big I” = ___ to be adjusted according to the following scale:
Do not feed the patient until Prelunch BS < 60 = ___ = (⅙ “Big I” dose) – 3% of “Big I.”
the blood sugar is below
61–90 = ___ = ⅙ “Big I” dose.
120 mg/dL.
91–120 = ___ = (⅙ “Big I” dose) + 3% of “Big I.”
>121 = ___ = (⅙ “Big I” dose) + 6% of “Big I.”
If today’s BS 1 hour after lunch is <110, then decrease tomorrow’s prebreakfast rapid-acting
insulin by 2 units.
If today’s BS 1 hour after lunch is 111–120, no change in tomorrow’s prebreakfast rapid-acting
insulin.
If today’s BS 1 hour after lunch is >121, then increase tomorrow’s prebreakfast rapid-acting
insulin by 2 units.
(Continued)
Insulin requirements 195
Table 23.1 (Continued) Insulin dosage regimen for diabetic pregnancy
1700 Predinner: NPH = “Big I” = ___

Do not feed the patient until Rapid-acting insulin is ⅙ “Big I” =: to be adjusted according to the following scale:
the blood sugar is below If yesterday’s prebedtime BS < 60, then decrease today’s dinner NPH by 2 units.
120 mg/dL.
If yesterday’s prebedtime BS 61–90, no change in today’s dinner NPH.
If yesterday’s prebedtime BS > 91, then increase today’s dinner NPH by 2 units.
Predinner BS < 60 = ___ = (⅙ “Big I” dose) – 3% of “Big I.”
61–90 = ____ = ⅙ “Big I” dose.
91–120 = ___ = (⅙ “Big I” dose) + 3% of “Big I.”
>121 = ____ = (⅙ “Big I” dose) + 6% of “Big I.”
If today’s BS 1 hour after dinner is <110, then decrease tomorrow’s dinner rapid-acting insulin
by 2 units.
If today’s BS 1 hour after dinner is 111–120, no change in tomorrow’s dinner rapid-acting
insulin.
If today’s BS 1 hour after dinner is >121, then increase tomorrow’s dinner rapid-acting insulin
by 2 units.
2400 Bedtime NPH: Give ⅙ “Big I” = ___
If today’s prebreakfast BS is <60, then decrease today’s bedtime NPH by 2 units.
If today’s prebreakfast BS is 61–90, no change in today’s bedtime NPH.
If today’s prebreakfast BS is >91, then check the 3 a.m. BS, and, if it is <70 (regardless of
today’s prebreakfast BS), decrease today’s bedtime NPH by 2 units.
If today’s prebreakfast BS is >91 and the 3 a.m. BS > 70, increase today’s bedtime NPH by
2 units.
Also, if the 3 a.m. BS is > 91, then call the doctor for 3 a.m. rapid-acting insulin scale equal
to the prelunch rapid-acting insulin scale.
8 hours, or at 8 a.m., 4 p.m., and 12 midnight; Table 23.1). the total daily insulin requirement. Clinicians should note
The other half of the total daily insulin dose should be a that hyperglycemia would occur if the patient used only
rapid-acting insulin (insulin lispro or insulin aspart) given insulin lispro or insulin aspart for the meal-related needs
before each meal to control postprandial glycemia (Table and if the woman goes a long time between meals. The dose
23.1). This dose of rapid-acting insulin can be given using of NPH insulin may not be sufficient to prevent an escape
the insulin infusion pump or by multiple doses of subcu- of the blood glucose before the next dose of insulin is given.
taneously injected insulin (Tables 23.1 through 23.3). The To prevent this escape of blood glucose when >3 hours
meal-related insulin dose (one-half of the total daily insu- elapse between injections of the rapid-acting insulin ana-
lin requirement [0.5I]) (Table 23.1). The exact division of logs of lispro or aspart, the patient should add 3% of her
this meal-related insulin dose depends on the size of the total daily insulin requirement as regular human insulin to
woman’s lunch versus her dinner. Compensatory doses to the lispro injection to extend the effectiveness of the rapid-
adjust for high or low glucose levels are calculated as 3% of acting component.
Table 23.2 Basal insulin pump program (using human regular, insulin lispro,
or insulin aspart)
Basal requirement (Ba)

Period (hourly infusion rate) Rationale
Midnight to 4 a.m. 50% less basal (B/24 × 0.5) Maternal cortisol at nadir
4–10 a.m. 50% more basal (B/24 × 1.5) Highest level of maternal cortisol
10 a.m. to noon Basal (B/24)
Note: To refine basal settings, have the patient perform self-monitoring blood glucose at the end of
each period to determine whether adjustments are needed. For instance, at the 4 a.m. test,
the blood glucose level should be 60–90 mg/dL; if blood glucose level is out of this range,
dial up or down insulin in increments of 0.10 unit/hour. (Consider using 0.1 unit/kg as NPH
insulin at bedtime to prevent diabetic ketoacidosis secondary to needle slippage; then
decrease the overnight basal from 4 to 10 a.m. by 0.02 unit/kg/hour.)
a B = 0.5I (the total daily insulin) or an hourly rate of B/24.
not be necessary to improve pregnancy outcome. Caruso

Table 23.3 Premeal sliding scale dose calculation
et al.63 treated 12 women with poorly controlled pregnant
using rapid-acting insulina
diabetes with insulin infusion pumps and showed that
glucose levels could be quickly normalized with a remark-
Premeal basal
able decrease in variation of glucose excursions. In addi-
glucose (mg/dL) Compensatory insulin
tion, they showed that insulin, glucose, and C-peptide
<60 Meal-related insulin dose minus 3% Ib levels in the amniotic fluid were normal and none of the
61–90 Meal-related insulin dose: no infants were macrosomic or had any neonatal problems.
adjustment necessary They concluded that insulin infusion pump therapy was
91–120 Meal-related insulin dose plus 3% Ib highly effective compared to intensified conventional
121 Meal-related insulin dose plus 6% Ib treatment. In a nested case- controlled study, Simmons
a Human regular insulin or insulin lispro or insulin aspart. et al.64 utilized insulin pumps in pregnancies complicated
by type 2 diabetes and GDM in a multiethnic community
and showed in 30 women that none experienced severe
hypoglycemia and 79% had improved glycemic control
within 1–4 weeks. Mothers using the pump had greater
insulin requirements and greater weight gain. Although
Insulin infusion pumps their infants were more likely to be admitted to the spe-
Insulin infusion pumps have been used for treatment for cial care unit, they were not heavier nor did they have
over two decades.57 However, the data on the safety and effi- more hypoglycemic events than control subjects. Jensen
cacy of insulin pumps during pregnancies complicated by et al.65 reported their series using insulin infusion pump
diabetes are still in its infancy. Kitzmiller et al.58 showed that therapy (n = 11) in the preconception treatment period in
insulin pump therapy did improve glucose control and mini- women with type 1 diabetes compared to women treated
mized clinically significant hypoglycemic events to 2.2/week. with conventional therapy (n = 9). Two fetuses born of
Coustan et al.59 then reported a randomized clinical trial of mothers treated with conventional therapy exhibited early
insulin pump therapy versus conventional therapy in preg- group delay, whereas all 11 fetuses born of mothers treated
nancies complicated by diabetes. They showed that there with pump therapy were normal; there were no malfor-
were no differences between the two treatment groups with mations in either group. Gabbe et al.66 published a series
respect to outpatients’ mean glucose levels, symptomatic of 24 patients with type 1 diabetes and reported no dif-
hypoglycemia, or HbA1C levels. They concluded that excel- ference, in the groups of women treated with pump com-
lent diabetes control can be achieved with both insulin infu- pared with those treated with intensified insulin therapy
sion pumps and with multiple injections of insulin. Carta for episodes of hypoglycemia, in the costs, complication,
et al.60 also performed a randomized trial of continuous glycemic control, or pregnancy outcome. The advantages
subcutaneous insulin infusion versus intensive conventional seen were all p ostpartum because those women who were
insulin therapy in pregnant women with type 1 and type 2 on the pump into their postpartum period sustained bet-
diabetes. They reported that there were not significant dif- ter glucose control than those who were on intensified
ferences in mean insulin requirements at the different stages insulin therapy during pregnancy.
of gestation and that perinatal outcome was satisfactory in The conclusion drawn from this review of the literature
both groups; however, in their study, control of fetal growth on using insulin pump therapy in pregnant women with
was better with interfiled convention therapy compared to diabetes is that pump therapy is not necessary in order to
fetal growth in the pump group. Mancuso et al.61 studied the achieve and maintain optimal control (27.4); however, 5 of
efficacy of the insulin pump in the home treatment of preg- the 10 papers suggest that insulin pump therapy has an
nant diabetic women. They reported that seven women with advantage over intensified multiple injections of insulin. In
type 1 diabetes, who used the pump, delivered term infants addition, only 155 patients were reported in these 10 trials,
and had no macrosomia or neonatal problems along with and in all but one paper, the women had their pump therapy
normal glucose tolerance tests at 2 years of life. started in the second or third trimester (Table 23.4).
Potter et al.57 studied continuous insulin infusion in
the third trimester of eight pregnancies complicated by
diabetes. Insulin algorithms for continuous
Compared to historical controls, they concluded that
diurnal variations of blood glucose concentrations were insulin infusion pump therapy in
dampened. Leveno et al.62 performed a case-controlled trial pregnancy
of insulin pump therapy versus intensified conventional
therapy and observed no significant differences between The basal need is usually 50% of the total daily insulin dose
the groups for glucose control, cost, and complications. (0.5I) and may be delivered using a constant infusion pump
They concluded that insulin pumps were not acceptable to (Table 23.2) or by multiple doses of intermediate-acting
all pregnant diabetic women and that such therapy may insulin (Table 23.1). When using a constant infusion pump,
Insulin and glucose requirements postpartum 197
Table 23.4 Review of the literature on using insulin infusion pump therapy in pregnancies
complicated by diabetes
Trimester or weeks of
Author (reference) No. on CSIIa gestation Type of DMb Type of trial/comments
Potter et al. 57 8 Third 1 Longitudinal/improved
Kitzmiller et al.58 24 5–10 1 Longitudinal/no difference
Coustan et al.59 22 Second and third 1 Randomized/no difference
Carta et al.60 14 First 1 Case control/improved
Mancuso et al.61 12 Third 1 Longitudinal/improved
Jensen et al.65 9 Before and all trimesters 1 Case control
Caruso et al.63 12 Third 1 Longitudinal/improved
Leveno et al.62 11 Second and third 1 Self-selection/no difference
Gabbe et al.66 23 Second, third, and postpartum 1 Only postpartum improved
Simmons et al.64 30 Second GDM and 2c Equal
a Continuous subcutaneous insulin infusion.
b Diabetes mellitus.
c Gestational diabetes mellitus.
the basal need is calculated as an hourly rate (Table 23.2) and dextrose infusion is begun at a rate of 2.55 mg/kg/minute.
is delivered such that the calculated rate (0.5I or total dose If labor is latent, normal saline is usually sufficient to main-
over 24 hours divided by 24) is given between 10 a.m. and tain normoglycemia until active labor begins, at which
midnight. The rate is cut in half (i.e., 0.5I divided by 24 times time dextrose is infused at 2.55 mg/kg/minute. Blood glu-
0.5) from midnight to 4 a.m. and increased by another 50% cose is then monitored hourly, and if it is <60 mg/dL, then
(i.e., 0.5I divided by 24 times 1.5) to counteract the morn- the infusion rate is doubled for the subsequent hour. If the
ing rise of cortisol levels that are potentiated during preg- blood glucose rises to >120 mg/dL, 2–4 units of regular
nancy. Also, low-dose NPH before bedtime has been used insulin are given i.v. each hour until the blood glucose level
by some clinicians to prevent the possible occurrence of dia- is 70–90 mg/dL. In the case of an elective cesarean section,
betic ketoacidosis if the needle slips out of position during the bedtime dose of NPH insulin is repeated at 8 a.m. on the
the overnight period. This dose of NPH insulin needs to be day of surgery and every 8 hours if the surgery is delayed. A
sufficient to provide protection from ketosis, or 0.1 unit of dextrose infusion may be started if the plasma glucose level
NPH times the weight of the women in kilograms. Then the falls to <60 mg/dL, and 2–4 units of regular insulin given
overnight basal insulin needs to be decreased to allow for i.v. every hour if the blood glucose rises to >120 mg/dL.67
the NPH dose. The 4–10 a.m. basal insulin should thus be
adjusted downward by 0.02 unit/kg/hour.
Insulin and glucose requirements
Insulin and glucose treatment postpartum
during labor Maternal insulin requirements usually drop precipitously
postpartum, possibly for 48–96 hours. Insulin require-
With improvement in antenatal care, intrapartum events ments should be recalculated at 0.6 unit/kg based on the
play an increasingly crucial role in the outcome of preg- postpartum weight and should be started when the 1-hour
nancy. The artificial beta cell may be used to maintain nor- postprandial plasma glucose value is >150 mg/dL or the
moglycemia during labor and delivery, but normoglycemia fasting glucose level is >100 mg/dL. The postpartum caloric
can be maintained easily by subcutaneous injections. Before requirements are 25 kcal/kg/day, based on the postpartum
active labor, insulin is required, and glucose infusion is not weight. For women who wish to breastfeed, the calculation is
necessary to maintain a blood glucose level of 70–90 mg/dL. 27 kcal/kg/day and insulin requirements are 0.6 unit/kg/day.
With the onset of active labor, insulin requirements decrease The insulin requirement during the night drops dramatically
to zero, and glucose requirements are relatively constant at during lactation, owing to the glucose siphoning into the
2.5 mg/kg/minute. From these data, a protocol for supply- breast milk. Thus, the majority of the insulin requirement
ing the glucose needs of labor has been developed.67 The is needed during the daytime to cover the increased caloric
goal is to maintain the maternal plasma glucose between needs of breastfeeding. Normoglycemia should especially
70 and 90 mg/dL. In cases of the onset of active sponta- be prescribed for nursing diabetic women, because hypergly-
neous labor, insulin is withheld and an intravenous (i.v.) cemia elevates milk glucose levels.68
Conclusions others, emphasize the need for preconception programs and

the need for support systems to facilitate the maintenance
With the advent of tools and techniques to maintain nor- of normoglycemia throughout pregnancy. The morbidity
moglycemia before, during, and between all pregnancies and subsequent development of the infant of the diabetic
complicated by diabetes, infants of diabetic mothers now mother is associated with hyperglycemia. Therefore, the goal
have the same chances of good health as those infants born of insulin therapy is to achieve and maintain normoglyce-
to the nondiabetic woman. Animal and human studies mia before, during, and after all pregnancies complicated by
clearly implicate glucose as the teratogen. These studies, and diabetes.
REFERENCES
1. Pedersen J. Course of diabetes during pregnancy. Acta Endocrinol 20. Jehle PM, Fussgaenger RD, Seibold A et al. Pharmacodynamics of
1952; 9: 342–347. insulin lispro in 2 patients with type II diabetes mellitus. Int J Clin
2. Mills JL, Knopp RH, Simpson JL et al. Lack of relation of increased Pharmacol Ther 1996; 34: 498–503.
malformation rates in infants of diabetic mothers to glycemic 21. Balsells M, Corcoy R, Mauricio D et al. Insulin antibody response
control during organogenesis. N Engl J Med 1988; 318: 671–676. to a short course of human insulin therapy in women with gesta-
3. Mills JL, Simpson JL, Driscoll SG et al. Incidence of spontaneous tional diabetes. Diabetes Care 1997; 20: 1172–1175.
abortion among normal women and insulin-dependent diabetic 22. Wyatt JW, Frias JL, Hoyme HE et al., IONS Study Group.
women whose pregnancies were identified within 21 days of Congenital anomaly rate in offspring of mothers with diabetes
conception. N Engl J Med 1988; 319: 1617–1623. treated with insulin lispro during pregnancy. Diabet Med 2005;
4. Mills JL, Fishl AR, Knopp RH et al. Malformations in infants of 22: 803–807.
diabetic mothers: Problems in study design. Prev Med 1983; 12: 23. Mathiesen ER, Kinsley B, Amiel SA et al., Insulin Aspart Pregnancy
274–286. Study Group. Maternal glycemic control and hypoglycemia in
5. Jovanovic L, Peterson CM, Saxena BB et al. Feasibility of main- type 1 diabetic pregnancy: A randomized trial of insulin aspart
taining normal glucose profiles in insulin-dependent pregnant versus human insulin in 322 pregnant women. Diabetes Care
diabetic women. Am J Med 1980; 68: 105–112. 2007; 30: 771–776.
6. DeVeciana M, Major CA, Morgan MA et al. Postprandial versus 24. Holstein A, Plaschke A, Egberts EH. Use of insulin glargine dur-
preprandial blood glucose monitoring in women with gestational ing embryogenesis in a pregnant woman with type 1 diabetes.
diabetes mellitus requiring insulin therapy. N Engl J Med 1995; Diabet Med 2003; 20: 779–780.
333: 1237–1241. 25. Devlin JT, Hothersall L, Wilkis JL. Use of insulin glargine during
7. Combs CA, Gunderson E, Kitzmiller JL et al. Relationship of fetal pregnancy in a type 1 diabetic woman. Diabetes Care 2002; 25:
macrosomia to maternal postprandial glucose control during 1095–1096.
pregnancy. Diabetes Care 1992; 15: 1251–1257. 26. Dolci M, Mori M, Baccetti F. Use of glargine insulin before and
8. Jovanovic L, Peterson CM, Reed GF et al. Maternal postprandial during pregnancy in a woman with type 1 diabetes and Addison’s
glucose levels and infant birth weight: The diabetes in early preg- disease. Diabetes Care 2005; 28: 2084–2085.
nancy study. Am J Obstet Gynecol 1991; 164: 103–111. 27. Price N, Bartlett C, Gillmer M. Use of insulin glargine during
9. Jovanovic L. What is so bad about a big baby? Diabetes Care pregnancy: A case–control pilot study. Br J Obstet Gynecol 2007;
2001; 24: 1317–1318. 114: 453–457.
10. Knopp RH. Hormone mediated changes in nutrient metabolism 28. Diamond T, Kormas N. Possible adverse fetal effects of insulin
in pregnancy: A physiological basis for normal fetal develop- lispro. N Engl J Med 1997; 337: 1009.
ment. In: Jacobsen MS, Rees JM, Golden NH, Irwin CE, eds., 29. Mills JL, Baker L, Goldman A. Malformations in infants of diabetic
Adolescent Nutritional Disorders: Prevention and Treatment, mothers occur before the seventh gestational week: Implications
Academy of Sciences: New York, 1997, pp. 251–71. for treatment. Diabetes 1979; 23: 292.
11. Knight G, Worth RC, Ward JD. Macrosomia despite well-controlled 30. Jovanovic L. Role of diet and insulin treatment of diabetes in
diabetic pregnancy. Lancet 1983; 2: 1431. pregnancy. Clin Obstet Gynecol 2000; 43: 46–51.
12. Visser GHA, van Ballegooie E, Slutter WJ. Macrosomia despite 31. Kucera J. Rate and type of congenital anomalies among offspring
well-controlled diabetic pregnancy. Lancet 1984; 1: 284–285. of diabetic mothers. J Reprod Med 1971; 7: 61–64.
13. Menon RK, Cohen RM, Sperling MA et al. Transplacental passage 32. Anderson J, Bastyr E, Wishner K. Response to diamond and kor-
of insulin in pregnant women with insulin-dependent diabetes mel- mas. N Engl J Med 1997; 337: 1009–1012.
litus. Its role in fetal macrosomia. N Engl J Med 1990; 323: 309–315. 33. Jovanovic L. Retinopathy risk: What is responsible? Hormones,
14. Jovanovic L, Kitzmiller JL, Peterson CM. Randomized trial of hyperglycemia, or humalog? Diabetes Care 199; 22: 846–850.
human versus animal species insulin in diabetic pregnant women: 34. Chew EY, Mills JL, Metzger BE et al.; National Institute of Child
Improved glycemic control, not fewer antibodies to insulin, influ- Health and Human Development. Metabolic control and pro-
ences birth weight. Am J Obstet Gynecol 1992; 167: 1325–1330. gression of retinopathy: The Diabetes in Early Pregnancy Study.
15. Anderson Jr. JH, Brunelle RL, Koivisto VA et al. Reduction of Diabetes Care 1995; 18: 631–637.
postprandial hyperglycemia and frequency of hypoglycemia in 35. Merimee TJ, Zapf J, Froesch ER. Insulin-like growth factors:
IDDM patients on insulin-analog treatment. Diabetes 1997; 46: Studies in diabetics with and without retinopathy. N Engl J Med
265–270. 1983; 309: 527–531.
16. Fineberg SE, Rathbun MJ, Hufferd S et al. Immunologic aspects of 36. Larinkari J, Laatikainen L, Ranta T. Metabolic control and serum
human proinsulin therapy. Diabetes 1988; 37: 276–280. hormone levels in relationship to retinopathy in diabetic preg-
17. Jovanovic L, Ilic S, Pettitt DJ et al. The metabolic and immuno- nancy. Diabetologia 1982; 22: 327–331.
logic effects of insulin lispro in gestational diabetes. Diabetes 37. Hill DJ, Clemmons DR, Riley SC et al. Immunohistochemical
Care 1999; 22: 1422–1426. localization of insulin like growth factors and IGF binding pro-
18. Fineberg NS, Fineberg SE, Anderson JH et al. Immunologic effects teins-1, -2, and -3 in human placenta and fetal membranes.
of insulin lispro [Lys(B23), Pro (B29)] human insulin in IDDM and Placenta 1993; 14: 1–12.
NIDDM patients previously treated with insulin. Diabetes 1996; 38. MacLeod JN, Worsley I, Ray Y et al. Human growth hor-
45: 1750–1754. mone variant is a biologically active somatogen and lactogen.
19. Jehle PM, Fussgaenger RD, Kunze U et al. The human insulin ana- Endocrinology 1991; 128: 1298–1302.
log insulin lispro improves insulin binding on circulating mono- 39. Gluckman PD. The endocrine regulation of fetal growth in late
cytes of intensively treated insulin dependent diabetes mellitus gestation: The role of insulin-like growth factors. J Clin Endocrinol
patients. J Clin Endocrinol Metab 1996; 81: 2319–2327. Metab 1995; 80: 1047–1050.
References 199
40. Holly JMP, Amiel SA, Sandhu RR et al. The role of growth hor- 55. Rosenn BM, Miodovnik M. Glycemic control in the diabetic
mone in diabetes mellitus. J Endocrinol 1988; 118: 353–364. pregnancy: Is tighter always better? J Matern Fetal Med 2000; 9:
41. Kitzmiller J, Main E, Ward B et al. Insulin lispro and the devel- 29–34.
opment of proliferative diabetic retinopathy during pregnancy. 56. Jovanovic L, Mills Jl, Knopp RH et al.; the National Institute of
Diabetes Care 1999; 22: 874. Child Health and Human Development—Diabetes in Early
42. DiMarchi RD, Chance RE, Long HB et al. Preparation of an insu- Pregnancy Study Group. Declining insulin requirement in the
lin with improved pharmacokinetics relative to human insulin late first trimester of diabetic pregnancy. Diabetes Care 2001;
through consideration of structural homology with insulin-like 24: 1130–1136.
growth factor-1. Horm Res 1994; 41(Suppl. 2): 93–96. 57. Potter JM, Beckless JP, Cullen DR. Subcutaneous insulin infusion
43. Llewelyn J, Slieker LJ, Zimmermann JL. Pre-clinical studies on control of blood glucose concentration in diabetics in third tri-
insulin lispro. Drugs Today 1998; 34(Suppl. C): 11–21. mester of pregnancy. Br Med J 1980; 26: 1099–1101.
44. Phelps RL, Sakol L, Metzger BE et al. Changes in diabetic reti- 58. Kitzmiller JL, Younger MD, Hare JW et al. Continuous subcuta-
nopathy during pregnancy: Correlations with regulation of hyper- neous insulin therapy during early pregnancy. Obstet Gynecol
glycemia. Arch Ophthalmol 1986; 104: 1806–1810. 1985; 66: 606–611.
45. Laatikainen L, Teramo K, Hieta-Heikurainen H et al. A controlled 59. Coustan DR, Reece A, Sherwin RS et al. A randomized clinical
study of the influence of continuous subcutaneous insulin infu- trial of the insulin pump versus intensive conventional therapy in
sion treatment on diabetic retinopathy during pregnancy. Acta diabetic pregnancies. J Am Med Assoc 1986; 255: 631–635.
Med Scand 1987; 221: 367–376. 60. Carta O, Meriggi E, Trossarelli GF et al. Continuous subcutane-
46. Lachin J, Clearly P; the DCCT Research Group. Pregnancy ous insulin infusion versus intensive conventional insulin therapy
increases the risk of complication in the DCCT. Diabetes 1998; in type 1 and type 2 diabetic pregnancy. Diabet Metab 1986;
47(Suppl. 1): 1091. 12: 121–129.
47. Hagay ZJ, Schachter M, Pollack A, Ley R. Development of pro- 61. Mancuso S, Caruso A, Lanzone A et al. Continuous subcutane-
liferative retinopathy in a gestational diabetes patient following ous insulin infusion in pregnant diabetic women. Acta Endocrinol
rapid metabolic control [Case report]. Eur J Obstet Gynecol 1986; 277(Suppl.): 112–116.
Reprod Biol 1994; 57: 211–213. 62. Leveno KJ, Fortunato SJ, Raskin P et al. Continuous subcutaneous
48. Omori Y, Minei S, Tamaki T et al. Current status of pregnancy insulin infusion during pregnancy. Diabet Res Clin Pract 1988;
in diabetic women. A comparison of pregnancy in IDDM and 4: 257–268.
NIDDM mothers. Diabet Res Clin Pract 1994; 24: S273–S278. 63. Caruso A, Lanzone V, Massidda M et al. Continuous subcutane-
49. Trujillo A. Insulin analogs in pregnancy. Diabet Spectr April ous insulin infusion in pregnant diabetic patients. Prenat Diagn
2007; 20: 94–100. 1987; 7: 41–50.
50. Mello G, Paretti E, Mecacci F et al. What degree of maternal met- 64. Simmons D, Thompson CF, Conroy C, Scott DJ. Use of insulin
abolic control in women with type 1 diabetes is associated with pumps in pregnancies complicated by type 2 diabetes and gesta-
normal body size and proportions in full-term infants? Diabetes tional diabetes in a multiethnic community. Diabetes Care 2002;
Care 2000; 23: 1494–1498. 24: 2078–2082.
51. Jovanovic L, Druzin M, Peterson CM. Effect of euglycemia on 65. Jensen BM, Kuhl C, Petersen LM et al. Preconceptional treatment
the outcome of pregnancy in insulin-dependent diabetic women with insulin infusion pumps in insulin dependent diabetic women
as compared with normal control subjects. Am J Med 1981; with particular reference to prevention of congenital malforma-
71: 921–927. tions. Acta Endocrinol 1986; 277(Suppl.): 81–85.
52. Langer O, Levy J, Brustman L, Anyaegbunam A, Merkatz R, Divon 66. Gabbe SG, Holing E, Temple P, Brown ZA. Benefits, risks, costs,
M. Glycemic control in gestational diabetes mellitus. How tight is and patient satisfaction associated with insulin pump therapy for
tight enough: Small for gestational age versus large for gestational the pregnancy complicated by type 1 diabetes mellitus. Am J
age? Am J Obstet Gynecol September 1989; 161(3): 646–653. Obstet Gynecol 2000; 182: 1283–1291.
53. Hellmuth E, Damm P, Molsted-Pedersen L, Bendtson I. 67. Jovanovic L, Peterson CM. Insulin and glucose requirements dur-
Prevalence of nocturnal hypoglycemia in the first trimester of ing the first stage of labor in insulin-dependent diabetic women.
pregnancy in patients with insulin treated diabetes mellitus. Acta Am J Med 1983; 75: 607–612.
Obstet Gynecol Scand 2000; 79: 3–5. 68. Jovanovic L. (ed.) Medical Management of Pregnancy Compli
54. Rosenn BM, Miodovnik M, Holcberg G et al. Hypoglycemia: cated by Diabetes. American Diabetes Associate Inc.: Alexandria,
The price of intensified insulin therapy for pregnant women with VA, 1993; revised 1995 and 2000.
insulin dependent diabetes mellitus. Obstet Gynecol 1995; 85:
417–422.
24 Use of oral hypoglycemic agents
in pregnancy
Oded Langer
from 1% to 19%. Among different ethnic groups, type 2

Introduction diabetes and GDM have been diagnosed in varying rates.6,7
Gestational diabetes mellitus (GDM) continues to be a Both are heterogeneous disorders whose pathophysiology
major public health problem for the mother and unborn is characterized by peripheral insulin resistance, impaired
fetus with an estimated incidence of 3%–10% depending regulation of hepatic glucose production, and declining
upon geographic location; it affects at least 105,000–350,000 beta-cell function. It has become the prime objective for
women annually in the United States. The cornerstone of the treatment of both pregnant and nonpregnant diabetic
treatment is a diabetic diet, and when dietary modifica- patients to optimize the glycemic profile.8–10
tions do not control maternal glycemia, pharmacological The controversy and historic ban on the use of oral hypo-
therapy is initiated. The administration of short- and long- glycemic agents in pregnancy has been based on scant evi-
acting insulin is required in 20%–60% of pregnancies that dence from case reports11,12 and one study in particular on
are complicated by GDM in order to maintain adequate fetal anomalies in 50 patients with poorly controlled diabetes
glycemic control. Based on its high efficacy, 50%–80% of prior to pregnancy13 begs the question: is it the drug or is it
patients will achieve the targeted level of glycemic control.1 the glucose? In the past decade, the use of oral hypoglycemic
Insulin is inconvenient and expensive, which makes it an agents in pregnancy have become widely used in the United
even greater concern in developing countries. Therefore, States following the recommendation of the American College
there has been an unremitting search for a safe and effec- of Obstetrics and Gynecologists (ACOG),14 evidence-based
tive alternative to insulin therapy especially in GDM and analysis of the data,15 and editorials and clinical opinions.16–20
type 2 diabetes. Despite the fact that oral hypoglycemic agents (first-generation
The prevalence of type 2 diabetes and GDM has increased sulfonylureas) have been historically prescribed, both in the
by 33% in the past decade in the United States.2 This may be United Kingdom and South Africa, presently the use of gly-
attributed to the increased rate of obesity in the general pop- buride (glipizide and metformin) is less widespread in Europe
ulation in all ethnic groups and the trend toward advanced even though there is no reported adverse side effects to the
maternal age in pregnancy. Because of the relative ease of fetus.21–26
administration and the low cost involved in overall therapy The purpose of this review is to provide the evidence and
with oral hypoglycemic agents, they have become the drug the foundation for understanding the use of oral hypoglyce-
of choice in the treatment of type 2 diabetes. One can assume mic drugs in pregnancy as an effective alternative to insulin
that their popularity will only increase in the future, espe- therapy in achieving glycemic control. The concerns of tera-
cially after confirmation from the large prospective study by togenicity due to possible placental transfer, neonatal and
the United Kingdom Prospective Diabetes Study (UKPDS) maternal outcome, and basic pharmacological benefits will
group. The results demonstrated that patients with type 2 be addressed.
diabetes can maintain the targeted level of glycemic control,
thereby lowering the risk for microvascular complications.3
The use of oral hypoglycemic agents in nonpregnant patients Oral hypoglycemic agents:
with type 2 diabetes has become the standard of care in the Classification and characteristics
United States.3,4 The randomized study of the use of oral
hypoglycemic agents demonstrated that glyburide is an effi- The oral hypoglycemic agents act, depending upon the spe-
cacious alternative to insulin in the treatment of diabetes in cific group, directly upon the beta cells to increase insulin
the pregnant subject.5 secretion and/or to decrease hepatic glucose production and
The prevalence of GDM varies in direct proportion to increase peripheral insulin sensitivity. The advantage of using
the prevalence of type 2 diabetes in a given population or these agents rather than administering exogenous insulin is
ethnic group. In the United States, the prevalence rate ranges their ability to have an impact by nutrient availability and
200
Oral hypoglycemic agents: Classification and characteristics 201
extrapancreatic effect and/or to increase insulin availability hyperglycemia. Enhanced insulin secretion diminishes glu-
through the physiological route. cose toxicity and improves insulin secretion after meals,
The reader should consider the following “drug compass” thus reducing postprandial hyperglycemia. The drugs can
when contemplating the use of oral hypoglycemic agents in also enhance peripheral tissue sensitivity to insulin.30,31
pregnancy: Sulfonylureas influence insulin secretion in direct propor-
tion to plasma glucose levels from 60 to 180 mg/dL: they do
1. Will the drug–drug interactions complicate its use with not stimulate insulin secretion when the plasma glucose is
the necessary and commonly administered drugs? <60 mg/dL.35,36 The mechanism of action of sulfonylureas is
2 . Can glycemic control be achieved by using the optimal to facilitate rapid insulin secretion in response to nutritional
dose? intake, which results in minimal to no lag time between the
3. After nutrient ingestion, can the drug reduce the time lag changes in plasma glucose and modification of the insulin
between the plasma glucose rise and insulin secretion? secretory rate.37,38
4 . Can serious postprandial and fasting hypoglycemia be Chlorpropamide has been available for more than
minimized because the drug duration of action is short 40 years and is a highly effective oral hypoglycemic agent
enough or its dependence on plasma glucose levels with a very long duration of action. The main side effect for
sufficient? nonpregnant patients with type 2 diabetes is a significantly
5. Are there any side effects that can reduce long-term high rate of severe and protracted hypoglycemia. The com-
beneficial effects? plication has not been reported to be a major concern for
pregnant patients in previous studies.21–25 The drug stimu-
A major consideration in the efficacy of the drug will be its lates antidiuretic hormone secretion, potentiating its effect
ability to cross the placenta and, if so, what toxicity, if at all, at the renal tubular level resulting in water retention and
can it cause to the developing fetus. Often, the fear of drug- hyponatremia. With the development of second-generation
induced adverse outcome, especially after the thalidomide sulfonylurea drugs that minimally cross the placenta, i.e.,
era of the 1960s, inhibits the physician’s ability to judge the glyburide, first-generation drugs should not be recom-
scientific rationale for using a drug and evaluating it using mended in pregnancy.27
evidence-based data instead of dogma. Very few medications Glyburide (also known as glibenclamide and glybenzcy-
have been shown not to cross the placental barrier. In fact, clamide) is one of the second-generation hypoglycemic sul-
a pregnant woman is often exposed to four or five prescrip- fonylureas; the group also includes glipizide, gliclazide, and
tion drugs during pregnancy for a variety of complaints. glimepiride. These are considerably more potent than the
Similar to other epithelial barriers, transfer of drugs across earlier agents. When given as a single agent, the peak plasma
the placenta is affected by several factors: molecular weight, level of glyburide occurs within 4 hours; the absorption of
pKa, lipophilicity, placental blood flow, blood protein bind- the drug is not affected by food digestion. Metabolism of gly-
ing, elimination half-life, and the specific placental trans- buride occurs in the liver and its metabolites are extracted
port system that affects the ability of drugs to enter the fetal in the bile and urine in equal proportions. Ten hours is the
compartment.27–29 approximate elimination half-life of glyburide. Adverse
effects of the drug are infrequent; they occur in less than
4% of patients receiving second-generation agents.27 In
Sulfonylureas 11%–38% of nonpregnant patients with type 2 diabetes, the
Sulfonylureas were discovered accidentally. Additional clin- main side effect of glyburide is hypoglycemia, with symp-
ical trials led to the discovery of tolbutamide in the 1950s, toms being dose related: the older patient is at greater risk of
and since that time, many drugs in this class have been a hypoglycemic episode.
developed, e.g., chlorpropamide. Second-generation sulfo- The patient most receptive to glyburide therapy is one
nylureas were subsequently developed that today include who has been hyperglycemic for less than 5 years, is willing
glyburide and glipizide. In 1997, the first drug in a new class to follow a dietary protocol, and is of either normal weight
of oral insulin secretagogues called meglitinides (benzoic or obese. Characteristic features of both type 2 diabetes
acid derivatives) was approved for clinical use. The agent, and GDM are beta-cell exhaustion and insulin resistance.
repaglinide, has gained acceptance as a fast-acting, premeal Most often, patients of both diabetic types are comparable
therapy to limit postprandial hyperglycemia.27 Sulfonylureas in obesity, are asymptomatic in the early stages of the dis-
have been used in the treatment of type 2 diabetes since 1942 ease, and have similar prevalence in the same ethnic group.
because of their capacity to cause hypoglycemia by stimulat- Given the similarity of the phenotypic features of these
ing insulin release from pancreatic beta cells. Sulfonylureas complications, it is safe to assume that the use of glyburide
bind to specific receptors on beta cells that force the closure may be beneficial in the prevention of maternal–fetal GDM
of potassium adenosine triphosphate channels and open complications.
calcium channels that cause an increase in cytoplasmic cal- Glimepiride is another sulfonylurea drug. Both this
cium stimulating insulin release. The major effect of these drug and glyburide displace one another from their respec-
drugs is to enhance insulin secretion.30–34 Sulfonylureas tive binding sites. Glimepiride has a 2.5–3-fold faster rate
may also further increase insulin levels by reducing hepatic of association and an 8–9-fold faster rate of dissociation
clearance of the hormone, the main contributor to fasting from the beta-cell sulfonylurea receptor (SUR) binding site
202 Use of oral hypoglycemic agents in pregnancy
than glyburide. This results in a more rapid release of insu- triglyceride and cholesterol levels as well as in promoting
lin and a shorter duration of insulin secretion. Glimepiride weight loss in obese diabetic patients. Hypoglycemia is not
significantly increases second-phase insulin secretion, an overt side effect of its use. Metformin does not stimulate
whole-body glucose uptake, and insulin sensitivity.4,39,40 The the fetal pancreas to oversecrete insulin.44 However, met-
increase in insulin sensitivity may be explained by studies formin in pregnancy crosses the placenta. Its concentration
demonstrating lower fasting plasma insulin and C-peptide in the fetal compartment results in umbilical vein levels
levels in patients using the drug compared to glyburide- t wofolds higher than the maternal serum. Its effects on the
treated patients with comparable levels of glycemic control. fetus are unknown.
It should be noted that, to date, glimepiride has not been
tested for use in pregnancy.4,39,40
Thiazolidinediones
Biguanides Thiazolidinediones were introduced in 1997. The first agent,
troglitazone, was reported to have a high rate of hepatic tox-
Biguanides were recognized as early as 1920 but received
icity and as a result was withdrawn from the market. Newer
clinical recognition in the United States only in the recent
agents in this class such as rosiglitazone and pioglitazone
past. Phenformin, the primary drug in this group, was with-
are widely used in clinical practice without reported hepatic
drawn from American and European markets because of the
toxicity. Thiazolidinediones are a class of drugs that may
side effects of lactic acidosis. Its replacement, metformin,
provide still another pharmacological alternative to insulin
although used extensively in Europe, has only been recog-
therapy, although to date there are no reported data on its
nized for use in the United States since 1995.27 Metformin is
use in pregnancy. Troglitazone, the first of these agents to
an oral antihyperglycemic agent that is chemically and phar-
be introduced, has been withdrawn from use because it was
macologically unrelated to the sulfonylureas. Metformin is
associated with severe hepatic toxicity, followed by a num-
a second-generation biguanide that was reintroduced and
ber of deaths. These oral hypoglycemic agents exert their
distributed in the United States after biguanide phenfor-
principal effects by lowering insulin resistance in peripheral
min was withdrawn from the market in the 1970s: both
tissue. A decrease in systemic and local tissue lipid availabil-
were introduced in 1957. Metformin has universally been
ity may also contribute to its positive effects in controlling
shown to be effective in improving the glycemic profile in
diabetes.
diabetic patients. Its mechanism of action is thought to
Rosiglitazone and pioglitazone, other oral agents in
include decreased hepatic glucose production and intestinal
this group, are more potent than troglitazone and claim to
absorption of glucose and increased peripheral uptake of
offer lower risk of hepatotoxicity. They are absorbed within
glucose and utilization. The two latter mechanisms result in
2 hours but the maximum clinical effect is not observed
improved insulin sensitivity, i.e., decreased insulin require-
for 6–12 weeks. Liver function should be measured before
ments.41–43 Importantly, metformin does not stimulate insu-
the start of therapy and monitored once initiated. Studies
lin secretion and, therefore, does not cause hypoglycemia
also report considerable weight gain with the drugs.45,46
either in diabetic or in control patients. The drug acts by
Similarities exist between rosiglitazone and glyburide in
causing the translocation of glucose transporters from the
pharmacological characteristics that may suggest that there
microsomal fraction to the plasma membrane of hepatic and
is a possibility that they minimally cross the placenta. If this
muscle cells.43
proves to be the case, rosiglitazone and pioglitazone, just as
Metformin has no significant effect on the secretion of
glyburide and metformin, may be ideal agents for the man-
glucagons, cortisol, growth hormone, or somatostatin. The
agement of GDM and type 2 diabetes in pregnancy, as a sin-
mechanism by which metformin reduces hepatic glucose
gle therapy or in combination with glyburide.
production is controversial, but the preponderance of data
indicates an effect on the reduction of gluconeogenesis.43 It Alpha-glucosidase inhibitors: Alpha-glucosidase inhibitors,
has a strong safety and efficacy record, with a frequency of which reduce intestinal absorption of starch and glu-
lactic acidosis one-tenth that of the parent drug. The inci- cose (acarbose), have now been introduced into clinical
dence of lactic acidosis with metformin is 0.03 cases/1000 practice.27Alpha-glucosidase inhibitors act by slowing the
patients annually. The elimination of plasma half-life time absorption of carbohydrates from the intestine, thereby
is 6 hours. Therefore, patients with renal compromise should reducing the postprandial rise in blood glucose. The
not receive metformin, since the risk of lactic acidosis postprandial rise is blunted in both normal and diabetic
increases with the degree of renal impairment and patient patients. Gastrointestinal side effects require gradual dos-
age. Metformin should be introduced gradually in 500 or age increments over time after initiation of therapy. This
850 mg increments to a maximum of 2000 mg daily.41,42 group of drugs may be considered a monotherapy in elderly
The peak plasma level when the drug is given as a sin- patients but are typically used in combination with other
gle agent occurs within 4 hours. The extent of absorption is oral antidiabetic agents and/or insulin. Acarbose is the oral
reduced with food intake, although it should be adminis- agent in this group currently in use; however, in pregnancy,
tered with meals to minimize gastrointestinal intolerance. it is less effective than glyburide in achieving targeted levels
Metformin is not metabolized and is eliminated unchanged of glycemia. Therefore, its optimal use is as a secondary and
in the urine. It has been effective in reducing plasma not a primary drug of choice.47,48
Congenital malformations: Is it the glucose or is it the oral hypoglycemic agent? 203
Rationale for the use of oral improved glycemic control and decrease in the complication
rates. Thus, intensified therapy can, alone, provide the pri-
hypoglycemic agents in the mary prevention for diabetic complications. Studies of preg-
management of gestational diabetes nant diabetic women, including a study of >2000 women
with GDM,1 demonstrated that intensified therapy results in
mellitus improvement in glycemic control and in perinatal outcomes
similar to those in the nondiabetic population.
The intensified insulin approach in the management of GDM
has been shown to result in perinatal outcomes comparable
to those in the general population; thus, it has become the
method of choice for the control of glycemia.1 A less invasive, Congenital malformations: Is it the
efficacious alternative that achieves similar perinatal out-
come while enhancing patient compliance has been a major glucose or is it the oral hypoglycemic
diabetes research goal. The underlying principle for the use agent?
of oral hypoglycemic agents in pregnancy has been moti-
vated by three factors, first of which is the similarity between The incidence of congenital anomalies in nondiabetic
type 2 diabetes and GDM. In addition to the insulin secre- women is 2%–3%: the rate increases to 7%–9% overall in
tion and resistance abnormalities found in both conditions, pregnant diabetic patients and is even higher in poorly con-
there is a loss of the first-phase insulin secretion with a strik- trolled diabetics and as the severity of the disease increases.
ing lag time between the postprandial rise in glucose and the An unanswered question remains: what is the toxic agent
presence of significant insulin at the peripheral sites.36,37 The that triggers the development of malformations—is it the
result is in an early increase in postprandial glucose values. glucose or is it the oral hypoglycemic agent? This dilemma
As discussed before, second-generation sulfonylurea agents has led to several investigations of animal species or tissue
are rapid in onset and have a short duration of action that cultures as a source for answers. These types of studies pro-
makes them ideal agents for treatment in GDM and possibly vide the conditions with which to test separately and together
in the very early stages of type 2 diabetes in pregnancy. the effect of different drug doses in conjunction with varying
Second, GDM and patients with impaired glucose tol- levels of glucose. However, we are well aware that the findings
erance (IGT) are characterized by mild hyperglycemia in of studies using laboratory mice do not automatically translate
comparison to type 2 diabetics. However, this mild hypergly- into use on humans.
cemia is significantly elevated in comparison to nondiabetic Smithberg and Runner52 studied different hypoglycemia-
women. As the disease progresses to type 2 diabetes, there inducing treatments, including insulin, tolbutamide, and fast-
is progressive loss of beta-cell function.49,50 In the presence ing of prepuberal mice, as well as combination treatments
of insulin resistance with obesity, pregnancy, and, especially, involving nicotinamide plus insulin or tolbutamide. They
GDM, insulin secretion will initially increase to compensate were all found to be potent teratogens in one or more inbred
for the impairment in insulin action. The ensuing decrease strains of mice. Teratogenic treatments, with the exception
in secretion over time (50%–80%) will, in turn, result in the of fasting, also cause a variable proportion of deaths. The
progression from normal glucose tolerance to GDM, from response of different strains of mice to individual treatments
there to IGT and to type 2 diabetes.50 Oral hypoglycemic relevant to teratogenicity or lethality was highly variable. It is
agents have been successfully used to decrease glycemic lev- the variability of response elicited from each strain of mouse
els in type 2 diabetic patients. Since GDM subjects have the as a group that may be the most pertinent finding in these
mildest form of the glucose tolerance abnormality, it is rea- experiments. Most noteworthy is the 3% mortality produced
sonable to assume that the use of oral hypoglycemic agents by insulin treatment in strain BALB/c as compared to 17%
in the treatment of GDM should be even more effective than in other strains. This example demonstrates the variability
its current use with type 2 diabetic patients. in study results reported in the literature. It also makes one
Third, the UKPDS on type 2 diabetes supported the effi- realize that it was the strain of mouse that was the deter-
cacy of the drugs and in particular the use of glyburide.3 mining factor in recommending or failing to recommend a
The study demonstrated that with the use of glyburide, 70% particular drug.
of patients achieved a targeted level of glucose control with First-generation sulfonylureas, such as tolbutamide and
the most favorable effect achieved within the first 5 years of chlorpropamide, were found to be associated with con-
therapy.3 The study also reported a decrease in microvascu- genital malformations in the majority of animal studies.
lar and macrovascular complications. Rather than crediting Adverse effects appear to be caused by the drugs and not by
a specific therapy as the factor responsible for the reduced the hypoglycemia they produce. Chlorpropamide appeared
risk of complications, the authors concluded that improve- to be embryotoxic in mouse embryos in culture.53–55 There
ment in glycemic control was the crucial factor in treating is paucity of studies that have been performed to evaluate
the disease. second- and third-generation sulfonylureas and their asso-
The results of the UKPDS3,49,50 and the Diabetes Control ciations with malformations.
and Complications Trial51 suggested that intensive ther- Denno and Sadler56 evaluated the effect of biguanides,
apy in patients with types 1 and 2 diabetes will result in using metformin and phenformin as embryotoxic agents at
concentrations equal to serum levels obtained in patients oral hypoglycemic or antihyperglycemic agents. Transport
treated with the agent clinically. They found that phenfor- and metabolism of glyburide across the human placenta
min is embryotoxic, whereas metformin is not, suggesting was investigated in which the following was found: (1)
that metformin is also the safer drug during pregnancy in Virtually there is no significant transport of glyburide in
patients with non-insulin-dependent diabetes mellitus. either the maternal-to-fetal or fetal-to-maternal directions,
However, it should be noted that in the present study, met- with an average transport of 0.26% at 2 hours. These lev-
formin was not without adverse effects since it produced a els are threefold to eightfold higher than the therapeutic
delay in neural tube closure and also reduced yolk sac pro- peak levels after a 5 mg oral dose in humans. In fact, when
tein values at two different concentrations. While delayed cord blood samples were tested using high-performance
closure of the neural tube may not have resulted in gross liquid chromatography (HPLC), glyburide was undetect-
morphological abnormalities, it was not possible to assess able in these samples despite maternal plasma levels of
subtler alterations that might result from such a delay using 50–150 ng/mL. (2) Increasing the glyburide concentration
the culture system. to 100 times therapeutic levels did not alter transport appre-
Shepard57 and Schardein58 reported that metformin did ciably. Equilibrium dialysis demonstrated that at least 98%
not appear to be a major teratogen because <0.5% of the rat of the glyburide was protein bound. (3) Glyburide is neither
fetuses in mothers administered 500–1000 mg/kg devel- metabolized nor sequestered by the placenta. (4) The results
oped anophthalmia and anencephaly. However, evidence of of the study suggest passive drug transfer.
embryo toxicity was evident with higher doses of the drug. Another set of studies conducted in 199470 and 199771
Since animal studies are not conclusive about the safety of demonstrated that second-generation oral hypoglycemic

the fetus regarding the association between drugs and mal- agents, especially glyburide, do not significantly cross the
formations, additional research approaches are needed to diabetic or nondiabetic placenta. Fetal concentrations reached
determine drug transfer across the placenta and/or tests of no more than 1%–2% of maternal concentrations. Although
fetal blood for evidence of the drug. glipizide crossed the placenta in small amounts, this was sig-
Unlike other species, the human placental barrier is com- nificantly higher than glyburide. In contrast, tolbutamide
posed of a single rate-limiting layer of multinucleated cells, diffused across the placenta most freely. Glyburide has not
the syncytiotrophoblasts. During the formation of the pla- been demonstrated to be teratogenic in animal studies and is
centa, fetal tissues erode the maternal blood vessels to attain thus classified as a category B or C depending on the source.
a closer proximity to the maternal circulation. Chorionic The recommendation not to use oral hypoglycemic or
villi that contain fetal blood vessels infiltrate the maternal antihyperglycemic agents because of an increased rate of
vessels and establish a sinusoid in which the villi are suffused anomalies was based on a retrospective study involving 20
by maternal blood.59–62 Animal studies addressing placen- patients with type 2 diabetes, all with hyperglycemia prior to
tal transfer (e.g., mice) will not necessarily be applicable in conception (glycosylated hemoglobin [HbA1C] levels >8%).13
humans.63 The existence of maternal hyperglycemia preconception
The characteristics of individual drugs will determine makes it impossible to determine if the increased rate of
their placental transfer capability. These factors will include anomalies in the study subjects was a result of the medication
molecular weight, pKa, lipophilicity, placental blood flow, or of the elevated glucose levels.
blood protein binding, and elimination half-life.62 Although In contrast, three studies have suggested that there is no
the cutoff for actual molecular weight passage across the association between oral hypoglycemic agents and congeni-
placental barrier has not yet been accurately defined, it is tal malformations. Towner et al.72 treated 332 patients with
generally agreed that molecular weights ≤1000 Da passively type 2 diabetes with oral hypoglycemic agents or insulin
permeate across the placental barrier with sustained mater- prior to pregnancy. They demonstrated, using a stepwise
nal blood concentrations.64,65 Glyburide has a relatively short logistic regression, that the mode of therapy did not have
elimination half-life time (6–8 hours) and an extremely high an adverse significant effect, while the level of glycemia and
protein binding (99.8%) that results in minimal placental maternal age were significant factors contributing to the rate
transfer. Furthermore, there is effluxing of glyburide by spe- of anomalies. Langer et al. demonstrated similar findings in
cific placental transporters (e.g., P-gP, MRP1, MRP3, MRP3, a retrospective analysis of 850 women with type 2 diabetes
and BCRP) even against concentration gradients from the exposed to different oral hypoglycemic agents, insulin, and
fetal to the maternal side.66 diet therapy prior to and during the first trimester of preg-
The recirculating single-cotyledon human placental nancy.73 Again, it was the blood glucose and not the mode
model is widely used to characterize the transport and of therapy that had the net effect on the rate of anomalies.
metabolism of numerous drugs and nutrients. It is a reli- Koren,74 at a National Institutes of Health and Food and
able in vitro model for human placental transfer since it Drug Administration conference, presented the results of
facilitates the study of intact human placenta independent eight studies and concluded that the use of oral hypoglyce-
of fetal metabolism. Experiments can be validated against mic agents have no effect on the rate of fetal anomalies due
known substances that freely cross the placenta.67–69 Using to a very narrow confidence interval (CI) (odds ratio [OR]
this model, we evaluated whether first- and second-gener- 1.0; 95% CI 1.05–1.85). Recently, two meta-analysis stud-
ation sulfonylureas cross the placenta.28,70,71 Glyburide’s ies reconfirmed that there is no association between gly-
molecular weight is 494 units (U); it is one of the largest buride or metformin therapies and fetal malformations.
Is there an association between oral agents and maternal/neonatal hypoglycemia? 205
In a meta-analysis (comprised of 10 studies), no relationship Are oral agents and insulin comparable in achieving
was found between first-trimester exposure and anoma- glycemic control?
lies, OR = 1.05 (0.65–1.70).75 In another meta-analysis that Since GDM is characterized by a milder abnormal glycemic
included 8 studies, metformin treatment was associated profile and occurs 2–10 years earlier than type 2 diabetes, the
with 57% protective effect (control 7.2% vs. 1.7% metformin use of oral hypoglycemic agents to treat GDM should prove
treated). There is no evidence for increased risk for major to be even more efficacious. In addition, it is reasonable to
malformations.76 expect that the success rate of therapy for patients with GDM
It remains unresolved if the treatment of type 2 diabe- should be ≥70%, as achieved in type 2 diabetes. Several ret-
tes with oral hypoglycemic agents will accelerate the rate of rospective studies and numerous randomized studies4,81–91
anomalies. On the other hand, is it an overreaction to con- evaluated the use of glyburide or metformin in comparison
demn these medications? With existing data, care provid- to insulin. All found no difference in success rate between
ers need to objectively present information to patients so the drug therapies. Notelowitz91 studied the utility of tol-
that issues are addressed and informed decisions are made. butamide, chlorpropamide, diet therapy, and insulin in a
Moreover, there should be diligence in separating data from randomized study with a small sample size with relatively
studies of type 2 diabetes when considering the treatment for low power (each of the 4 arms contained 50 patients). There
GDM. In the case of GDM, the issue of anomalies is simpler. was no significant difference in perinatal mortality or con-
Patients with GDM are diagnosed and enter therapy after the genital anomalies. Good glycemic control was defined
first trimester (after the organogenesis period). There then as a blood glucose level <150 mg/dL. The percentages of
remains concern about potential fetal hypoglycemia and the patients who achieved the targeted glycemic category
stimulation for macrosomia if the drug crosses the placenta. (i.e., <150 mg/dL) are 80% of those using oral hypoglycemic
However, as previously discussed, glyburide does not cross agents or diet therapy and 36% of those treated with insu-
the placenta in appreciable amounts, and metformin that lin. Eighty percent of subjects treated with oral agents or in
does cross the placenta is not clinically associated with fetal diet therapy alone maintained targeted blood glucose levels
malformation. Finally, we demonstrated4 that patients enter- <150 mg/dL. In contrast, only 38% of the patients with insu-
ing therapy after the first trimester have rate of anomalies lin treatment were able to achieve this level, probably due to
comparable to patients treated with insulin and glyburide and poor compliance.
similar to the rate reported in the nondiabetic general popu- In Langer’s study,4 glyburide was as effective as insulin,
lation. Others reported that in women with type 2 diabetes with 82% of the patients treated with glyburide and 88% of
that were treated with oral agents prior and during organo- the patients treated with insulin achieving targeted levels of
genesis, no increased rate of fetal anomalies was found. control. Four hundred and forty women in between 11 and
33 weeks gestation, with a singleton pregnancy, having GDM
requiring treatment (a failed oral glucose tolerance test and
Is there an association between a fasting plasma glucose level 95–140 mg/dL) were enrolled.
The blood glucose profile characteristics prior to initiation
oral agents and maternal/neonatal of therapy were comparable for the glyburide- and the insu-
hypoglycemia? lin-treated groups (114 ± 9 versus 116 ± 22 mg/dL, respec-
tively). Patients were randomly assigned to receive either
An example of a study used to generate the recommenda- glyburide (n = 201; initial dose 2.5 mg orally, increasing by
tion that there is an increased risk for neonatal hypoglyce- 5 mg/week up to a total of 20 mg) or insulin (n = 203; initial
mia with the use of these drugs was a case report of three dose 0.7 U/kg subcutaneously three times daily, increasing
infants whose mothers received chlorpropamide and another each week as necessary) for glycemic control. Patients were
mother of an infant given acetohexamide; another case report required to measure glucose values seven times daily. The tar-
reported prolonged symptomatic neonatal hypoglycemia.11,12 gets for glycemic control were a mean blood glucose level of
Studies evaluating the existence of hypoglycemia using data 90–105 mg/dL, a fasting blood glucose level of 60–90 mg/dL,
from continuous blood glucose during a 72-hour monitoring a preprandial blood glucose level of 80–95 mg/dL, and a post-
period defined hypoglycemia as blood glucose <50 mg/dL prandial blood glucose level of <120 mg/dL.
for 30 minutes. In the insulin-treated patients, 4.2 ± 2.1/day Both treatments caused significant reductions in blood
episodes and, for glyburide-treated subjects, 2.1 ± 1.1/day glucose levels compared with levels measured at home for
episodes (p = 0.03) were reported. Asymptomatic hypogly- 1 week prior to initiation of treatment. Mean blood glu-
cemia was identified in 63% of insulin-treated women and cose levels in the glyburide group decreased from 114 to
28% for the glyburide-treated subjects. Overall, 84% of the 105 mg/dL, whereas those in the insulin group decreased
hypoglycemic episodes were nocturnal.77 Another study from 116 to 105 mg/dL. The percentages of the subjects who
using self-monitoring blood glucose revealed that the rate of were able to achieve targeted levels of glycemia are 82%
maternal hypoglycemia does not increase with the use of gly- of those treated with glyburide and 88% of those treated
buride. In fact, women in diet therapy and with insulin treat- with insulin. However, eight glyburide-treated women (4%)
ment had higher rates of hypoglycemic episodes compared to failed to achieve the desired level of control early in the third
women treated with glyburide.4,78 These findings were reiter- trimester and were transferred to insulin therapy. None of
ated in two meta-analysis studies.79,80 the patients developed severe symptoms of hypoglycemia.
However, in the insulin-treated group, a significantly more efficacious compared to metformin as a treatment
higher rate of subjects had 1%–6% of self-monitoring blood modality in GDM in both controlling the level of glycemia
glucose determination values <40 mg/dL compared to the and decreasing complications related to glycemia. In one
glyburide subjects. The glyburide and insulin groups had study, 26 the failure of metformin to achieve targeted levels
similar rates of preeclampsia (6%) and cesarean sections of glycemic control was 53.8% for established diabetics and
(23%–24%). Neonatal outcomes did not differ significantly 28.6% in patients with GDM. Apart from a high incidence
between the two groups. The glyburide and insulin groups of neonatal jaundice requiring phototherapy, the infant
had a similar incidence of large-for-gestational-age (LGA) morbidity in the metformin group was low. The rate of LGA
infants (12% vs. 13%), macrosomia (7% vs. 4%), lung com- infants was double the rate found in the authors’ general
plications (8% vs. 6%), hypoglycemia (9% vs. 6%), admis- population. However, the LGA rate was comparable in the
sion to a neonatal intensive care unit (6% vs. 7%), and fetal metformin- and insulin-treated patients, approaching 20%.
anomalies (2% vs. 2%). Finally, in this study, 26 the mothers of the three infants
In another study of the use of glyburide in pregnancy, with congenital malformations in the metformin group
the level of compliance in women using glyburide therapy initiated therapy in the third trimester. Another study by
to achieve glycemic control was reexamined. The success of the same authors23 compared patients treated with metfor-
glyburide therapy was defined as maintaining fasting plasma min and glibenclamide alone, and the combination of diet,
level <90 mg/dL and 1-hour postprandial level <130 mg/dL. metformin, and glibenclamide. Patients who failed the oral
Approximately, 83% of the women achieved these goals with hypoglycemic agent therapy were transferred to insulin
universal satisfaction with the mode of therapy.92 therapy. The incidence of LGA neonates (>90th percentile)
Success in achieving targeted levels of glycemia will was 15% (metformin), 27% (glibenclamide), 33% (com-
vary from study to study because of different doses, admin- bined therapy group), and 41% (failed oral insulin–treated
istration algorithms, length of therapy, type of patient group). The relative increase in the rate of LGA infants must
(severity and ethnicity), and noncomparable groups (com- be explained by the severity of the disease and the higher
pliant versus noncompliant subjects). To date, there is no rate of poorly controlled subjects in the group with the
evidence that a diabetic medication will be able to main- combination approach and the group treated with insulin.
tain targeted levels of glycemic control in all patients. Neonatal hypoglycemia is defined as <25 mg/dL: the overall
Several studies93–96 determined that 75%–84% success was rate of neonatal hypoglycemia was 11.5%, with the highest
achieved with the use of glyburide as a primary therapy. rates for the patients treated with glibenclamide, 27% and
When patients on glyburide failed to achieve targeted lev- combination therapy (glibenclamide and metformin), 18%,
els of glycemic control and transferred to insulin therapy, and the lowest rate in metformin-treated patients, 5%. The
failure rate was 54%–67%. Physician intervention can pre- high rate of neonatal hypoglycemia corresponds with a rate
vent avoidable failure by adequately adjusting drug dose of LGA infants reported in Langer’s study, suggesting that
or prescribing alternative treatment when the original a significant number of their patients were in suboptimal
therapy has failed; however, other factors are unavoidable glucose control.
and not subject to physician intervention, e.g., race, obe-
sity, and disease severity.
The success rate of metformin compared to insulin or Effect of obesity and disease severity on achievement
glyburide in achieving targeted levels of glycemic control of targeted level of control
is less well defined. Several randomized studies82–86,90,97–99 Although glyburide and probably metformin are compa-
evaluated the need for insulin supplements in metformin- rable to insulin in the success of achieving targeted levels
treated women. In these studies, the range of metformin of glycemic control, the question remains if this compari-
failure was 14%–46%. In the Rowen study,86 the range of son will hold true along the severity spectrum of GDM. We
insulin dose required was 22–81 units. Predictors of the evaluated this issue by stratifying women with GDM based
need for supplemental insulin include higher BMI and fast- on fasting plasma glucose. We demonstrated that there
ing glucose on the OGTT and earlier gestational age at diag- was no difference in the rate of success for insulin- and
nosis. A randomized small-scale study (46 patients in each glyburide-treated patients. This relationship was obtained
arm) found lower mean blood glucose in the metformin vs. in over 85% of patients with a glyburide dose ≤15 mg. For
the insulin group.90 Silva82,83 compared the success rates of both drugs, the success rate decreased with the increase
glyburide compared to metformin in other small random- in severity category of the disease. The incidence of LGA
ized studies. They concluded that treatment of GDM with infants increased in direct relation to the severity of the
metformin or glyburide was comparable for both women disease for both drugs.100 In regard to metformin, the rela-
and newborns. In contrast, Moore84,85 in a randomized tionship between severity level and success in achieving
study of metformin vs. glyburide found that the failure rate targeted levels of control is less clear. Some studies have
to achieve the glycemic target was 2.1-fold higher for the demonstrated comparable levels of success, while in oth-
metformin group (34.7% vs. 16.2%, 95% CI 1.2–3.9). Finally, ers, approximately 46% of subjects needed the addition of
a higher rate of fetal macrosomia and metabolic complica- insulin to the metformin therapy.86
tions was found in the metformin study86 compared to the Women with Type 2 diabetes, by definition, are in a
glyburide studies.4 The data suggest that glyburide may be higher disease severity level than patients with GDM
Conclusions 207
(fasting plasma glucose ≥126 mg/dL). The pregnancy exac- and insulin, metformin with a glucagon-like peptide 1
erbates the disease (insulin resistance) making it even receptor agonist and insulin, and metformin and insulin.
more difficult to achieve targeted levels of glycemic con- They found that all treatment regimens yielded comparable
trol. Therefore, these patients will benefit from immediate life years and QALYs, regardless of glycemic control tar-
assignment to insulin therapy rather than diet alone or oral get. However, the regimen in which sulfonylurea was the
agent treatment.101 However, when we compared success in second-line treatment accrued significantly lower cost per
achieving glucose targets and adverse perinatal outcome, QALYs and had the longest time to insulin dependence. For
we found no difference between patients treated with gly- all treatment approaches, an HbA1C goal of 7% resulted in
buride or insulin.102 higher QALYs vs. a goal of 8%. The study took into account
concerns regarding severe hypoglycemia in association
with sulfonylurea use. The study data suggest that a glyce-
Association between oral agents and breastfeeding mic control goal of 6.5% or 7% (~135–150 mg/dL) will be
Breastfeeding has, in the past decade, gained popularity sufficient to prevent hypoglycemia. However, this thresh-
in the United States where it was traditionally less popu- old is significantly higher than the threshold of <100 mg/
lar than in Europe. Women are encouraged and motivated dL needed to optimize outcome of pregnancy.
to breastfeed because of purported health benefits to the To date, a single study sought to evaluate the cost of
newborn and the psychosocial variable of mother–child insulin vs. glyburide therapy. The author included the cost
bonding. of insulin, syringes, needles, alcohol pads, and the teach-
From the pharmaceutical perspective, only when the ing required for patient insulin administration. For the
drug level is ≥10% is there a need for concern. Studies have glyburide patients, they calculated the cost of the medica-
shown that the infant metformin level was 0.28% of the tion and the cost of instructing failed glyburide patients in
weight normalized to maternal dose while glyburide showed insulin administration. The outcome of pregnancy was not
no trace in breast milk. Therefore, both drugs are safe to use included in the calculation since the outcome was com-
during breastfeeding.103–105 parable between the two modalities. They concluded that
glyburide is significantly less costly than insulin therapy
resulting in an average savings of at least $165.84/patient
Cost-effectiveness of insulin vs. glyburide (based on wholesale drug costs).107
In nonpregnant subjects with type 2 diabetes, the use of a
sulfonylurea drug as a second-line agent appears to yield
glycemic control and quality-adjusted life years (QALYs) Conclusions
similar to those provided by similar agents, but at a lower
cost, according to recent findings. In the study,106 the Different oral hypoglycemic agents act via diverse mecha-
researchers created a population-based glycemic control nisms of action. These drug characteristics provide a more
Markov model to mimic natural variations in HbA1C pro- physiological approach to the treatment of type 2 diabe-
gression. The study’s outcome measures were life years, tes and GDM. Furthermore, combination therapies will
QALYs, mean time to insulin dependence, and anticipated enhance the effect of these drugs on glucose metabolism
costs of medication per QALY from diagnosis, onset of dia- (Figure 24.1).
betes complications, or death. They evaluated the following In addition, a pill will generally be more acceptable than
four second-line intensification regimens: metformin, sul- an injection. In addition, the use of oral hypoglycemic agents
fonylurea, and insulin, metformin with a DPP-IV inhibitor is more convenient, less costly to the patient in supplies,
Insulin resistance
Diagnosis If type 2 in pregnancy
obesity
of GDM start insulin
other risk factor
Diet and exercise

fasting = 95 mg/dL
Oral agents
glyburide metformin?
Combination therapy
with oral agents or insulin
Insulin
Figure 24.1 Usual sequence of interventions.

and less expensive in office infrastructure and staff. Insulin adherence, etc.) will determine the overall success in manag-
therapy involves daily injections that do not always result in ing the disease and maximizing the quality of perinatal out-
optimal compliance by many women, and women in many come. Thus, as food for thought, we need to remember the
developing countries cannot afford insulin therapy. Hippocratic Oath (Orko V) that physicians swear to uphold:
Today, sufficient data exist generated by several random- Epi dhlhsei de kai adikihi eirxein—“first, do no harm or
ized and well-designed prospective and retrospective stud- cause injustice….” Withholding potential efficacious treat-
ies to support the use of oral hypoglycemic agents since they ment because of “conventional wisdom” does not uphold the
may have an even greater therapeutic effect in controlling Hippocratic Oath. We need to remember that often it is not
glycemic levels. However, regardless of the mode of therapy, a question of right or wrong—it is more often a question of
whole-patient care (glucose monitoring, education, diet what is more right.
REFERENCES
1. Langer O, Rodriguez DA, Xenakis EMJ et al. Intensified vs. con- 20. Gabbe S, Graves C. Management of diabetes mellitus complicat-
ventional management of gestational diabetes. Am J Obstet ing pregnancy. Obstet Gynecol 2003; 102(4): 857–868.
Gynecol 1994; 170: 1036–1047. 21. Sutherland HW, Bewsher PD, Cormack JD et al. Effect of moder-
2. Mokdad A, Ford E, Bowman B et al. Prevalence of obesity, diabe- ate dosage of chlorpropamide in pregnancy on fetal outcome.
tes, and obesity-related health risk factors. J Am Med Assoc 2003; Arch Dis Child 1974; 49: 283–291.
289(1): 76–79. 22. Coetzee EJ, Jackson WPU. Oral hypoglycemics in the first trimes-
3. American Diabetes Association. Implications of the United ter and fetal outcome. South Afr Med J 1984; 65: 635–637.
Kingdom prospective diabetes study. Diabetes Care 2000; 23. Coetzee EJ, Jackson WPU. Pregnancy in established non-insulin-
23(Suppl. 2): S27–S31. dependent diabetics. S Afr Med J 1980; 61: 795–802.
4. Lebovitz HE. Insulin secretagogues: Old and new. Diabet Rev 24. Notelovitz M. Oral hypoglycemic therapy in diabetic pregnan-
1999; 7: 139–153. cies. Lancet 1974; ii: 902–903.
5. Langer O, Conway DL, Berkus MD et al. A comparison of gly- 25. Coetzee EJ, Jackson WPU. The management of non-insulin-
buride vs. insulin in women with gestational diabetes mellitus. dependent diabetes during pregnancy. Diabet Res Clin Pract
N Engl J Med 2000; 343: 1134–1138. 1986; 1: 281–287.
6. Langer O, Umans J, Miodovnik M. Perspectives on the proposed 26. Coetzee EJ, Jackson WPU. Metformin in management of pregnant
gestational diabetes mellitus diagnostic criteria. Obstet Gynecol insulin-independent diabetics. Diabetologia 1979; 16: 241–245.
2013; 121: 1–7. 27. In: Hardmons J, Limbird, eds., Goodman and Gillman’s: The
7. Langer O, Umans J, Miodovnik M. The proposed GDM diagnos- Pharmacologic Basis of Therapeutics, 9th ed., McGraw Hill:
tic criteria: A difference to be a difference, must make a differ- New York, 1996, pp. 1712–1792.
ence. J Matern Fetal Neonatal Med 2013; 26(2): 111–115. 28. Elliot B, Langer O, Schenker S, Jonhson RF. Insignificant trans-
8. Freinkel N, Metzger BE, Phelps RL et al. Gestational diabetes fer of glyburide occurs across the human placenta. Am J Obstet
mellitus: Heterogeneity of maternal age, weight, insulin secre- Gynecol 1991; 165: 807–812.
tion, HLA antigens, and islet cell antibodies and the impact 29. Koren G. Glyburide and fetal safety: Transplacental pharmacoki-
of maternal metabolism on pancreatic B-cell function and netic considerations. Reprod Toxicol 2001; 15: 225–229.
somatic growth in the offspring. Diabetes 1995; 34(Suppl. 2): 30. Rossetti L, Giaaccari A, DeFronzo RA. Glucose toxicity. Diabetes
1–7. Care 1990; 13: 610–630.
9. Catalano PM, Virgo KM, Bernstein IM, Amine SB. Incidence 31. Simonson DC, Farrannini E, Bevilacqua S et al. Mechanism of
and risk factors associated with abnormal glucose tolerance in improvement in glucose metabolism after chronic glyburide ther-
women with gestational diabetes. Am J Obstet Gynecol 1991; apy. Diabetes Care 1984; 33: 838–845.
165: 914–919. 32. Groop L, Luzi L, Melanger A et al. Different effects of glyburide
10. Catalano PM, Tyzbir ED, Wolfe RR et al. Carbohydrate metabo- and glipizide on insulin secretion and hepatic glucose production
lism during pregnancy in control subjects and women with ges- in normal and NIDDM subjects. Diabetes 1987; 36: 1320–1328.
tational diabetes. Am J Physiol 1993; 264: E60–E67. 33. Groop LC, Barzilai N, Ratheiser K et al. Dose-dependent effects
11. Kemball ML, McIvert C, Milner RDG et al. Neonatal hypoglyce- of glyburide on insulin secretion and glucose uptake in humans.
mia in infants of diabetic mothers given sulphonylurea drugs in Diabetes Care 1991; 14: 724–727.
pregnancy. Arch Dis Child 1970; 45: 696–701. 34. DeFronzo RA, Simonson DC. Oral sulfonylurea agents suppress
12. Zucker P, Simon G. Prolonged symptomatic neonatal hypoglyce- hepatic glucose production in non-insulin-dependent diabetic
mia associated with maternal chlorpropamide therapy. Pediatrics individuals. Diabetes Care 1984; 7: 72–80.
1968; 42: 824–825. 35. Kahn SE, McCulloch DK, Porte Jr D. Insulin secretion in normal
13. Piacquadio K, Hollingsworth DR, Murphy H. Effects of in- and diabetic humans. In: Alberti KGMM, Zimmer P, DeFronzo
utero exposure to oral hypoglycemic drugs. Lancet 1991; 338: RA, Keen H, eds., International Textbook of Diabetes Mellitus,
866–869. 2nd ed., Wiley: Chichester, U.K., 1997, pp. 337–354.
14. American College Obstetricians and Gynecologists. Practice 36. Mitrakou A, Kelley D, Mokan M et al. Role of reduced suppres-
Bulletin: Gestational diabetes mellitus. Obstet Gynecol 2013; sion of glucose production and early insulin release in impaired
122(2, Pt. 1): 406–416. glucose tolerance. N Engl J Med 1992; 326: 22–29.
15. Nicholson W, Bolen S, Witkop C et al. Benefits and risks of oral 37. Leahy JL. Natural history of beta cell dysfunction in NIDDM.
diabetes agents compared with insulin in women with gesta- Diabetes Care 1990; 13: 992–1010.
tional diabetes: A systemic review. Obstet Gynecol 2009; 113(1): 38. Polansky KS, Given BD, Hirsch I et al. Abnormal patterns of insu-
193–205. lin secretion in non-insulin dependent diabetes mellitus. N Engl J
16. Greene M. Oral hypoglycemic drugs for gestational diabetes. Med 1988; 318: 1231–1239.
N Engl J Med 2000; 343: 1178–1179. 39. Clark HE, Matthews DR. The effect of glimepiride on pancreatic
17. Cefalo R. A comparison of glyburide and insulin in women with beta-cell function under hyperglycemic clamp and hyperinsu-
gestational diabetes mellitus. Obstet Gynecol Survey 2001; 56: linaemic, euglycaemic clamp conditions in non-insulin depen-
126–127. dent diabetes mellitus. Horm Metab Res 1996; 28: 445–450.
18. Saade G. Gestational diabetes mellitus: A pill or a shot? Obstet 40. van der Wal PS, Draeger KE, van Iperen AM et al. Beta cell response
Gynecol 2005; 105(3): 456–457. to oral glimepiride administration during and following a hypergly-
19. Koren G. The use of glyburide in gestational diabetes: An ideal caemic clamp in NIDDM patients. Diabet Med 1997; 14: 556–563.
example of “bench to bedside.” Pediatr Res 2001; 49: 734. 41. Product information. Glucophage. Bristol-Myers Squibb, 1997.
References 209
42. Klepser TB, Kelly MW. Metformin hydrochloride: An anti- 70. Elliot B, Schenker S, Langer O et al. Comparative placental trans-
hyperglycemic agent. Am J Health Syst Pharm 1997; 54: port of oral hypoglycemic agents. A model of human placental
893–903. drug transfer. Am J Obstet Gynecol 1994; 171: 653–660.
43. Stumvoll M, Nurjhan N, Perriello G et al. Metabolic effects of 71. Elliot B, Langer O, Schussling F. A model of human placental drug
metformin in non-insulin dependent diabetes mellitus. N Engl J transfer. Am J Obstet Gynecol 1997; 176: 527–530.
Med 1995; 333: 550–554. 72. Towner D, Kjos SL, Montoro MM et al. Congenital malformations
44. Vanky E, Zahlsen K, Spigset O et al. Placental passage of metfor- in pregnancies complicated by NIDDM. Diabetes Care 1995; 18:
min in women. Fertil Steril 2005; 83: 1575–1578. 1446–1451.
45. Buckingham RE, Al-Barazanji KA, Toseland N et al. Peroxisome 73. Langer O, Conway D, Berkus M, Xenakis EMJ. There is no asso-
proliferator-activated receptor-gamma agonist, rosiglitazone, ciation between hypoglycemic use and fetal anomalies. Am J
protects against nephropathy and pancreatic islet abnormalities Obstet Gynecol 1999; 180: S38.
in Zucker fatty rats. Diabetes 1998; 47: 1326–1334. 74. Koren G. Proceedings of the NIH/FDA Toxicology in Pregnancy
46. Lebovitz HE. Thiazolidinediones. In: Lebovitz HE, ed., Therapy Conference, Toronto, Ontario, Canada, 2000.
for Diabetes Mellitus and Related Disorders, 3rd ed., American 75. Gutzin S, Kozer E, Magee L et al. The safety of oral hypoglycemic
Diabetes Association: Alexandria, VA, 1998, pp. 181–185. agents: A meta-analysis. Can J Clin Pharmacol 2003; 10: 179–183.
47. Coniff RF, Seaton TB, Shjapiro JA et al. Reduction of glycosyl- 76. Gilbert C, Valois M, Koren G. Pregnancy outcome after first-
ated hemoglobin and postprandial hyperglycemia by acarbose in trimester exposure to metformin: A meta-analysis. Fertil Steril
patients with NIDDM. Diabetes Care 1995; 18: 817–820. 2006; 86(3): 658–663.
48. Lebovitz HE. Alpha-glucosidase inhibitors. Endocrinol Metab 77. Yogev Y, Ben-Haroush A, Chen R. Undiagnosed asymptomatic
Clin North Am 1997; 26: 539–551. hypoglycemia: Diet, insulin, and glyburide for gestational dia-
49. UK Prospective Diabetes Study Group V. Characteristics of newly betic pregnancy. Obstet Gynecol 2004; 104(1): 88–93.
presenting type 2 diabetic patients: Estimates of insulin sensitivity 78. Brustman L, Langer O, Scarpelli S et al. Hypoglycemia in
and islet beta-cell function. Diabet Med 1988; 5: 444–448. glyburide-treated diabetes: Is it dose dependent? Obstet Gynecol
50. UK Prospective Diabetes Study Group 16. Overview of 6 years’ 2011; 117: 349–353.
therapy of type 2 diabetes: A progressive disease. Diabetes 1995; 79. Moretti M, Rezvani M, Koren G. Safety of glyburide for gesta-
44: 1249–1258. tional diabetes: A meta-analysis of pregnancy outcomes. Ann
51. The Diabetes Control and Complications Trial Research Group. Pharmacother 2008; 42(4): 483–490.
The effect of intensive treatment of diabetes on the development 80. Dhulkotia J, Ola B, Fraser R et al. Oral hypoglycemic agents
and progression of long-term complications in insulin-dependent vs. insulin in management of gestational diabetes: A systematic
diabetes mellitus. N Engl J Med 1993; 329: 977–986. review and meta-analysis. Am J Obstet Gynecol 2010; 203(5): 457.
52. Smithberg M, Runner MN. Teratogenic effects of hypoglyce- 81. Silva J, Bertini A, Taborda W et al. Glibenclamide in the treatment
mic treatments in inbred strains of mice. Am J Anat 1963; 113: for gestational diabetes mellitus in a compared study to insulin.
479–489. Arq Bras Endocrinol Metabol 2007; 51(4): 541–546.
53. Sivan E, Feldman B, Dolitzki M et al. Glyburide crosses the pla- 82. Silva J, Fachin D, Coral M et al. Perinatal impact of the use of
centa in vivo in pregnant rats. Diabetologia 1995; 38: 753–756. metformin and glyburide for the treatment of gestational diabetes
54. Smoak IW. Teratogenic effects of chlorpropamide in mouse mellitus. J Perinat Med 2012; 40(3): 225–228.
embryos in vitro. Teratology 1992; 45: 474. 83. Silva J, Pacheco C, Bizato J et al. Metformin compared with gly-
55. Smoak IW, Sadler TW. Embryopathic effects of short-term expo- buride for the management of gestational diabetes. Int J Gynecol
sure to hypoglycemia in the mouse. Am J Obstet Gynecol 1990; Obstet 2010; 111(1): 37–40.
163: 619–624. 84. Moore L, Clokey D, Rappaport V et al. Metformin compared with
56. Denno KM, Sadler TW. Effects of the biguanide class of oral glyburide in gestational diabetes: A randomized controlled trial.
hypoglycemic agents on mouse embryogenesis. Teratology 1994; Obstet Gynecol 2010; 115(1): 55–59.
49: 260–266. 85. Moore L, Briery C, Clikey D et al. Metformin and insulin in the
57. Shepard TH. Catalog of Teratogenic Agents, 8th ed., John management of gestational diabetes mellitus: Preliminary results
Hopkins University Press: Baltimore, MD, 1995, p. 270. of a comparison. J Reprod Med 2007; 52(11): 1011–1015.
58. Schardein JL. Chemically Induced Birth Defects, 2nd ed., Marcel 86. Rowen J, Hague W, Wanzhen G et al. Metformin versus insu-
Dekker: New York, 1993, pp. 417–418. lin for the treatment of gestational diabetes. N Eng J Med 2008;
59. Audus KL. Controlling drug delivery across the placenta. Eur J 358(119): 2003–2015.
Pharmacol Sci 1999; 8: 161–165. 87. Ogunyemi D, Jesse M, Davidson M. Comparison of glyburide
60. Dancis J. Placental physiology. In: Kretchmer N, Quilligan EJ, versus insulin in management of gestational diabetes mellitus.
Johnson JD, eds., Prenatal and Perinatal Biology and Medicine, Endocrinol Pract 2007; 13(4): 427–428.
Physiology and Growth, vol. 1, Harwood Academic Publishers: 88. Anjalakshi C, Balaji V, Balaji M et al. A prospective study compar-
Chur, Switzerland, 1987, pp. 1–33. ing insulin and glibenclamide in gestational diabetes mellitus in
61. Enders AC, Blakenship TN. Comparative placental structure. Adv Asian Indian women. Diabet Res Clin Pract 2007; 76(3): 474–475.
Drug Deliv Rev 1999; 38: 3–16. 89. Bertini A, Silva J, Taborda W et al. Perinatal outcomes and the use
62. Ala-Kokko TL, Vahakangas K, Pelkonen O. Placental function of oral hypoglycemic agents. J Perinat Med 2005; 33(6): 519–523.
and principles of drug transfer. Acta Anaesth Scand 1993; 37: 90. Spaulonci C, Bernardes L, Trindade T et al. Randomized trial of
47–49. metformin vs. insulin in the management of gestational diabetes.
63. Sibley CP. Mechanism of ion transfer by the rat placenta: A model Am J Obstet Gynecol 2013; 34–36.
for the human placenta? Placenta 1994; 15: 675–691. 91. Notelowitz M. Sulfonylurea therapy in the treatment of the preg-
64. Willis DM, O’Grady JP, Faber JJ, Thornburg KL. Diffusion perme- nant diabetic. S Afr Med J 1971; 45: 226–229.
ability of cyanocobalamin in human placenta. Am J Physiol 1986; 92. Chmait R, Dinise R, Moore T. Prospective observational study to
250: R459–R464. establish predictors of glyburide success in women with gesta-
65. Malek A, Blann E, Mattison DR. Human placental transport of tional diabetes mellitus. J Perinatol 2004; 24: 617–622.
oxytocin. J Matern Fetal Med 1996; 5: 245–255. 93. Conway D, Gonzales O, Skiver D. Use of glyburide for the
66. Pollex E, Feig D, Koren G. Oral hypoglycemic therapy: treatment of gestational diabetes: The San Antonio experience.
Understanding the mechanisms of transplacental transfer. J Matern Fetal Neonatal Med 2004; 15(1): 51–55.
J Matern Fetal Neonatal Med 2010; 23(3): 224–228. 94. Yogev Y, Melamed N, Chen R et al. Glyburide in gestational
67. Brandes JM, Travoloni N, Potter JB et al. A new recycling tech- diabetes—Prediction of treatment failure. J Matern Fetal Neonatal
nique for human placental cotyledon perfusion: Application to Med 2011; (6): 842–846.
studies of the fetomaternal transfer of glucose, insulin, and anti- 95. Kahn B, Davies J, Lynch A et al. Predictors of glyburide failure
pyrine. Am J Obstet Gynecol 1983; 146: 800–806. in the treatment of gestational diabetes. Obstet Gynecol 2006;
68. Schenker S, Johnson R, Hays S et al. Effects of nicotine and nico- 107(6): 1303–1309.
tine/ethanol on human placental amino acid transfer. Alcohol 96. Rochon M, Rand L, Roth L et al. Glyburide for the management of
1989; 6: 289–296. gestational diabetes: Risk factors predictive of failure and associ-
69. Schenker S, Dicke J, Johnson R et al. Human placental transport ated pregnancy outcomes. Am J Obstet Gynecol 2006; 195(4):
of cimetidine. J Clin Invest 1987; 80: 1428–1434. 1090–1094.
97. Ijas H, Vaarasmaki M, Morin-Papunen L et al. Metformin should 102. Langer O, Monga S, Most O. Obese gestational diabetic women:
be considered in the treatment of gestational diabetes: A prospec- Comparison of insulin and glyburide therapies. Am J Obstet
tive randomized study. Br J Obstet Gynaecol 2011; 118: 880–885. Gynecol 2006; 195(6): 136S (Abstract).
98. Niromanesh S, Alavi A, Sharbaj F et al. Metformin compared with 103. Feig D, Briggs G, Koren G. Oral antidiabetic agents in pregnancy
insulin in the management of gestational diabetes mellitus: A ran- and lactation: A paradigm shift? Ann Pharmacother 2007; 41:
domized clinical trial. Diabet Res Clin Pract 2012; 98(3): 422–429. 1174–1180.
99. Tertti K, Ekblad W, Koskinen P et al. Metformin vs. insulin in ges- 104. Hale T, Kristensen J, Hackett L et al. Transfer of metformin into
tational diabetes: A randomized study characterizing metformin human milk. Diabetologia 2002; 45: 1509–1514.
patients needing additional insulin. Diabet Obes Metab 2013; 105. Gardiner S, Kirkpatrick C, Begg E et al. Transfer of metformin into
15(3): 246–251. human milk. Clin Pharmacol Ther 2003; 73: 71–77.
100. Langer O, Yogev Y, Xenakis E et al. Insulin and glyburide therapy: 106. Zhang Y, McCoy R, Mason J et al. Second-line agents for gly-
Dosage, severity level of gestational diabetes, and pregnancy cemic control for type 2 diabetes: Are newer agents better?
outcome. Am J Obstet Gynecol January 2005;192(1): 134–139. Diabetes Care 2014; 37: 1338–1345.
101. Langer O, Yogev Y, Xenakis E, Brustman L. Overweight and obese 107. Goetz L, Wilkins I. Glyburide compared to insulin for the treat-
in gestational diabetes: The impact on pregnancy outcome. Am J ment of gestational diabetes mellitus: A cost analysis. J Perinatol
Obstet Gynecol June 2005; 192(6): 1768–1776. 2002; 22(5): 403–406.
25 The drug dilemma of oral
antidiabetic agents in pregnancy:
Metformin
Yoel Toledano, Moshe Zloczower, and Nicky Lieberman
AMP-activated protein kinase (AMPK). By this enzyme

Introduction action, metformin probably lowers the blood glucose and
Metformin is the first-line oral treatment for type 2 diabe- lipid levels.8–10 Instead, metformin may inhibit gluconeogen-
tes mellitus (T2DM). A 2012 consensus statement from the esis by alternative pathways.
American Diabetes Association (ADA) and the European
Association for the Study of Diabetes (EASD) suggested
that both lifestyle intervention and metformin should be Efficacy
initiated as soon as diabetes is detected.1,2
Monotherapy
Metformin is the first-line medical treatment in most patients
Mechanism of action with T2DM. Glycemic response to metformin is comparable
to that achieved with sulfonylureas.3,4,6,7,11 Hence, glycated
As a general rule, the effect of metformin is evident only in hemoglobin (A1C) and fasting blood glucose concentrations
the presence of insulin. Moreover, it has several important decrease by approximately 1.5% and 20%, respectively.
effects: Compared with other oral agents, metformin has several
advantages:
1. Decreases hepatic glucose output.3,4 This is its major
effect. 1. Has a benefit of modest weight loss or is at least weight
2 . Increases insulin-mediated glucose utilization in periph- neutral in obese patients.4,6,11 Conversely, many other
eral tissues (such as muscle and liver), particularly after antidiabetic medications, e.g., sulfonylureas and insulin,
meals. are related to weight gain.6,11
3. Has an antilipolytic effect that lowers serum free fatty 2 . Has no hypoglycemia risk generally.
acid concentrations. Correspondingly, substrate avail- 3. Has lipid-lowering activity. Accordingly, serum triglycer-
ability for gluconeogenesis is diminished.3–5 Overall, ide and free fatty acid concentrations are lowered.3,4,7
serum insulin concentrations may be mildly reduced in 4. Is related to a decreased rate of myocardial infarction and
parallel with improved glycemic control.6,7 all-cause mortality12 in overweight and obese patients
4. Increases intestinal glucose utilization. This effect, with diabetes.
demonstrated in experimental animals, is mediated by
nonoxidative metabolism.4 This process generates lac-
tate that can be metabolized in the liver as a substrate for Combination therapy
gluconeogenesis.3 Consequently, hypoglycemia can be Optimal glycemic control is frequently achieved with
avoided, at least partly. combinations of drugs. Metformin can be given in combi-
nation with sulfonylureas, glinides, dipeptidyl peptidase-IV
Exactly how metformin acts at the molecular level is not fully (DPP-IV) inhibitors, alpha-glucosidase inhibitors, thiazoli-
known, but might depend on the activation of the enzyme dinediones, insulin, and exenatide.
211
212 The drug dilemma of oral antidiabetic agents in pregnancy
Side effects of reproductive age.22–24 Increased insulin resistance, i.e.

subnormal biological response to normal insulin concentra-
1. Gastrointestinal reactions are the most common. tions, and compensatory hyperinsulinemia play a key role
Generally, 5% of patients stop metformin due to gastro- in the pathogenesis of PCOS. At least 40% of women with
intestinal side effects. They include anorexia, nausea, PCOS are obese25 and are more insulin resistant than weight-
abdominal pain, diarrhea, and a metallic taste in the matched individuals with normal ovaries. On the other
mouth.4 These reactions are typically mild and transitory. hand, women with insulin resistance commonly present
After discontinuation or dose reduction of the medica- with hyperandrogenism and reproductive abnormalities.26
tion, the effects may reverse. Moreover, obesity, mainly central or abdominal, as indicated
2. Decreases intestinal absorption of vitamin B12 in one- by an increased waist to hip ratio, is correlated with reduced
third of patients. Hence, serum vitamin B12 concentra- fecundity, menstrual disorders, and hyperinsulinemia.22
tions may be reduced. Megaloblastic anemia only rarely
develops.13
3. Lactic acidosis: Incidence is very rare,14–16 up to 9 cases Diagnosis
per 100,000 person-years of exposure.14 The mortality The diagnosis of classical PCOS in women of reproductive
rate is very high. Symptoms are nonspecific. They include age is usually based on the Rotterdam criteria.27 These cri-
lethargy, hypotension, hyperventilation, anorexia, nau- teria were accepted by the American Endocrine Society.28
sea, vomiting, and abdominal pain. The basis for diagnosis is the presence of at least two of
the following three criteria: androgen excess (usually hir-
Predisposing factors are usually parallel conditions that sutism), ovulatory dysfunction (practically anovulation
are relative or absolute contraindications to the metformin and infertility), and polycystic ovaries (by ultrasound).
therapy. The most significant predisposing factor is renal Congenital adrenal hyperplasia should be excluded before
insufficiency. This category includes patients who are poten- diagnosing primary PCOS29 using basal or stimulated serum
tially at risk for developing contrast-induced renal failure, 17-hydroxyprogesterone.
i.e., after receiving intravenous iodinated contrast material.
Other risk factors include liver disease, alcohol abuse, past
history of lactic acidosis, and, more rarely, heart failure. Physiology
Simultaneously, this grave complication is generally pre- Hyperinsulinemia and hyperandrogenism represent a tan-
sented in patients with shock or tissue hypoxia.4,17,18 gled web that reaches its greatest complexity in women with
Taken together, patients should be recommended to PCOS. The basis for the association between insulin resis-
stop metformin once they develop a predisposing factor for tance and ovarian hyperandrogenism is not known. Theories
lactic acidosis. Hence, the risk of metformin-induced lac- have proposed that hyperinsulinemia causes hyperan-
tic acidosis can be significantly reduced. For example, it is drogenism, or vice versa, or that some unknown third defect
recommended to discontinue metformin immediately prior is responsible for both phenomena.30 In addition, there is no
to and for 48 hours after any exposure to parenteral iodin- clear correlation between the severities of hyperinsulinemia
ated contrast material administered as part of a radiologic and hyperandrogenism.31 While excess androgen can cause
procedure. a modest reduction of insulin sensitivity, the primary abnor-
mality is probably insulin resistance, which in some way
causes the ovarian abnormality. Insulin receptors, and the
Dosing closely related receptors for IGF-1, are present in ovarian
The excretion of metformin is rapid in the urine.4,19 It is avail- cells. Stimulation of these receptors increases androgen
able as 500, 850, or 1000 mg tablets. Metformin should be production.32 The administration of metformin or a thia-
administered with meals. The recommended starting dose zolidinedione, to reduce insulin resistance, lowers serum
is 500 mg once daily with the evening meal. The dose can free testosterone concentrations and reduces cytochrome
be titrated and increased slowly per the glycemic response.4 P450c17 (17-hydroxylase) activity in the ovaries.33
The usual effective dose is 1500–2000 mg/day. The glycemic However, there is currently no validated test for measur-
control is not significantly better with the maximum dose ing insulin resistance in a clinical setting. Serum triglycer-
of 2550 mg/day (850 mg three times daily).1 A long-acting ide concentration, the ratio of triglyceride to high-density
preparation, an extended release tablet, is also available.20 lipoprotein (HDL)-cholesterol concentrations, and fasting
insulin concentration are useful markers for identifying
those who may be insulin resistant.34 Optimal cut-points
PCOS and metformin were identified as 130 mg/dL (1.47 mmol/L), 3.0 (1.8 SI
units), and 109 pmol/L for triglycerides, triglyceride-to-
Background HDL ratio, and insulin, respectively.35 Sensitivity and speci-
Polycystic ovaries syndrome (PCOS) was first identified ficity for the cut-points were 67%, 64%, and 57%, and 71%,
and described by Stein and Leventhal as a combination of 68%, and 85%, respectively. In large population epidemiol-
hyperandrogenism, menstrual disturbances, anovulatory ogy studies, simple ratios derived from fasting insulin and
infertility, and obesity.21 PCOS affects up to 15% of women glucose (e.g., glucose to insulin ratios, homeostasis model
PCOS and metformin 213
assessment of insulin resistance [HOMA-IR or HOMA]) as a more convenient screening tool.46 The A1C test should
have been extensively used. There are limitations to their be performed using a method that is certified by the National
use, including changes in beta cell function over time, lack Glycohemoglobin Standardization Program (NGSP) and
of a standardized universal insulin assay, and lack of data standardized or traceable to the Diabetes Control and
demonstrating that markers of insulin resistance predict Complications Trial (DCCT) reference assay.46
response to treatment. As a result, although indexes such as
HOMA and QUICKI (quantitative insulin sensitivity check
index) have been proposed and cut-points identified,36 none Treatment
are recommended for routine assessment of insulin resis- Although it might be preferable to treat only insulin-resistant
tance in the clinic. women with PCOS, there is currently no readily available
Hyperandrogenemia can be stimulated by several mecha- and reliable test to evaluate the presence of insulin resistance.
nisms. First, abdominal obesity is associated with reduced Thus, therapeutic interventions may be directed to specific
levels of serum sex hormone–binding globulin (SHBG) and symptoms of PCOS as control of hyperandrogenism and
increased bioavailable serum androgens, which include free restoration of menstrual cyclicity and ovulation. Increasing
and albumin-bound androgens.37 Second, obesity is associ- insulin sensitivity and reducing insulin levels are consid-
ated with an increased testosterone production rate and a ered as a more holistic therapeutic action.22,30 The treatments
non-SHBG-bound androgen production rate of dehydro- are either favorable or with no impact regarding the glucose
epiandrosterone and androstenedione.38 Third, the com- homeostasis.
bination of insulin resistance and increased LH secretion
appears to stimulate ovarian androgen production.39 Finally, Therapeutic lifestyle change
elevated insulin levels inhibit hepatic synthesis of SHBG.40 Weight loss strategies with calorie-restricted diets and physi-
Furthermore, estrogen levels, particularly estrone, may cal activity are the key to clinical improvement in overweight
also be higher in PCOS,41 although not participating in the or obese PCOS patients. Indeed, they are likely beneficial
hyperandrogenic state. for reproductive and metabolic dysfunction. There is no evi-
The molecular defects that cause insulin resistance have dence that one type of diet is superior for weight loss. Weight
not been fully determined. In some patients, insulin resis- loss is probably insufficient as a treatment for normal-weight
tance appears related to mutations in the insulin receptor PCOS women.28
gene that result in abnormal function of the insulin recep-
tor.42 Accordingly, genetic defects leading to extreme levels Hormonal contraceptives
of insulin resistance in women (e.g., those resulting from Combined oral contraceptives are considered as the first-
mutations in the insulin receptor gene) are typically accom- line management for the menstrual abnormalities and hir-
panied by manifestations of PCOS. sutism and/or acne of PCOS.28
Dysregulation of local follicle regulatory systems by
androgens and other factors impedes normal follicular Metformin
growth, resulting in follicular arrest at the 4–8 mm diam- Treatment of T2DM is the only indication for metformin
eter size.37 A dominant follicle (i.e., 18–25 mm in diameter) administration approved by the FDA. According to the 2014
does not develop, and therefore, ovulation does not occur. ADA recommendations, metformin therapy for the preven-
Thus, the combination of elevated LH, hyperinsulinemia, tion of T2DM may also be considered. It may be adminis-
ovarian androgen overproduction, and disruption of follicle tered to patients with IGT, impaired fasting glucose (IFG),
growth produces the PCOS phenotype of oligoovulation and or an A1C of 5.7%–6.4%, especially for those with a BMI
hyperandrogenism.43 ≥35 kg/m2 and women with prior GDM.46 Nevertheless, it
has been used “off-label” to treat or prevent several clinical
problems associated with PCOS47: oligomenorrhea, hirsut-
Associated morbidity ism, infertility, obesity, and early miscarriage. Next, we pres-
Abdominal obesity is the main metabolic abnormality in ent the different therapeutic effects of metformin in patients
PCOS women. Women with PCOS are at increased risk for with PCOS.
developing T2DM or impaired glucose tolerance (IGT). In
a study of 122 obese women with PCOS, 45% had either Metabolic effects: Metformin, compared with placebo,
IGT (35%) or T2DM (10%) by age 40.44 Actually, T2DM reduced plasma insulin by improving insulin sensitiv-
and gestational diabetes (GDM) are two different variants ity as measured by glucose clamp studies. Likewise, it
for the same disease in different settings. They are consid- reduced serum free testosterone and increased mean
ered as direct outcomes of the insulin resistance metabolic serum HDL cholesterol. Both the clinical and bio-
abnormality in PCOS. Moreover, they can lead to increased chemical alterations were not correlated to changes
cardiovascular risk,45 obstructive sleep apnea, nonalcoholic in body weight. Also, they were sustained over a
fatty liver disease, and nonalcoholic steatohepatitis. mean duration of treatment of 11 months in a follow-
Taken together, women with PCOS should be actively up observational study. However, there was no sig-
screened with fasting and 75 g oral glucose load for IGT nificant reduction in hirsutism despite the lower free
and T2DM. A hemoglobin A1C (HbA1c) test is also accepted serum testosterone concentration.34 A meta-analysis
of 13 trials reported that metformin is associated with believe that women with PCOS who do not respond at
a reduction in blood pressure, LDL, and fasting insu- these doses should be treated with 2000 mg daily.64
lin levels, when compared with placebo.48 Spontaneous abortion: The spontaneous abortion rate in
Gestational diabetes: The incidence of GDM is several- women with PCOS is 20%–40% higher than the baseline
fold higher in women with PCOS than in the general in the general obstetric population.65 Three studies in
obstetric population.49,50 Two observational studies women with PCOS reported miscarriage rates of 62%–
suggested that metformin use was associated with a 73% in pregnancies when metformin was not taken,
lower risk of GDM. However, this was not confirmed and only 9%–36% of pregnancies in the same women
in a trial of 273 pregnancies among 257 women with under metformin.49,66,67 By contrast, in a meta-analysis
PCOS who were randomly assigned to receive metfor- incorporating 17 trials in women with PCOS taking
min (2000 mg/day) or placebo from the first trimester metformin, with or without clomiphene, for ovulation
until delivery. No significant difference in the preva- induction, no effect of metformin was detected on the
lence of GDM was detected between the groups (met- pregnancy loss risk (secondary end point50).
formin 17.6% vs. placebo 16.9%).51 Hirsutism: Although metformin may reduce serum
Obesity: At least 40% of women with PCOS are obese.25 androgen concentrations, it has limited benefit for the
Metformin may be useful for achieving weight loss or treatment of hirsutism.68–74
potentiating weight loss by caloric restriction. A large
clinical trial randomly assigned 150 obese women In summary, the Endocrine Society recommended using
to receive weight reduction drugs (sibutramine or metformin28 in women with PCOS who have T2DM or IGT
orlistat) or metformin (850 mg twice daily).52 After and failed lifestyle modification. For women with PCOS
6 months of treatment, all three groups had 9%–14% and menstrual irregularity who cannot take or do not tol-
reductions in BMI. Thus, metformin treatment was erate oral contraceptives, metformin should be offered as a
associated with as much weight reduction as other second-line therapy. By contrast, the use of metformin as a
anti-obesity medications. In another study involving first-line treatment of cutaneous manifestations, as preven-
obese women with and without PCOS, metformin (850 tion of pregnancy complications, or for the treatment of
mg twice daily) plus a low-calorie diet (1200–1400 kcal obesity is not recommended. In addition, treatment of infer-
daily) was superior to a low-calorie diet alone in facili- tility should be initiated with clomiphene citrate (or compa-
tating weight loss.53 Other investigators have reported rable estrogen modulators such as letrozole) as the first-line
similar results.54,55 One trial reported that higher- treatment of anovulatory infertility in women with PCOS.
dose metformin (2550 mg/day) was more effective for Metformin usage is an adjuvant therapy for infertility and
weight loss than lower-dose metformin (1500 mg/day) for OHSS prevention in IVF patients.
in obese women with PCOS.56 The dose-dependent
effect was observed in women with a BMI between 30
and 37 kg/m2, but not in those whose BMI was greater GDM and metformin
than 37 kg/m2. Metformin appears to decrease caloric
intake by suppressing appetite.57 This effect may be Among the antidiabetic agents, metformin and human
independent of its gastrointestinal side effects.58 insulin are classified as category B (no evidence of risk in
Infertility: Metformin has not been proven to be endo- humans). Yet, despite the inconvenience of injections, met-
metrial protective but will restore ovulatory men- formin is barely used in pregnancy. The American College
ses in approximately 50% of women with PCOS, 34,59 of Obstetricians and Gynecologists (ACOG)75 and the
although some studies report ovulatory rates from Endocrine Society76 endorsed the use of metformin in GDM.
23% to 90% in normoandrogenic anovulatory Importantly, metformin crosses the placenta and could affect
women.60 This drug improves pregnancy rates when fetal physiology directly.77 This fact is not considered as a
compared with placebo. Metformin and clomiphene disadvantage if fetal harm is excluded. Actually, a drug that
together also improve rates when compared with improves insulin sensitivity in the mother and fetus has a the-
clomiphene alone, but not when metformin is com- oretical advantage. This was the rationale of the Metformin
pared directly with clomiphene. Metformin does in Gestational Diabetes (MiG) trial, as discussed later.78
not improve live births whether it is used alone or in The drug has not been specifically approved by the FDA
combination with clomiphene.61,62 Thus, the benefit of for treatment during pregnancy. Also, it was not recom-
metformin appears limited regarding the improve- mended by the ADA.79 Second- and third-trimester met-
ment of reproductive outcomes in women with PCOS. formin treatment of GDM appears to be safe. One small
However, it may prevent ovarian hyperstimulation retrospective cohort study80 showed increased rates of
syndrome (OHSS) in women with PCOS undergoing perinatal loss and preeclampsia as compared with insulin
in vitro fertilization (IVF).63 treatment. Many other studies reported that metformin use
Dose and side effects: The optimum metformin dose for during pregnancy is safe and effective,81 even in the first
restoring ovulatory menses has not been determined. The trimester.51,65,67,82–90 The only disadvantage may be limited
two best studied metformin regimens are 500 mg three effectiveness. Between one-quarter and one-half of women
times daily and 850 mg twice daily. Some authorities will need insulin to achieve glycemic targets.
Postpartum prevention of T2DM in women with prior GDM 215
The largest trial assessing metformin use in GDM is long-term safety, patients who are prescribed metformin
the Australian MiG trial.78 Seven hundred and fifty-one should be informed of the existing uncertainties related to
Australian women with GDM at 20–33 weeks of gestation transplacental passage.
were randomly assigned to receive metformin (with insu-
lin supplementation if needed) or insulin therapy. Perinatal
morbidity was reported as the composite rate of neonatal
hypoglycemia, respiratory distress, need for phototherapy,
Postpartum prevention of T2DM
birth trauma, prematurity, and low Apgar score. One-third in women with prior GDM
of women in each group experienced composite morbid-
ity. There were no serious adverse effects related to metfor- T2DM is characterized by hyperglycemia, insulin resistance,
min treatment; however, almost half of the women using it and relative impairment in insulin secretion. Abnormal glu-
needed supplemental insulin to achieve glycemic control. cose metabolism can be documented years before the onset
Another two small randomized prospective, open- of overt diabetes. Although the natural history of IFG and
labeled, Finnish studies in GDM pregnancies examined the IGT is variable, approximately 25% of subjects with either
effectiveness of metformin compared to insulin in the pre- will progress to diabetes over 3–5 years.99 Individuals with
vention of fetal macrosomy as the primary outcome. Both additional diabetes clinical risk factors, including obesity and
concluded that metformin is possibly suitable for the pre- family history, are more likely to develop diabetes. The risk
vention of fetal macrosomy, especially in lean or moderately for T2DM is higher in women who have had GDM.100–103
overweight women developing GDM during late gestation. These women have defects in both insulin secretion and
However, 20.9%–31.9% of the women randomized to the insulin action, the severity of which correlate with the future
metformin arm needed supplemental insulin, more likely risk of diabetes.100,101
among older, those with considerable obesity, high fasting As GDM is a predictor of developing T2DM, it is also a
blood glucose or higher serum fructosamine concentra- predictor for developing cardiac disease and metabolic syn-
tion at diagnosis, early GDM diagnosis, and earlier need for drome in both the mothers and the offspring. For example,
pharmacological treatment.91,92 higher birth weight (>4.0 kg) may be associated with an
The latest published study evaluated glycemic control in increased risk of diabetes.104 Long-term follow-up by the
women receiving metformin or insulin for GDM. Lower family physicians and efforts to prevent T2DM are sug-
mean glucose levels were observed in the metformin group.93 gested by ACOG,75 the ADA,105 and the Fifth International
Women using metformin developed less weight gain and a Workshop Conference on Gestational Diabetes.106
lower frequency of neonatal hypoglycemia. In the metfor- The goals of diabetes prevention are delaying the onset
min group, 26.08% of women required supplemental insulin of diabetes, preserving beta cell function, and preventing or
for glycemic control. Early gestational age at diagnosis and delaying microvascular and cardiovascular complications.
higher mean pretreatment glucose levels were identified as In order to prevent the development of T2DM, the post-
predictors of insulin need.93 GDM patient should be perceived as a “prediabetic” patient.
The main principles for managing these patients are as
follows:
Metformin for GDM prevention
Although observational studies of metformin use in women 1. Long-term intensive follow-up, including timely testing
with PCOS supported this hypothesis,89,90 a randomized trial for fasting glucose and HgA1C.
among women with PCOS did not.51 In this trial, as in many 2. Promotion of lifestyle changes, including healthy diet,
other trials, women in the metformin group gained less weight control, and regular exercise.
weight during pregnancy than those in the placebo group. 3. Encouraging breastfeeding.
4. For those women in which lifestyle changes fail to improve
Long-term outcome glycemic indices, it is suggested to use medical treatment,
Metformin crosses the placenta.77 Indeed, in one study, mainly metformin.
cord arterial levels were twice as high as maternal venous
levels.94 Long-term outcome of fetal exposure to an insulin- All patients should be counseled on the benefits of weight
sensitizing agent such as metformin is unknown. Notably, loss and increasing physical activity. Different programs for
it is not even known whether the exposure is beneficial or successful compliance, including health coaches and psy-
harmful. Thus, caution is still warranted in its use during chological support, are recommended. Patients should also
pregnancy. A 2-year follow-up study of the offspring in the be encouraged to stop smoking.
MiG trial reported no difference in total fat or central adipos- Importantly, changes in lifestyle, including diet modifi-
ity between metformin-exposed and nonexposed o ffspring. cation, weight loss, and exercise, slow progression of IGT to
However, there was an increase in subcutaneous fat deposi- overt diabetes.107 The Mediterranean diet appears to be associ-
tion.95 Whether this is beneficial or harmful remains to be ated with several health benefits, including diabetes preven-
determined.96,97 Because offspring of diabetic mothers may tion.108,109 There is no single definition of a Mediterranean diet,
not manifest obesity until the age of 5–7 years,98 longer- but typically it is rich in fruits, vegetables, whole grains, beans,
term studies are needed. Until such studies demonstrate nuts, and seeds; includes olive oil as an important source of
monounsaturated fat; and allows low to moderate wine con- Pharmacologic therapy may be helpful in preventing
sumption. There are typically low to moderate amounts of T2DM in high-risk patients (GDM) for whom lifestyle
fish, poultry, and dairy products, with little red meat. interventions fail or are not sustainable.120,121 Metformin is
The benefit of exercise in preventing diabetes has been a pivotal therapy in this setting. To the best of our knowl-
demonstrated in several studies.110–115 Weight reduction edge, there is only one randomized controlled trial of
resulting from lifestyle intervention, i.e., combined diet and pharmacological therapy in women with prior GDM that
exercise aimed at weight loss and increasing activity levels, considered diabetes prevention as the primary end point
can improve glucose tolerance and prevent progression from (the Troglitazone in Prevention of Diabetes [TRIPOD]
IGT to T2DM.107,116 Moreover, breastfeeding is correlated study122). In this study, T2DM incidence reduced from
with reduced blood glucose levels and incidence of T2DM 45% to 20% (RR 0.45) with troglitazone versus placebo.
among both women with a history of GDM and women in Troglitazone is not marketed anymore due to the risk of
the general population.117–119 fatal liver failure.
REFERENCES
1. Nathan DM, Buse JB, Davidson MB et al. Management of 16. Salpeter SR, Greyber E, Pasternak GA, Salpeter EE. Risk of fatal
hyperglycemia in type 2 diabetes: A consensus algorithm for and nonfatal lactic acidosis with metformin use in type 2 diabetes
the initiation and adjustment of therapy: A consensus statement mellitus. Cochrane Database Syst Rev 2010; 14(4): CD002967.
from the American Diabetes Association and the European 17. Gan SC, Barr J, Arieff AI et al. Biguanide-associated lactic aci-
Association for the Study of Diabetes. Diabetes Care 2006; 29: dosis. Case report and review of the literature. Arch Intern Med
1963–1972. 1992; 152: 2333–2336.
2. Nathan DM, Buse JB, Davidson MB et al. Medical management 18. Sirtori CR, Pasik C. Re-evaluation of a biguanide, metformin:
of hyperglycemia in type 2 diabetes: A consensus algorithm Mechanism of action and tolerability. Pharmacol Res 1994; 30:
for the initiation and adjustment of therapy: A consensus state- 187–228.
ment of the American Diabetes Association and the European 19. Schäfer G. Biguanides. A review of history, pharmacodynamics
Association for the Study of Diabetes. Diabetes Care 2009; 32: and therapy. Diabete Metab 1983; 9: 148–163.
193–203. 20. Direct-to-consumer advertisements for Glucophage XR. Med Lett
3. Bailey CJ. Biguanides and NIDDM. Diabetes Care 1992; 15: Drugs Ther 2001; 43: 25–26.
755–772. 21. Stein I, Leventhal M. Amenorrhea associated with bilateral poly-
4. Bailey CJ, Turner RC. Metformin. N Engl J Med 1996; 334: cystic ovaries. Am J Obstet Gynecol 1935; 29: 181–185.
574–579. 22. Misso ML, Costello MF, Garrubba M et al. Metformin versus clo-
5. Stumvoll M, Nurjhan N, Perriello G et al. Metabolic effects of miphene citrate for infertility in non-obese women with polycys-
metformin in non-insulin-dependent diabetes mellitus. N Engl J tic ovary syndrome: A systematic review and meta-analysis. Hum
Med 1995; 333: 550–554. Reprod Update 2013; 19(1): 2–11.
6. United Kingdom Prospective Diabetes Study (UKPDS). 13: 23. Azziz R, Woods KS, Reyna R et al. The prevalence and features
Relative efficacy of randomly allocated diet, sulphonylurea, insu- of the polycystic ovary syndrome in an unselected population. J
lin, or metformin in patients with newly diagnosed non-insulin Clin Endocrinol Metab 2004; 89: 2745–2749.
dependent diabetes followed for three years. Br Med J 1995; 310: 24. Ehrmann DA. Polycystic ovary syndrome. N Engl J Med 2005;
83–88. 352: 1223–1236.
7. Wu MS, Johnston P, Sheu WH et al. Effect of metformin on 25. Yildiz BO, Knochenhauer ES, Azziz R. Impact of obesity on
carbohydrate and lipoprotein metabolism in NIDDM patients. the risk for polycystic ovary syndrome. J Clin Endocrinol Metab
Diabetes Care 1990; 13: 1–8. 2008; 93: 162–168.
8. Zhou G, Myers R, Li Y et al. Role of AMP-activated protein 26. Poretsky L, Piper B. Insulin resistance, hypersecretion of LH, and
kinase in mechanism of metformin action. J Clin Invest 2001; 108: a dual-defect hypothesis for the pathogenesis of polycystic ovary
1167–1174. syndrome. Obstet Gynecol 1994; 84(4): 613–621.
9. Fryer LG, Parbu-Patel A, Carling D. The Anti-diabetic drugs rosi- 27. The Rotterdam ESHRE/ASRM-Sponsored PCOS consensus work-
glitazone and metformin stimulate AMP-activated protein kinase shop group. Revised 2003 consensus on diagnostic criteria and
through distinct signaling pathways. J Biol Chem 2002; 277: long-term health risks related to polycystic ovary syndrome
25226–25232. (PCOS). Hum Reprod 2004; 19: 41–47.
10. Hawley SA, Gadalla AE, Olsen GS et al. The antidiabetic drug 28. Legro RS, Arslanian SA, Ehrmann DA et al. Diagnosis and treat-
metformin activates the AMP-activated protein kinase cascade ment of polycystic ovary syndrome: An Endocrine Society clinical
via an adenine nucleotide-independent mechanism. Diabetes practice guideline. J Clin Endocrinol Metab 2013; 98: 2013–2050.
2002; 51: 2420–2425. 29. Martin KA, Chang RJ, Ehrmann DA et al. Evaluation and treatment of
11. Hermann LS, Scherstén B, Bitzén PO et al. Therapeutic compari- hirsutism in premenopausal women: An Endocrine Society clinical
son of metformin and sulfonylurea, alone and in various combi- practice guideline. J Clin Endocrinol Metab 2008; 93: 1105–1120.
nations. A double-blind controlled study. Diabetes Care 1994; 17: 30. Legro R. Hyperandrogenism and Hyperinsulinemia. Glob.
1100–1109. libr. women’s med., (ISSN: 1756-2228) 2008; DOI 10.3843/
12. Saenz A, Fernandez-Esteban I, Mataix A et al. Metformin mono- GLOWM.10303.
therapy for type 2 diabetes mellitus. Cochrane Database Syst Rev 31. Dunaif A, Xia J, Book CB et al. Excessive insulin receptor serine
2005; 20(3): CD002966. phosphorylation in cultured fibroblasts and in skeletal muscle. A
13. de Jager J, Kooy A, Lehert P et al. Long term treatment with met- potential mechanism for insulin resistance in the polycystic ovary
formin in patients with type 2 diabetes and risk of vitamin B-12 syndrome. J Clin Invest 1995; 96: 801–810.
deficiency: Randomised placebo controlled trial. Br Med J 2010; 32. Barbieri RL, Smith S, Ryan KJ. The role of hyperinsulinemia in the
340: c2181. pathogenesis of ovarian hyperandrogenism. Fertil Steril 1988; 50:
14. Stang M, Wysowski DK, Butler-Jones D. Incidence of lactic aci- 197–212.
dosis in metformin users. Diabetes Care 1999; 22: 925–927. 33. Nestler JE, Jakubowicz DJ. Decreases in ovarian cytochrome
15. Cryer DR, Nicholas SP, Henry DH et al. Comparative outcomes P450c17 alpha activity and serum free testosterone after reduc-
study of metformin intervention versus conventional approach tion of insulin secretion in polycystic ovary syndrome. N Engl J
the COSMIC Approach Study. Diabetes Care 2005; 28: 539–543. Med 1996; 335: 617–623.
References 217
34. Moghetti P, Castello R, Negri C et al. Metformin effects on 56. Harborne LR, Sattar N, Norman JE et al. Metformin and weight loss
clinical features, endocrine and metabolic profiles and insulin in obese women with polycystic ovary syndrome: Comparison of
sensitivity in polycystic ovary syndrome: A randomized, dou- doses. J Clin Endocrinol Metab 2005; 90: 4593–4598.
ble-blind, placebo-controlled 6-month trial, followed by open, 57. Yki-Järvinen H, NikkiläK, Mäkimattila S. Metformin prevents
long-term clinical evaluation. J Clin Endocrinol Metab 2000; 85: weight gain by reducing dietary intake during insulin therapy in
139–146. patients with type 2 diabetes mellitus. Drugs 1999; 58(Suppl. 1):
35. McLaughlin T, Abbasi F, Cheal K et al. Use of metabolic markers 53–54.
to identify overweight individuals who are insulin resistant. Ann 58. Haupt E, Knick B, Koschinsky T et al. Oral antidiabetic combina-
Intern Med 2003; 139: 802–809. tion therapy with sulphonylureas and metformin. Diabete Metab
36. Ascaso JF, Pardo S, Real JT et al. Diagnosing insulin resistance 1991; 17(1 Pt. 2): 224–231.
by simple quantitative methods in subjects with normal glucose 59. Unlühizarci K, Keleştimur F, Bayram F et al. The effects of metfor-
metabolism. Diabetes Care 2003; 26: 3320–3325. min on insulin resistance and ovarian steroidogenesis in women
37. Gambineri A, Pelusi C, Vicennati V et al. Obesity and the poly- with polycystic ovary syndrome. Clin Endocrinol 1999; 51(2):
cystic ovary syndrome. Int J Obes Relat Metab Disord 2002; 26: 231–236.
883–896. 60. Carmina E, Lobo RA. Does metformin induce ovulation in nor-
38. Pasquali R. Obesity and androgens: Facts and perspectives. Fertil moandrogenic anovulatory women? Am J Obstet Gynecol 2004;
Steril 2006; 85: 1319–1340. 191: 1580–1584.
39. Taylor AE, McCourt B, Martin KA et al. Determinants of abnor- 61. Tang T, Lord M, Norman RJ et al. Insulin-sensitising drugs
mal gonadotropin secretion in clinically defined women with (metformin, rosiglitazone, pioglitazone, D-chiro-inositol)
polycystic ovary syndrome. J Clin Endocrinol Metab 1997; 82: for women with polycystic ovary syndrome, oligo amenor-
2248–2256. rhoea and subfertility. Cochrane Database Syst Rev 2012; 5:
40. Nestler JE, Powers LP, Matt DW et al. A direct effect of hyperinsu- CD003053.
linemia on serum sex hormone-binding globulin levels in obese 62. Tang T, Balen AH. Use of metformin for women with polycystic
women with the polycystic ovary syndrome. J Clin Endocrinol ovary syndrome. Hum Reprod Update 2013; 19: 1.
Metab 1991; 72: 83–89. 63. Tso LO, Costello MF, Albuquerque LE et al. Metformin treat-
41. Stener-Victorin E, Holm G, Labrie F et al. Are there any sensitive ment before and during IVF or ICSI in women with polycys-
and specific sex steroid markers for polycystic ovary syndrome? tic ovary syndrome. Cochrane Database Syst Rev 2009; 2:
J Clin Endocrinol Metab 2010; 95: 810–819. CD006105.
42. Legro RS, Strauss JF. Molecular progress in infertility: Polycystic 64. Nestler JE, Stovall D, Akhter N et al. Strategies for the use of
ovary syndrome. Fertil Steril 2002; 78: 569–576. insulin-sensitizing drugs to treat infertility in women with poly-
43. Fujiwara T, Sidis Y, Welt C et al. Dynamics of inhibin subunit cystic ovary syndrome. Fertil Steril 2002; 77: 209–215.
and follistatin mRNA during development of normal and polycys- 65. Glueck CJ, Wang P, Goldenberg N et al. Pregnancy outcomes
tic ovary syndrome follicles. J Clin Endocrinol Metab 2001; 86: among women with polycystic ovary syndrome treated with
4206–4215. metformin. Hum Reprod 2002; 17: 2858–2864.
44. Ehrmann DA, Barnes RB, Rosenfield RL et al. Prevalence of 66. Jakubowicz DJ, Iuorno MJ, Jakubowicz S. Effects of metformin
impaired glucose tolerance and diabetes in women with polycys- on early pregnancy loss in the polycystic ovary syndrome.
tic ovary syndrome. Diabetes Care 1999; 22: 141–146. J Clin Endocrinol Metab 2002; 87: 524–529.
45. Rosenzweig JL, Ferrannini E, Grundy SM et al. Primary preven- 67. Glueck CJ, Phillips H, Cameron D et al. Continuing metfor-
tion of cardiovascular disease and type 2 diabetes in patients at min throughout pregnancy in women with polycystic ovary
metabolic risk: An Endocrine Society clinical practice guideline. syndrome appears to safely reduce first-trimester spontaneous
J Clin Endocrinol Metab 2008; 93: 3671–3689. abortion: A pilot study. Fertil Steril 2001; 75: 46–52.
46. American Diabetes Association. Standards of medical care in 68. Ibáñez L, Valls C, Ferrer A et al. Additive effects of insulin-
diabetes 2014. Diabetes Care 2014; 37(Suppl. 1): 14–80. sensitizing and anti-androgen treatment in young, nonobese
47. Nestler JE. Metformin for the treatment of the polycystic ovary women with hyperinsulinism, hyperandrogenism, dyslip-
syndrome. N Engl J Med 2008; 358: 47–54. idemia, and anovulation. J Clin Endocrinol Metab 2002; 87:
48. Lord JM, Flight IH, Norman RJ. Insulin-sensitising drugs (metfor- 2870–2874.
min, troglitazone, rosiglitazone, pioglitazone, D-chiro-inositol) 69. Ibáñez L, Valls C, Potau N et al. Sensitization to insulin in ado-
for polycystic ovary syndrome. Cochrane Database Syst Rev lescent girls to normalize hirsutism, hyperandrogenism, oligo-
2003; (3): CD003053. menorrhea, dyslipidemia, and hyperinsulinism after precocious
49. Thatcher SS, Jackson EM. Pregnancy outcome in infertile patients pubarche. J Clin Endocrinol Metab 2000; 85: 3526–3530.
with polycystic ovary syndrome who were treated with metfor- 70. Pasquali R, Antenucci D, Casimirri F et al. Clinical and hormonal
min. Fertil Steril 2006; 85: 1002–1009. characteristics of obese amenorrheic hyperandrogenic women
50. Palomba S, Falbo A, Orio F et al. Effect of preconceptional met- before and after weight loss. J Clin Endocrinol Metab 1989; 68:
formin on abortion risk in polycystic ovary syndrome: A system- 173–179.
atic review and meta-analysis of randomized controlled trials. 71. Kolodziejczyk B, Duleba AJ, Spaczynski RZ et al. Metformin
Fertil Steril 2009; 92: 1646–1658. therapy decreases hyperandrogenism and hyperinsulinemia in
51. Vanky E, Stridsklev S, Heimstad R et al. Metformin versus placebo women with polycystic ovary syndrome. Fertil Steril 2000; 73:
from first trimester to delivery in polycystic ovary syndrome: 1149–1154.
A randomized, controlled multicenter study. J Clin Endocrinol 72. Kelly CJ, Gordon D. The effect of metformin on hirsutism in poly-
Metab 2010; 95: E448–E455. cystic ovary syndrome. Eur J Endocrinol 2002; 147: 217–221.
52. Gokcel A, Gumurdulu Y, Karakose H et al. Evaluation of the 73. Harborne L, Fleming R, Lyall H et al. Metformin or antiandrogen
safety and efficacy of sibutramine, orlistat and metformin in the in the treatment of hirsutism in polycystic ovary syndrome. J Clin
treatment of obesity. Diabetes Obes Metab 2002; 4: 49–55. Endocrinol Metab 2003; 88: 4116–4123.
53. Pasquali R, Gambineri A, Biscotti D et al. Effect of long-term 74. Morin-Papunen LC, Vauhkonen I, Koivunen RM et al. Endocrine
treatment with metformin added to hypocaloric diet on body and metabolic effects of metformin versus ethinyl estradiol-
composition, fat distribution, and androgen and insulin lev- cyproterone acetate in obese women with polycystic ovary syn-
els in abdominally obese women with and without the poly- drome: A randomized study. J Clin Endocrinol Metab 2000; 85:
cystic ovary syndrome. J Clin Endocrinol Metab 2000; 85: 3161–3168.
2767–2774. 75. Committee on Practice Bulletins—Obstetrics. Practice Bulletin
54. Kay JP, Alemzadeh R, Langley G et al. Beneficial effects of metfor- No. 137: Gestational diabetes mellitus. Obstet Gynecol 2013;
min in normoglycemic morbidly obese adolescents. Metabolism 122: 406–416.
2001; 50: 1457–1461. 76. Blumer I, Hadar E, Hadden DR et al. Diabetes and pregnancy: An
55. Freemark M, Bursey D. The effects of metformin on body mass Endocrine Society clinical practice guideline. J Clin Endocrinol
index and glucose tolerance in obese adolescents with fast- Metab 2013; 98: 4227–4249.
ing hyperinsulinemia and a family history of type 2 diabetes. 77. Charles B, Norris R, Xiao X et al. Population pharmacokinetics of
Pediatrics 2001; 107: E55. metformin in late pregnancy. Ther Drug Monit 2006; 28: 67–72.
78. Rowan JA, Hague WM, Gao W et al. Metformin versus insulin for 102. Feig DS, Zinman B, Wang X et al. Risk of development of diabetes
the treatment of gestational diabetes. N Engl J Med 2008; 358: mellitus after diagnosis of gestational diabetes. Can Med Assoc J
2003–2015. 2008; 179: 229–234.
79. American Diabetes Association. Preconception care of women 103. Kim C, Newton KM, Knopp RH. Gestational diabetes and the
with diabetes. Diabetes Care 2004; 27(Suppl. 1): S76–S78. incidence of type 2 diabetes: A systematic review. Diabetes Care
80. Hellmuth E, Damm P, Molsted-Pederson L. Oral hypoglycae- 2002; 25: 1862–1868.
mic agents in 118 diabetic pregnancies. Diabet Med 2000; 17: 104. Dyck RF, Klomp H, Tan L. From “thrifty genotype” to “hefty fetal
507–511. phenotype”: The relationship between high birthweight and dia-
81. Coetzee EJ, Jackson WP. Pregnancy in established non-insulin- betes in Saskatchewan Registered Indians. Can J Public Health
dependent diabetics. S Afr Med J 1980; 58: 795–802. 2001; 92: 340–344.
82. Coetzee EJ, Jackson WP. Oral hypoglycaemics in the first trimes- 105. American Diabetes Association. Standards of medical care in
ter and fetal outcome. S Afr Med J 1984; 65: 635–637. diabetes—2014. Diabetes Care 2014; 37(Suppl. 1): S14–S80.
83. Coetzee EJ, Jackson WP. Metformin in management of pregnant 106. Metzger BE, Buchanan TA, Coustan DR et al. Summary and rec-
insulin-independent diabetics. Diabetologia 1979; 16: 241–245. ommendations of the Fifth International Workshop-Conference
84. Coetzee EJ, Jackson WP. The management of non-insulin-depen- on Gestational Diabetes Mellitus. Diabetes Care 2007; 30(Suppl.
dent diabetes during pregnancy. Diabetes Res Clin Pract 1985– 2): S251–S260.
1986; 1: 281–287. 107. Orozco LJ, Buchleitner AM, Gimenez-Perez G et al. Exercise
85. Gilbert C, Valois M, Koren G. Pregnancy outcome after first-tri- or exercise and diet for preventing type 2 diabetes mellitus.
mester exposure to metformin: A meta-analysis. Fertil Steril 2006; Cochrane Database Syst Rev 2008; 16(3): CD003054.
86: 658–663. 108. Salas-Salvadó J, Bulló M, Babio N et al. Reduction in the inci-
86. Glueck CJ, Bornovali S, Pranikoff J et al. Metformin, pre-eclampsia, dence of type 2 diabetes with the Mediterranean diet: Results
and pregnancy outcomes in women with polycystic ovary of the PREDIMED-Reus nutrition intervention randomized trial.
syndrome. Diabet Med 2004; 21: 829–836. Diabetes Care 2011; 34: 14–19.
87. Gutzin SJ, Kozer E, Magee LA et al. The safety of oral hypogly- 109. Salas-Salvadó J, Bulló M, Estruch R et al. Prevention of diabetes
cemic agents in the first trimester of pregnancy: A meta-analysis. with Mediterranean diets: A subgroup analysis of a randomized
Can J Clin Pharmacol 2003; 10: 179–183. trial. Ann Intern Med 2014; 160: 1–10.
88. Hughes RCE, Rowan JA. Pregnancy in women with Type 2 diabe- 110. Helmrich SP, Ragland DR, Leung RW et al. Physical activity and
tes: Who takes metformin and what is the outcome? Diabet Med reduced occurrence of non-insulin-dependent diabetes mellitus.
2006; 23: 318–322. N Engl J Med 1991; 325: 147–152.
89. Glueck CJ, Wang P, Kobayashi S et al. Metformin therapy through- 111. Lynch J, Helmrich SP, Lakka TA et al. Moderately intense physical
out pregnancy reduces the development of gestational diabetes activities and high levels of cardiorespiratory fitness reduce the
in women with polycystic ovary syndrome. Fertil Steril 2002; 77: risk of non-insulin-dependent diabetes mellitus in middle-aged
520–525. men. Arch Intern Med 1996; 156: 1307–1314.
90. Glueck CJ, Pranikoff J, Aregawi D et al. Prevention of gestational 112. Hu FB, Sigal RJ, Rich-Edwards JW et al. Walking compared
diabetes by metformin plus diet in patients with polycystic ovary with vigorous physical activity and risk of type 2 diabetes in
syndrome. Fertil Steril 2008; 89: 625–634. women: A prospective study. J Am Med Assoc 1999; 282:
91. Ijäs H, Vääräsmäki M, Morin-Papunen L et al. Metformin should 1433–1439.
be considered in the treatment of gestational diabetes: A pro- 113. Eriksson KF, Lindgärde F. Prevention of type 2 (non-insulin-
spective randomised study. Br J Obstet Gynaecol 2011; 118: dependent) diabetes mellitus by diet and physical exercise.
880–885. The 6-year Malmö feasibility study. Diabetologia 1991; 34:
92. Tertti K, Ekblad U, Koskinen P et al. Metformin vs. insulin in ges- 891–898.
tational diabetes. A randomized study characterizing metformin 114. Knowler WC, Barrett-Connor E, Fowler SE et al. Reduction in the
patients needing additional insulin. Diabetes Obes Metab 2013; incidence of type 2 diabetes with lifestyle intervention or metfor-
15: 246–251. min. N Engl J Med 2002; 346: 393–403.
93. Spaulonci CP, Bernardes LS, Trindade TC et al. Randomized trial 115. Hu G, Lindström J, Valle TT et al. Physical activity, body mass
of metformin vs insulin in the management of gestational diabe- index, and risk of type 2 diabetes in patients with normal or
tes. Am J Obstet Gynecol 2013; 209: 34.e1–34.e7. impaired glucose regulation. Arch Intern Med 2004; 164:
94. Vanky E, Zahlsen K, Spigset O et al. Placental passage of metfor- 892–896.
min in women with polycystic ovary syndrome. Fertil Steril 2005; 116. Gillett M, Royle P, Snaith A et al. Non-pharmacological interven-
83: 1575–1578. tions to reduce the risk of diabetes in people with impaired glu-
95. Rowan JA, Rush EC, Obolonkin V et al. Metformin in gestational cose regulation: A systematic review and economic evaluation.
diabetes: The offspring follow-up (MiG TOFU): Body composi- Health Technol Assess 2012; 16: 1–236.
tion at 2 years of age. Diabetes Care 2011; 34: 2279–2284. 117. Kjos SL, Henry O, Lee RM et al. The effect of lactation on glucose
96. Barbour LA, Van Pelt RE, Brumbaugh DE et al. Comment on: and lipid metabolism in women with recent gestational diabetes.
Rowan et al. Metformin in Gestational diabetes: The Offspring Obstet Gynecol 1993; 82: 451–455.
Follow-Up (MiG TOFU): Body composition at 2 years of age. 118. Taylor JS, Kacmar JE, Nothnagle M et al. A systematic review of
Diabetes Care 2011; 34: 2279–2284; Diabetes Care 2012; 35: the literature associating breastfeeding with type 2 diabetes and
e28; author reply e30. gestational diabetes. J Am Coll Nutr 2005; 24: 320–326.
97. Feig DS, Moses RG. Metformin therapy during pregnancy: Good 119. Gunderson EP, Hedderson MM, Chiang V et al. Lactation inten-
for the goose and good for the gosling too? Diabetes Care 2011; sity and postpartum maternal glucose tolerance and insulin resis-
34: 2329–2330. tance in women with recent GDM: The SWIFT cohort. Diabetes
98. Silverman BL, Rizzo T, Green OC et al. Long-term prospec- Care January 2012; 35(1): 50–56.
tive evaluation of offspring of diabetic mothers. Diabetes 1991; 120. Padwal R, Majumdar SR, Johnson JA et al. A systematic review of
40(Suppl. 2): 121–125. drug therapy to delay or prevent type 2 diabetes. Diabetes Care
99. Nathan DM, Davidson MB, DeFronzo RA et al. Impaired fasting 2005; 28: 736–744.
glucose and impaired glucose tolerance: Implications for care. 121. Gillies CL, Abrams KR, Lambert PC et al. Pharmacological and
Diabetes Care 2007; 30: 753–759. lifestyle interventions to prevent or delay type 2 diabetes in
100. Ryan EA, Imes S, Liu D et al. Defects in insulin secretion and people with impaired glucose tolerance: Systematic review and
action in women with a history of gestational diabetes. Diabetes meta-analysis. Br Med J 2007; 334: 299.
1995; 44: 506–512. 122. Buchanan TA, Xiang AH, Peters RK et al. Preservation of pancre-
101. Damm P, Kühl C, Hornnes P et al. A longitudinal study of plasma atic B-cell function and prevention of type 2 diabetes by phar-
insulin and glucagon in women with previous gestational diabe- macological treatment of insulin resistance in high-risk Hispanic
tes. Diabetes Care 1995; 18: 654–665. women. Diabetes 2002; 51: 2796–2803.
26 Facing noncommunicable
diseases’ global epidemic:
The battle of prevention starts
in utero—The FIGO challenge
Luis Cabero and Sabaratnam Arulkumaran
about 80% of NCD deaths.1 It must be emphasized that all

Introduction of them share four common risk factors: tobacco consump-
In the last few decades, different health organizations have tion, insufficient physical activity, excessive use of alcohol,
highlighted the growing importance of noncommunicable and unhealthy diet.
diseases (NCDs) on population health. Noncommunicable In terms of attributable deaths, the main NCD risk fac-
diseases, also known as chronic diseases, are not transmit- tor globally is high blood pressure (attributed to 16.5% of
ted from person to person, are not acute, and cannot be deaths worldwide1), followed by consumption of tobacco (9%),
treated in a short period but generally evolve slowly and tend increased blood glucose (6%), insufficient physical activity
to remain lifelong. The four main NCDs are cardiovascular (6%), and excessive weight gain and obesity (5%).1 The rapid
disease (including heart attacks and stroke), cancer, chronic increase in the number of overweight children is observed in
respiratory diseases (such as chronic obstructive pulmonary the countries with low and medium incomes.
disease and asthma), and diabetes. NCDs affect all age groups and all regions. Children,
adults, and seniors are all vulnerable to the risk factors
predisposing to NCD, such as unhealthy diets, physical
NCDs’ impact on life expectancy inactivity, exposure to tobacco smoke, or excessive use of
alcohol. These diseases are also favored by factors such as
In its latest report, the WHO acknowledged the impact of aging, rapid and unplanned urbanization seen in certain
deaths due to NCDs. NCDs were responsible for 36 mil- countries, and the globalization of unhealthy lifestyles. An
lion deaths every year. Of this death toll, nearly 80% (29 example of such factor can be the unhealthy diet, which can
million) occurs in low- and median-income countries.1 It be responsible for high blood pressure, increased blood glu-
was considered that NCDs are the leading causes of death cose, hyperlipidemia, and excessive weight gain and obesity.
in all regions except Africa, but according to the current Those factors are called “intermediate risk factors,” and they
estimates, by 2020 the largest increase in mortality from can lead to cardiovascular disease, one of the NCDs.
NCDs will be in Africa. Deaths from NCDs in the African
countries are expected to exceed the sum of those caused
by communicable and nutritional diseases and maternal
and perinatal mortality. NCD would be the most common Impact of NCDs on national
cause of death in 2030.1 and global economy
In addition, more than nine million deaths attributed
to NCDs occur in people under 60 years of age, and 90% of NCDs threaten the progress toward the target of the
these “premature” deaths occur in low- and medium-income Millennium Development Goals (MDGs) of the United
countries. Cardiovascular diseases are responsible for most Nations (UN).1 Poverty is closely related to NCDs. It is antic-
NCD deaths (17.3 million annually), followed by cancer ipated that the rapid rise of these diseases will hinder efforts
(7.6 million), respiratory diseases (4.2 million), and diabe- to reduce poverty in low-income countries, because of the
tes (1.3 million). These four groups of diseases account for increased family expenses for healthcare. Vulnerable and
219
220 Facing noncommunicable diseases’ global epidemic
socially disadvantaged people get sick and die earlier than needs of people suffering from these diseases. However, one
people of higher social status, mainly because of the higher must acknowledge that low-income countries tend to have a
risk of exposure to harmful products, such as tobacco or low capacity for the prevention and control of NCDs.9
unhealthy foods, and limited access to health services. 2,3
In resource-poor settings, the healthcare expenses for car-
diovascular disease, cancer, diabetes, and chronic lung
disease, as stated, can quickly exhaust the resources of Developmental origins of health
families and force them into poverty. The exorbitant costs
of NCDs, particularly prolonged and expensive treatment,
and disease
or the demise of the breadwinner, are pushing millions of In recent decades, various research areas have suggested
people into poverty each year, preventing the progress of that events involved in normal fetal development could have
the country and the society. National economies are suffer- long-term effects and influence health in adulthood.10,11 It
ing considerable losses due to premature death or disability appears that metabolic changes in utero can influence the
to work resulting from heart disease, stroke, and diabetes. physiological and structural patterns that “program” long-
For example, it is expected that India and China will lose term health in adulthood.12,13 Early studies by Barker et al.14,15
about $558 billion from their gross domestic production in the 1980s established that the prevalence of some diseases
(GDP) between 2005 and 2015 due to premature deaths.4–6 in adulthood, such as atherosclerosis, high blood pressure
In 2005, heart disease, stroke, and diabetes caused an esti- (hypertension), stroke, type 2 diabetes mellitus, and dyslip-
mated loss of $9 billion in India and $3 billion in Brazil from idemia, is related to the intrauterine environment (“Barker
the GDP.7 A study commissioned by the World Economic hypothesis”). Moreover, the association has been reported
Forum concluded that the world will sustain a cumulative between low weight and height at birth, with increased risk
output loss of $47 trillion between 2011 and 2030 because of of subsequent development of diseases such as hypertension,
NCDs and mental illnesses, about $30 trillion of which will metabolic syndrome, and stroke.16–18 Other studies have con-
be attributable to cardiovascular diseases, cancers, chronic firmed this relationship and currently are trying to reveal
pulmonary diseases, and diabetes.8 some of the mechanisms behind it.19,20 On an experimental
level, for example, nutritional restriction during pregnancy
has been shown to irreversibly affect the structure, metab-
Combating NCDs olism, and function in some organs, “programming” the
development of certain diseases in offspring.21 The result of
In many countries, the excessive use of alcohol and con- this research is that today we can talk about the fetal origins
sumption of unhealthy foods are affecting both high- of many adult diseases (Developmental Origins of Health and
income and low-income groups. However, the former can Disease [DOHaD]).21 Animal experiments and epidemiolog-
access products and services that protect them from the ical observations in humans suggest that nutrition received
major risks, while low-income groups are often unable to in the intrauterine environment modulates the metabolic
afford it. activity of various tissues in postnatal life.22,23 An important
To reduce the impact of NCDs on individuals and the consequence of intrauterine caloric and nutrient restriction
society, we should apply a comprehensive approach that rein- is the accelerated growth in the postnatal period.24
forces all sectors, including those related to health, finance, In fact, during periods of development, certain epigenetic
foreign affairs, education, and agriculture and planning, to modifications are taking place that establish the parameters
work together to reduce the risks associated with NCDs, and within which the tissue functions. It should be mentioned
to promote strategies and intervention to prevent and con- that the functional changes made by epigenetic influences
trol them. need not be phenotypically (physically) visible in order to
An important tactic to reduce NCDs is to reduce the be significant; however, they usually have long-term health
risk factors associated with these diseases. Effective and consequences. These functional changes cannot be reliably
low-cost solutions exist to modify common risk factors detected by conventional testing, are observed only when
(mainly tobacco consumption, unhealthy diets and physical they match certain environmental contexts, and are likely to
inactivity, and excessive use of alcohol) and to map the epi- be initially masked by systemic effects.
demic of NCDs and their risk factors.1 This developmental plasticity is more important during
Another option to combat NCDs is to introduce some periods in which the cells are differentiating and forming
essential high-impact interventions with the focus on pri- specific tissues, as happens mainly during pregnancy (for
mary care in order to improve their early detection and timely both mother and child) and during early childhood, puberty,
treatment. Evidence shows that such interventions are also and menopause. DOHaD hypothesizes that environmental
excellent financial investment, since if introduced early they exposures during critical periods of development may cause
can reduce the need for more expensive treatments. These subtle changes in certain biological functions, although,
measures may also be applied in situations with different lev- practically invisible, they can increase the risk of disease and
els of resources. For maximum effect, the public policies have dysfunction later in life.25
to focus on healthy lifestyles that promote prevention and Some of the most striking evidence of developmental
control of NCDs and reorienting health systems to meet the plasticity that were first discovered came from the analysis
Epigenetic programming 221
of the consequences of prolonged periods of famine. Dutch governs the cell machinery for the production of a specific pro-
conceived in the time of famine during the last two years tein. The addition or removal of a methyl group to a gene is a
of World War II were more likely to develop metabolic way of influencing this process, if a certain gene can be tran-
syndrome in adulthood.26 Subsequent research has also
scribed into RNA and, therefore, whether or not to produce the
confirmed these findings in Chinese famine victims.27 In both corresponding protein. Thus, methylation in a group of genes
groups, the offspring were more likely to have hypertension, may cause changes in many biological processes occurring in
glucose intolerance, and excessive weight gain. the body, and affect the metabolism, causing energy expendi-
ture or conservation. This process could also be responsible for
more specific aspects such as eye or skin color.
Epigenetic programming The discovery of noncoding small RNAs has opened a
new field of microRNAs, which also play an important role in
The discovery of epigenetic programming has provided a epigenetics by modifying gene expression posttranscription-
plausible explanation why and how nutritional characteris- ally.32 MicroRNAs indirectly affect translation of key factors
tics during critical periods of life can have health manifesta- or enzymes required for epigenetic processes in the nucleus.
tion much later. At present, there are a number of examples in What is also remarkable is that these changes in program-
animal models of how changes in the diet can directly influ- ming are largely functional. They include alterations in gene
ence the epigenetic machinery by inhibition of enzymes that expression, protein levels, cell metabolism and differentia-
catalyze the DNA methylation, by histone modifications, or tion, and the number of cells and their location. Functional
by altering the availability of substrates for enzymatic reac- changes are not necessarily identifiable as pathological, but
tions. These epigenetic alterations cause the expression or due to the changes in gene expression, these can lead to dys-
suppression of certain genes that can alter the phenotype, function and disease in adulthood. In such case, they can be
depending on the nature of the affected biological process. considered as markers of increased risk of NCDs. As already
Epigenetic changes can both be transient28 or persist for long mentioned, the functional changes are not necessarily
periods of time.29 apparent at the time of birth and, in some cases, may require
Epigenetic phenomena are fundamental features of mam- a particular environmental or physiological trigger in order
malian development that causes persistent changes in gene to be revealed. For example, certain subtle changes in breast
expression without altering the DNA sequence. In this way, tissue can significantly increase the risk of tumor growth
epigenetic mechanisms shape the phenotype of a cell within middle-aged patients without children.33 A clear positive
out changing the genotype. Although causality has not been association between breast cancer and higher birth weight,
fully established, epigenetics provides information on the birth length, and placental weight was reported. Further
possible in utero that may cause predisposition to diseases in epidemiologic studies have confirmed these findings. For
adult life. During development, epigenetic mechanisms may females who weighed 3500–3999 g at birth, the adjusted odds
undergo substantial changes caused by various factors. These ratio for breast cancer is 0.86. In comparison, in females
changes affect genes that are essential for both the develop- with birth weight less than 2500 g, the adjusted odds ratio
ment early in life as well as later physiological functions in is reduced to 0.5.34,35 Another example (as found in famine
adult life. An important fact is that epigenetic modifications situations) is that certain changes in some specific metabolic
are stable during cell division and can be transmitted to the set points can only be established when some types of diet
next generation. There is increasing evidence suggesting that are predominantly used (this could be the reason that fast
exposure to nutritional imbalances or environmental con- food, modern lifestyle, and certain chemical exposures are
taminants—including metals, pesticides, persistent organic the perfect “recipe” for obesity).
pollutants, and chemicals in drinking water, such as triethyl- The essential concept of “gestational programming”
tin, chloroform, and trihalomethanes—can cause changes in means that the nutritional, hormonal, and metabolic envi-
epigenetic mechanisms, and this could provide explanation ronment provided by the mother to the fetus could per-
for their effects on adult health.30,31 manently alter the structure and the cellular responses of
Epigenetics, in this sense, can also be understood as a various organs and alter the expression of certain genes
mechanism by which organisms can survive in an unpredict- that ultimately affect the metabolism and physiology of
able and changing environment. The fastest adaptation to the the offspring. Moreover, these effects vary according to the
environment comes from the immediate demand to maintain period of development, and as such, fetuses and newborns
homeostasis—the processes by which the body maintains a (rapidly growing) are the most vulnerable.
constant internal environment in response to external changes, The effects on programming can be immediate, for
often increasing energy expenditure. These processes are gov- example, the deterioration of organ growth at a critical stage,
erned mainly by hormonal and neural signals. The other end while other effects are delayed and not apparent until later.
of the spectrum of evolutionary change is adaptive, i.e., long- In this case, the question is how the memory of the early
term adaptations of species that are driven by genetic change events is stored and how it is expressed later, despite the con-
and therefore resulting in survival or demise. tinuous cell replication and replacement. This again could be
Epigenetics can insert influence by turning on and off the mediated through an epigenetic control of gene expression,
gene transcription. Transcription is the process by which which involves modifying the genome without altering the
an RNA copy of a gene is made, and this RNA copy then proper sequence of DNA.36
In addition, there is currently some debate on whether fetal epigenetic programming, is probably the most effective
certain nutrients can also modify the programming of the strategy to control the expansion of this great pandemic of
immune system. 37 Certain dietary patterns increase the NCDs we mentioned earlier.
risk of altering the immune mechanisms along with meta- The International Federation of Gynecology and
bolic dysregulation and increase the risk of a wide range Obstetrics (FIGO) is a professional organization dedicated
of NCDs. Besides changes in the nutrient profile and their to the improvement of women’s health and rights and
caloric load, diet-induced changes in the intestinal micro- to the reduction of disparities in healthcare available to
flora may also be involved in the pathogenesis and may women and newborns, as well as to advancing the science
increase susceptibility to many chronic diseases. In this and practice of obstetrics and gynecology. The organization
context, it is highly relevant that allergic diseases are one of shall pursue its mission through advocacy, programmatic
the first NCDs that may appear. Preventive strategies can, activities, capacity strengthening of member associations,
in this sense, enhance immune and metabolic health. education, and training. The FIGO has a vision that women
It must be acknowledged that, to date, most of the of the world achieve the highest possible standards of
research on epigenetic mechanisms induced by nutrients physical, mental, reproductive, and sexual health and well-
has focused on the identification of single specific genes being throughout their lives. The values of the organization
that somehow limit the complete picture of nutritional are those of innovative leadership, integrity, transparency,
effects on the entire epigenetic landscape. Using a genome- professionalism, respect for cultural diversity, and high
wide approach has the advantage and potential ability to scientific and ethical standards.
discover the targets of certain novel epigenetic mecha- Traditionally, the FIGO’s focus is on high maternal
nisms, which are sensitive to a specific diet modification. mortality ratio (MMR) in low-resource countries and
As observed in other studies of epigenomic associa- measures to reduce it. Recently, there has been a reduction
tions, epigenetic mechanisms of allergic diseases consti- in maternal mortality to 350,000 maternal deaths annu-
tute a major challenge to be studied in greater depth, i.e., ally, which represents a 34% decline. However, reduction
whether epigenetic variation is the cause or consequence of maternal mortality in sub-Saharan Africa is far below
of the disease. However, clarification of the potential link optimum for achieving MDG 5 target A: to reduce the
between epigenetic changes in allergic diseases could offer MMR by three quarters. The Countdown to 2015 initiative
new opportunities for diagnosis and/or therapy. (www.countdown2015mnch.org) has shown that only
Also, our biggest challenge is to identify the stages in the 5 of the 68 countries are on track to achieve this target.
life cycle where the epigenetic machinery is more sensitive to The leading causes of maternal mortality are preventable
a particular dietary factor in relation to subsequent health and include postpartum hemorrhage (PPH), preeclamp-
outcomes, so that prevention strategies based on changes sia, obstructed labor, infections, and other causes such
in the diet could produce maximum benefit. It is important as undernutrition, anemia, and unsafe abortion. NCDs
to note that multiple and complex multisystem interactions such as diabetes, obesity, undernutrition, and anemia are
require the implementation of an interdisciplinary approach responsible for maternal conditions that have an important
to overcome the growing burden of NCDs. impact on maternal mortality. 39 Diabetes (pregestational
and gestational) can cause macrosomia, obstructed labor,
PPH, and neonatal mortality (prematurity, respiratory
distress syndrome, hypoglycemia, etc.). Ensuring optimal
NCDs, maternal morbidity, and health during pregnancy and in the early childhood years
MDG 5: FIGO’s role not only provides the best chance for a healthy start but
also reduces suffering and the cost to society of chronic dis-
All this current knowledge allows us to conclude that the eases over decades of life. In this sense, the preconceptional
perinatal period is a golden opportunity for preventive state can be crucial. Many parental effects on the develop-
measures aimed at reducing the impact of the epigenetic ing offspring occur even before pregnancy.
preconditioning of NCDs and thereby reducing the likeli- The FIGO is developing different programs to address
hood of developing the aforementioned diseases when the these concerns. For example, the FIGO Saving Mothers
newborns reach adulthood. However, the critical window and Newborns Initiative aims to increase women’s access
of development that offers some plasticity (or pluripotency) to new, cost-effective, and evidence-based technology for
may vary from organ to organ with differing effects, depend- the reduction of maternal and newborn mortality in 10
ing on when the insult occurs and the stage of development low-resource countries in Asia, Africa, Latin America,
of a particular organ.38 Focusing on the preventive care of and Eastern Europe. The FIGO works with communi-
pregnant women has the potential to modify the epigenetic ties and promotes utilization of interventions to reduce
environment of the fetus. Prevention or the optimal treat- maternal and newborn morbidity and mortality. Another
ment of obesity, diabetes, and chronic hypertension in future example is the FIGO Adolescent Sexual and Reproductive
mothers interrupts the vicious cycle of epigenetic program- Health (ASRH) Initiative, which focuses on strength-
ming of the fetus not only in the current pregnancy but also ening the capacity of FIGO member associations at the
in future pregnancies. Hence, the effort to provide adequate national level and promotes access to information and
prenatal care, preventing or treating conditions affecting quality services in ASRH. The Leadership in Obstetrics
Conclusion 223
and Gynecology for Impact and Change in Maternal and In this sense, there are three elements of great impact:
Newborn Health Initiative is another program that has (1) The detection and treatment of anemia is one element
an impact on the diagnosis and treatment of NCDs. The (note that in certain areas the prevalence in anemia in the
FIGO received a substantial grant from the Bill & Melinda pregnant women reaches 80%, resulting in maternal mor-
Gates Foundation for capacity building of member asso- tality from PPH, as well as fetal morbidity from IUGR and
ciations in eight low-resource countries in Asia and prematurity, not to mention the long-term effect on the
Africa. The project will enable member associations to possible emergence of NCDs). (2) The detection and treat-
play a leadership role and influence policy and practices ment of diabetes is another element (in certain countries,
in maternal and newborn health. Improving maternal and it is prevalent in 25% of pregnant women, with its impact
newborn health in low-resource countries by strengthen- on fetal growth as well as the increase in maternal mortal-
ing the role of national obstetric and gynecologic associa- ity by obstructed childbirth). Recall the increased risk for
tions will help to achieve MDGs 4 and 5 to reduce child obesity, diabetes, and metabolic syndrome that these chil-
and maternal mortality and morbidity. dren present, if predisposing factors are not controlled in the
The close link between child and maternal health and early postnatal life. (3) Maternal nutrition, as already cited,
the importance of early-life origin of NCDs requires that is an essential element of prevention, since a change in the
preventive and healthcare interventions related to such availability of nutrients causes a hormonal adjustment in the
diseases are integrated into reproductive, maternal, and fetus to reset its metabolism, so that the newborn is prepared
child health programs, especially at the primary health- for further malnutrition. However, if not properly fed during
care level.40 Many of these preventive strategies are effec- the postnatal age, this can lead to metabolic alterations that
tive for reducing both maternal and infant mortality and make him or her susceptible to certain diseases in adulthood.
later chronic diseases. In addition, it will also be impor- It has been demonstrated that low birth weight, followed by
tant to demonstrate that cost-effective interventions are an accelerated growth during childhood, is a risk factor for
possible. It may be helpful to emphasize the benefits of an cardiovascular disease and/or type 2 diabetes mellitus.
alliance bringing together new advocates to each cause. In
general, the NCD field is not addressing maternal health,
and the maternal health field has not established a place Conclusion
for itself in the push for greater attention paid to NCDs.
We have a unique opportunity for a change and we should The knowledge that certain processes in utero and in early
not miss it. Getting attention to the subject of DOHaD on childhood can affect the risk of developing NCDs provides
the global scene requires buy-in from a global voice. The an opportunity to enforce interventions during this critical
FIGO is such a voice.41 time, when they have the greatest effect. Using appropriate
Professional training is accompanied by an improvement protocols, healthcare providers can educate mothers about
in performance indicators. There are many areas, especially the risks of certain nutritional and environmental exposures
in low-resource countries, in which training different lev- and integrate health promotion on the agenda, as part of the
els of women’s healthcare professionals can be improved, social and economic development, and all this could moti-
so that better outcomes can be achieved, especially with vate a substantial reduction in the risk of NCDs.42
respect to maternal and neonatal morbidity and mortality. At the high-level meeting of the UN on the prevention
Strengthening communication with and among member and control of NDCs held in New York, United States, in
associations and building the capabilities and capacity of September 2011, a four-by-four strategy was proposed.
those from low-resource countries through strengthening Priority efforts to prevent NCDs (diabetes, cardiovascular
leadership, management, good practice, and the promotion disease, cancer, and chronic obstructive pulmonary disease)
of policy dialogues will enable societies to play a pivotal role mainly focus on four risk factors in adults: poor diet, physi-
in the development and implementation of projects and poli- cal inactivity, smoking, and alcohol consumption. Although
cies aimed at the improvement of care available to women paragraphs 26 and 28 of the Political Declaration of the
and their babies. United Nations relate to certain aspects of prenatal nutri-
The FIGO, aware of that responsibility and joining efforts tion, maternal diseases, and air pollution in households,
with other agencies and companies, is making a big effort so these paragraphs describe only partially the full scope of
that prenatal care programs are appropriate for prevention the problem and the opportunities for intervention.43,44 This
and treatment of the most common preconditioning factors, global health challenge has left no country untouched and
especially in low-income countries. The detection and treat- bridges the division between rich and poor countries. We are
ment of anemia, gestational diabetes, nutritional deficiencies, all affected by the rising prevalence of chronic NCDs, wher-
decreased alcohol and tobacco dependence, consumption of ever one takes a closer look.45
addictive substances, etc. are measures of high therapeutic The concept that the foundation for lifelong risk and
value and not only could decrease the immediate adverse susceptibility to numerous diseases begins in the womb
effects on pregnancy (prematurity, macrosomia, PPH, intra- and in early life—now referred to by the term develop-
uterine growth retardation [IUGR], etc.) but also have effects mental origins of health and disease—is not entirely new,
in the medium and long term, with the decline of the NCDs but has only recently gained acceptance from the broader
and their corresponding impact on population health. scientific community because of a swathe of good research
producing hard evidence of the link. The biological evi- the DOHaD link presents another compelling reason to
dence is there, but the knowledge has remained confined do so with more vigor. Evidence suggests that much more
to the academic environment and needs wider dissemi- attention is needed on early life interventions to optimize
nation. To move from evidence to policy change requires nutrition and reduce exposure to toxic substances in order
a broad evidence based on proven solutions, including to reduce the increasing prevalence of NCDs.
assessments of how much it costs, over how long a period Current and future health challenges demand new
of time, and what health benefits are expected. But policy and changing competencies to form the basis for edu-
is never dictated by hard evidence alone. It requires advo- cation, training development, and workforce planning.
cacy, persistence, and resourceful arguments. This is par- International developments in health promotion and evi-
ticularly important in the current global economic setting dence-based practice provide the context for developing
where there is fierce competition for limited resources, health promotion competencies, standards, quality assur-
and the focus on maximizing health impact for the money ance, and accountability in professional preparation and
spent is even stronger. Economic arguments make focus- practice. In addition to filling the training and develop-
ing on the link between maternal health and future health ment gap, there is a need to develop a comprehensive sys-
burden even more relevant and attractive. We know that tem for competency-based standards and accreditation to
provision of good services for maternal and child health is strengthen global capacity in health promotion, which is a
needed to stimulate development and reduce high rates of critical element in achieving the goals for the improvement
maternal and child morbidity and mortality; addressing of global health.
REFERENCES
1. WHO. Global status report on noncommunicable diseases 2010, 16. Cutfield WS, Hofman PL, Vickers M, Breier B, Blum WF,
World Health Organization: Geneva, Switzerland, 2011. Robinson EM. IGFs and binding proteins in short children with
2. Blas E, Kurup AS. Equity, Social Determinants and Public Health intrauterine growth retardation. J Clin Endocrinol Metab 2002;
Programmes, World Health Organization: Geneva, Switzerland, 87: 235–239.
2010. 17. Hofman PL, Cutfield WS, Robinson EM. Insulin resistance in short
3. WHO. Closing the gap in generation health equality through children with intrauterine growth retardation. J Clin Endocrinol
action on the social determinants of health. Commission on Metab 1997; 82: 402–406.
Social Determinants of Health Final Report, World Health 18. Leon DA, Lithell H, Vagero D, McKeigue P, Koupilova I, Mohsen R.
Organization: Geneva, Switzerland, 2008. Reduced fetal growth rate and increased risk of death from isch-
4. Stuckler D. Population causes and consequences of leading aemic heart disease: Cohort study of 15000 Swedish men and
chronic diseases: A comparative analysis of prevailing explana- women born 1915–29. Br Med J 1998; 317: 241–245.
tions. Milbank Q 2008; 86: 273–326. 19. Fall CH, Stein CE, Kumaran K, Cox V, Osmond C, Barker DJ. Size
5. Fuster V, Kellz BB eds., Promoting Cardiovascular Health in at birth, maternal weight, and type 2 diabetes in South India.
the Developing World, Institute of Medicine: Washington, DC, Diabet Med 1998; 15: 220–227.
2010. 20. Bertram CE, Hanson MA. Animal models and programming of
6. Abegunde D, Stanicole A. An estimation of the economic the metabolic syndrome. Br Med Bull 2001; 60: 103–121.
impact of chronic noncommunicable diseases in selected coun- 21. The International Society for Developmental Origins of Health
tries (Working Paper), World Health Organization: Geneva, and Disease. Available from: http://www.dohadsoc.org
Switzerland, 2006. 22. Boney CM, Verma A, Tucker R, Vohr BR. Metabolic syndrome in
7. WHO. Preventing chronic diseases: A vital investment, World childhood: Association with birth weight, maternal obesity, and
Health Organization: Geneva, Switzerland, 2005. gestational diabetes mellitus. Pediatrics 2005; 115: e290–e296.
8. Bloom DE, Cafiero ET, Jané-Llopis E et al. The global eco- 23. Fernandez-Twinn DS, Ozanne SE. Mechanisms by which poor
nomic burden of noncommunicable diseases. World Economic early growth programs type-2 diabetes, obesity and the meta-
Forum, Geneva, Switzerland, 2011. Available from: http:// bolic syndrome. Physiol Behav 2006; 88: 234–243.
www.weforum.org /repor ts/global-economicburden-non- 24. Barker DJ. The developmental origins of insulin resistance. Horm
communicable-diseases. Res 2005; 64(Suppl. 3): 2–7.
9. Lim SS, Vos T, Flaxman AD et al. A comparative risk assessment 25. Barouki R, Gluckman PD, Grandjean P, Hanson M, Heindel JJ.
of burden of disease and injury attributable to 67 risk factors and Developmental origins of non-communicable disease:
risk factor clusters in 21 regions, 1990–2010: A systematic analy- Implications for research and public health. Environ Health June
sis for the Global Burden of Disease Study 2010. Lancet 2012; 27, 2012; 11: 42.
380(9859): 2224–2260. 26. de Rooij SR, Painter RC, Holleman F, Bossuyt PMM, Roseboom TJ.
10. Martínez de Villarreal L. Programación fetal de enfermedades The metabolic syndrome in adults prenatally exposed to the Dutch
expresadas en la etapa adulta. Med Univer 2008; 10: 108–113. famine1,2,3. Am J Clin Nutr October 2007; 86(4): 1219–1224.
11. Barker DJP, Osmond C, Winter PD, Margetts B, Simmonds SJ. 27. Li Y, Jaddoe VW, Qi L, He Y, Wang D, Lai J, Zhang J, Fu P, Yang X,
Weight in infancy and death from ischemic heart disease. Lancet Hu FB. Exposure to the Chinese famine in early life and the risk
1989; 2: 577–580. of metabolic syndrome in adulthood. Diabetes Care April 2011;
12. Vieau D. Perinatal nutritional programming of health and meta- 34(4): 1014–1018.
bolic adult disease. World J Diabetes 2011; 2: 116–133. 28. Barth TK, Imhof A. Fast signals and slow marks: The dynamics of
13. Bloomfield FH. Epigenetic modifications may play a role in the histone modifications. Trends Biochem Sci 2010; 35(11): 618–626.
developmental consequences of early life events. J Neurodev 29. Talens RP, Boomsma DI, Tobi EW, Kremer D, Jukema JW,
Disord 2011; 3: 348–355. Willemsen G, Putter H, Slagboom PE, Heijmans BT. Variation,
14. Barker DPJ. Fetal Programming: Influences on Development and patterns, and temporal stability of DNA methylation:
Disease in Later Life, NIH Monograph Series, Marcel Dekker: Considerations for epigenetic epidemiology. FASEB J 2010; 24(9):
New York, 2000. 3135–3144.
15. Koletzko B, Dodds P, Akerblom H, Ashwell M. Early Nutrition 30. Hou L, Zang X, Wang D, Baccarelli A. Environmental chemical
and Its Later Consequences: New Opportunities, Perinatal exposures and human epigenetics. Int J Epidemiol 2012; 41: 79–105.
Programming of Adult Health—EC Supported Research, Springer 31. Perera F, Herbstman J. Prenatal environmental exposures, epi-
Science + Business Media, Inc: New York, 2005. genetics, and disease. Reprod Toxicol 2011; 31: 363–373.
References 225
32. Chuang JC, Jones PA. Epigenetics and microRNAs. Pediatr Res 39. Keeling A. Halting the rise of non-communicable diseases: An
2007; 61: 24R. urgent priority for women’s health. Diabetes Res Clin Pract 2011;
33. Ahlgren M, Melbye M, Wohlfahrt J et al. Growth patterns and the 92(1): 143–144.
risk of breast cancer in women. N Engl J Med 2004; 351: 1619. 40. World Health Organization. 2008–2013 Action plan for the global
34. Michels KB, Trichopoulos D, Robins JM et al. Birthweight as a risk strategy for the prevention and control of non-communicable
factor for breast cancer. Lancet 1996; 348: 1542. diseases. WHO: Geneva, Switzerland, 2008. Available from:
35. dos Santos Silva I, De Stavola BL, Hardy RJ, Kuh DJ, http://www.who.int/nmh/publications/9789241597418/en/index.
McCormack VA, Wadsworth ME. Is the association of birth html
weight with premenopausal breast cáncer risk mediated through 41. Serour GI, Cabero Roura L. FIGO—A professional nonprofit
childhood growth? Br J Cancer 2004; 91: 519–524. organization: Reproductive, maternal, and child health policy
36. Ross MG, Disai M. Developmental origins of adult health and and programs to address noncommunicable childhood disease.
disease. In: Gabbe SG, Niebyl JR, Simpson JL, Landon MB, Int J Gynecol Obstet 2011; 115(Suppl. 1): S48–S49.
Galan HL, Jauniaux ERM, Driscoll DA, eds., Normal and Problem 42. Balbus JM, Barouki R, Bimbaum LS et al. Early life prevention of
Pregnancies, 6th ed., Saunders: Philadelphia, PA, 2012, pp. 83–98. non-communicable diseases. Lancet 2013; 381: 3–4.
37. Azad MB, Coneys JG, Kozyrskyj AN, Field CJ, Ramsey CD, 43. United Nations. Millennium development goals report
Friesen AC, Abou-Setta AM, Zarychanski R. Probiotic supple- 2011, 2011. Available from: http://www.unhcr.org/refworld/
mentation during pregnancy or infancy for the prevention of docid/4e42118b2.html.
asthma and wheeze: Systematic review and meta-analysis. 44. United Nations. NCDs and MDGs—Success in synergy,
Br Med J 2013; 347: f6471. 2011. Available from: http://www. who.int/nmh/events/
38. Godfrey KM, Lillycrop KA, Burdge GC et al. Epigenetic mecha- un_ncd_summit2011/ncd_mdg.pdf.
nisms and the mismatch concept of the developmental origins 45. Roura LC, Hod M, Kapur A, Lohse N. From biology to policy:
of the developmental origins of health and disease. Pediatr Res The link between maternal health and current and future bur-
2007; 61: 5R. den of chronic noncommunicable disease. Int J Gynaecol Obstet
November 2011; 115(Suppl. 1): S1–S2.
27 Links between maternal health
and noncommunicable diseases
Anil Kapur
intervening? Are there synergies between maternal health

Introduction and NCD prevention that can be leveraged to the advantage
The enormity of the health and economic challenges of both?
from noncommunicable diseases (NCDs) is now being
understood as evidenced by the adoption of the Political
Declaration on NCDs at the high-level meeting of the UN NCDs impact maternal health
General Assembly in September 2011.1 NCDs account for
Adaptation of the Millennium Development Goals in 2000
36 million deaths annually (63% of global deaths), which
brought about justifiable increased attention on mater-
is expected to rise to 52 million by 2030. Four out of five
nal and child health (MCH) programs, especially in the
(79%) of these deaths occur prematurely in low- or middle-
developing world, given the pathetic state of MCH and its
income countries. 2 In 2008, more than 1.4 billion adults
impact on human development. To best utilize resources,
aged ≥20 years (35%) were overweight; of these, 11% were
MCH programs have only focused on factors that directly
obese (200 million men and nearly 300 million women). 3
impact maternal, neonatal, and infant mortality, result-
Since 1980, the prevalence of obesity has nearly doubled
ing in improved access to maternity services and survival
worldwide. Over a billion people live with high blood
of “at-risk” mothers and their offsprings in many low- and
pressure. In 2008, the global prevalence of hypertension
middle-income countries. Unfortunately, this narrow short-
in adults aged ≥25 years was around 40%.4 Approximately
term biomedical focus has failed to address the root causes
700 million people have dysglycemia (diabetes mellitus
and social determinants, and the very individuals saved
and impaired glucose tolerance [IGT]), which is projected
continue to be vulnerable and are at highest risk of poor
to cross the billion mark by 2035. 5 The World Economic
health and lower longevity due to NCDs striking in early
Forum has, for 2 years in a row, rated chronic diseases
adult life. In addition, NCDs, particularly diabetes, obe-
among the five top threats to the global economy includ-
sity, and hypertension, have significant adverse impacts
ing in the low- and middle-income countries.6 According
on maternal health and pregnancy outcomes, and through
to the World Health Organization’s (WHO) report on
intrauterine programming, the cycle of vulnerability to
Women and Health, high blood pressure, high blood glu-
NCDs is repeated with increasing risk accumulation in sub-
cose, and overweight and obesity are the leading risk fac-
sequent generations, as will be explained in greater detail
tors of death from chronic conditions in women >20 years
later. To improve both the short-term MCH outcomes and
of age, accounting for 25%, 39%, and 35% of deaths in
the long-term population health, NCDs must be addressed
low-income, middle-income, and high-income countries,
simultaneously alongside MCH.
respectively.7
Undernutrition, overweight and obesity, hypertension,
Based on well-designed studies with interventions tar-
and hyperglycemia are commonly associated with preg-
geting adults at high risk, it is suggested that up to 80% of
nancy; apart from poor pregnancy outcomes and caus-
the NCD burden can be prevented by addressing the com-
ing considerable maternal morbidity and mortality, these
mon risk factors of tobacco use and unhealthy diet including
result in further escalating the NCD burden through fetal
excessive use of alcohol and physical inactivity—a strategy
programming.
proven in small initiatives but fraught with implementa-
tion difficulties at the population-wide level.8 When chal-
lenged with a problem of this magnitude, there is a need to Maternal nutrition
look at innovative ways to address it. It is now well known
that risk exposure to NCDs begins early—as early as the Worldwide, almost 870 million people suffer from chronic
intrauterine life—and accumulates over many years. Does undernourishment9; 60% of these are girls or women.10
pregnancy and early life offer a window of opportunity for In most developing countries, maternal undernutrition is
226
Overweight and obesity 227
endemic, contributing significantly to maternal morbidity, lower risk of preterm births, stillbirths, neonatal deaths,
mortality, and poor birth outcomes including low birth LBWs, predelivery anemia in the mother, or low predeliv-
weight (LBW), neonatal mortality, and subsequent child- ery red cell folate, although predelivery serum folate levels
hood malnutrition. Annually, 13 million children are born were improved. The review also did not show any impact
with intrauterine growth retardation (IUGR), 112 million are of folate supplementation on improving mean birth weight
underweight, and 178 million children ≤5 years suffer from and mothers’ mean Hb levels during pregnancy compared
stunting. Combined together, severe wasting, stunting, and with placebo treatment. The evidence from the review did
IUGR-LBW are responsible for 2.1 million deaths and 91.0 not show any overall benefit of folic acid supplementation
million disability-adjusted life years.11 The effect of undernu- throughout pregnancy. Most of the studies included in the
trition during pregnancy goes beyond one or two generations review were old (conducted over 30–45 years ago).
because of fetal programming and has tremendous public Vitamin B12 deficiency in women has been shown to
health and economic consequences. Balanced protein–energy increase the risk of infertility or recurrent spontaneous
supplementation in undernourished mothers during preg- abortions. Starting pregnancy with inadequate B12 level
nancy results in 34% and 38% risk reduction for small-for- may increase the risk of birth defects such as NTD and may
gestational-age (SGA) babies and stillbirths, respectively.12 contribute to preterm delivery, although this needs further
Anemia, defined as hemoglobin (Hb) concentration evaluation.29 B12 deficiency in pregnancy is associated with
<110 g/L, affects more than 56 million pregnant women higher insulin resistance and higher incidence of GDM, as
globally; two-thirds of these are from Asia.13 Nutritional well as higher prevalence of type 2 diabetes (T2DM) at 5 years.
iron-deficiency anemia is the most common reason and Among B12-deficient women, the incidence of GDM increases
results in increased maternal and perinatal morbidity and with rising folate concentration.30 Low levels of circulating
mortality and long-term adverse effects on newborns.14 B12 in mothers who are folate replete has been shown to be
These include significantly higher risk of LBW, stillbirth, associated with “thin fat” offsprings with high prevalence of
and p reterm birth.15,16 Iron supplementation lowers the insulin resistance, suggesting a future risk for T2DM.31
incidence of LBW (RR [relative risk] 0.80) but has no effect
on the incidence of preterm or SGA birth.17
Severe anemia is associated with higher risk of preeclamp- Overweight and obesity
sia compared to women with no anemia.16,18 Based on the data
from the WHO Global Survey on Maternal and Perinatal Complications of overweight and obesity during pregnancy
Health, Zhang et al.18 concluded that both nulliparous and include hypertensive disorders, coagulopathies, GDM,
multiparous women with severe anemia had significant associ- respiratory problems, and fetal complications such as large-
ation with preeclampsia/eclampsia (aOR [adjusted odds ratio] for-gestational-age (LGA) babies, congenital malformations,
3.55 [95% CI [confidence interval] 2.87, 4.41] and aOR 3.94 stillbirth, and shoulder dystocia. Women being overweight
[95% CI 3.05, 5.09], respectively), whereas only multiparous in early pregnancy have a twofold to threefold increased risk
women with severe anemia were at increased risk of gesta- for preeclampsia.32 Obesity is associated with increased risk
tional hypertension (aOR 1.58 [95% CI 1.15, 2.19]). of preeclampsia (aOR 4.46), induction of labor (aOR 1.97),
On the other hand, high iron intake in pregnancy postpartum hemorrhage (aOR 3.04), intensive care admis-
increases the risk of gestational diabetes mellitus (GDM) sion (aOR 3.86), GDM (aOR 7.89), thrombosis (aOR infinity),
especially in nonanemic women, and routine iron supple- shoulder dystocia (aOR 1.89), C-section (aOR 3.50),33 mater-
mentation should be reconsidered in this group of women.19 nal infection (aOR 3.35), prolonged hospital stay (aOR 2.84),
Higher prepregnancy intake of dietary heme iron20,21 and instrumental delivery (aOR 1.17).34
and raised serum ferritin level22–24 are associated with an Maternal overweight and obesity (BMI >25 kg/m²) is the
increased risk of GDM. most important modifiable risk factor for stillbirths in high-
Studies from around the world show high rates of vitamin income countries, contributing to around 8000 stillbirths
D deficiency among women of reproductive age and during (≥22 weeks gestation) annually.35 In developing countries,
pregnancy.25 A systematic review in 2010 of first-trimester it is associated with a twofold to threefold increased risk of
25(OH) vitamin D level and adverse pregnancy outcomes macrosomia, requiring institutional and assisted delivery,
concluded that the evidence of the association between the lack of which results in significantly increased maternal
vitamin D levels and pregnancy complications such as pre- morbidity and mortality.36 A meta-analysis and systematic
eclampsia and diabetes is inconclusive.26 A recent meta- review of cohort studies of maternal BMI and risk of fetal
analysis and systematic review including some new studies death, stillbirth, or infant death concluded that even the
concluded that vitamin D insufficiency is associated with modest increases in maternal BMI lead to significantly
an increased risk of GDM, preeclampsia, and SGA and increased risk.37 The number of women in their reproductive
LBW infants.27 age who are overweight now exceeds the number of under-
Folic acid supplementation (450 µg) before conception weight women,38 and if this trend continues unchecked, it
and throughout the first 12 weeks of pregnancy reduces may again reverse the recent nebulous gains in improved
the risk of neural tube defects (NTDs) in the offsprings. maternal health outcomes. Focusing on quality nutrition,
A Cochrane review in 201328 concluded that folate sup- physical activity, and general health of women especially
plementation during pregnancy was not associated with during the reproductive years is an important public health
228 Links between maternal health and noncommunicable diseases
investment. Raising public awareness on the hazard of unexplained stillbirth, and spontaneous abortion in previ-
overweight and obesity, particularly among low- and middle- ous pregnancies; excessive weight gain; and the presence of
income communities undergoing rapid economic transition, polycystic ovary syndrome, metabolic syndrome, polyhy-
is the need of the hour in relation to NCD risk and from a dramnios, and suspected macrosomia during current preg-
maternal health risk perspective. nancy have been described to clinically identify women with
high risk of GDM.46 In practice, they fail to correctly identify
more than half the women with GDM47–50; thus universal
Hyperglycemia screening for hyperglycemia during pregnancy must be the
standard practice.
According to the International Diabetes Federation, there Hemorrhage, hypertensive disorders, obstructed labor,
are now an estimated 382 million people (184 million and infection/sepsis are among the leading global causes
women) with diabetes; in addition, about 316 million peo- of maternal mortality.51 High blood pressure and gesta-
ple have impaired glucose tolerance (IGT).5 By 2035, this tional hyperglycemia are linked directly or indirectly to all
number is likely to grow to over 592 million with diabetes of them; yet women are not routinely screened for hypergly-
and 471 million with IGT. The Asia Pacific region accounts cemia during pregnancy, and the diagnosis of GDM is often
for about half the global burden; China, India, Indonesia, missed. Maternal mortality and morbidity attributable to
Pakistan, and Bangladesh account for 185 million people diabetes in women may, therefore, actually be higher than
with diabetes and figure among the top 10 countries with current estimates.
the highest number of people with diabetes.5 These very five Diabetes in pregnancy is associated with serious com-
countries also account for over half of the global live births plications for both the mother and child. It has been shown
(66 million/year). that the negative consequences on the fetus and the mother
Worldwide, one in six pregnancies may be associated increase linearly with increasing maternal blood glucose.52
with hyperglycemia, 84% of which involve GDM.5 In 2013, Infants of mothers with pregestational diabetes have higher
16.8% of live births (21.4 of 127 million) were associated rates of malformation53–55; good blood glucose control before
with hyperglycemia in pregnancy and 16% of these were conception and throughout pregnancy reduces these risks
due to overt diabetes in pregnancy. This does not account substantially.56,57 The major problems related to hyper
for pregnancies ending in spontaneous abortions, still- glycemia during pregnancy are shown in Table 27.1.
births, or intrauterine deaths that may have been associ- A meta-analysis58 shows that, overall, women with GDM
ated with hyperglycemia, proven or otherwise. In high-risk have an increased risk of developing T2DM (RR 7.43, [95%
groups, up to 30% of pregnancies may involve diabetes.39,40 CI 4.79, 11.51]). Within 5 years of the index pregnancy, the
The age-adjusted prevalence of GDM in the United States is relative risk is 4.69, which more than doubles to 9.34, 5 years
higher than those of Asian or Pacific Island–origin women postpartum. The risk can be reduced or the onset of diabetes
but more so (almost threefold compared to non-Hispanic considerably delayed through preventive actions in terms of
whites) for migrant women born in the country of their ori- postpartum weight loss and adopting a healthy lifestyle.59
gin.41 In Southeast Asia, one in four live births may occur in Women with history of GDM also have higher prevalence
the setting of maternal hyperglycemia.5 A fact generally not of the metabolic syndrome and increased risk of cardiovas-
known to or recognized by public health experts and policy cular disease (CVD).60 Over a median follow-up of 12 years,
makers is that 91.6% of cases of hyperglycemia in pregnancy women with GDM had higher risk of CVD (adjusted hazard
occur in low- and middle-income countries.5 With limited ratio 1.66 [95% CI 1.30, 2.13], p < 0.001),61 more noninvasive
access to maternal care, this may have major consequences
for maternal health and a future burden of NCDs.
Increasing age is associated with higher prevalence of Table 27.1 Risks associated with hyperglycemia
hyperglycemia in pregnancy, which is the highest (47.7%) in pregnancy
among women >45 years. In general, the age of onset of dia-
betes is declining; at the same time, the age of marriage and Fetal risks Maternal risks
childbearing is increasing; as a consequence, we may, in the Spontaneous abortion, Polyhydramnios
future, see more women entering pregnancy with preexist- intrauterine death, and
ing diabetes.42,43 Between 1999 and 2005, the prevalence of stillbirth
pregestational diabetes among pregnant women in south- Lethal or handicapping Pregnancy-induced
ern California doubled.44 In 2010, there were an estimated congenital hypertension and
22 million women with diabetes in the reproductive age malformation preeclampsia
group of 20–39 years; an additional 54 million in this age Shoulder dystocia and Prolonged labor, obstructed
birth injuries labor, assisted delivery,
group had IGT with the potential to develop GDM if they and C-section
become pregnant.45 Thus, over 76 million women were at risk Neonatal hypoglycemia Uterine atony and postpartum
of their pregnancies being complicated with pregestational hemorrhage
(overt) diabetes or GDM. Infant respiratory distress Infections
Several markers such as age; race/ethnicity; BMI; his- syndrome Progression of retinopathy
tory of T2DM in first-degree relatives, GDM, macrosomia,
Maternal health impacts future burden of NCDs 229
cardiac diagnostic procedures (OR1.8 [95% CI 1.4–2.2]), sim- Although most cases of preeclampsia can be managed
ple cardiovascular events (OR 2.7 [95% CI 2.4–3.1]), and total successfully in well-resourced settings, severe preeclamp-
cardiovascular hospitalizations (OR 2.3 [95% CI 2.0–2.5]) sia is a life-threatening multisystem disease associated with
over a 10-year follow-up, after adjusting for age, ethnicity, eclampsia, HELLP syndrome (hemolysis, elevated liver
and comorbidities such as preeclampsia and obesity.62 enzymes, low platelets), acute kidney injury, pulmonary
Fetal environment contributes significantly to higher risk edema, placental abruption, and intrauterine fetal death, all
for diabetes than can be explained by genetic inheritance of which are difficult to handle in poor-resourced settings.80
alone. Offsprings of mothers with GDM are four to eight Several risk factors are associated with higher predilec-
times more likely to develop diabetes63,64 compared to sib- tion for preeclampsia; these include nulliparity (RR 2.38;
lings born to the same parents in a non-GDM pregnancy. 95% CI 2.28–2.49), multiple pregnancy (RR 2.10; 95% CI
Children born to obese or diabetic mothers are at higher risk 1.90–2.32), history of chronic hypertension (RR 1.99; 95%
of obesity and/or insulin resistance and T2DM during child- CI 1.78–2.22), GDM (RR 1.93; 95% CI 1.66–2.25), maternal
hood,65 adolescence,66 and early adulthood.64 Almost half age ≥35 years (RR 1.67; 95% CI 1.58–1.77), fetal malforma-
(47.2%) of diabetes and obesity in the youth can be attributed tion (RR 1.26; 95% CI 1.16–1.37), and mother not living with
to maternal GDM and obesity.67 Population attributable risk infant’s father (RR 1.21; 95% CI 1.15–1.26).81 Preeclampsia
for T2DM from GDM in certain populations may be as high risk increases with increasing prepregnancy BMI.81
as 19%–30%.68 Maternal pregravid obesity combined with HPD has long-term consequences for the offspring and
GDM leads to newborn hyperinsulinemia and increased fat the mother. Women with previous HPD have higher glu-
mass until 6 weeks, whereas pregravid obesity alone does cose, insulin, triglyceride, and total cholesterol levels after
not, suggesting the pivotal role of GDM.69 GDM creates a pregnancy.82 Women with HPD are at increased risk of
vicious cycle in which diabetes begets diabetes. cardiovascular and metabolic disorders, including a twofold
In view of the dramatic increases in obesity and diabetes, increased risk for hypertension, a threefold increased risk
we should accept that screening, diagnosing, and treating for T2DM, and a 1.3-fold increased risk for dyslipidemia,83
GDM is worthwhile.70 Skeptics, however, continue to ques- and these women may benefit from close postpartum moni-
tion whether screening women for GDM is cost-effective. toring, timely implementation of lifestyle modifications,
Most of the cost-effectiveness analyses in the past have not and preventive measures for cardiovascular and metabolic
included long-term benefits71 or have been conducted in risk reduction.82,83
populations with a relatively lower burden of GDM; none Offsprings of mothers with preeclampsia have higher
of these were in low-income countries. A few recent studies blood pressure during childhood and young adulthood.84–86
including modeling studies show that GDM screening asso- The mechanism for the higher risk is not clear and may be
ciated with postpartum lifestyle interventions for T2DM genetic and epigenetic—a consequence of vascular or meta-
prevention is cost-effective in both high-income (United bolic programming—and have shared family risks or a com-
States, Israel) and low-income (India) countries.72–74 bination of these.
Hypertension Maternal health impacts future

Worldwide, high blood pressure with or without proteinuria burden of NCDs
is a major cause of maternal morbidity and mortality,75 and
hypertensive pregnancy disorders (HPDs) account for 10% Prenatal and early-life development through epigenetic pro-
and 15% of maternal deaths in low- and middle-income coun- gramming influences the risks of NCD in later life,87–91 and
tries, respectively,76–78 as well as increased perinatal morbid- this might be especially relevant to low-resource countries.91–94
ity and mortality as a consequence of prematurity and poor The parent’s health, particularly the mother’s body com-
fetal growth. Although the incidence varies in different parts position and nutritional and metabolic status during preg-
of the world, overall nearly 10% of previously normotensive nancy, determines fetal environment and affects risk for later
women experience abnormally elevated blood pressure at NCDs.95,96 Ensuring a healthy pregnancy and a disease-free
some point during pregnancy. There is no consensus about early childhood may be the most effective means of attaining
the definition of HPDs and several classifications have been the best future health and preventing NCDs. Fetal environ-
proposed. The generally accepted broad categories are (1) ment represented by the mother’s periconceptional and gesta-
gestational hypertension or pregnancy-induced hyperten- tional health determines whether one starts life with a health
sion (hypertension without proteinuria), (2) preeclampsia “advantage” or “handicap,” and it is on this “foundation” that
(hypertension with proteinuria), (3) chronic hypertension NCD risk factors play out in later life. People starting life with
or essential hypertension (preexisting hypertension), and a “health handicap” may be less able to withstand lifestyle risks
(4) chronic hypertension with superimposed preeclampsia. and may be vulnerable to diseases early, compared to those
Preeclampsia/eclampsia have the highest impact on mortal- starting with a “health advantage.” Similarly, lifestyle inter-
ity and morbidity; these include renal or liver failure, clot- ventions in adult life to prevent diseases may have variable
ting disorders, stroke, preterm delivery, stillbirth or neonatal effects based on early-life programming.97 The impact of life
death,79 and C-section, especially emergency C-section. conditions on health—the social determinants of health—is
high on the global development agenda, and therefore, it is had a prepregnancy BMI of <19. Undiagnosed or poorly man-
relevant to ponder that perhaps these social determinants aged GDM sets off a cycle of future obesity and T2DM in their
get “hard wired” into the next generations’ genome through offsprings, and the cycle may repeat in subsequent genera-
epigenetic changes. The recognition that early-life influences tions with ever-growing risk accumulation until interrupted
play an important role in the causation of chronic diseases through appropriate preventive actions in the pre- and peri-
does not imply an absolute deterministic process that cannot conceptual period, during pregnancy, and in the postpartum
be overcome by later-life intervention; it only helps to point period of the next generation.
that the task becomes more difficult and expensive with lower The concept of fetal programming and its consequences
chance of success when initiated later in life. is paradigm changing; it highlights that pregnancy offers a
Focusing on short-term survival in terms of lowered window of opportunity to provide maternal care services
maternal and perinatal morbidity and mortality to assess not only to reduce the traditionally known maternal and
maternal health programs as has been the routine practice so perinatal morbidity and mortality indicators but also for
far does not capture outcomes that have longer-term impli- intergenerational prevention of several chronic diseases.40
cations for adult health, life expectancy, quality of life, and There are several barriers in achieving these objectives.
accumulation of human capital.97 Moreover, recommenda- These barriers related to GDM, for example, have been
tions for nutritional interventions are frequently based on recently described in a systematic review and are briefly
increasing birth weight, focusing on survival, gains in stat- described.112 Knowledge of and adherence to existing guide-
ure, or micronutrient status in the short term.98 Longer-term lines and procedures for screening and diagnosis seem
follow-up data confirm the existence of only a narrow win- suboptimal at best and arbitrary at worst, with no clear or
dow of opportunity for interventions up to 24 months of age, consistent correlation to the health provider, health system,
and only limited benefit, or even harm, of feeding strategies or client characteristics. Most women express commitment
thereafter.99,100 Small babies continuing to be malnourished and motivation for behavior change to protect the health of
and stunted during childhood and early adult life remain at their unborn babies, but knowledge about how to effectively
relatively low risk for NCDs as long as they have sufficient make changes is missing. Compliance to recommended
sustenance for which they were programmed. With changes treatment and advice is seen as difficult and challenging, and
in living conditions as a consequence of economic develop- precious little is known about health system or societal fac-
ment or urban migration, these individuals manifest dia- tors that hinder compliance and what can be done to improve
betes, hypertension, and other NCDs at much lower BMI them. When properly informed, immediately following a
and central adiposity threshold.101,102 Studies on survivors of GDM pregnancy, many women express desire and inten-
the Dutch103,104 and Chinese105 famine show that individu- tion to adapt healthy lifestyles to prevent future diabetes,
als exposed to intrauterine undernutrition had significantly but find the effort challenging. Adherence to recommended
higher rates of diabetes in adult life, and the risk was high- postpartum screening and continued lifestyle modifications
est in the subgroup that were relatively well-off in adult life. seems even lower. Here, some healthcare providers, health
Similarly, LGA babies born to overweight/obese women with systems, and client-related determinants and barriers have
or without GDM are at high risk of obesity, diabetes, and been identified.112 Studies reveal that a sense of self-efficacy
metabolic syndrome in early adult life.63–69 and social support is important for making and adhering to
Developmental effects operate through a gamut of subtle the changes. Noncompliance to screening or nonacceptance
influences that provide the fetus the cues (via the intrauter- of diagnosis of GDM may be due to poor understanding
ine environment) to predict the external environment it will of the consequences, or fear of stigmatization, and one needs
be born into, as well as the flexibility to adjust its growth to be careful how to address this in public health campaigns to
trajectory to match that environment. These cues, such as avoid creating another platform for women to be blamed for
maternal under- or overnutrition (pregnancy weight gain), adverse effects on their children’s future health.
maternal obesity or GDM,106,107 or other maternal health
insults like stress, or infections (malaria, tuberculosis, HIV/
AIDS, etc.),108,109 create multigenerational cycles of dis- Future direction for action and
ease110 through epigenetic changes. The mismatch between research
the predicted environment for survival programming and
the actual environment in adult life may be a critical factor Having saved a mother with GDM and preeclampsia from
driving the NCD epidemic primarily through its effects on dying of obstructed labor or postpartum hemorrhage and
weight, blood pressure, and blood glucose. her macrosomic baby, or a mother with severe malnutrition
In young women born small themselves, the physiological and anemia and her LBW baby, what can be done to ensure
effects of pregnancy-induced weight gain, insulin resistance, their future good health and to prevent or significantly
and increased insulin demand may be exaggerated by the pre- delay the onset of hypertension or T2DM? What should be
existing insulin resistance and the lower ability to produce done to ensure that a girl child born of such pregnancy is
insulin as a consequence of their early-life programming, given due antenatal attention to prevent further intergen-
resulting in higher rates of GDM and/or pregnancy-induced erational risk transfer? This requires transforming policy
hypertension. Seshiah et al.111 reported GDM prevalence rates to integrate services for MCH, NCD care and prevention,
of 8%–10% among women of low socioeconomic status who and health promotion. It will also require cost-effective
References 231
investments in information technology, to identify and health system response to jointly address the challenge of
track these high-risk individuals to enlighten, empower, improving maternal health and addressing prevention of
and encourage them to adopt healthy living throughout NCDs and developing scalable programs in real-life set-
life as well as empower local health workers to support and tings. In addition, developing tools and incentives to engage,
follow their progress. Enrolling, monitoring, and track- enlighten, empower, and encourage “at-risk” mother-and-
ing women during and after pregnancy and their offspring child pair and study of the impact of these actions over
using information technology may be the most appropriate time on population health are required. Given the size of
place to begin this health system transformation to break the problem, there is also a need to develop point-of-care
the ever-rising curve of chronic NCDs.113 easy-to-use diagnostic and prognostic tests to more accu-
There is a great need to carry out operational research rately identify individuals with greater risk so as to deploy
to understand the facilitators and barriers to an integrated resources appropriately.
REFERENCES
1. United Nations. Prevention and control of non-communicable 19. Helin A, Kinnunen TI, Raitanen J et al. Iron intake, haemoglobin
diseases. Report of the Secretary General. UN General Assembly, and risk of gestational diabetes: A prospective cohort study. BMJ
May 19, 2011. Available from: http://www.un.org/ga/search/ Open 2012; 2: e001730.
view_doc.asp?symbol=A /66/83&referer=/english/&Lang=E, 20. Raj Pathak S, Ma J, Manson J et al. Iron intake and the risk of type
Retrieved on March 2014. 2 diabetes in women: A prospective cohort study. Diabetes Care
2. WHO. Global status report on noncommunicable diseases 2006; 29: 1370–1376.
2010. Available from: www.who.int. http://www.who.int/nmh/ 21. Bowers K, Yeung E, Williams MA et al. A prospective study of pre
publications/ncd_report2010/en/, Retrieved on March 20, 2014. pregnancy dietary iron intake and risk for gestational diabetes
3. Obesity and overweight. Fact sheet No. 311, WHO. Available mellitus. Diabetes Care 2011; 34: 1557–1563.
from: http://www.who.int/mediacentre/factsheets/fs311/en/, 22. Islam N, Chowdhury SB. Serum ferritin and gestational diabetes
Retrieved on March 20, 2014. mellitus: A case control study. Ibrahim Cardiac Med J 2013; 1:
4. Raised blood pressure. Global Health Observatory (GHO). 15–19.
Available from: http://www.who.int/gho/ncd/risk_factors/blood_ 23. Tarim E, Kilicdag E, Bagis T et al. High maternal hemoglobin
pressure_prevalence_text/en/, Retrieved March 20, 2014. and ferritin values as risk factors for gestational diabetes. Int J
5. International Diabetes Federation. IDF Atlas, 6th ed., International Gynaecol Obstet 2004; 84: 259–261.
Diabetes Federation: Brussels, Belgium, 2013. 24. Lao TT, Tam KF. Maternal serum ferritin and gestational impaired
6. Bloom DE, Cafierio ET, Jane-Llopis E et al. The Global Economic glucose tolerance. Diabetes Care 1997; 20: 1368–1369.
Burden of Non Communicable Diseases, World Economic 25. Urrutia RP, Thorp JM. Vitamin D in pregnancy: Current concepts.
Forum: Geneva, Switzerland, 2011. Curr Opin Obstet Gynecol 2012; 24: 57–64.
7. World Health Organization. Women and Health: Today’s Evidence, 26. Nassar N, Halligan GH, Roberts CL et al. Systematic review of
Tomorrow’s Agenda, WHO: Geneva, Switzerland, 2009. first-trimester vitamin D normative levels and outcomes of preg-
8. Hanson MA, Gluckman PD. Developmental origins of noncom- nancy. Am J Obstet Gynecol 2011; 205: 208.e1–208.e7.
municable disease: Population and public health implications. 27. Aghajafari F, Nagulesapillai T, Ronksley PE et al. Association
Am J Clin Nutr 2011; 94: 1754S–1758S. between maternal serum 25-hydroxyvitamin D level and preg-
9. Food and Agriculture Organisation. Available from: http://www. nancy and neonatal outcomes: Systematic review and meta-
fao.org/news/story/en/item/161819/icode, Retrieved on March analysis of observational studies. Br Med J 2013; 346: f1169.
20, 2014. 28. Lassi ZS, Salam RA, Haider BA et al. Folic acid supplementation
10. World Food Programme (WFP). WFP Gender Policy and Strategy: during pregnancy for maternal health and pregnancy outcomes.
Promoting Gender Equality and the Empowerment of Women in Cochrane Database Syst Rev 2013; 3: CD006896.
Addressing Food and Nutrition Challenges, WFP: Rome, Italy, 2009. 29. Molloy AM, Kirke PN, Brody LC et al. Effects of folate and vita-
11. Black RE, Allen LH, Bhutta ZA et al. for the Maternal and Child min B12 deficiencies during pregnancy on fetal, infant, and child
Under nutrition Study Group. Maternal and child under nutrition: development. Food Nutr Bull 2008; 29: S101–S111.
Global and regional exposures and health consequences. Lancet 30. Krishnaveni GV, Hill JC, Veena SR et al. Low plasma vitamin B12
2008; 371: 243–260. in pregnancy is associated with gestational ‘diabesity’ and later
12. Imdad A, Bhutta ZA. Effect of balanced protein energy supple- diabetes. Diabetologia 2009; 52: 2350–2358.
mentation during pregnancy on birth outcomes. BMC Public 31. Yajnik CS, Deshpande SS, Jackson AA et al. Vitamin B12 and
Health 2011; 11: S17. folate concentrations during pregnancy and insulin resistance in
13. de Benoist B, McLean E, Egli I, Cogswell M, eds., Worldwide the offspring: The Pune Maternal Nutrition Study. Diabetologia
Prevalence of Anaemia 1993–2005, World Health Organization: 2008; 51: 29–38.
Geneva, Switzerland, 2008. WHO Global Database on Anaemia. 32. Sibai BM, Gordon T, Thom E et al. Risk factors for preeclampsia
Available from: http://www.who.int/vmnis/publications/anaemia_ in healthy nulliparous women: A prospective multicenter study.
prevalence/en/index.html, Retrieved on March 20, 2014. The National Institute of Child Health and Human Development
14. Goonewardene M, Shehata M, Hamad A. Anaemia in pregnancy. Network of Maternal-Fetal Medicine Units. Am J Obstet Gynecol
Best Pract Res Clin Obstet Gynaecol 2012; 26: 3–24. 1995; 172: 642–648.
15. Haider BA, Olofin I, Wang M et al. Anaemia, prenatal iron use, 33. Heslehurst N, Simpson H, Ells LJ et al. The impact of maternal BMI
and risk of adverse pregnancy outcomes: Systematic review and status on pregnancy outcomes with immediate short-term obstetric
meta-analysis. Br Med J 2013; 346: f3443. resource implications: A meta-analysis. Obes Rev 2008; 9: 635–683.
16. Ali AA, Rayis DA, Abdallah TM et al. Severe anaemia is associ- 34. Knight M, Kurinczuk JJ, Spark P et al. Extreme obesity in pregnancy
ated with a higher risk for preeclampsia and poor perinatal out- in the United Kingdom. Obstet Gynecol 2010; 115: 989–997.
comes in Kassala hospital, eastern Sudan. BMC Res Notes 2011; 35. Flenady V, Koopmans L, Middleton P et al. Major risk factors
4: 311. for stillbirth in high-income countries: A systematic review and
17. Imdad A, Bhutta ZA. Routine iron/folate supplementation dur- meta-analysis. Lancet 2011; 377: 1331–1340.
ing pregnancy: Effect on maternal anaemia and birth outcomes. 36. Koyanagi A, Zhang J, Dagvadorj A et al. Macrosomia in 23 devel-
Paediatr Perinat Epidemiol 2012; 26: 168–177. oping countries: An analysis of a multicountry, facility-based,
18. Zhang JJ, Grewal J, Roosen-Runge M et al. OS029. Severe anae- cross-sectional survey. Lancet 2013; 381: 476–483.
mia, sickle cell disease, and thalassemia as risk factors for pre- 37. Aune D, Saugstad OD, henriksen T, Tonstad S. Maternal body mass
eclampsia in developing countries. Pregnancy Hypertens 2012; index and risk of fetal death, stillbirth and infant death: A systematic
2: 191–192. review and meta-analysis. J Am Med Assoc 2014; 311: 1536–1546.
38. Mendez MA, Monteiro CA, Popkin BM. Overweight exceeds 64. Clausen TD, Mathiesen ER, Hansen T et al. High prevalence of
underweight among women in most developing countries. Am J type 2 diabetes and pre-diabetes in adult offspring of women
Clin Nutr 2005; 81: 714–721. with gestational diabetes mellitus or type 1 diabetes: The role of
39. Jiwani A, Marseille E, Lohse N et al. Gestational diabetes mel- intrauterine hyperglycemia. Diabetes Care 2007; 31: 340–346.
litus: Results from a survey of country prevalence and practices. 65. Catalano PM, Farell K, Thomas A, Houston-Presley L, Mencin
J Matern Fetal Neonat Med 2012; 25: 600–610. P, Mouzon SHD, Amini SB. Perinatal risk factors for childhood
40. World Diabetes Foundation, Global Alliance for Women’s obesity and metabolic dysregulation. Am J Clin Nutr 2009; 90:
Health. Diabetes, women, and development. Meeting, expert 1303–1313.
recommendations for policy action, conclusions, and follow-up 66. Gillman MW, Rifas-Shiman S, Berkey CS, Field AE, Colditz GA.
actions. Int J Gynecol Obstet 2009; 104: S46–S50. Maternal gestational diabetes, birth weight and adolescent obe-
41. Henderson MM, Darbinian JA, Ferara A. Disparities in the risk sity. Pediatrics 2003; 111: e221–e226.
of gestational diabetes by race-ethnicity and country of birth. 67. Dabelea D, Mayer-Davis EJ, Lamichhane AP et al. Association
Paediatr Perinat Epidemiol 2010; 24: 441–448. of intrauterine exposure to maternal diabetes and obesity with
42. World Health Organization. Global Status Report on Non- type 2 diabetes in youth. The SEARCH Case-Control Study.
Communicable Diseases, World Health Organization: Geneva, Diabetes Care 2008; 31: 1422–1426.
Switzerland, 2011. 68. Osgood ND, Dyck RF, Grassmann WK. The inter and intra gen-
43. Matyka KA. Type 2 diabetes in childhood: Epidemiological and erational impact of gestational diabetes on the epidemic of
clinical aspects. Br Med Bull 2008; 86: 59–75. type 2 diabetes. Am J Public Health 2011; 101: 173–179.
44. Lawrence JM, Chen W, Contreras R et al. Trends in the preva- 69. Ubel K, Pusch K, Gedrich K, Schneider KT, Hauner H, Bader BL.
lence of pre-existing diabetes and gestational diabetes melli- Effect of maternal obesity with and without gestational diabe-
tus among a racially/ethnically diverse population of pregnant tes on offspring subcutaneous and preperitoneal adipose tissue
women,1999–2005. Diabetes Care 2008; 31: 899–904. development from birth up to year-1. BMC Pregnancy Childbirth
45. International Diabetes Federation. Available from: http://www. 2014; 14(1): 138.
diabetesatlas.org/, Retrieved on June 10, 2011. 70. Meltzer SJ. Treatment of gestational diabetes—The question is
46. Berger H, Crane J, Farine D et al. Screening for gestational diabe- not whether to treat, but how and who? Br Med J 2010; 340:
tes mellitus. J Obstet Gynaecol Can 2002; 24: 894–912. 770–771.
47. Lavin JP Jr. Screening of high-risk and general populations for ges- 71. Poncet B, Touzet S, Rocher L et al. Cost-effectiveness analysis
tational diabetes: Clinical application and cost analysis. Diabetes of gestational diabetes mellitus screening in France. Eur J Obstet
1985; 34: S24–S27. Gynecol Reprod Biol 2002; 103: 122–129.
48. Coustan DR, Nelson C, Carpenter MW et al. Maternal age and 72. Marseille E, Lohse N, Jiwani A et al. The cost-effectiveness of
screening for gestational diabetes: A population-based study. gestational diabetes screening including prevention of type 2 dia-
Obstet Gynecol 1989; 73: 557–561. betes: Application of a new model in India and Israel. J Matern
49. Moses RG, Moses J, Davis WS. Gestational diabetes: Do lean Fetal Neonatal Med 2013; 26: 802–810.
young Caucasian women need to be tested? Diabetes Care 1998; 73. Li R, Ping Z, Barker LE et al. Cost-effectiveness of interventions
21: 1803–1806. to prevent and control diabetes mellitus: A systematic review.
50. Neilsen KK, De Courten M, Kapur A. Urgent need for a univer- Diabetes Care 2010; 33: 1872–1894.
sally applicable, simple point of care diagnostic test for gesta- 74. Werner EF, Pettker CM, Zuckerwise L et al. Screening for gesta-
tional diabetes mellitus (GDM). Glob Health Action 2012, 5: tional diabetes mellitus: Are the criteria proposed by the inter-
17277. national association of the diabetes and pregnancy study groups
51. Khan KS, Wojdlya D, Say L et al. WHO analysis of causes cost-effective? Diabetes Care 2012; 35: 529–535.
of maternal death: A systematic review. Lancet 2006; 367: 75. Department of Health Welsh Office, Scottish Office Home and
1066–1074. Health Department, Department of Health and Social Security,
52. Metzger BE, Lowe LP, Dyer AR et al. The HAPO Study Northern Ireland. Report on confidential enquiries into maternal
Cooperative Research Group. Hyperglycemia and adverse preg- deaths in the United Kingdom 1988–1990. HMSO: London, U.K.,
nancy outcomes. N Engl J Med 2008; 358: 1991–2002. 1994.
53. Cundy T. Pregnancy loss and neonatal death in women with 76. Moodley J. Maternal deaths due to hypertensive disorders in
type 1 or type 2 diabetes mellitus. Insulin 2008; 3: 167–175. pregnancy: Saving Mothers report 2002–2004. Cardiovasc J Afr
54. Women and Diabetes. Available from: http://www.idf.org/ 2007; 18: 358–361.
women-and-diabetes, Retrieved August 31, 2011. 77. Moodley J. Maternal deaths due to hypertensive disorders in preg-
55. Fetita LS, Sobngwi E, Serradas P et al. Consequences of fetal nancy. Best Pract Res Clin Obstet Gynaecol 2008; 22: 559–567.
exposure to maternal diabetes in offspring. J Clin Endocrinol 78. Chhabra S, Kakani A. Maternal mortality due to eclamptic and
Metab 2006; 91: 18–24. non-eclamptic hypertensive disorders: A challenge. J Obstet
56. Crowther CA, Hiller JE, Moss JR et al. Effect of treatment of ges- Gynaecol 2007; 27: 25–29.
tational diabetes mellitus on pregnancy outcomes. N Engl J Med 79. Druzin ML, Charles B, Johnson AL. Editorial summary of sympo-
2005; 352: 2477–2486. sium on hypertensive disorders of pregnancy. Curr Opin Obstet
57. Landon MB, Spong CY, Thom E et al. A multicenter, random- Gynecol 2008; 20: 91–95.
ized trial of treatment for mild gestational diabetes. N Engl J Med 80. Arulkumaran N, Lightstone L. Severe pre-eclampsia and hyperten-
2009; 361: 1339–1348. sive crises. Best Pract Res Clin Obstet Gynaecol. 2013; 27: 877–884.
58. Bellamy L, Casas JP, Hingorani AD et al. Type 2 diabetes mellitus 81. Conde-Agudelo A, Belizán JM. Risk factors for pre-eclampsia in
after gestational diabetes: A systematic review and meta-analysis. a large cohort of Latin American and Caribbean women. Br J
Lancet 2009; 373: 1773–1779. Obstet Gynaecol 2000; 107: 75–83.
59. Ratner RE, Christophi CA, Metzger BE. Prevention of diabetes in 82. Hermes W, Ket JC, van Pampus MG et al. Biochemical cardio-
women with a history of gestational diabetes: Effects of metfor- vascular risk factors after hypertensive pregnancy disorders:
min and lifestyle interventions. J Clin Endocrinol Metab 2008, 93: A systematic review and meta-analysis. Obstet Gynecol Surv
4774–4779. 2012; 67: 793–809.
60. Retnakaran R. Glucose tolerance status in pregnancy: A window 83. Hashemi S, Ramezani Tehrani F, Mehrabi Y et al. Hypertensive
to the future risk of diabetes and cardiovascular disease in young pregnancy disorders as a risk factor for future cardiovascular and
women. Curr Diabetes Rev 2009; 4: 239–244. metabolic disorders (Tehran Lipid and Glucose Study). J Obstet
61. Retnakaran R, Shah BR. Mild glucose intolerance in pregnancy Gynaecol Res 2013; 39: 891–897.
and risk of cardiovascular disease: A population-based cohort 84. Fraser A, Nelson SM, Macdonald-Wallis C et al. Hypertensive
study. Can Med Assoc J 2009; 181: 371–376. disorders of pregnancy and cardiometabolic health in adolescent
62. Kessous R, Shoham-Vardi I, Pariente G et al. An association offspring. Hypertension 2013; 62: 614–620.
between gestational diabetes mellitus and long-term maternal 85. Geelhoed JJ, Fraser A, Tilling K et al. Preeclampsia and gesta-
cardiovascular morbidity. Heart 2013; 99: 1118–1121. tional hypertension are associated with childhood blood pres-
63. Damm P. Future risk of diabetes in mother and child after ges- sure independently of family adiposity measures: The Avon
tational diabetes mellitus. Int J Gynecol Obstet 2009; 104: Longitudinal Study of Parents and Children. Circulation 2010;
S25–S26. 122: 1192–1199.
References 233
86. Lawlor DA, Macdonald-Wallis C, Fraser A et al. Cardiovascular 100. Stein AD, Wang M, Ramirez-Zea M et al. Exposure to a nutri-
biomarkers and vascular function during childhood in the off- tion supplementation intervention in early childhood and risk
spring of mothers with hypertensive disorders of pregnancy: factors for cardiovascular disease in adulthood: Evidence from
Findings from the Avon Longitudinal Study of Parents and Guatemala. Am J Epidemiol 2006; 164: 1160–1170.
Children. Eur Heart J 2012; 33: 335–345. 101. Ramachandran A, Snehalatha C, Baskar AS et al. Temporal
87. Gluckman PD, Hanson MA, Beedle AS. Non-genomic transgen- changes in prevalence of diabetes and impaired glucose toler-
erational inheritance of disease risk. Bioessays 2007; 29: 149–154. ance associated with lifestyle transition occurring in the rural
88. Gluckman PD, Hanson MA, Spencer HG et al. Environmental population in India. Diabetologia 2004; 47: 860–865.
influences during development and their later consequences for 102. Snehalatha C, Ramachandran A. Cardiovascular risk factors in
health and disease: Implications for the interpretation of empiri- the normoglycaemic Asian-Indian population—Influence of
cal studies. Proc Biol Sci 2005; 272: 671–677. urbanisation. Diabetologia 2009; 52: 596–599.
89. Godfrey KM. Maternal regulation of fetal development and 103. Ravelli AC, van der Meulen JH, Michels RP et al. Glucose toler-
health in adult life. Eur J Obstet Gynecol Reprod Biol 1998; 78: ance in adults after prenatal exposure to famine. Lancet 1998;
141–150. 351: 173–177.
90. McCance DR, Pettitt DJ, Hanson RL et al. Birth weight and non- 104. de Rooij SR, Painter RC, Roseboom TJ et al. Glucose tolerance
insulin dependent diabetes: Thrifty genotype, thrifty phenotype, at age 58 and the decline of glucose tolerance in comparison
or surviving small baby genotype? Br Med J 1994; 308: 942–945. with age 50 in people prenatally exposed to the Dutch famine.
91. Gluckman PD, Hanson MA, Cooper C et al. Effect of in utero and Diabetologia 2006; 49: 637–643.
early life conditions on adult health and disease. N Engl J Med 105. Li Y, He Y, Qi L et al. Exposure to the Chinese famine in early life
2008; 359: 61–73. and the risk of hyperglycemia and type 2 diabetes in adulthood.
92. Yajnik CS, Deshmukh US. Maternal nutrition, intrauterine pro- Diabetes 2010; 59: 2400–2406.
gramming and consequential risks in the off spring. Rev Endocr 106. Armitage JA, Poston L, Taylor PD. Developmental origins of obe-
Metab Disord 2008; 9: 203–211. sity and the metabolic syndrome: The role of maternal obesity.
93. Ma RC, Chan JCN. Pregnancy and diabetes scenario around the Front Horm Res 2008; 36: 73–84.
world: China. Int J Gynecol Obstet 2009; 104: S42–S45. 107. Gluckman PD, Hanson MA, Beedle AS et al. Fetal and neonatal
94. Tam WH, Ma RC, Yang X et al. Glucose intolerance and cardio- pathways to obesity. Front Horm Res 2008; 36: 61–72.
metabolic risk in children exposed to maternal gestational diabe- 108. Goldberg AD, Allis CD, Bernstein E. Epigenetics: A landscape
tes mellitus in utero. Pediatrics 2008; 122: 1229–1234. takes shape. Cell 2007; 128: 635–638.
95. Bateson P. Fetal experience and good adult design. Int J Epidemiol 109. Burdge GC, Hanson MA, Slater-Jeffries JL et al. Epigenetic regu-
2001; 30: 928–934. lation of transcription: A mechanism for inducing variations in
96. Kuzawa CW. Fetal origins of developmental plasticity: Are fetal phenotype (fetal programming) by differences in nutrition during
cues reliable predictors of future nutritional environments? Am J early life? Br J Nutr 2007; 97: 1036–1046.
Hum Biol 2005; 17: 5–21. 110. Fall CHD. Non-industrialised countries and affluence:
97. Gluckman PD, Hanson MA, Bateson P et al. Towards a new Relationship with type 2 diabetes. Br Med Bull 2001; 60: 33–50.
developmental synthesis: Adaptive developmental plasticity and 111. Seshiah V, Balaji V, Balaji MS et al. Pregnancy and diabetes scenario
human disease. Lancet 2009; 373: 1654–1657. around the world: India. Int J Gynecol Obstet 2009; 104: S35–S38.
98. Rivera JA, Sotres-Alvarez D, Habicht JP et al. Impact of the Mexican 112. Nielsen KK, Kapur A, Damm P et al. From screening to post-
program for education, health, and nutrition (Progresa) on rates of partum follow-up—The determinants and barriers for gestational
growth and anemia in infants and young children: A randomized diabetes mellitus (GDM) services, a systematic review. BMC
effectiveness study. J Am Med Assoc 2004; 291: 2563–2570. Pregnancy Childbirth 2014, 14: 41.
99. Victora CG, Adair L, Fall C et al. Maternal and child under nutri- 113. Kapur A. Pregnancy a window of opportunity for improving
tion: Consequences for adult health and human capital. Lancet current and future health. Intl J Gynecol Obstet 2011; 115:
2008; 371: 340–357. S50–S51.
28 Diabetic pregnancy in the
developing world
Eran Hadar, Eran Ashwal, and Moshe Hod
Introduction Epidemiology of hyperglycemia in

Diabetes and obesity, as components of metabolic syndrome pregnancy in the developing world
and representatives of the noncommunicable disease spec-
The epidemic of diabetes has a profound and universal effect
trum, are incorrectly perceived as a unique problem of the
on pregnant women, both prior and during pregnancy—
developed world. In fact, they are emerging as a major health
type 1 and type 2 diabetes mellitus (T1DM and T2DM)
challenge in developing countries as well. Low-resource
and gestational diabetes mellitus (GDM), respectively. This
countries are experiencing a surge in the incidence of such
holds true not only for high-income countries, which is well
diseases, traditionally associated with the lifestyle of high-
established in extensive research, but even more so in low-
resource countries, while limited funds are available to over-
resource countries. Some low- and middle-resourced coun-
come this dilemma.
tries have the highest worldwide incidence of diabetes while
The issue of diabetes in the developing world is a rela-
in others the reported incidence is still relatively low. These
tively new problem. Hence, available data on the epidemi-
data must be interpreted with caution, as partial and inappro-
ology, diagnosis, management, and outcome of diabetes in
priate screening and diagnosis may lead to bias. This occurs
pregnancy, in such settings, are scarce. Although hyper-
due to lack of recourses and facilities, especially in rural
glycemia is a well-established cause for adverse pregnancy
areas, leading to a significant underdiagnosis and underre-
outcome, studies of the racial and ethnic distribution
porting bias, of the true incidence of T2DM and GDM.
of hyperglycemia in pregnancy have shown significant
variation in its epidemiology and have not been system-
atically and thoroughly investigated in other aspects, such Pregestational type 2 diabetes mellitus
as screening, diagnosis, management, and maternal and The overall prevalence of T2DM is highly deviated in dif-
perinatal outcome. Paucity of adequate data is the result of ferent ethnic groups and geographical regions. It is lowest
several factors: underdiagnosis due to lack of resource and in Africa (2.4%) and highest in Europe and North America
facilities and multiple strategies for screening and diagno- (7.89%).1 Even in the face of such low numbers, the incidence
sis, of which some are poorly implemented and incompati- of T2DM in Africa has increased almost twofold from 1997
ble to low-resource rural areas along with poor compliance to 2010.2 Recently published estimations anticipate that
to diagnosis and follow-up. This results not only in an the number of individuals with diabetes will again double
epidemiological bias but also leads to lack of proper diag- by 2030, reaching 366 million people worldwide, of them
nosis, inappropriate treatment, and an increased risk 61 million in Africa and 190 million in Asia.3
for diabetes-related adverse pregnancy outcome, for the The World Health Organization (WHO) Ad-Hoc Diabetes
mother and her infant (Figure 28.1). Moreover, the tre- Reporting Group reported on the prevalence of diabetes and
mendous therapeutic progress in diabetes in general and impaired glucose tolerance (IGT) in a diverse patient popu-
during pregnancy in particular, in the developed world, lation of reproductive aged women.4 They noted the lowest
during the last century has not reached to its full extent in rates of diabetes (<1%) in Bantu (Tanzania), Chinese, rural
low-resource countries, where resources for diabetes man- Indian, Sri Lankan, and some Pacific populations. Low rates
agement are often lacking, resembling a situation reminis- of diabetes (3%–5%) were reported in Italian women and
cent of the preinsulin era. in white, black, and Hispanic women in the United States.
In this chapter, we shall review current information on Rural Fijian Indian and Aboriginal Australian women had
the relevant issues of diabetes and hyperglycemia before, high rates of diabetes, reaching a prevalence of 7%. The high-
during, and following pregnancy, with an emphasis and est rates were found in Pima, Papago, and Nauruan Indians
focus on the special consideration in the developing world. (14%–22%). The prevalence of IGT was found to be under 3%
234
Epidemiology of hyperglycemia in pregnancy in the developing world 235
Stigma and
Orphanage
expulsion from
in affected family
local community
Birth defects or
congenital Increased risk for
Lower life Risk of malformations hysterectomy
expectancy maternal death in newborn during delivery
Maternal impact
Gestational diabetes
Impact on offspring
Higher risk of Birth defects or Higher risk for diab-

diabetes in later life congenital etes in pregnancy
malformations in later life in the
in newborn female offspring
Lower life
expectancy
Stigma and
expulsion from
local community
Figure 28.1 Maternal and neonatal consequences of gestational diabetes mellitus.
in Chinese and Malays and over 10% in black and Hispanic The reported prevalence of GDM in the United States, as a
women in the United States, urban Indian, Tanzania, Pima, representation of high-resource countries, ranges from 1%
and Nauruan Indians and in other Pacific communities. The to 14%, with 2%–5% being the most commonly reported rate.7
combined age-stratified rate of both diabetes and IGT ranged Dooley et al.8 studied 3744 pregnant women who under-
from 0% to 36%, with over 10% prevalence in one-third of the went universal screening for GDM. The patient population
populations and beyond 30% prevalence in Pima and Nauruan included 39.1% white, 37.7% black, 19.8% Hispanic, and 3.4%
Indians. Importantly, in several populations, the majority of Oriental and others. The adjusted relative risk was increased
cases diagnosed with diabetes were in fact undiagnosed prior in black (1.81, 95% confidence interval [CI] 1.13–2.89) and
to the survey. Thus, a significant proportion of patients with in Hispanic women (2.45, 95% CI 1.48–4.04). These findings
abnormal glucose tolerance will be missed without screening. were supported by others, showing that Asian women were
This fact warrants special consideration for pregnancy plan- more likely to have GDM than Caucasian women (31.7%
ning, as it implies that a significant proportion of high-risk vs. 14%, p < 0.02) despite their lower body mass index and
women will become pregnant without proper glycemic control socioeconomic status.9,10 Recent studies supported race as
prior to pregnancy, thus exposing themselves and their fetus a risk factor, with similar findings of a high rate of GDM
to adverse outcome associated with pregestational diabetes, among Asians, Hispanics, Native Americans, and African
specifically congenital anomalies and pregnancy loss. Americans as compared to non-Hispanic whites.11
Although closely related to race and ethnicity, from a
geographical point of view, Guariguata et al.12 in an exten-
Gestational diabetes mellitus sive review concluded that 90% of the global GDM burden
The prevalence of GDM is linearly and positively correlated occurs in low- and middle-income countries, with a global
with that of T2DM, in any given population. Therefore, in prevalence of 16.9% and a disease burden of 21.4 million
correlation with T2DM, the incidence of GDM has also live births affected annually. The highest rate of GDM was
increased dramatically in the past decade in all racial and found in Southeast Asia (23.1%) and the Middle East and
ethnic groups.5 The overall rate of GDM ranges vastly from North Africa (22.3%). In contrast to many countries in Asia
5% up to 25% of pregnant women, depending upon multi- and Latin America, where screening programs quantified
ple risk factors, such as race and ethnicity, familial history, the present problem, the majority of countries in Africa,
weight, height, age, and parity.6 mostly sub-Saharan Africa, lack the actual prevalence data
236 Diabetic pregnancy in the developing world
Table 28.1 Prevalence of gestational diabetes mellitus in selected

African countries
Country Description Date Prevalence (%)

Ethiopia Rural, Northern Ethiopia 1999 3.7
Mozambique Urban, Maputo 2002 7.3–11.0
Nigeria Urban, six studies 2004–2012 1.0–13.9
South Africa Urban and rural, four studies 1979–2011 0–8.8
Tanzania Rural 2011 0
Source: Macaulay S, et al., PLOS ONE, 9(6), e97871, 2014. With permission.
of diabetes in pregnancy.13 Currently available epidemio- probability of intrauterine fetal death, macrosomia, shoulder
logical data from this region are summarized in Table 28.1, dystocia, uterine rupture, and postpartum hemorrhage.17
showing not only the wide prevalence but also the pau- Finally, it is well established that untreated GDM causes
city of data coming from the 48 countries in sub-Saharan late-onset diabetes, leading to related complications in later
Africa. stages of life and negatively impacting women’s life expec-
The wide distribution variations in the prevalence of tancy, morbidity, and mortality. GDM leads to a higher
GDM by race and geography may be related to intricate prevalence of diabetes in the offspring of mothers who were
associations between genetic factors affecting insulin resis- affected by diabetes in pregnancy and, in turn, burdens
tance, lifestyle, diet, sociocultural factors, and healthcare future female generations.18
access and utilization. As such, and due to the remarkably From an environmental perspective, there is a complex
varied approaches used, different methods of screenings, interplay of obesity and malnutrition. Biological and envi-
various oral and intravenous glucose loads, and different ronmental factors in rural areas of developing countries,
diagnostic criteria, it remains unclear if this marked diver- in combination with malnutrition and infections, cause
sity represents true differences in the prevalence of GDM, metabolic alterations distinct from those seen in developed
or is simply biased. Understanding the racial, ethnic, and countries.19 Malnutrition concomitant with physical work
geographical disparities in GDM prevalence is a substan- and iron and protein insufficiencies contributes to the high
tial first step toward improving maternal and child health, prevalence of low-birth-weight newborns.20 Also, the pan-
through improved diagnosis and better treatment. creas appears to be sensitive to maternal anemia and infec-
tious diseases. In both in vitro studies and human studies,
a reduced number of pancreatic beta cells have been noted
Pathophysiology of hyperglycemia in in children who were small for gestational age at birth, sug-
pregnancy in the developing world gesting a reduced vascularization and replication rate of
beta pancreatic cells.21 Fetal and early postnatal undernutri-
Several factors come into play, when attesting to explain tion can provoke metabolic adaptations that affect variables
disparities in the prevalence and outcome of diabetes in such as glucose uptake,21 insulin signaling,22 leptin levels,23
low-resource countries. This is a combination of both genetic and endocrine adaptations that affect the hypothalamic–
and environmental influences. pituitary–adrenal axis.24 These changes mount to health out-
From a sociocultural perspective, gender studies pres- comes such as coronary heart disease, stroke, T2DM, and
ent the East African woman as a complex and dominant the metabolic syndrome, all of which have been shown to be
individual with a central role in society as mothers, work- increased in low-birth-weight neonates.25
ers, and members of families and communities.15 Women On the contrary, obesity has a profound impact on
are focused and educated on their central role in reproduc- hyperglycemia. The global trend of an increased prevalence
tion. Accordingly, being childless leads to stigmatization of diabetes in low-resource populations, with the subse-
and expulsion from the community.16 As a consequence, quent increase of GDM, is closely linked to higher rates of
childbearing even in later stages of life is common, and this obesity and is related to lifestyle and dietary changes.26,27
impacts the rate of diabetes in pregnancy. In recent decades, the incidence of T2DM has dramatically
Moreover, the impact of diabetes draws various social increased worldwide due to changes in lifestyle, from those
implications. In cases of diabetes prior to pregnancy, chil- of hunters, gatherers, and farmers to a contemporary pat-
dren may be born with birth defects, becoming customarily tern mainly characterized by sedentary occupations and
a female responsibility, with all the attendant implications. high-energy fuel resource intake.28 It has been demonstrated
In addition, poor obstetric outcomes in such pregnancies that Africans, especially in urban settings, are experiencing
may lead to a desire for “replacement” children. This, in turn, rapid demographic and epidemiological transitions, chang-
motivates high parity and its adverse sequel on the rate of ing diets and lifestyles encouraging urban drift and transna-
diabetes. Both GDM and pregestational T2DM increase the tional migration. This is characterized by an enormous rise
Diagnosis of hyperglycemia in pregnancy in the developing world 237
in noncommunicable diseases—diabetes, hypertension, and programming, exposing them to a higher rate of diabetes
cardiovascular diseases.29 Currently, the highest prevalence and the metabolic syndrome.
of T2DM is found in populations that have undergone rapid
alterations in lifestyle, such as Australian Aborigines, Native
Americans, Pacific Islanders, and some migrant populations Diagnosis of hyperglycemia in
such as the Asian Indians.30 pregnancy in the developing world
Finally, there is an important role for genetic alterations.
In 1962, Neel et al. proposed the “thrifty genotype” hypoth- Screening and diagnosis of diabetes in pregnancy appears
esis as a preliminary explanation for these observations.31 to be more complex in low-resource countries, especially in
In order to persist through centuries of evolution and in rural areas. Several aspects need to be addressed in relation
the face of the obvious and strong genetic selection against to the diagnosis of GDM and T2DM, which include finan-
this condition, the diabetogenic genes must have conferred cial, technical, and cultural barriers to appropriate diagnosis.
a survival advantage in times of nutritional deprivation;
yet they were detrimental at times of adequate or overnu- Screening and diagnosis
trition. In the 1980s, Barker et al.32 observed that the geo- The dilemma of when and how to screen and diagnose
graphical distribution of heart disease was closely related diabetes, worldwide, as well as in developing countries—is
to a person’s place and birth weight. This suggests that continuously debated.46,47 In the developing world, incon-
early life events can cause permanent changes in physiol- sistency in detection mechanisms and screening methods
ogy that, depending on the environment, may later entail of diabetes in pregnancy is being further magnified by the
a predisposition to disease. In the 1990s, the “thrifty phe- fact that most of the screening programs do no distinguish
notype” hypothesis was suggested by Hales and Barker as between GDM and undiagnosed T2DM—this may be solved
a probable etiology of T2DM.33 They claimed that adverse with the adoption of the International Association of the
mechanisms of nutritional thrift on the growing individual Diabetes and Pregnancy Study Groups (IADPSG) principles
are imposed by poor fetal and early infant nutrition. The for diabetes testing during pregnancy.48 Moreover, simpler,
thrifty phenotype hypothesis proposes that subsequent cheaper, and accessible strategies for diagnosis of hyper-
development of T2DM and the metabolic syndrome results glycemia in pregnancy need to be implemented uniquely
from poor nutrition in fetal and early life, which produces for the low-resource countries:
permanent changes in glucose and insulin metabolism.
During the period that the individual persists in the under- Diagnostic thresholds—Defining cutoff glucose levels that
nourished state, the physiological state is appropriately are adequate for the particular setting of a develop-
maintained. However, when the early adaptations that have ing country, considering the demographic shifts and
been invoked to survive under restricted nutrient supply local genotypes as a background determinant, remains
are later faced with conditions of affluence, the physiol- a nonresearched challenge. Both diabetes in preg-
ogy is disturbed, sequentially leading to the occurrence nancy and infectious diseases being highly prevalent
of disease.34 This phenomenon is particularly apparent in in low-resource countries, along with malnutrition,
countries undergoing a rapid economic and nutritional potentially influence maternal metabolism, induce
transition, or in peoples migrating from poor to wealthy a catabolic state in the mother, and increase the risk
countries.35 Observations in support of this assumption of diabetic ketoacidosis and its complications dur-
involve, for example, the increased susceptibility of Asians ing pregnancy.19 Basal glucose levels are most likely
to diabetes36 and of Africans to hypertension. 37 lower in pregnant women in low-income countries in
Yajnik et al. studied the body size and cord blood leptin comparison with their counterparts in high-income
and insulin concentrations in urban newborn (Pune, countries. Hence, in developing countries where a
India) and white Caucasians (London, England). The high prevalence of diabetes in pregnancy is assumed,
Indian babies were much smaller than white Caucasian outcomes could be more favorable when thresholds are
babies in all aspects apart from measurements of body fat set low, as more near-hyperglycemic pregnant women
and particularly central fat as judged by the subscapular will be diagnosed with diabetes. However, taking into
skinfold thickness. 38,39 They described these babies as hav- account financial constraints that such countries are
ing the “thin-fat” baby syndrome claiming it showed that facing and their need to focus on cost-effective, imme-
the excess visceral adiposity in most Asian adults can be diate lifesaving healthcare interventions, the cost
traced back to the neonatal phase. Also, cord blood insulin implications of setting a low cutoff level may not be
levels appeared to be higher, in Indian babies versus white feasible, nor efficacious in most low-income countries.
Caucasian babies, and were correlated with subscapular In many developing countries, specific target levels
skinfold thickness. Later in childhood, these “thin-fat” for the diagnosis and therapy of GDM and the conse-
Indian babies have substantial impaired insulin sensitiv- quences of their variation on the obstetrical outcome
ity, which inversely correlated with birth weight.40,41 These have not yet been formally evaluated.
studies, with subsequent support,42–45 provide solid data Selective versus universal—Selective screening based
to confirm the claim that people living in developing on risk factors may be a cheaper alternative to labo-
countries are prone to suffer the greatest effects of early ratory testing. However, such schemes score poorly
in predicting GDM.49 When selective screening is developing world,56 which include lack of standard proto-
applied, it is estimated that approximately 16% of cols, facilities, equipment, and the overall financing of health
GDM women will be undetected.50 The American services; lack of appropriately trained healthcare providers,
Diabetes Association recommended selective screen- particularly female doctors, nurses, and diabetic educators;
ing for women belonging to ethnic groups with low poor compliance to suggested care due to distance to health
prevalence of GDM. However, the lack of appropriate facilities, misperceptions of body size, weight gain/loss dur-
data on the prevalence of GDM in many low-income ing pregnancy, and cultural issues of health negligence; lack
countries makes this selective screening nearly impos- of decision-making power among women; and superstitions
sible. Universal screening for GDM in comparison to as to healthcare in general and diabetes in particular.
selective screening detects more cases and improves
maternal and offspring prognosis.51 The cost analysis
of universal screening when compared with risk factor Antenatal care and follow-up
screening showed only a negligible difference.49 Thus, Adherence to care, a lack of compliance because of poor
universal screening appears to be the most reliable and availability, religious beliefs or superstitions, and fear of
desired method for the detection of GDM, even in low- stigma can be expected. As families residing in rural areas do
income countries. not have the financial means to pay for appropriate screening
Methods for detection—Universal screening as such must and/or treatment. Hence, most pregnant women will opt for
be simple and cost effective. Currently, the two-step a traditional pregnancy and birth attendant. In some rural
procedure of screening with the 50 g glucose chal- areas, the patient with diabetes is considered as bewitched
lenge test (GCT) and then diagnosing GDM based on and as such might first consult a traditional healer or tra-
the oral glucose tolerance test (OGTT) is not feasible ditional birth attendant rather than seeking care from the
in low-resource countries because it places a tremen- medical health personnel.57 Women residing in rural areas
dous burden on the gravid women and jeopardizes are much less likely to receive antenatal care than their
their compliance. The two-step procedure requires the urban counterparts. For instance, in the developing world
pregnant women to visit the antenatal clinic twice, and as a whole, only one-third of rural women receive four or
at least three to five blood samples need to be drawn. more antenatal care visits (WHO recommendation on the
Alternatively, a single-step diagnostic procedure may number of antenatal visits), in comparison to two-thirds of
be easier to implement. Even though the National urban women.58
Institutes of Health (NIH) consensus concluded Therefore, as for screening and diagnosis, low-cost and
that there is insufficient evidence to adopt a one-step patient-friendly alternatives must be sought. Maiti et al.59
approach, concerning that the adoption of a one-step from Kolkata, India, studied 50 GDM patients versus 31 con-
procedure would increase the prevalence of GDM and trols employing a minimalistic approach for diabetes care.
the corresponding costs and interventions.52 They included 2 weeks of nutritional care followed, if neces-
sary, by twice daily regular human insulin, along with three
Some investigators have suggested low-cost alternatives for fortnight glucose measurements at a laboratory and visits by
GDM diagnosis. Such innovative algorithms can certainly be the patient or family member, every 2 weeks. No differences
used as alternatives when addressing diabetes in pregnancy were found between the two groups including perinatal
in the developing world.53 Agarwal et al.54 evaluated fasting mortality, macrosomia, shoulder dystocia, and birth injury.
plasma glucose (FPG) as a screening tool for diabetes among Although underpowered, this study shows that a simple and
1644 multiethnic high-risk population. They concluded that minimal approach may be adequate enough to ensure good
FPG below 4.4 mmol/L rules out GDM with 94.7% sensitiv- pregnancy outcome. This approach reached 100% compli-
ity and that FPG ≥5.3 mmol/L rules in GDM with compa- ance versus lower rates of 60% compliance in other studies.
rable sensitivity. Similar cutoffs of 5.1 and 4.4 mmol/L were
tested on the hyperglycemia and adverse pregnancy outcome
(HAPO) study population with similar results.55 Such algo- Diet and pharmacological treatment
rithms eliminate the need for OGTT in approximately 50% Even in cases where elevated blood sugar levels are actually
of the population, thus offering a simple low-cost algorithm detected, the lack of financial resources poses a main barrier
for diagnosis, saving money, and reducing the need for a for adequate treatment, such as insulin or oral hypoglycemic
cumbersome laboratory procedure. agents.
Lifestyle changes such as adopting appropriate eating
Obstacles to treatment of habits and engaging in regular physical exercise, which have
already been proven to be cost effective in diabetes care, play
hyperglycemia in pregnancy in an important role during and even before pregnancy. Such a
the developing world strict diet may be impossible to implement in a low-resource
setting, not to mention the impossibility to counsel for phys-
Following diagnosis of GDM, proper pregnancy care, treat- ical activity and weight loss during pregnancy. Obesity may
ment, and follow-up should commence. There are several be considered a desirable state in many low-income coun-
recognized obstacles to ensure such high-quality care in the tries as it signals strength, wealth, affluence, and even beauty
Summary 239
in women. Being thin might be associated with being seri- varied between countries, from 8.8% in the Maldives to an
ously ill, and therefore, women who are overweight might be increase of 5.5% in Zimbabwe. In 2008, more than 50% of all
reluctant to act on advice to lose weight to avoid diabetes or maternal deaths occurred in only six countries, all low-income
improve their diabetes control in pregnancy.60 countries (India, Nigeria, Pakistan, Afghanistan, Ethiopia,
One of the major problems regarding diabetes care is the and the Democratic Republic of the Congo).64
insulin itself. The high cost of insulin might not be affordable Coetzee et al.,53 from their departmental reports, found
for some, especially for pregnant women from rural areas an overall PNM rate of 30/1000 births, while women with
where health insurance schemes are absent. Furthermore, GDM and T2DM had PNM of 14:1000 and 70:1000, respec-
the appropriate storage of insulin supplies may be impos- tively. If left untreated, the PNM of women with GDM was
sible, as temperatures in some tropical low-income countries 264:1000, and of women with T2DM it soared to 313:1000.
can reach and go beyond 40°C, and refrigerators are not For comparison, nondiabetic mothers who had received
available. no prenatal care had a PNM of 94:1000. Hence, it is obvious
Oral glucose-lowering agents are an alternative in low- that untreated diabetes poses a greater threat of PNM than
resource settings, where patients might have a poor level having no prenatal care.
of education and a lack of means to use insulin injections In addition to common diabetic complications such as
resulting in poor diabetic control.61 However, their use has retinopathy, nephropathy, and neuropathy in women with
not been studied in low-resource settings. pregestational diabetes, infectious morbidity is significantly
more frequent than in healthy women. There is a correlation
between the frequency of infections and poor metabolic
Technological availability control.65 Preeclampsia, which occurs in 20% of diabetic
Women living in a rural low-resource environment are less pregnant women—three times more frequent than in healthy
likely to remember their last menstrual period. Thus, the pregnant women—significantly increases perinatal and
estimation of gestational age is inaccurate and may not be maternal mortality and morbidity.66 The major cause of mater-
dated properly due to lack of access to pregnancy care dur- nal complications and subsequent deaths in low-resource
ing the first trimester. Even when early pregnancy care is countries as those in East Africa seems to be the delay in
sought and reached, sonography is not available as a tool for arrival at healthcare centers and hospitals.67 Apart from the
pregnancy dating and for later follow-up and management. lack of financial means and living far from health centers,
The lack of ultrasound even in referral centers throughout the low educational level often prevents a timely and poten-
countries such as in East Africa poses tremendous difficul- tially lifesaving presentation at health facilities especially for
ties for adequate pregnancy dating, monitoring of fetal well- women in the rural areas.68
being, and estimated growth throughout pregnancy.62 Also,
the lack of access to even more basic tools, such as capillary
glucose meters, remains difficult in rural areas. Summary
There is a paucity of data available on the prevalence of
Adverse outcomes due to GDM and pregestational diabetes in low-resource countries.
Furthermore, the available data are derived mainly from
hyperglycemia in pregnancy the more accessible urban population while less data exist
in the developing world regarding populations in rural areas. In addition, the criteria
for diagnosis are inconsistent, making comparisons diffi-
Hyperglycemia in pregnancy has a profound impact on cult. The management of diabetes in pregnancy is a tremen-
maternal and neonatal outcome (Figure 28.1). Adequate dously challenging task especially where existing resources
information about maternal and perinatal morbidity and are few, infrastructures are very weak, and the educational
mortality (PNM) as a consequence of diabetes in pregnancy levels are low. More newborn infants are being exposed to
is scarcely reported. the metabolic environment of diabetes during intrauterine
In 2004, the WHO estimated that 529,000 maternal deaths development each year as a result of the changing incidence
occurred worldwide—251,000 in Africa (47%), 253,000 in and demographics of diabetes and pregnancy. National gov-
Asia (48%), 22,000 (4%) in Latin America and the Caribbean, ernments along with the general public in the developing
and approximately 2,500 (Less than 1%) in the developed world must constantly be aware of the problems related to
countries.63 In terms of maternal mortality rate (MMR), the diabetes in pregnancy in order to establish appropriate pro-
world figure is estimated to be 4 to 1000 live births. Stratified grams including an approach to the diagnosis and therapy of
by region, the MMR was highest in Africa (8.3:1000); followed diabetes in pregnancy. Within such programs and policies,
by Asia (3.3:1000); Oceania (2.4:1000), Latin America, and the the prevention and issues of accessibility and affordability
Caribbean (1.9:1000); and the developed countries (0.2:1000). of care have to be addressed. Simultaneously, education for
A follow-up report estimated that maternal deaths worldwide healthcare workers must be implemented as they are able
decreased to 342,900 in 2008, with global MMR of 2.5:1000. to identify groups at a high risk of diabetes and to contact
The global MMR declined at a rate of 1.3% since 1990. During women and educate them of the existence of diabetes in
the years 1990–2008, the rates of yearly decline in the MMR pregnancy and its impact on their offspring.
Awareness of diabetes at the primary care level, assisted care. Improved healthcare knowledge will be associated
by access to obstetrical and medical intervention when with better family planning, better nutrition, and a better
necessary, will decrease perinatal and maternal compli- understanding of the threat that diabetes poses to preg-
cations and diabetes-specific complications. Efforts to nant women.
identify primary prevention strategies should be sought The development of programs and policies for the diag-
as secondary prevention as it is much more difficult to nosis and treatment of diabetes in pregnancy is a necessary
apply in low-resource countries. Women of reproductive key issue in attempting to prevent the diabetes complications
age must be educated and empowered to seek antenatal in both the mother and her fetus.
REFERENCES
1. International Diabetes Federation. Global burden of diabetes. 21. Hoet JJ. Influence of dietary changes on the development of the
In: Gan D, ed., Diabetes Atlas, International Diabetes Federation: fetal pancrease consequences later in life. Isr J Med Sci 1991;
Brussels, Belgium, 2003, p. 15. Available from: www.idf.org/e-atlas/. 27(8–9): 423–424.
Accessed on November 8, 2014. 22. Ozanne SE, Nave BT, Wang CL et al. Poor fetal nutrition causes
2. Amos AF, McCarty DJ, Zimmet P. The rising global burden of long-term changes in expression of insulin signaling components
diabetes and its complications: Estimates and projections to the in adipocytes. Am J Physiol 1997; 273: E46–E51.
year 2010. Diabet Med 1997; 14(Suppl. 5): S1–S85. 23. Phillips DI, Fall CH, Cooper C et al. Size at birth and plasma
3. Wild S, Roglic G, Green A et al. Global prevalence of diabetes: leptin concentrations in adult life. Int J Obes Relat Metab Disord
Estimates for the year 2000 and projections for 2030. Diabetes 1999; 23(10): 1025–1029.
Care 2004; 27(5): 1047–1053. 24. Seckl JR, Nyirenda MJ, Walker BR et al. Glucocorticoids and fetal
4. WHO Ad Hoc Diabetes Reporting Group. Diabetes and impaired programming. Biochem Soc Trans 1999; 27(2): 74–78.
glucose tolerance in women aged 20–39 years. World Health Stat 25. Barker DJP. Mothers, Babies and Diseases in Later Life, BMJ
1992; 45: 25–29. Publishing Group: London, U.K., 1994.
5. King H. Epidemiology of glucose intolerance and gestational 26. Sobngwi E, Mauvais-Jarvis F, Vexiau P et al. Diabetes in Africans.
diabetes in women of childbearing age. Diabetes Care 1998; Part 1: Epidemiology and clinical specificities. Diabetes Metab
21(Suppl. 2): B9–B13. 2001; 27(6): 628–634.
6. American Diabetes Association. Gestational diabetes mellitus. 27. Sobngwi E, Mauvais-Jarvis F, Vexiau P et al. Diabetes in Africans.
Diabetes Care 2003; 26(Suppl. 1): S103–S105. Part 2: Ketosis-prone atypical diabetes mellitus. Diabetes Metab
7. Coustan DR. Gestational diabetes. In: National Institute of 2002; 28(1): 5–12.
Diabetes and Digestive and Kidney Diseases, ed., Diabetes in 28. Dowse GK, Zimmet PZ, Finch CF et al. Decline in incidence of
America, 2nd ed., NIH Publication No. O95-1468, (NIDDK): epidemic glucose intolerance in Nauruans: Implications for the
Bethesda, MD, 1995, pp. 703–717. “thrifty genotype.” Am J Epidemiol 1991; 133(11): 1093–1104.
8. Dooley SL, Metzger BE, Cho NH. Gestational diabetes mellitus. 29. Chukwuma Sr C, Tuomilehto J. The ‘thrifty’ hypotheses: Clinical
Influence of race on disease prevalence and perinatal outcome in and epidemiological significance for non-insulin depen-
a U.S. population. Diabetes 1991; 40(Suppl. 2): 25–29. dent diabetes mellitus and cardiovascular disease risk factors.
9. Gunton JE, Hitchman R, McElduff A. Effects of ethnicity on glu- J Cardiovasc Risk 1998; 5(1): 11–23.
cose tolerance, insulin resistance and beta cell function in 223 30. McCance DR, Pettitt DJ, Hanson RL et al. Birth weight and non-
women with an abnormal glucose challenge test during preg- insulin dependent diabetes: Thrifty genotype, thrifty phenotype,
nancy. Aust N Z J Obstet Gynaecol 2001; 41(2): 182–186. or surviving small baby genotype? Br Med J 1994; 308(6934):
10. Hunsberger M, Rosenberg KD, Donatelle RJ. Racial/ethnic dis- 942–945.
parities in gestational diabetes mellitus: Findings from a popula- 31. Neel JV. Diabetes mellitus: A “thrifty” genotype rendered detri-
tion-based survey. Womens Health Issues 2010; 20(5): 323–328. mental by “progress”? Am J Hum Genet 1962; 14: 353–362.
11. Beckles GLA, Thompson-Reid PE. Diabetes and Women’s 32. Barker DJ, Osmond C. Infant mortality, childhood nutrition, and
Health across the Life Stages, Centers for Disease Control and ischaemic heart disease in England and Wales. Lancet 1986;
Prevention: Atlanta, GA, 2006. 1(8489): 1077–1081.
12. Guariguata L, Linnenkamp U, Beagley J, Whiting DR, Cho NH. 33. Hales CN, Barker DJ. Type 2 (non-insulin-dependent) diabetes
Global estimates of the prevalence of hyperglycaemia in preg- mellitus: The thrifty phenotype hypothesis. Diabetologia 1992;
nancy. Diabetes Res Clin Pract 2014; 103(2): 176–185. 35(7): 595–601.
13. Seshiah V, Balaji V, Balaji MS et al. Pregnancy and diabetes 34. Adair LS, Prentice AM. A critical evaluation of the fetal origins
scenario around the world: India. Int J Gynaecol Obstet 2009; hypothesis and its implications for developing countries. J Nutr
104(Suppl. 1): S35–S38. 2004; 134(1): 191–193.
14. Macaulay S, Dunger DB, Norris SA. Gestational diabetes mellitus 35. Fall CH. Non-industrialised countries and affluence. Br Med Bull
in Africa: A systematic review. PLOS ONE 2014; 9(6): e97871. 2001; 60: 33–50.
15. Vlassoff C, Garcia Moreno C. Placing gender at the centre of 36. Yajnik CS. Early life origins of insulin resistance and type 2 dia-
health programming: Challenges and limitations. Soc Sci Med betes in India and other Asian countries. J Nutr 2004; 134(1):
2002; 54(11): 1713–1723. 205–210.
16. Holloway M. Trends in women’s health: A global view. Sci Am 37. McKeigue PM, Lithell HO, Leon DA. Glucose tolerance and resis-
1994; 271: 77. tance to insulin-stimulated glucose uptake in men aged 70 years
17. Lucchini R, Barba G, Giampietro S et al. Macrosomic infants: in relation to size at birth. Diabetologia 1998; 41(10): 1133–1138.
Clinical problems at birth and afterward. Minerva Pediatr 2010; 38. Yajnik CS, Lubree HG, Rege SS et al. Adiposity and hyperinsu-
62(3 Suppl. 1): 65–66. linemia in Indians are present at birth. J Clin Endocrinol Metab
18. Weiss PAM, Walcher W, Scholz HS. Der vernachl€assigte 2002; 87(12): 5575–5580.
Gestationsdiabetes: Risiken und Folgen. Geburtshilfe Frauenheilkd 39. Yajnik CS, Fall CH, Coyaji KJ et al. Neonatal anthropometry: The
1999; 59: 535–544. thin-fat Indian baby. The Pune Maternal Nutrition Study. Int J
19. Weiss PAM, ed. Diabetes und Schwangerschaft, Springer: Obes Relat Metab Disord 2003; 27(2): 173–180.
New York, 2000, p. 148. 40. Bavdekar A, Yajnik CS, Fall CH et al. Insulin resistance syndrome
20. Ebrahim G. Influence on maternal nutrition on pregnancy out- in 8-year-old Indian children: Small at birth, big at 8 years, or
come. Maternity services in the developing world. What the both? Diabetes 1999; 48(12): 2422–2429.
community needs. In: Philpott RH, ed. Maternity services in the 41. Yajnik C. Interactions of perturbations in intrauterine growth and
developing world: what the community needs. London, Royal growth during childhood on the risk of adult-onset disease. Proc
College of Obstetricians and Gynaecologists, 1979, 228–232. Nutr Soc 2000; 59(2): 257–265.
References 241
42. Singhal A, Cole TJ, Fewtrell M et al. Breastmilk feeding and 56. Nielsen KK, de Courten M, Kapur A. Health system and soci-
lipoprotein profile in adolescents born preterm: Follow-up etal barriers for gestational diabetes mellitus (GDM) services—
of a prospective randomised study. Lancet 2004; 363(9421): Lessons from World Diabetes Foundation supported GDM
1571–1578. projects. BMC Int Health Hum Rights December 5, 2012; 12: 33.
43. Prentice AM. Intrauterine factors, adiposity, and hyperinsulinae- 57. Peltzer K, Khoza LB, Lekhuleni ME et al. Concepts and treatment
mia. Br Med J 2003; 327(7420): 880–881. for diabetes among traditional and faith healers in the northern
44. Krishnaveni GV, Hill JC, Veena SR et al. Truncal adiposity is province, South Africa. Curationis 2001; 24(2): 42–47.
present at birth and in early childhood in South Indian children. 58. UNICEF Progress for Children. Achieving the MDGs with Equity,
Indian Pediatr 2005; 42(6): 527–538. No. 9, UNICEF: New York, 2010, p. 30.
45. Prentice AM, Moore SE. Early programming of adult diseases in 59. Maiti A, Nandi K, Chatterjee S. Management of GDM in a public
resource poor countries. Arch Dis Child 2005; 90(4): 429–432. hospital setting in India: Lessons from a minimalistic approach.
46. Homko CJ, Reece EA. To screen or not to screen for gestational Diabetes Res Clin Pract February 2012; 95(2): e34–e36.
diabetes mellitus. The clinical quagmire. Clin Perinatol 2001; 60. Ahmed AM, Ahmed NH. Diabetes mellitus in Sudan: The size of
28(2): 407–417. the problem and the possibilities of efficient care. Prac Diabetes
47. Zeck W, McIntyre HD. Gestational diabetes in rural East Africa: Int 2001; 18(9): 324–327.
A call to action. J Womens Health 2008; 17(3): 403–411. 61. Rowan JA, Hague WM, Gao W et al. Metformin versus insulin for
48. International Association of Diabetes and Pregnancy Study the treatment of gestational diabetes. N Engl J Med 2008; 358(19):
Groups Consensus Panel, Metzger BE. International Association 2003–2015.
of Diabetes and Pregnancy Study Groups recommendations on 62. WHO. World Diabetes Day. 2006. Available from: www.who.
the diagnosis and classification of hyperglycemia in pregnancy. int /mediacentre/events/2006/world_diabetes_day/en/index.
Diabetes Care 2010; 33: 676–682. html. Accessed on November 11, 2014.
49. Shamsuddin K, Mahdy ZA, Siti Rafiaah I et al. Risk factor screen- 63. Maternal Mortality in 2000. Estimates developed by WHO,
ing for abnormal glucose tolerance in pregnancy. Int J Gynaecol UNICEF and UNFPA. Department of Reproductive Health and
Obstet 2001; 75(1): 27–32. Research, World Health Organization: Geneva, Switzerland,
50. Soonthornpun S, Soonthornpun K, Aksonteing J et al. A com- 2004.
parison between a 75-g and 100-g oral glucose tolerance test in 64. Hogan MC, Foreman KJ, Naghavi M et al. Maternal mortality
pregnant women. Int J Gynaecol Obstet 2003; 81(2): 169–173. for 181 countries, 1980–2008: A systematic analysis of prog-
51. Cosson E, Benchimol M, Carbillon L et al. Universal rather than ress towards Millennium Development Goal 5. Lancet 2010;
selective screening for gestational diabetes mellitus may improve 375(9726): 1609–1623.
fetal outcomes. Diabetes Metab 2006; 32(2): 140–146. 65. Cousins L. Pregnancy complications among diabetic women:
52. Vandorsten JP, Dodson WC, Espeland MA et al. NIH consensus Review 1965–1985. Obstet Gynecol Surv 1987; 42(3): 140–149.
development conference: Diagnosing gestational diabetes mel- 66. Garner PR, D’Alton ME, Dudley DK et al. Preeclampsia in dia-
litus. NIH Consens State Sci Statements 2013; 29(1): 1–31. betic pregnancies. Am J Obstet Gynecol 1990; 163(2): 505–508.
53. Coetzee EJ, Jackson WP. The management of non-insulin-depen- 67. ZeckW, Kelters I, Lang U et al. Lessons learned from maternal
dent diabetes during pregnancy. Diabetes Res Clin Pract 1985; deaths at an East African health center. J Perinat Med 2006; 34(6):
1(5): 281–287. 459–461.
54. Agarwal M. Fasting plasma glucose as a screening test for GDM. 68. UNICEF Kinderhilfswerk der Vereinten Nationen. Todesursache:
Diabet Med 2000; 17: 720–726. Schwangerschaft und Geburt UIzM, Bereich Grundsatz und
55. Agarwal MM, Weigl B, Hod M. GDM: The low cost algorithm. Int Information. Available from: www.unicef.de/download/i_0087.
J Gynaecol Obstet 2011; 115(Suppl. 1): S30–S33. pdf. Accessed on November 11, 2014.
29 Managing diabetic pregnancy
in China
Huixia Yang, Weiwei Zhu, and Rina Su
Shanxi, Heilongjiang, and Guangdong and attended the

Introduction national obstetrics conference. Nine hundred of 1190 ques-
Following the publication of the Hyperglycemia and tionnaires were collected as valid wherein the recovery rate
Adverse Pregnancy Outcome (HAPO) study,1 the Inter was 75.6% (119 from Tianjin, 81 from Shanxi, 399 from
national Association of Diabetes in Pregnancy Study Heilongjiang, 169 from the national conference). Four hun-
Groups (IADPSG)2 formulated consensus guidelines for dred and sixty-seven HCPs worked in tertiary hospitals
the testing and diagnosis of gestational diabetes mellitus (51.9%), 58 HCPs had a master’s degree (23.1%), and 543 had
(GDM). These have been adopted by the American Diabetes a bachelor’s degree (60.3%).
Association (ADA).3 The Ministry of Health (MOH) of
China also recommended new testing and diagnostic crite-
ria based on the IADPSG guidelines4, which became effec- Use of the new GDM diagnostic criteria in the
tive on December 1, 2011. The Chinese criteria agree with beginning
the ADA in 2013 that a diagnostic 75 g oral glucose toler- The criteria published by MOH (WS331-2011) adopted 75 g
ance test (OGTT) should be performed between the 24th OGTT at 24–28 gestational weeks, and one of the three
and 28th weeks of gestation for all pregnant women not values matched the standard to get a GDM diagnosis. The
previously known to have diabetes. A diagnosis of GDM percentage of HCPs who adopted the new diagnostic cri-
can be made if one or more of the following glucose levels teria was unbalanced in these survey regions. It was high
are elevated: fasting ≥5.1 mmol/L, 1-hour ≥10.0 mmol/L, in Shanxi and Guangdong, which were 63.0% and 53.1%,
and 2-hour ≥8.5 mmol/L. but Heilongjiang showed a very low percentage of 23.5%
This is the one-step approach that the MOH has recom- (Table 29.1).
mended to well-resourced medical institutions. However, Meanwhile, few HCPs adopted FPG 7.0 mmol/L in the
in China, a formal OGTT may be difficult to implement first trimester to diagnose pregestational diabetes mellitus
in low-resourced rural areas. Therefore, and different from (PGDM), which was 33.3%.
the ADA in 2013, the MOH has recommended a two-step At the hospital level, 205 HCPs from tertiary hospitals
approach under these circumstances. An fasting plasma (46.6%) adopted the new criteria, which was higher than
glucose (FPG) can be used as a screening tool to reduce HCPs from other clinic instructions (194, 3.3%), but there
the number of OGTTs required. If the FPG ≥5.1 mmol/L was no statistical difference (χ2 = 0.97, p > 0.05).
after the 24th week of gestation, GDM can be diagnosed,
and if <4.4 mmol/L, GDM is unlikely. Women with FPG
≥4.4 mmol/L and ≤5.0 mmol/L will still require a OGTT. Management of GDM
This approach will approximately halve the number of When a pregnant woman is diagnosed with GDM, medical
OGTTs required.5 nutrition therapy (MNT) and excise should be the priority
Though MOH in China established the new diagnostic treatment. Most HCPs chose this option, with the highest
criteria of GDM (WS331-2011), standardized management to rate in Guangdong (93.5%) and the lowest in Heilongjiang
GDM has not spread in every area. A questionnaire survey (77.8%). Though most HCPs chose MNT and exercise as
was carried out in some regions to get the situation of the use priority, only about 20% doctors mastered MNT and could
of the new criteria and management to GDM at the begin- teach patients the knowledge of the food exchange method
ning of the new criteria published (July–October, 2011). (Heilongjiang 19.8%, Guangdong 21.3%). Over 60% doctors
The study subjects were the healthcare practitioners heard of it but could not practice. About 80% HCPs would
(HCPs) who attended the GDM training seminars in Tianjin, use insulin to treat GDM (Table 29.2).
242
Management of GDM in China 243
Knowledge of postpartum follow-up

Table 29.1 Use of new GDM diagnostic criteria
in the beginning In the survey, more than 95% of HCPs thought mothers with
GDM and their offspring needed long-term follow-up. But in
reality, only about 60% of them suggested the patients come
n Constitution rate (%)
for follow-up (Heilongjiang 50.6%, Guangdong 67.9%, and
Tianjin 50 37.9 national 64.5%).
Shanxi 75 63.0 The suggested sections for follow-up were mostly the
Heilongjiang 19 23.5 endocrinology department (about 60%) and obstetrical
Guangdong 212 53.1 department (about 20%); see Table 29.5.
National 51 30.2
Total 407 45.2
Management of GDM in China
The historical diagnostic criteria for GDM derived from the
Health education for GDM patients study of O’Sullivan et al.7 50 years ago, which evolved into
The women with GDM has increased risk of gestational the national diabetes date group (NDDG) criteria and ADA
hypertension, preeclampsia, polyhydramnios, and urinary criteria and is continuously used till now. The IADPSG rec-
system infection in the short term and type 2 diabetes in ommended the new criteria based on the HAPO study while
the long term.6 If GDM in women is not well controlled, the ADA accepted it in 2011. The MOH of China established the
fetus is at high risk for hyperinsulinemia, macrosomia, and same new criteria on July 1, 2011. There is still some con-
birth hypoglycemia in the short term. The offspring also are troversy about the new GDM diagnostic criteria. Walsh
at high risk for metabolic disorder, obesity, and early-onset et al.8 found that the new diagnostic criteria for GDM would
type 2 diabetes in the long term. result in an increased rate of diagnosis of GDM and there-
There were 20.7%, 16.3%, and 15.9% of HCPs who did not fore increased surveillance of these patients. However, this
realize the three long-term impacts separately. treatment will not improve perinatal outcome, and we may
Most HCPs would tell patients that GDM had influence now be diagnosing and treating a group of women without
to the mother and offspring in both the short term and long sufficient evidence of clinical benefit, while in this condition,
term. Meanwhile most of HCPs thought intervention would the spread of the new criteria is still necessary.
decrease the complications of mother and offpring. But about Other nations also made a similar survey. For example,
10% doctors still considered intervention would have no Akinci et al.9 made a study in Turkey and suggested that
effect on the complication of GDM (see Tables 29.3 and 29.4). there is considerable variation in the clinical practice patterns
Table 29.2 Management of GDM
HCPs chose MNT and exercise Constitution rate HCPs chose insulin to treat GDM Constitution rate
as priority (n) (%) (n) (%)
Tianjin 115 87.1 98 74.2
Shanxi 96 80.7 84 70.6
Heilongjiang 63 77.8 70 86.4
Guangdong 373 93.5 322 80.7
National 142 84.0 155 91.7
Total 789 87.7 729 81.0
Table 29.3 Patient education about the short- and long-term effects of GDM
About effect to the mother Constitution rate (%) About effect to the offspring Constitution rate (%)
Tianjin 126 95.5 126 95.5
Shanxi 114 95.8 114 95.8
Guangdong 382 95.7 367 92.0
National 156 92.3 154 91.1
Total 851 94.6 833 92.6
244 Managing diabetic pregnancy in China
Poorly mastered medical nutrition therapy

Table 29.4 Admit intervention effective to the
complications of mother and offspring The treatment of GDM is effective in reducing macrosomia
(high-quality evidence), preeclampsia, and shoulder dys-
tocia.10 MNT is the preferred treatment for GDM, wherein
n Constitution rate (%)
the majority of GDM patients control their blood glucose,
Tianjin 123 93.2 thus making the doctor’s guidance particularly important
Shanxi 118 99.2 to patients. About 20% of doctors mastered MNT and could
Heilongjiang 69 85.2 teach patients the knowledge of the food exchange method,
Guangdong 371 93.0 while over 60% of doctors heard of it but could not operate,
National 145 85.8 which reminds us that the most important thing to improve
Total 826 91.8 GDM management is to train HCPs, in every possible way,
to master MNT.
of physicians, and they considered that an education pro- Very low long-term postpartum follow-up rate
gram to enhance the clinical aptitude of physicians, particu- Carson et al.11 collected 307 citations to identify system-
larly family physicians, in the medical management of GDM based factors that could explain the broad variation in
should be designed throughout the country. We conducted postpartum GDM testing rates and found that proactively
WDF 10-517 to train HCPs and patients to improve their contacting patients via phone calls, education programs, or
knowledge regarding GDM. At the beginning of the project, postal reminders was associated with higher postpartum
the survey showed the following. testing rates. Rather than working to identifying individ-
ual demographic factors, system-based approaches were
associated with a larger potential impact and appear easily
Unbalanced use of the new GDM standard in generalizable. Clinicians should think beyond individual
different areas habits and consider systematic approaches to improving
Over 50% of hospitals in Guangdong and Shanxi adopted testing rates.
the new standard, while only 30% did in Heilongjiang and In this study, lack of GDM knowledge prevented the
Tianjin. A lot of hospitals still performed 50 g GCT. Some doctors to educate and treat patients and value the impor-
hospitals used the new standards only in high-risk pregnant tance of follow-up. In reality, they seldom asked the patients
women. For example, 31.8% of HCPs in Guangdong did not to come back for follow-up. Most doctors considered that
perform routine FPG test to pregnant women in the first the endocrinologists should be responsible for the follow-up
trimester, while this percentage was 18.5% in Heilongjiang. of GDM patients, which reminds us that follow-up requires
This may be caused by less knowledge of doctors. collaboration of both obstetricians and endocrinologists.
The pregnancy managements were unbalanced in dif-
ferent areas in China; thus we should make an appropri-
Update of GDM knowledge lagged behind seriously ate management strategy based on local conditions. Maiti
The majority of doctors relied on textbooks or guidelines et al.12 concluded that it may not be possible to provide
issued by the government. Fewer doctors obtained GDM access to Western standards of care to a vast majority of
knowledge through literature, such as in Tianjin and Shanxi, GDM patients in resource-constrained settings in devel-
with only two and one doctor, respectively. Over 40% doc- oping countries, yet with minimal care delivered in a con-
tors had never heard of the HAPO study. Many doctors still sistent fashion, it is possible to obtain excellent maternal
rely on their knowledge of GDM during college, which is and neonatal outcome. In China, for example, in Tianjin,
outdated. Less than 40% doctors knew the weight gain scope pregnancy management in the first and second trimester
of diabetes during pregnancy. Update of GDM knowledge is in community medical instructions; therefore, in this
lagged behind seriously. area, community capability should be strengthened when
Table 29.5 Chosen postpartum section for follow-up
Endocrinology Constitution rate (%) Obstetrical Constitution rate (%)

Tianjin 28 21.2 95 72.0
Shanxi 27 22.7 79 66.4
Guangdong 147 36.8 224 56.1
National 32 18.9 97 57.4
Total 249 27.7 556 61.8
References 245
GDM management is promoted. In other situations, train- built. GDM management is promoted in a systemic pattern
ing HCPs from tertiary hospitals should be considered, and where 45 experienced tertiary hospitals are unified in man-
tertiary hospitals should be in charge of improving their aging locals with GDM.
medical instructions, thereby developing a pattern where As the incidence of diabetes and overweight in pregnancy
experienced tertiary hospitals are unified in managing has increased in China, a new project (WDF14-908) for dia-
locals with GDM. betes management beyond pregnancy funded by WDF has
been carried out from 2015. This project has mainly focused
on prepregnancy, pregnancy care, and postpartum care at
the hospital and community health service centers (CHSC)
Development of GDM management level for women with prediabetes and high risks of GDM.
in China Like the FIGO initiative on GDM,15 the CHSC can support
postpartum follow-up through child vaccination.
Since the MOH criteria gained acceptance throughout Patient education is important in the management of
China in 2011, the incidence of GDM has been estimated GDM. In China, a day-care outpatient clinic has been
at 17.5%.13 The updated criteria and management in China launched, where patients with GDM spend whole day and
have been published in 2014.14 receive basic information on GDM, medical nutritional
With the support of the World Diabetes Foundation therapy, exercise advice, and self-monitoring of blood glu-
(WDF 10-517), national standardized trainings for obste- cose, guided by professional nurses, clinical nutritionists,
tricians, professional nurses, and clinical nutritionists have and physicians. The concepts of healthy lifestyle and blood
been conducted in 16 different regions in China, totally glucose self-management promoted in the manual aim to
19,754 doctors and nurses have been trained, and 45 centers prevent or postpone the development of type 2 diabetes later
that focus on GDM screening and perinatal care have been in life.16
REFERENCES
1. HAPO Collaborative Research Group. Hyperglycemia and 9. Akinci B, Tosun P, Bekci E, Yener S, Demir T, Yesil S. Management
adverse pregnancy outcomes. N Engl J Med 2008; 358: of gestational diabetes by physicians in Turkey. Prim Care
1991–2002. Diabetes 2010; 4(3): 173–180.
2. IADPSG Consensus Panel. International Association of Diabetes 10. Falavigna M, Schmidt MI, Trujillo J et al. Effectiveness of gesta-
and Pregnancy Study Groups recommendations on the diagnosis tional diabetes treatment: A systematic review with quality of evi-
and classification of hyperglycemia in pregnancy. Diabetes Care dence assessment. Diabetes Res Clin Pract 2012; 98(3): 396–405.
2010; 33: 676–682. 11. Carson MP, Frank MI, Keely E. Postpartum testing rates among
3. American Diabetes Association. Standards of medical care in women with a history of gestational diabetes—Systematic review.
diabetes. Diabetes Care 2011; 34(Suppl. 1): S11–S61. Prim Care Diabetes 2013; 7(3): 177–186.
4. Yang HX, Medical Service Specialty Standard Committee of 12. Maiti A, Nandi K, Chatterjee S. Management of gestational diabe-
Ministry of Health. Diagnostic criteria for gestational diabetes tes mellitus in a public hospital setting in India: Lessons from a min-
mellitus (WS 331–2011). Chin Med J (Engl) 2012; 125: 1212–1213. imalist approach. Diabetes Res Clin Pract 2012; 95(2): e34–e36.
5. Zhu WW, Fan L, Yang HX et al. Fasting plasma glucose at 24–28 13. Zhu WW, Yang HX, Wei YM et al. Evaluation of the value of fast-
weeks to screen gestational diabetes—New evidence from ing plasma glucose in the first prenatal visit to diagnose gestational
China. Diabetes Care 2013; 36: 2038–2040. diabetes mellitus in china. Diabetes Care 2013; 36: 586–590.
6. Hedderson MM, Ferrara A, Sacks DA. Gestational diabetes mel- 14. Chinese Medical Association, Branch of Obstetrics and
litus and lesser degrees of pregnancy hyperglycemia: Association Gynaecology, Obstetric Group, Chinese Society of Perinatal
with increased risk of spontaneous preterm birth. Obstet Gynecol Medicine, Diabetic Pregnancy Study Group. Guide for diagno-
2003; 102(4): 850–856. sis and treatment of diabetes beyond pregnancy. Chin J Obstetr
7. O’Sullivan JB, Mahan CM, Charles D et al. Screening criteria Gynecol 2014; 49(8): 561–569.
-for high-risk gestational diabetic patients. Am J Obstet Gynecol 15. Hod M, Kapur A, Sacks DA et al. The International Federation
1973; 116: 895–900. of Gynecology and Obstetrics (FIGO) initiative on gestational
8. Walsh C, Khan M, Kinsley B, Daly S. Have the new diagnostic diabetes mellitus: A pragmatic guide for diagnosis, management,
criteria for gestational diabetes mellitus (GDM) impacted on peri- and care. Int J Gynecol Obstetr 2015; 131(Suppl. 1): S173.
natal maternal and fetal outcomes? Am J Obstet Gynecol 2013; 16. Yan J, Yang H. Gestational diabetes in China: Challenges and cop-
208(Suppl. 1): S114. ing strategies. Lancet Diabetes Endocrinol 2014; 2(12): 930–931.
30 Gestational diabetes mellitus,
obesity, and pregnancy
outcomes
Harold David McIntyre, Marloes Dekker-Nitert, Helen Lorraine Graham Barrett,
and Leonie Kaye Callaway
recognized that women identified in pregnancy as “GDMs”

Introduction carry a substantial long-term risk of permanent (generally
Diabetes and obesity are frequently referred to as the type 2) diabetes (T2DM), and indeed, this risk formed the
“twin epidemics” of the modern world. The prevalence basis for the original diagnostic thresholds proposed by
of both conditions is increasing worldwide, across high-, O’Sullivan and Mahan in 1964.5
middle-, and lower-income countries.1 Concurrently, In the modern context, the population prevalence of
the trend toward women having children at later ages, T2DM and lesser forms of dysglycemia including impaired
particularly in many developed countries, has led to an fasting glucose (IFG) and impaired glucose tolerance
increasing prevalence of gestational and preexisting dia- (IGT) has increased dramatically among women of child-
betes in pregnancy. 2 Gestational diabetes mellitus (GDM) bearing age.6 Therefore, especially in high T2DM preva-
and maternal obesity are both associated with a similar lence countries, many of the cases previously described as
spectrum of risk for a broad range of pregnancy com- GDM under older definitions that spoke of “any form of
plications. 3 This chapter aims to consider these risks hyperglycemia first identified in pregnancy” 7 would likely
together, analyzing their contribution to overall preg- have had abnormalities of glucose metabolism antedating
nancy outcomes from several perspectives: (1) underly- pregnancy but only identified through pregnancy-related
ing pathophysiology (common and discrete mechanisms), testing. Further, women with prepregnancy-undiagnosed
(2) combined clinical effects of obesity and GDM, (3) the T2DM have a much higher risk of severe complications
current evidence base for various modalities of treatment in pregnancy 8 and require urgent and intensive medical
with emphasis on the findings of randomized controlled attention.
trials, (4) population risks including the potential for pre- For this reason and for the purposes of the current chap-
vention, and (5) current evidence gaps, evidence–practice
ter, we have adopted the recently developed International
gaps, and areas for future research. While recogniz-
Association of Diabetes in Pregnancy Study Groups
ing that intergenerational effects are potentially of great
(IADPSG)9 or World Health Organization (WHO)10 defini-
importance, for reasons of space we are unable to consider
tions of GDM, which classify those women with more severe
these in detail in the current review.
disturbances of glucose metabolism (which would be con-
sistent with diabetes if noted outside pregnancy) into sepa-
rate categories described as “overt diabetes” (IADPSG) or
more simply “diabetes in pregnancy” (WHO). The relevant
Definitions thresholds for glycemic measures defining GDM are shown
The definition of GDM has been controversial for almost in Table 30.1.
the entire period since the term was first coined by The definition of obesity is less controversial, with most
Carrington.4 It is widely recognized that the hormonal authorities recognizing the WHO classification11 while
changes of pregnancy can convert some women from a acknowledging both the imperfect precision of body mass
state of normoglycemia (in the fasting or post–glucose index (BMI) as a measure of underlying adiposity and the
load state) to varying degrees of dysglycemia. However, fact that identical BMI thresholds are not uniformly appli-
from the very beginning of the study of GDM, it has been cable across all ethnic groups.
246
Metabolism in healthy pregnancy 247
Table 30.1 Definitions of gestational diabetes and (overt) diabetes in pregnancy
Category IADPSG WHO

“Overt diabetes”—IADPSG Fasting plasma glucose ≥7.0 mmol/L Fasting plasma glucose ≥7.0 mmol/L
“Diabetes in HbA1c ≥6.5% 2-hour plasma glucose ≥11.1 mmol/L
pregnancy”—WHO Random plasma glucose ≥11.1 mmol/L after 75 g glucose load
(with confirmation) Random plasma glucose ≥11.1 mmol/L
with symptoms
Gestational diabetes Fasting plasma glucose—5.1–6.9 mmol/L Fasting plasma glucose—5.1–6.9 mmol/L
Values obtained with a 75 g oral 1-hour plasma glucose ≥10.0 mmol/L 1-hour plasma glucose ≥10.0 mmol/L
glucose tolerance test 2-hour plasma glucose ≥8.5 mmol/L 2-hour plasma glucose—8.5–11.0 mmol/L
Abbreviations: IADPSG, International Association of Diabetes in Pregnancy Study Groups; WHO, World Health Organization; HbA1c,
glycosylated hemoglobin; mmol/L, millimoles per liter.
Pathophysiology of GDM and obesity circulating levels of glucose, insulin, cortisol, and leptin all
increase. The physiological increase in maternal insulin resis-
in pregnancy tance is an example of allostatic regulation: an adaptation of the
maternal physiology to a changing environment (Figure 30.1).
Pregnancy can be described as an allostatic state. Allostasis
is the process by which the body responds to stressors in
order to regain homeostasis or as achieving stability through Metabolism in healthy pregnancy
change. Where homeostasis is the regulation of physiology
around a set point, allostasis is the regulation of physiology In early pregnancy, maternal metabolism is anabolic and
to create a new set point. In an allostatic state, there is chronic results in maternal fat deposition. In healthy pregnancy,
activation of regulatory systems either due to some form of maternal insulin resistance increases throughout the second
dysregulation/dysfunction of physiological processes or due trimester with a peak in the third trimester due to secretion
to conflicting, competing, or opposing demands.12 The allo- of placental lactogen. In normal women, insulin sensitiv-
static load is the strain on physiology and regulatory capacity ity decreases by 50%–60% from prepregnancy to late preg-
due to sustained activation of regulatory systems. This could nancy.13 This decrease in insulin sensitivity (or increase in
also be defined as the regulatory load or metabolic load. insulin resistance) is usually overcome by an amplification
However, continuous activation of any regulatory system or of insulin secretion from the pancreatic beta cells, ensuring
activation to a level outside of its norm will eventually lead normoglycemia throughout pregnancy.
to failure of the system. Both obesity and pregnancy can be The rise in insulin resistance contributes to shifting
considered allostatic states. In combination, they may lead to maternal metabolism from an anabolic state to a catabolic
allostatic overload in which the compensatory systems may state. In the catabolic state, maternal metabolism becomes
fail resulting in adverse pregnancy outcomes. more reliant on lipids and ketones. This ensures an adequate
Outside of pregnancy, the brain is the main site for regulatory glucose supply to the developing fetus. Insulin sensitivity is
control in humans. In pregnancy, regulatory control is funda- inversely correlated to maternal plasma free fatty acid lev-
mentally altered due to the need for regulation of two organisms els.14 In addition to an increase in free fatty acid levels, other
simultaneously. Additionally, regulatory control is partially aspects of maternal lipid metabolism are also altered during
allocated to the placenta in addition to the brain. In pregnancy, pregnancy. Fat oxidation is significantly higher and a marked
Gestational Gestational Gut microbiome

hormones weight gain changes Placenta
Gut
microbiome
Late gestation
Genetics gulation
Allostatic re
Exercise Prepregnancy
Level of insulin resistance
Diet
BMI
Figure 30.1 Factors influencing insulin resistance before and during pregnancy. Allostatic regulation maintains euglycemia in
normal pregnancy.
248 Gestational diabetes mellitus, obesity, and pregnancy outcomes
hyperlipidemia occurs in late gestation. This includes an Inflammation in obesity and GDM
increase in the triglyceride content of very-low-density lipo-
protein (VLDL), low-density lipoprotein, and high-density in pregnancy
lipoprotein (HDL). Maternal plasma cholesterol also rises.
In recent decades, the perception of adipose tissue has shifted
The metabolic adaptations of healthy pregnancy are altered
from being regarded as a passive organ dedicated to fat stor-
in pregnancies affected by GDM and/or maternal obesity.15
age to a metabolically active organ contributing to the regu-
lation of whole-body metabolism through the synthesis and
Glucose metabolism in GDM secretion of adipokines, cytokines, and hormones. Obesity
and pregnancy are both states of chronic low-grade inflam-
and obesity mation sometimes termed metabolic inflammation or “meta-
inflammation.”18 In obese pregnancy, there is an increase in
In women with prepregnancy obesity, the pregnancy-
circulating proinflammatory cytokines, e.g., tumor necrosis
associated rise in insulin resistance further exacerbates
factor alpha (TNFα), interleukin-6 (IL-6), and C-reactive
the high level of insulin resistance present in obesity. For
protein.20,21 In obese pregnant women, the balance between
many obese women, the additional rise in insulin resistance
proinflammatory and anti-inflammatory cytokines is altered,
can still be compensated for by increased insulin secretion
and this may contribute to increased maternal insulin resis-
from the pancreatic beta cells. However, when the pancre-
tance, for example, by differentially regulating phosphoryla-
atic beta cells cannot muster this compensatory increase in
tion of the insulin receptor and its substrates.22
insulin secretion, hyperglycemia ensues and GDM is diag-
The placenta is an active endocrine organ, which contrib-
nosed. GDM is more prevalent in women with an increased
utes to the regulation of metabolism in both the mother and
prepregnancy BMI but can also develop in women with a
the developing fetus. The placenta synthesizes and secretes a
prepregnancy BMI in the normal range. Additionally, the
large array of hormones, cytokines, and metabolic signaling
examination of maternal glucose profiles using continuous
molecules. Microarray studies of placentas from overweight
glucose monitoring devices showed that maternal glucose
women with GDM have shown increased gene expression for
levels, while frequently below the diagnostic cutoff for GDM,
genes involved with inflammation and lipid metabolism but
are increased in obese women.16 Additionally, in obese
not glucose metabolism.23 In obese pregnancy, the placenta
women, adipokines including leptin and adiponectin, which
and the adipose tissue both regulate maternal metabolism
are involved in lipid and glucose metabolism, are altered.
although their regulation is not coordinated.
As noted earlier, pregnancy is also a state of low-grade
Lipid and adipose tissue metabolism inflammation with the expression of proinflammatory
cytokines from the placenta and the uterine epithelium.24
in GDM and obese pregnancies The triggers for this inflammatory response are not well
understood, but placental debris in the form of microparticles
Decreased insulin sensitivity not only leads to increased
known as syncytiotrophoblast membrane microparticles
availability of glucose but also of lipids. Obese women and
(STBM)25 or exosomes26 as well as placental-derived signal-
women with GDM have higher triglycerides and VLDL cho-
ing molecules may be implicated. Over time, the balance
lesterol and lower HDL cholesterol levels than normal-weight
between proinflammatory and anti-inflammatory cytokines
normoglycemic women from early pregnancy onward.17,18 The
shifts toward the anti-inflammatory cytokines in normal
subcutaneous adipose tissue of women with GDM shows lower
healthy pregnancy.27 In normal-weight women with GDM,
protein expression for the insulin receptor substrate 1 com-
levels of TNFα but not IL-6 are increased when compared
pared with pregnant women without GDM, whereas the levels
to matched controls.28 Furthermore, increased leukocyte
of insulin receptor substrate 2 are increased in fasting preg-
counts in early pregnancy are predictive of the development
nant women compared with nonpregnant women indepen-
of GDM, which is independent of BMI.29
dent of their glucose tolerance.14 Furthermore, it was recently
reported that in adipose tissue from obese women with and
without GDM, gene expression of many genes involved in Infant weight and body composition
fatty acid metabolism was decreased.19 This included fatty
acid uptake and intracellular transport, triglyceride synthe- The Pedersen hypothesis, first proposed in 1952, states that
sis, lipogenesis, and lipolysis. In late pregnancy, inhibition of macrosomia (e.g., a birthweight over 4000 g) results from
lipolysis by insulin through inhibition of hormone-sensitive fetal hyperglycemia and hyperinsulinism due to hypertro-
lipase is less effective. Gene expression for transcription fac- phy of fetal islets of Langerhans, which is the result of mater-
tors, including PPARγ, that regulate lipid metabolism is also nal hyperglycemia.30 Macrosomia can occur even when
reduced in obese pregnant women with GDM.14,19 These stud- maternal glucose control appears satisfactory and may be
ies show that insulin signaling in adipose tissue in women a result of increased maternal triglycerides and other lip-
with GDM is altered and could contribute to the excess insu- ids.15,17 Maternal obesity is associated with higher fetal fat
lin resistance seen in GDM. Lower expression of PPARγ in mass31 as well as fetal insulin resistance.32 Maternal obe-
adipose tissue in pregnancy may be a reflection of the acceler- sity is a predictor for fetal fat mass but not for lean mass.31
ated starvation with fasting late in pregnancy.14 These results suggest that maternal obesity specifically
Evidence base for treatment of GDM and obesity 249
affects fetal adiposity rather than overall fetal growth. GDM glycemia on pregnancy outcomes were similar. Both were
unaccompanied by obesity is also associated with fetal adi- associated with increased rates of the primary outcomes
posity.33 Birthweight and fat-free mass are both correlated (LGA babies, primary cesarean section, clinical neonatal
with insulin sensitivity in late gestation.13 hypoglycemia, and neonatal hyperinsulinemia) and impor-
tant secondary outcomes including fetal adiposity and pre-
eclampsia. In general, the associations of maternal BMI
Long-term influences with these outcomes tended to “plateau” in the highest cat-
egories, whereas those of glucose did not show this trend.45
Maternal obesity doubles the risk for childhood obesity34,35
It is important to remember that, for ethical/safety reasons,
and is associated with metabolic syndrome in the offspring.36
women were unblinded from the HAPO study if their glu-
Similarly, children born large for gestational age (LGA) in
cose levels exceeded the predefined thresholds, whereas no
the presence or absence of GDM also have increased risks
such limits were enforced for BMI.
of developing metabolic syndrome.36,37 In glucose-tolerant
The HAPO study group have also reported the combined
Pima Indian mothers, maternal glucose levels in the third
associations of BMI and GDM with adverse pregnancy out-
trimester were strongly associated with increased risk of
comes.3 Across the HAPO study, obesity was present in 13.7%
type 2 diabetes in the offspring.38
of participants and GDM by the IADPSG criteria9 was pres-
These increased risks may partially be determined by a
ent in 16.1% of those who remained blinded within the study.
genetic background predisposing to obesity as well as by the
Only 25% of the women with GDM were obese. Compared
postnatal environment related to the lifestyle of the family.
to women with neither GDM nor obesity, the adjusted odds
However, some of the increase in risk may result from an
ratios for most pregnancy complications were increased both
altered intrauterine environment programming the offspring
in women with obesity alone and in those with GDM alone.
for later disease, as suggested by the Developmental Origins
Preeclampsia appeared to be more prevalent in the “obesity
of Health and Disease (DOHaD) theory. Animal models
alone group,” while excess fetal growth and fetal hyperinsu-
have implicated altered epigenetic regulation affecting many
linemia were slightly more common in the “GDM alone group”
regulatory and metabolic organs in the offspring includ-
than in the “obesity alone group.” The combination of GDM
ing the brain, liver, pancreas,39 and adrenals. In humans,
and obesity was clearly associated with a marked increase in
long-term effects of the intrauterine environment have been
the risk of pregnancy complications. These data are presented
reported with maternal undernutrition in the Dutch hunger
graphically in Figure 30.2. The HAPO study also presented data
winter study.40 Placental function, which is a crucial deter-
dividing their cohort firstly by BMI into normal and under-
minant of the intrauterine environment, may therefore affect
weight/overweight/obese, with similar categorization applied to
the health of the offspring long beyond its own lifespan.
the composite oral glucose tolerance test “z” scores that were
GDM for the mother is associated with higher risks of
categorized as normal/intermediate/GDM. The definition of
hypertension, IGT, and hyperlipidemia, which are all com-
“intermediate” glycemia was selected to approximate the fre-
ponents of the metabolic syndrome.37 These increased risks
quency of overweight in the HAPO participants. As shown in
are especially pronounced in women who were obese prior
Figure 30.3, the frequency of LGA (adjusted birthweight >90th
to falling pregnant.37
centile) increased across both axes, with the combination of
GDM and obesity clearly carrying the highest risk.
GDM and obesity in combination In summary, the HAPO data clearly show that maternal
BMI and glycemia have independent and essentially additive
Many reports have attempted to dissect the relative clini- effects in terms of their associations with adverse pregnancy
cal and population health importance of GDM and obesity outcomes. In view of this, the relative “importance” of these
in terms of their effects on pregnancy outcomes and later factors is heavily influenced by the potential costs and ben-
maternal and infant health. Separation of their associations efits of preventative or treatment strategies.
is difficult, especially as the two conditions frequently coex-
ist and obesity commonly lies on the casual pathway toward
hyperglycemia. In some reports, women were not uniformly Evidence base for treatment
screened for hyperglycemia in pregnancy.41,42 In studies that of GDM and obesity
provide evidence of glucose screening of their entire cohort,
detailed analyses of the prevalence of GDM in overweight There are now two well-designed large prospective ran-
and obese women were not reported.43,44 In the vast major- domized controlled studies that confirm that diagnosis and
ity of studies, treatment of GDM (but generally not spe- treatment of gestational diabetes has short-term benefits
cific treatment of obesity) is also likely to have confounded for both mother and baby.46,47 Women in the Australian
reported results. Fortunately, these problems are less of an (Crowther) study had early pregnancy BMIs ranging from
issue when considering data from the Hyperglycemia and 22.9 to 31.2 kg/m2, and women in the intervention arm of
Adverse Pregnancy Outcome (HAPO) study,3,45 as this study the study had lower weight gain during pregnancy, with
was blinded in terms of glycemia and did not provide spe- less macrosomia, less LGA, and lower rates of preeclamp-
cific interventions for obesity. The HAPO study showed that sia.46 In the U.S. (Landon) study, the BMI at recruitment of
the associations of increasing maternal BMI and increasing women in the treatment arm was 30.1 ± 5 kg and control
7
No GDM, no obesity Obesity, no GDM
6
GDM, no obesity GDM + obesity
5
Adjusted odds ratio
4
0
BW > 90th Cord c-peptide Primary CS Preeclampsia NN body fat Shoulder
> 90th > 90th dystocia
Pregnancy outcome
Figure 30.2 Fully adjusted odds ratios for selected pregnancy complications in the Hyperglycemia and Adverse Pregnancy
Outcome study. The “No GDM, No Obesity” category served as the referent group for all comparisons. The other categories, as
labeled, refer to GDM alone, obesity alone, and the combination of both factors. Odds ratios refer to “Model II” as explained in
detail in the source publication, with full adjustment for potential confounders. GDM, gestational diabetes mellitus.
6 and obese women, and the results of this study were disap-
Glucose pointing.48 Essentially this study aimed to limit weight gain
5 Normal Intermediate GDM in overweight and obese pregnant women through lifestyle
intervention. This study demonstrated that despite three face-
Adjusted or for LGA
4 to-face visits (including two visits with a research dietitian),

three phone calls to assist in lifestyle modification, detailed
3 individualized dietary advice, and a program tailored to
stages of change theory, with detailed analysis of barriers to
2 change, and individualized assistance, there was no difference
in weight gain and no difference in any maternal or neonatal
1 outcome (except for a reduction in infants weighing over 4 kg,
although no difference in infants born LGA).
0 There have been numerous small randomized controlled
Normal/Uweight Over weight Obese
trials of lifestyle and other interventions in overweight and
BMI
obese women that have demonstrated reduced weight gain
Figure 30.3 Fully adjusted odds ratios for delivery of a in pregnancy, and some have demonstrated reduced neona-
large-for-gestational-age (LGA) infant, characterized as tal weight. However, other maternal and neonatal outcomes
birthweight >90th centile in the Hyperglycemia and Adverse have not been examined. These studies have been summa-
Pregnancy Outcome study participants. The group with rized in meta-analyses.49,50 To date, the benefits seen in these
normal glucose levels and normal or underweight range BMI
served as the referent group. The intermediate glucose group small studies were not demonstrated in practical interven-
was defined according to their mean standard deviation tions that are affordable at a population level.48
(“z”) scores for the fasting, 1-hour, and 2-hour glucose levels At present, it would appear that diagnosing and treating
during the diagnostic oral glucose tolerance test (OGTT). gestational diabetes in overweight and obese women has the
The values used to define this category were chosen to best evidence with regard to limiting weight gain, prevent-
achieve a frequency of intermediate glucose equivalent to
the frequency of overweight in the HAPO study cohort. BMI ing maternal adverse outcomes, and preventing neonatal
was measured at the time of the diagnostic OGTT and con- adverse outcomes. However, it is important to note that the
verted to equivalent WHO categories by regression analysis. long-term impacts of any of these interventions on the health
Uweight, underweight; GDM, gestational diabetes mellitus; of adult offspring are as yet unproven and will need to be
HAPO, hyperglycemia and adverse pregnancy outcome. carefully examined.
arm 30.2 ± 5.1 kg. Once again, in this study, women in the

intervention arm had lower weight gain, there were lower Population risks and prevention
rates of LGA and macrosomia in infants, and women had
lower rates of preeclampsia.47 Another consideration in addressing obesity and GDM is their
In contrast, there is only one completed trial (the “LIMIT” relative importance across the entire population of women
study) that is sufficiently powered to examine maternal and undergoing pregnancy. A number of studies have attempted
perinatal outcomes after a lifestyle intervention in overweight to address this issue. In this context, ascertainment bias
References 251
regarding glycemia in pregnancy and treatment confounding excessive fetal growth and adiposity and to hypertensive com-
due to active intervention for GDM remain major issues. plications of pregnancy. GDM and obesity commonly, but not
A recent report by Black et al.51 regarding a cohort of 9835 inevitably, coexist, and the development of common preventa-
women from Southern California with a high prevalence of tive and therapeutic strategies seems an attractive prospect.
overweight (32%) and obesity (28%) reported that, on a pop- Despite the epidemiologic data showing associations of
ulation basis, overweight and obesity accounted for 21.6% of GDM and obesity with adverse pregnancy outcomes, our
LGA in women without and 23.3% in women with GDM. understanding of the mechanisms underlying these associ-
In their cohort, 75% of women who developed GDM were ations remains limited. The Pedersen hypothesis has served
overweight or obese. This study also emphasized the impor- as a useful framework for considering the pathogenesis of
tance of gestational weight gain as a determinant of LGA. hyperglycemia in pregnancy,18 but even our most energetic
These conclusions depend heavily on the background popu- approaches to achieving normoglycemia have not normal-
lation, as in the global HAPO study cohort, the prevalences ized pregnancy outcomes, particularly with the addition of
of overweight (22%) and obesity (14%) were much lower.52 obesity as a common comorbid condition. Specific therapy
Similar findings have been noted in other studies. Kim targeted a pregnancy dyslipidemia in GDM and obesity
et al. reported on a cohort from Florida.53 Twenty percent of seems attractive in principle but difficult in practice due to
this cohort were obese and 24% overweight. These authors safety concerns.15 The exploration of metabolic inflamma-
concluded that gestational weight gain was the major poten- tion and its consequences in pregnancy may also improve
tially modifiable factor in the prevention of LGA births. our mechanistic understanding,18,24,55,56 but again the ther-
Alberico et al. reported on a cohort from Italy54 and con- apeutic options with current agents appear to be limited.
cluded that maternal obesity, GDM, and gestational weight As previously noted, interventions commenced during
gain all warranted careful assessment during pregnancy. pregnancy appear to have limited efficacy in preventing
In summary, it is clear that, especially in populations with the complications of obesity. This aligns with the observa-
a high prevalence of overweight and obesity, these factors tion that, in obese women (in contrast to women of normal
constitute a large fraction of the population risk of LGA. weight), prepregnancy weight is more closely associated with
However, as noted previously, therapeutic strategies for neonatal adiposity than weight gain during pregnancy.57
addressing obesity in pregnancy have proved disappointing, Thus, preventative and therapeutic measures may need to be
in contrast to the positive results noted for the “glucocentric” initiated preconception to reap positive benefits.
treatment of GDM. By contrast, the efficacy of therapy for GDM is well
demonstrated. Effective implementation of GDM diag-
nostic and therapeutic strategies58–60 appears to have the
greatest potential for proven benefit at the current time.
Evidence gaps and evidence: Practice Future research should concentrate on delineation of
gaps and future research the pathogenesis of obesity and GDM through basic and
clinical studies; the development of effective strategies
From the evidence presented earlier, it is clear that both GDM to target both conditions before, during, and after preg-
and maternal obesity are associated with a similar spectrum nancy; and the effective implementation of treatments that
of adverse pregnancy outcomes, especially those related to are known to be effective but are currently underutilized.
REFERENCES
1. Hanson MA, Gluckman PD, Ma RC, Matzen P, Biesma RG. Early 8. Omori Y, Jovanovic L. Proposal for the reconsideration of the defini-
life opportunities for prevention of diabetes in low and middle tion of gestational diabetes. Diabetes Care 2005; 28: 2592–2593.
income countries. BMC Public Health 2012; 12: 1025. 9. Metzger BE, Gabbe SG, Persson B et al. International association
2. Guariguata L, Linnenkamp U, Beagley J, Whiting DR, Cho NH. of diabetes and pregnancy study groups recommendations on
Global estimates of the prevalence of hyperglycaemia in preg- the diagnosis and classification of hyperglycemia in pregnancy.
nancy. Diabetes Res Clin Pract 2014; 103: 176–185. Diabetes Care 2010; 33: 676–682.
3. Catalano PM, McIntyre HD, Cruickshank JK et al. The 10. Colagiuri S, Falavigna M, Agarwal MM et al. Strategies for imple-
Hyperglycemia and Adverse Pregnancy Outcome Study: menting the WHO diagnostic criteria and classification of hyper-
Associations of GDM and obesity with pregnancy outcomes. glycaemia first detected in pregnancy. Diabetes Res Clin Pract
Diabetes Care 2012; 35: 780–786. 2014; 103: 364–372.
4. Carrington ER, Shuman CR, Reardon HS. Evaluation of the 11. WHO. Obesity: Preventing and managing the global epidemic.
prediabetic state during pregnancy. Obstet Gynecol 1957; 9: Report of a WHO consultation. World Health Organ Tech Rep
664–669. Ser 2000; 894: i–xii, 1–253.
5. O’Sullivan JB, Mahan CM. Criteria for the oral glucose tolerance 12. Power ML, Schulkin J. Maternal obesity, metabolic disease, and
test in pregnancy. Diabetes 1964; 13: 278–285. allostatic load. Physiol Behav 2012; 106: 22–28.
6. McIntyre HD. Diagnosing gestational diabetes mellitus: 13. Catalano PM, Ehrenberg HM. The short- and long-term impli-
Rationed or rationally related to risk? Diabetes Care 2013; 36: cations of maternal obesity on the mother and her offspring.
2879–2880. BJOG 2006; 113: 1126–1133.
7. Metzger BE, Buchanan TA, Coustan DR et al. Summary and rec- 14. Catalano PM, Nizielski SE, Shao J, Preston L, Qiao L, Friedman
ommendations of the Fifth International Workshop-Conference JE. Downregulated IRS-1 and PPARg in obese women with ges-
on Gestational Diabetes Mellitus. Diabetes Care 2007; 30(Suppl. 2): tational diabetes: Relationship to FFA during pregnancy. Am J
S251–S260. Physiol Endocrinol Metab 2002; 282: E522–E533.
15. Barrett HL, Dekker Nitert M, McIntyre HD, Callaway LK. 40. Roseboom TJ, Painter RC, van Abeelen AFM, Veenendaal
Normalizing metabolism in diabetic pregnancy: Is it time to MVE, de Rooij SR. Hungry in the womb: What are the conse-
target lipids? Diabetes Care 2014; 37: 1484–1493. quences? Lessons from the Dutch famine. Maturitas 2011; 70:
16. Harmon KA, Gerard L, Jensen DR et al. Continuous glucose 141–145.
profiles in obese and normal-weight pregnant women on a con- 41. Callaway LK, Prins JB, Chang AM, McIntyre HD. The prevalence
trolled diet: Metabolic determinants of fetal growth. Diabetes and impact of overweight and obesity in an Australian obstetric
Care 2011; 34: 2198–2204. population. Med J Aust 2006; 184: 56–59.
17. Barrett HL, Gatford KL, Houda CM et al. Maternal and neona- 42. Sebire NJ, Jolly M, Harris JP et al. Maternal obesity and preg-
tal circulating markers of metabolic and cardiovascular risk in nancy outcome: A study of 287,213 pregnancies in London. Int J
the Metformin in Gestational Diabetes (MiG) Trial: Responses Obes Relat Metab Disord 2001; 25: 1175–1182.
to maternal metformin versus insulin treatment. Diabetes Care 43. Cnattingius S, Bergstrom R, Lipworth L, Kramer MS. Prepregnancy
2013; 36: 529–536. weight and the risk of adverse pregnancy outcomes. N Engl J Med
18. Catalano PM, Hauguel-De Mouzon S. Is it time to revisit the 1998; 338: 147–152.
Pedersen hypothesis in the face of the obesity epidemic? Am J 44. Sermer M, Naylor CD, Farine D et al. The Toronto Tri-Hospital
Obstet Gynecol 2011; 204: 479–487. Gestational Diabetes Project. A preliminary review. Diabetes
19. Lappas M. Effect of pre-existing maternal obesity, gestational dia- Care 1998; 21(Suppl. 2): B33–B42.
betes and adipokines on the expression of genes involved in lipid 45. HAPO Study Cooperative Research Group. Hyperglycaemia and
metabolism in adipose tissue. Metabolism 2014; 63: 250–262. Adverse Pregnancy Outcome (HAPO) Study: Associations with
20. Richardson AC, Carpenter MW. Inflammatory mediators in maternal body mass index. BJOG 2010; 117: 575–584.
gestational diabetes mellitus. Obstet Gynecol Clin North Am 46. Crowther CA, Hiller JE, Moss JR, McPhee AJ, Jeffries WS,
2007; 34: 213–224, viii. Robinson JS. Effect of treatment of gestational diabetes mellitus
21. Mastorakos G, Valsamakis G, Papatheodorou DC et al. The on pregnancy outcomes. N Engl J Med 2005; 352: 2477–2486.
role of adipocytokines in insulin resistance in normal preg- 47. Landon MB, Spong CY, Thom E et al. A multicenter, random-
nancy: Visfatin concentrations in early pregnancy predict insu- ized trial of treatment for mild gestational diabetes. N Engl J Med
lin sensitivity. Clin Chem 2007; 53: 1477–1483. 2009; 361: 1339–1348.
22. Grimble RF. Inflammatory status and insulin resistance. Curr 48. Dodd JM, Turnbull D, McPhee AJ et al. Antenatal lifestyle advice
Opin Clin Nutr Metab Care 2002; 5: 551–559. for women who are overweight or obese: LIMIT randomised trial.
23. Radaelli T, Varastehpour A, Catalano P, Hauguel-de Mouzon S. Br Med J 2014; 348: g1285.
Gestational diabetes induces placental genes for chronic stress 49. Dodd JM, Grivell RM, Crowther CA, Robinson JS. Antenatal
and inflammatory pathways. Diabetes 2003; 52: 2951–2958. interventions for overweight or obese pregnant women: A
24. Sacks G, Sargent I, Redman C. An innate view of human preg- systematic review of randomised trials. BJOG 2010; 117:
nancy. Immunol Today 1999; 20: 114–118. 1316–1326.
25. Messerli M, May K, Hansson SR et al. Feto-maternal interac- 50. Oteng-Ntim E, Varma R, Croker H, Poston L, Doyle P. Lifestyle
tions in pregnancies: Placental microparticles activate periph- interventions for overweight and obese pregnant women to
eral blood monocytes. Placenta 2010; 31: 106–112. improve pregnancy outcome: Systematic review and meta-analy-
26. Clifton VL, Stark MJ, Osei-Kumah A, Hodyl NA. Review: The sis. BMC Med 2012; 10: 47.
feto-placental unit, pregnancy pathology and impact on long 51. Black MH, Sacks DA, Xiang AH, Lawrence JM. The relative con-
term maternal health. Placenta 2012; 33(Suppl.): S37–S41. tribution of prepregnancy overweight and obesity, gestational
27. Denney JM, Nelson EL, Wadhwa PD et al. Longitudinal modu- weight gain, and IADPSG-defined gestational diabetes mellitus
lation of immune system cytokine profile during pregnancy. to fetal overgrowth. Diabetes Care 2013; 36: 56–62.
Cytokine 2011; 53: 170–177. 52. McIntyre HD, Catalano PM. Comment on: Black et al. The rela-
28. Lopez-Tinoco C, Roca M, Fernandez-Deudero A et al. Cytokine tive contribution of prepregnancy overweight and obesity, ges-
profile, metabolic syndrome and cardiovascular disease risk in tational weight gain, and IADPSG-defined gestational diabetes
women with late-onset gestational diabetes mellitus. Cytokine mellitus to fetal overgrowth. Diabetes Care 2013; 36: 56–62.
2012; 58: 14–19. [Diabetes Care 2013; 36: e127].
29. Wolf M, Sauk J, Shah A et al. Inflammation and glucose intol- 53. Kim SY, Sharma AJ, Sappenfield W, Wilson HG, Salihu HM.
erance: A prospective study of gestational diabetes mellitus. Association of maternal body mass index, excessive weight gain,
Diabetes Care 2004; 27: 21–27. and gestational diabetes mellitus with large-for-gestational-age
30. Pedersen J. Diabetes and pregnancy; blood sugar of newborn births. Obstet Gynecol 2014; 123: 737–744.
infants during fasting and glucose administration. Nord Med 54. Alberico S, Montico M, Barresi V et al. The role of gestational
1952; 47: 1049. diabetes, pre-pregnancy body mass index and gestational
31. Sewell MF, Huston-Presley L, Super DM, Catalano P. Increased weight gain on the risk of newborn macrosomia: Results from a
neonatal fat mass, not lean body mass, is associated with prospective multicentre study. BMC Pregnancy Childbirth 2014;
maternal obesity. Am J Obstet Gynecol 2006; 195: 1100–1103. 14: 23.
32. Catalano PM, Presley L, Minium J, Hauguel-de Mouzon S. 55. Challier JC, Basu S, Bintein T et al. Obesity in pregnancy stimu-
Fetuses of obese mothers develop insulin resistance in utero. lates macrophage accumulation and inflammation in the pla-
Diabetes Care 2009; 32: 1076–1080. centa. Placenta 2008; 29(3): 274–281.
33. Catalano PM, Thomas A, Huston-Presley L, Amini SB. Increased 56. Desoye G, Hauguel-de Mouzon S. The human placenta in ges-
fetal adiposity: A very sensitive marker of abnormal in utero tational diabetes mellitus. The insulin and cytokine network.
development. Am J Obstet Gynecol 2003; 189: 1698–1704. Diabetes Care 2007; 30(Suppl. 2): S120–S126.
34. Whitaker RC. Predicting preschooler obesity at birth: The role of 57. Waters TP, Huston-Presley L, Catalano PM. Neonatal body
maternal obesity in early pregnancy. Pediatrics 2004; 114: e29–e36. composition according to the revised institute of medicine
35. Reilly JJ, Armstrong J, Dorosty AR et al. Early life risk factors for recommendations for maternal weight gain. J Clin Endocrinol
obesity in childhood: Cohort study. Br Med J 2005; 330: 1357. Metab 2012; 97: 3648–3654.
36. Boney CM, Verma A, Tucker R, Vohr BR. Metabolic syndrome in 58. Marseille E, Lohse N, Jiwani A et al. The cost-effectiveness of
childhood: Association with birth weight, maternal obesity, and gestational diabetes screening including prevention of type
gestational diabetes mellitus. Pediatrics 2005; 115: e290–e296. 2 diabetes: Application of a new model in India and Israel.
37. Vohr BR, Boney CM. Gestational diabetes: The forerunner for the J Matern Fetal Neonatal Med 2013; 14: 14.
development of maternal and childhood obesity and metabolic 59. Jiwani A, Marseille E, Lohse N, Damm P, Hod M, Kahn JG.
syndrome? J Matern Fetal Neonatal Med 2008; 21: 149–157. Gestational diabetes mellitus: Results from a survey of country
38. Franks PW, Looker HC, Kobes S et al. Gestational glucose toler- prevalence and practices. J Matern Fetal Neonatal Med 2012; 25:
ance and risk of type 2 diabetes in young Pima Indian offspring. 600–610.
Diabetes 2006; 55: 460–465. 60. Lohse N, Marseille E, Kahn JG. Development of a model to
39. Sandovici I, Smith NH, Nitert MD et al. Maternal diet and aging assess the cost-effectiveness of gestational diabetes melli-
alter the epigenetic control of a promoter-enhancer interaction at tus screening and lifestyle change for the prevention of type
the Hnf4a gene in rat pancreatic islets. Proc Natl Acad Sci USA 2 diabetes mellitus. Int J Gynaecol Obstet 2011; 115(Suppl. 1):
2011; 108: 5449–5454. S20–S25.
31 Obesity versus glycemic control:
Which contributes more to
adverse pregnancy outcome?
Amir Aviram and Yariv Yogev
Introduction Weight gain, obesity, and pregnancy

Obesity has long been recognized as a global health con- The amount of weight gain recommended in pregnancy is
cern, be it among adults, adolescents, or children, of both controversial. Historically, obstetricians used to restrict
sexes. Over the past three decades, the prevalence of overweight gain to 15 lb, regardless of race, ethnicity, or prepreg-
weight and obese adults has more than doubled, with over nancy weight. In the 1970s, a more lax approach to weight
a third of women aged 20 years and above being overweight gain was followed, allowing weight gain for up to 20–27 lb.
and 14% obese.1 According to the latest data from the In 1990, the Institute of Medicine (IOM) published guide-
Organization for Economic Cooperation and Development lines based on the effects of weight gain on fetal size. The
(OCED), more than half of the adult population in 20 out new recommendations were based on prepregnancy BMI:
of 34 OECD countries are overweight or obese.2 The World weight gain of 28–40 lb for BMI of 19.8 and lower, 25–35 lb
Health organization (WHO) and the National Institute for for BMI of 19.9–26, and 15–25 lb for BMI of 26.1 and greater.
Health (NIH) define overweight as a body mass index (BMI) The IOM stated that the effect of weight gain on fetal size
of 25–29.9 and obesity as a BMI of 30 or greater. Obesity diminishes as the mother’s prepregnancy BMI increases.7
is also subcategorized into three subgroups: class I (BMI of This approach considered only immediate fetal outcomes
30–34.9), class II (BMI of 35–39.9), and class III (BMI of 40 and disregarded long-term maternal and neonatal effects.
or greater).3 This concept was challenged in the past decades when stud-
As obesity becomes an ever-growing concern, the num- ies evaluated the association between maternal weight gain,
ber of women of reproductive age who are overweight or obesity, pregnancy outcome, and future development of dia-
obese increases, and the incidence of obesity among preg- betes in the mother and the child.
nant women is now estimated between 18.5% and 38.3%.4 Rooney et al.8 evaluated a cohort of 540 women who had
Maternal overweight is now a known risk factor that affects documented weight over a 5-year postpartum period. They
the vast continuum of pregnancy. Fertility and fecundity concluded that excess weight gain and failure to lose weight
rates are lower among overweight and obese women, in after pregnancy are important and identifiable predictors
spontaneous conception as well as in artificial reproduc- of long-term obesity. Edwards9 reported that obese patients
tive techniques.4 During pregnancy, these women are more gained an average of 5 kg less during pregnancy and were
susceptible to pregnancy hypertensive disorders, gestational more likely to lose weight or not gain weight at all com-
diabetes mellitus (GDM), respiratory complications, and pared with normal-weight women. Obese women who lost
thromboembolic events.4–6 Numerous studies have exam- or did not gain any weight had lower mean birth weights
ined the impact of obesity on labor-related complications. and higher rates of small-for-gestational-age (SGA) infants
As delivery approaches, overweight women have slower in comparison to obese women who gained 1 lb (0.45 kg) or
labor progression rate, higher rate of cesarean deliveries, and more. The incidence of macrosomic fetuses increased sig-
more surgery-related complications such as difficult spinal, nificantly only in the group that gained 12 kg or higher. No
epidural, or general anesthesia, wound infection, and endo- weight gain and weight gain up to 11.5 kg were associated
myometritis.4–6 As for the fetus and neonate, complications with a macrosomia rate of 12.5%–13.3% compared with a
include congenital malformations, large-for-gestational-age background rate of 10% in nonobese women. Therefore, they
(LGA) infants, stillbirths, shoulder dystocia, and adolescent recommend weight gains of 15–25 lb for obese women and
complications such as obesity and diabetes.4–6 25–35 lb for normal-weight women to optimize fetal growth.
253
254 Obesity versus glycemic control
Bianco et al.10 reported that weight gain of more than mothers were fed with high fat diet during pregnancy and
25 lb was strongly associated with birth of LGA infants. lactation, as well as others fed with high fat diet from wean-
However, poor weight gain did not appear to increase ing up through pregnancy, were hypoleptinemic and hypo-
the risk of low-birth-weight infants. Luke and associates insulinemic at birth, compared with controls. As adults
reported that for every kilogram of gestational weight nourished with postweaning high fat diet, these rats were
gained, birth weight increased by 44.9 g for underweight significantly more obese than controls and characterized
women, 22.9 g for normal-weight women, and 11.9 g for with hyperinsulinemia and hyperleptinemia. It is of inter-
overweight women. For every kilogram of retained weight, est that a lifetime of high fat nutrition produced a pheno-
birth weight was increased by 35.6 g for underweight type similar to high fat nutrition restricted to pregnancy and
women, 15.9 g for normal-weight women, and 5.1 g for lactation alone, thus suggesting that the postnatal sequelae
overweight women.11 These findings suggest that beyond a of maternal high fat nutrition occurs independently of pre-
certain level of weight gain, there is an increase in birth conceptional diet. Laitinen and colleagues19 found that
weight at the expense of increasing maternal postpartum children of overweight or obese women had higher BMI at
obesity for women who have gained an excessive amount of different ages until 31 years old than did children of normal-
weight during pregnancy. or underweight mothers. At the age of 31, measures of cir-
As the prevalence of obesity increased and evidences cumferences of the waist and hip and waist-to-hip ratio were
regarding adverse pregnancy outcomes in obese pregnant higher in those born to mothers with higher BMI. Forsen
women accumulated, the IOM issued new guidelines for et al.20 showed that Finish men whose mothers had a high
pregnancy weight gain in 2009, taking into account both BMI in pregnancy had a 1.24 hazard ratio for coronary heart
maternal and fetal health.12 The recommended weight disease for every standard deviation increase in the moth-
gain for underweight women (BMI < 18.5) is 12.5–18 kg, ers’ BMI. In south India, glucose and insulin metabolism in
for normal-weight women (BMI 18.5–24.9) 11.5–16 kg, 506 men and women aged 39–60 were tested. The research-
for overweight women (BMI 25.0–29.9) 7–11.5 kg, and for ers noticed that those with non-insulin-dependent diabetes
obese women (BMI 30 and greater) 5–9 kg. Hull and col- mellitus were born to mothers who tended to be heavier than
leagues examined whether differences exist in infant body average during pregnancy.21
composition based on the new IOM guidelines.13 It was
found that infants of obese mothers had greater percent fat
compared with infants of normal-weight and overweight
mothers. Within the excessive weight gain group, infants of Obesity and gestational diabetes
normal-weight mothers had less percent fat than infants of mellitus
obese or overweight mothers. Another study by Vesco et al.
found that obese women gaining weight above the IOM The association of obesity and carbohydrate intolerance in
2009 recommendations did not decrease the risk for SGA pregnancy is well established. Bianco performed a retrospec-
but increased the risk for delivering LGA or macrosomic tive study comparing 613 patients with BMI of >35 with more
infants, and that obese women gaining weight below the than 11,000 women with normal BMI, and found that the
IOM recommendations had higher risk for delivering SGA rates of type 2 diabetes and GDM were 4.5-fold and 3.3-fold
and lower risk for LGA infants.14 Bodnar and associates higher, respectively, among the obese group.10 Kumari et al.
demonstrated that the prevalence of excessive gestational examined pregnancy outcomes among women with BMIs of
weight gain (GWG) declined, and weight loss increased, as >40 and found a GDM rate of 24.5% versus 2.2% among nor-
obesity became more severe.15 Generally, weight loss was mal BMI patients.22 In one of the largest studies concerning
associated with an elevated risk of SGA, medically indi- obesity in pregnancy, including more than 287,000 patients
cated and spontaneous preterm delivery, and high weight (of which approximately 79,000 are overweight patients and
gain tended to increase the risk of LGA and medically indi- 31,000 are obese), the odds ratio (OR) for developing GDM
cated preterm delivery. was 1.68 for overweight patients and 3.6 for obese patients.23
Getahun and colleagues examined if BMI changes between Weiss et al.24 published the results of their 16,102 patients
two consecutive pregnancies are associated with increased and demonstrated that a BMI of 30–34.9 was associated with
risk for LGA in the second pregnancy.16 They found that an OR of 2.6 for developing GDM, and that a BMI of 35 and
women who were overweight or obese in both pregnancies greater held an OR of 4.0 for GDM. Other studies of different
had increased risk for LGA infant, and that any decrease in cohort sizes supported these observations.25,26 Super obesity
BMI attenuated the risk. They also concluded that 17.1% of (BMI equal to or greater than 50) was found to be associated
LGA infants of underweight mothers, 13.2% of LGA infants with a higher rate of GDM compared with normal-weight or
of normal-weight mothers, and 7.6% of LGA infants of over- other obesity classes’ women.27 A meta-analysis of 20 studies
weight mothers could be prevented if BMI had not increased concluded that the OR for developing GDM are 2.15 (95%
between pregnancies. CI 1.82–2.53), 3.56 (3.05–4.21), and 8.56 (5.07–16.04) among
Obesity and its late sequelae may prove to be one of the overweight, obese, and severely obese gravidas, respectively.28
greatest threats to adult health in the developing world in Another study sought to evaluate the percentage of GDM
this century.17 Different experiments on animals led to some attributable to overweight and obesity. The percentages of
interesting results. Howie et al.18 found that rats whose GDM attributable to BMI were calculated to be 15.4%, 9.7%,
Glycemic control 255
and 21.1% for overweight, obese, and severely obese, respec- Glycemic control
tively. The authors concluded that if all overweight and obese
women had the same GDM risk as normal-weight women, The second question raised earlier was whether glycemic
nearly half of GDM cases would be prevented.29 control goals of obese and overweight patients different from
Apart from the well-established correlation between those of lean GDM patients and what treatment adjustments
type 2 diabetes among nonpregnant women and obesity, are needed for achieving those goals.
attempts have been made to further elucidate the associa- Early in 1987, Leikin and colleagues tried to solve this
tion between obesity and GDM. Colomiere and associates dilemma. They conducted a prospective study in which
published their work about insulin signaling and glucose they divided GDM patients into two groups. The first group
transporter expression in maternal skeletal muscle and had fasting serum glucose levels of 95 mg/dL or less on oral
subcutaneous adipose tissue. They found that both skeletal glucose tolerance test and was treated with diet alone, and
muscle and adipose tissue from obese women demonstrated those with higher levels were treated with diet and insulin.
lower GLUT-1 and GLUT-3 mRNA expression, leading them The diet-treated group and lean GDM patients receiving
to conclude that this may be the cause of GDM.30 insulin treatment had the same rate of macrosomia as the
Not until long ago, GDM was not proved to result in nondiabetic group. However, obese GDM patients treated
adverse pregnancy outcome. Only in 2005 did the pioneer- with insulin demonstrated a higher rate of macrosomia.41
ing work of Crowther and associates31 as well as Langer Another study aimed at answering this complex question
and associates32 demonstrated that the treatment of GDM was performed by Schaefer-Graf et al.42 They evaluated dif-
reduces perinatal complications. Later, Landon et al. reiter- ferent parameters among GDM patients and found that while
ate these results.33 In fact, it was also shown that mild car- mean level of glycemic control was similar among mothers
bohydrate intolerance, in the absence of confirmed GDM, of macrosomic and nonmacrosomic neonates (except for
results in various adverse outcomes.34,35 fasting glucose levels at 32–35 weeks), macrosomia rates were
Since obesity is a substantial risk factor for GDM, and higher among the obese patients. They concluded that the
since GDM may lead to adverse pregnancy outcomes, association between glycemic control and macrosomia was
a series of dilemmas arise: is there a difference between limited, while obesity appeared to be a strong risk factor for
overweight or obese and normal-weight GDM patient, and macrosomia throughout pregnancies with GDM. In their
if there is, are glycemic control goals of obese and over- words, “in obese women, the high fetal macrosomia rate did
weight patients different from those of lean GDM patients not appear to be normalized by therapy based on maternal
and what treatment adjustments are needed for achieving euglycemia.”
those goals? Yogev and associates sought to determine whether preg-
Several investigators sought the answer for the first ques- nancy outcomes differ between obese (BMI above 30 but less
tion. Ben-Haroush et al. found that the prevalence of LGA than 35) and morbidly obese (BMI of 35 or greater) GDM
neonates was higher in obese (BMI of ≥30) GDM patients patients by analyzing retrospectively a cohort of 4830 patients
compared with normal-weight GDM patients.36 Another with GDM. The rates of obesity and morbid obesity were
study compared four groups of patients: reference group 15.7% and 11.6%, respectively. Similar rates of cesarean section,
(normal weight, no GDM), normal-weight GDM, obese fetal macrosomia, shoulder dystocia, composite outcome, and
GDM, and obese with no GDM. It was found that obese metabolic complications were noted. Of note, no differences
GDM patients had higher rates of cesarean deliveries, pre- were found in the rate of desired glycemic control, although a
eclampsia, macrosomia, and higher mean birth weights than higher rate of insulin treatment was needed among the mor-
lean GDM patients. Additionally, after accounting for con- bidly obese.43 Another study by Joy examined the impact of
founders, obese GDM patients had a two- to threefold higher obesity on pregnancy outcomes among GDM patients treated
risk for macrosomia and cesarean deliveries, while the lean by diet, oral hypoglycemic agent, or insulin. Although gly-
GDM patients had no such increased risk.37 A subanalysis cemic control data were missing, in each group, obesity was
of the results of the Hyperglycemia and Adverse Pregnancy associated with adverse outcome, suggesting its pivotal role.39
Outcome (HAPO) study among more than 23,000 patients Langer et al. performed a study comparing four groups of
revealed similar results. The adjusted OR for LGA neonate, patients (diet treated with good glycemic control, diet treated
cesarean delivery, preeclampsia, and shoulder dystocia were with poor glycemic control, insulin treated with good glyce-
calculated for the same four groups as depicted earlier. The mic control, and insulin treated with poor glycemic control)
results demonstrated an OR of 3.3 for LGA neonate among stratified by weight. It was found that obese and overweight
obese GDM (compared with 1.99 for lean GDM), an OR of GDM patients treated with insulin who achieved good glyce-
2.3 for cesarean delivery (compared with OR 1.5), an OR mic control showed no increased risk of composite outcome,
5.3 for preeclampsia (compared with 1.7), and an OR of 1.7 macrosomia, and LGA in comparison with normal-weight
for shoulder dystocia (compared with nonsignificant OR).38 GDM patients.44 In contrast, obese patients with good glyce-
Furthermore, obesity was found to worsen maternal and mic control achieved by diet still demonstrated higher risk
perinatal outcome regardless of treatment modality.39 It was for composite outcome, LGA, and macrosomia. Poorly con-
suggested that one mechanism responsible for LGA neonates trolled overweight and obese patients, regardless of treatment
among obese GDM women is hypertriglyceridemia, despite modality, had significantly higher rates of composite outcome,
glycemic control.40 metabolic complications, macrosomia, and LGA. It may be
that although obesity, in and of itself, translates into potential Intervention during pregnancy is also available. Harrison
adverse outcomes in pregnancy, GDM women treated with and colleagues randomized 228 women at risk of developing
insulin and possibly oral antidiabetic drugs who achieve tar- GDM to a four-session lifestyle program aimed at acquir-
geted levels of glycemic control will have pregnancy outcomes ing adequate GWG or controls. They found that by 28 weeks,
comparable to those of normal-weight women. The improved GWG was significantly lower among women in the inter-
outcome in the insulin-treated overweight and obese women vention group, and a trend was noticed toward less GDM.51
may be due to an unidentified effect of insulin itself on the Another study examined the effect of gestational weight
fetus or the activation of other metabolic fuel pathways. loss among overweight and obese women after the diag-
Most and Langer later published their results concerning nosis of GDM was established. Gestational weight loss was
the impact of maternal weight gain on fetal growth among associated with less macrosomia, NICU admissions, and
GDM patients in relation to treatment modality, BMI, and cesarean deliveries, yet rates of SGA neonates and deliver-
glycemic control. Among obese or overweight diet-treated ies <34 weeks increased.52 A study of similar design demon-
women, the rate of LGA was fourfold higher, compared with strated lower mean birth weight among overweight and class
insulin-treated women.45 I obese women with gestational weight loss.53 Park and asso-
Olmos and associates studied the impact of obesity and ciates found that overweight/obese GDM patients gaining
glycemic control on rates of macrosomia.46 They defined less weight than recommended by the IOM 2009 guidelines
glycemic control using mean blood glucose and HbA1c had better neonatal outcomes and better glycemic control
levels. They reported that the prevalence of macrosomia and lower rates of insulin treatment.54
was 14.9% among their patients with good glycemic con- A randomized trial of weight-bearing exercise (versus
trol versus 28.1% in those with lesser glycemic control. They no exercise) from 8 weeks of pregnancy in women who did
also reported macrosomia rates of 10.4% in the normal- not normally exercise resulted in significantly higher birth
weight group compared with 24.6% in the overweight group, weights in offspring of women randomized to exercise.55 In
although both groups had similar glycemic control levels another study, exercise two times/week or less was associ-
and GWG. Their conclusion was that good glycemic control ated with low birth weight (adjusted OR, 2.64 95% CI 1.29–
was not sufficient to reduce macrosomia rates in the presence 5.39) relative to women who exercised three to four times/
of overweight or obesity. week after adjustment for potential confounders including
social class.56 Dye et al.57 demonstrated that women who
were obese at the time of conception but exercised regularly
had lower rates of GDM. A meta-analysis published in 2011
Potential management and concluded that prepregnancy physical activity yielded a
intervention pooled OR of 0.45 (95% CI 0.28–0.75) for developing GDM,
and that exercise in early pregnancy was also protective
In obese women, a modification of risk factors prior to or (OR 0.76, 95% CI 0.70–0.83). Several studies showed that
early in pregnancy is recommended. Regarding prepregnancy moderate exercise in pregnancy did not appear to be harm-
interventions aimed at reducing the risk for GDM, Villamor ful and that there was no association with premature labor
and Cnattingius found that compared with women whose or poor Apgar scores, thus reassuring the safety of exercise
BMI changed between −1.0 and 0.9 units, women who gained during pregnancy.58,59
three or more units had increased risk of preeclampsia, gesta-
tional hypertension, GDM, cesarean delivery, stillbirth, and
LGA birth.47 The associations were linearly related to weight Summary
change and were also noted in women who had a healthy pre-
pregnancy BMI in both pregnancies. Ehrlich and colleagues In conclusion, obesity is a risk factor for adverse pregnancy
sought to estimate the association between interpregnancy outcomes, including GDM. The hazardous combination
change in BMI and the risk of GDM in a second pregnancy. of overweight and obesity with GDM increases the risks
They found that compared with women who had no change involved with GDM or obesity alone. It seems that while gly-
in BMI, those whose BMI increased had an increased risk cemic control is of utmost importance in GDM, the role of
of GDM. Conversely, those whose BMI decreased had lower obesity may be even more critical than previously believed.
risk of GDM only if they were overweight or obese during the We may hypothesize that the presence of obesity or over-
first pregnancy and lost at least two BMI units.48 weight in a GDM patient alters metabolic pathways, thus
Dell’Agnolo and colleagues49 found that women who mandating tighter and stricter glycemic control. Therefore,
underwent bariatric surgery had less obesity-related comor- the desired and adequate levels of glycemic control among
bidities such as diabetes mellitus and hypertension. Their overweight and obese patients need to be established.
infants were more likely to be appropriate-for-gestational- Nonetheless, the importance of insulin therapy may lie
age (AGA) and be born at term. A systematic review pub- beyond merely achieving glycemic control, but in regulating
lished by Maggard et al. included 75 articles, and concluded these aberrant metabolic pathways among overweight and
that less maternal complications occurred following bariat- obese patients. Prepregnancy and early pregnancy inter-
ric surgery (such as GDM and preeclampsia), and that neo- ventions aimed at weight loss are efficacious in reducing
natal outcomes were better than in obese controls.50 overweight-, obesity-, and GDM-related complications.
References 257
REFERENCES
1. Finucane MM, Stevens GA, Cowan MJ et al. National, regional, 25. Cedergren MI. Maternal morbid obesity and the risk of adverse
and global trends in body-mass index since 1980: Systematic pregnancy outcome. Obstet Gynecol 2004; 103: 219–224.
analysis of health examination surveys and epidemiological stud- 26. Baeten JM, Bukusi EA, Lambe M. Pregnancy complications and
ies with 960 country-years and 9·1 million participants. Lancet outcomes among overweight and obese nulliparous women. Am
2011; 377: 557–567. J Public Health 2001; 91: 436–440.
2. OECD. Health at a Glance 2013: OECD Indicators, OECD 27. Alanis MC, Goodnight WH, Hill EG, Robinson CJ, Villers MS,
Publishing: Paris, France, 2013. Johnson DD. Maternal super-obesity (body mass index > or = 50)
3. World Health Organization. Obesity: Preventing and Managing and adverse pregnancy outcomes. Acta Obstet Gynecol Scand
the Global Epidemic. WHO Technical Report Series 894, World 2010; 89: 924–930.
Health Organization: Geneva, Switzerland, 2000. 28. Chu SY, Callaghan WM, Kim SY et al. Maternal obesity and
4. Yogev Y, Catalano PM. Pregnancy and obesity. Obstet Gynecol risk of gestational diabetes mellitus. Diabetes Care 2007; 30:
Clin North Am 2009; 36: 285–300, viii. 2070–2076.
5. Reece EA. Perspectives on obesity, pregnancy and birth out- 29. Kim SY, England L, Wilson HG, Bish C, Satten GA, Dietz
comes in the United States: The scope of the problem. Am J P. Percentage of gestational diabetes mellitus attributable
Obstet Gynecol 2008; 198: 23–27. to overweight and obesity. Am J Public Health 2010; 100:
6. Langer O. Management of obesity in GDM: Old habits die hard. 1047–1052.
J Matern Fetal Neonatal Med 2008; 21: 165–171. 30. Colomiere M, Permezel M, Lappas M. Diabetes and obesity
7. Institute of Medicine. Nutrition during Pregnancy. Part I, Weight during pregnancy alter insulin signalling and glucose transporter
Gain; Part II: Nutrient Supplements, Washington, DC: National expression in maternal skeletal muscle and subcutaneous adi-
Academy Press, 1990. pose tissue. J Mol Endocrinol 2010; 44: 213–223.
8. Rooney BL, Schauberger CW. Excess pregnancy weight gain and 31. Crowther CA, Hiller JE, Moss JR, McPhee AJ, Jeffries WS,
long-term obesity: One decade later. Obstet Gynecol 2002; 100: Robinson JS. Effect of treatment of gestational diabetes mellitus
245–252. on pregnancy outcomes. N Engl J Med 2005; 352: 2477–2486.
9. Edwards LE, Hellerstedt WL, Alton IR, Story M, Himes JH. 32. Langer O, Yogev Y, Most O, Xenakis EMJ. Gestational diabetes:
Pregnancy complications and birth outcomes in obese and The consequences of not treating. Am J Obstet Gynecol 2005;
normal-weight women: Effects of gestational weight change.
192: 989–997.
Obstet Gynecol 1996; 87: 389–394. 33. Landon MB, Spong CY, Thom E et al. A multicenter, random-
10. Bianco AT, Smilen SW, Davis Y, Lopez S, Lapinski R, Lockwood ized trial of treatment for mild gestational diabetes. N Engl J Med
CJ. Pregnancy outcome and weight gain recommendations for 2009; 361: 1339–1348.
the morbidly obese woman. Obstet Gynecol 1998; 91: 97–102. 34. Jensen DM, Damm P, Sørensen B et al. Clinical impact of mild
11. Luke B, Hediger ML, Scholl TO. Point of diminishing returns: carbohydrate intolerance in pregnancy: A study of 2904 non-
When does gestational weight gain cease benefiting birthweight diabetic Danish women with risk factors for gestational diabetes
and begin adding to maternal obesity? J Matern Fetal Med 1996; mellitus. Am J Obstet Gynecol 2001; 185: 413–419.
5: 168–173. 35. Sermer M, Naylor CD, Gare DJ et al. Impact of increasing carbo-
12. Rasmussen KM, Yaktine AL, eds. Weight Gain during Pregnancy: hydrate intolerance on maternal-fetal outcomes in 3637 women
Reexamining the Guidelines, The National Academies Press: without gestational diabetes. The Toronto Tri-Hospital Gestational
Washington, DC, 2009. Diabetes Project. Am J Obstet Gynecol 1995; 173: 146–156.
13. Hull HR, Thornton JC, Ji Y et al. Higher infant body fat with 36. Ben-Haroush A, Hadar E, Chen R, Hod M, Yogev Y. Maternal
excessive gestational weight gain in overweight women. Am J obesity is a major risk factor for large-for-gestational-infants in
Obstet Gynecol 2011; 205: 211.e1–211.e7. pregnancies complicated by gestational diabetes. Arch Gynecol
14. Vesco KK, Sharma AJ, Dietz PM et al. Newborn size among obese Obstet 2009; 279: 539–543.
women with weight gain outside the 2009 Institute of Medicine 37. Wahabi HA, Fayed AA, Alzeidan RA, Mandil AA. The indepen-
recommendation. Obstet Gynecol 2011; 117: 812–818. dent effects of maternal obesity and gestational diabetes on the
15. Bodnar LM, Siega-Riz AM, Simhan HN, Himes KP, Abrams B. pregnancy outcomes. BMC Endocr Disord 2014; 14: 47.
Severe obesity, gestational weight gain, and adverse birth out- 38. Catalano PM, McIntyre HD, Cruickshank JK et al. The hyper-
comes. Am J Clin Nutr 2010; 91: 1642–1648. glycemia and adverse pregnancy outcome study: Associations
16. Getahun D, Ananth CV, Peltier MR, Salihu HM, Scorza WE. of GDM and obesity with pregnancy outcomes. Diabetes Care
Changes in prepregnancy body mass index between the first and 2012; 35: 780–786.
second pregnancies and risk of large-for-gestational-age birth. 39. Joy S, Roman A, Istwan N et al. The effect of maternal obesity on
Am J Obstet Gynecol 2007; 196: 530.e1–530.e8. pregnancy outcomes of women with gestational diabetes con-
17. Armitage JA, Poston L, Taylor PD. Developmental origins of obe- trolled with diet only, glyburide, or insulin. Am J Perinatol 2012; 29:
sity and the metabolic syndrome: The role of maternal obesity. 643–648.
Front Horm Res 2008; 36: 73–84. 40. Olmos PR, Rigotti A, Busso D et al. Maternal hypertriglyceride-
18. Howie GJ, Sloboda DM, Kamal T, Vickers MH. Maternal nutri- mia: A link between maternal overweight-obesity and macro-
tional history predicts obesity in adult offspring independent of somia in gestational diabetes. Obesity 2014; 22: 2156–2163.
postnatal diet. J Physiol 2009; 587: 905–915. 41. Leikin E, Jenkins JH, Graves WL. Prophylactic insulin in gesta-
19. Laitinen J, Power C, Järvelin MR. Family social class, maternal body tional diabetes. Obstet Gynecol 1987; 70: 587–592.
mass index, childhood body mass index, and age at menarche as 42. Schaefer-Graf UM, Heuer R, Kilavuz O, Pandura A, Henrich
predictors of adult obesity. Am J Clin Nutr 2001; 74: 287–294. W, Vetter K. Maternal obesity not maternal glucose values cor-
20. Forsén T, Eriksson JG, Tuomilehto J, Teramo K, Osmond C, Barker relates best with high rates of fetal macrosomia in pregnancies
DJ. Mother’s weight in pregnancy and coronary heart disease in complicated by gestational diabetes. J Perinat Med 2002; 30:
a cohort of Finnish men: Follow up study. Br Med J 1997; 315: 313–321.
837–840. 43. Yogev Y, Langer O. Pregnancy outcome in obese and morbidly
21. Fall CH, Stein CE, Kumaran K et al. Size at birth, maternal weight, obese gestational diabetic women. Eur J Obstet Gynecol Reprod
and type 2 diabetes in South India. Diabet Med 1998; 15: Biol 2008; 137: 21–26.
220–227. 44. Langer O, Yogev Y, Xenakis EMJ, Brustman L. Overweight and
22. Kumari AS. Pregnancy outcome in women with morbid obesity. obese in gestational diabetes: The impact on pregnancy outcome.
Int J Gynaecol Obstet 2001; 73: 101–107. Am J Obstet Gynecol 2005; 192: 1768–1776.
23. Sebire NJ, Jolly M, Harris JP et al. Maternal obesity and preg- 45. Most O, Langer O. Gestational diabetes: Maternal weight gain in
nancy outcome: A study of 287,213 pregnancies in London. Int J relation to fetal growth, treatment modality, BMI and glycemic con-
Obes Relat Metab Disord 2001; 25: 1175–1182. trol. J Matern Fetal Neonatal Med 2012; 25: 2458–2463.
24. Weiss JL, Malone FD, Emig D et al. Obesity, obstetrical compli- 46. Olmos PR, Borzone GR, Olmos RI et al. Gestational diabetes and
cations and cesarean delivery rate-a population-based screening pre-pregnancy overweight: Possible factors involved in newborn
study. Am J Obstet Gynecol 2004; 190: 1091–1097. macrosomia. J Obstet Gynaecol Res 2012; 38: 208–214.
47. Villamor E, Cnattingius S. Interpregnancy weight change and risk 53. Katon J, Reiber G, Williams MA, Yanez D, Miller E. Weight loss
of adverse pregnancy outcomes: A population-based study. Lancet after diagnosis with gestational diabetes and birth weight among
2006; 368: 1164–1170. overweight and obese women. Matern Child Health J 2013; 17:
48. Ehrlich SF, Hedderson MM, Feng J, Davenport ER, Gunderson 374–383.
EP, Ferrara A. Change in body mass index between pregnancies 54. Park JE, Park S, Daily JW, Kim S-H. Low gestational weight gain
and the risk of gestational diabetes in a second pregnancy. Obstet improves infant and maternal pregnancy outcomes in overweight
Gynecol 2011; 117: 1323–1330. and obese Korean women with gestational diabetes mellitus.
49. Dell’Agnolo CM, Carvalho MD de B, Pelloso SM. Pregnancy after Gynecol Endocrinol 2011; 27: 775–781.
bariatric surgery: Implications for mother and newborn. Obes Surg 55. Clapp JF, Kim H, Burciu B, Lopez B. Beginning regular exercise
2011; 21: 699–706. in early pregnancy: Effect on fetoplacental growth. Am J Obstet
50. Maggard MA, Yermilov I, Li Z et al. Pregnancy and fertility follow- Gynecol 2000; 183: 1484–1488.
ing bariatric surgery: A systematic review. J Am Med Assoc 2008; 56. Campbell MK, Mottola MF. Recreational exercise and occupa-
300: 2286–2296. tional activity during pregnancy and birth weight: A case-control
51. Harrison CL, Lombard CB, Strauss BJ, Teede HJ. Optimizing study. Am J Obstet Gynecol 2001; 184: 403–408.
healthy gestational weight gain in women at high risk of ges- 57. Dye TD, Knox KL, Artal R, Aubry RH, Wojtowycz MA. Physical
tational diabetes: A randomized controlled trial. Obesity 2013; activity, obesity, and diabetes in pregnancy. Am J Epidemiol
21: 904–909. 1997; 146: 961–965.
52. Yee LM, Cheng YW, Inturrisi M, Caughey AB. Gestational weight 58. Carpenter MW. The role of exercise in pregnant women with dia-
loss and perinatal outcomes in overweight and obese women betes mellitus. Clin Obstet Gynecol 2000; 43: 56–64.
subsequent to diagnosis of gestational diabetes mellitus. Obesity 59. Kardel KR, Kase T. Training in pregnant women: Effects on fetal
2013; 21: E770–E774. development and birth. Am J Obstet Gynecol 1998; 178: 280–286.
32 Pharmacological treatment for
the obese gestational diabetes
mellitus patient
Fiona C. Denison and Rebecca M. Reynolds
she was severely obese (BMI > 40 kg/m2).3,4 Recent studies
Pandemic of obesity, diabetes, and suggest that the association between maternal obesity and
diabetes in pregnancy GDM may be even higher with a population-based record
linkage study reporting adjusted odds ratios for d
eveloping
The World Health Organization has described the ris- GDM of 11.90 (7.54–18.79) and 67.40 (37.84–120.03) in obese
ing prevalence of overweight (defined as body mass index and severely obese women, respectively.5
[BMI] > 25 kg/m2) and obesity (BMI > 30 kg/m2) in devel-
oping and developed countries as a global pandemic.1
Between 1980 and 2013, the proportion of women with a
BMI of 25 kg/m2 or more increased worldwide from 29.8% Obesity, diabetes, and rationale for
(29.3–30.2) to 38.0% (37.5–38.5), and in at least seven treatment
countries including Kuwait, Libya, Qatar, and Samoa, the
estimated prevalence of obesity in women now exceeds 50%. Obese women are at increased risk of developing GDM
In developed countries, although the prevalence is much because they are more likely to enter pregnancy with pre-
lower, the overall pattern and rise in rates of overweight and existing insulin resistance. To maintain euglycemia in the
obesity parallels that of developing countries. third trimester, a 200%–250% increase in insulin synthesis
The peak age prevalence of obesity in women in developed is required to compensate for the 50% reduction in insulin
countries is at 60 years, which is beyond the natural reproduc- sensitivity.6,7 In the obese, this physiological adaptation either
tive age. However, although obesity rates in younger women are fails to occur or is insufficient to compensate for the reduc-
lower than in older women, they are also rising in women of tion in insulin sensitivity with resulting hyperglycemia.8,9
reproductive age.2 As a consequence, pregravid maternal obe- Both maternal obesity and diabetes in pregnancy are
sity is now the most common comorbidity of pregnancy, with associated with increased risk of adverse outcomes for both
overweight and obesity affecting up to 40% and 20% of the ante- mother and fetus.5,10–12 For the mother, these include increased
natal population, respectively. Recent evidence suggests that the risk of pregnancy-induced hypertension, preeclampsia, med-
peak age prevalence of obesity has started to fall in developed ically indicated preterm delivery, and cesarean section. For
countries.1 This is likely to further increase the proportion of the fetus, these range from increased risk of miscarriage,
women of reproductive age who are overweight and obese. congenital anomaly, macrosomia, and shoulder dystocia
In parallel with the obesity epidemic, there is also a dia- through to neonatal death and stillbirth. In addition, both
betes pandemic with increasing numbers of obese pregnant maternal obesity and diabetes are recognized to have long-
women having preexisting diabetes (typically type 2 but also term adverse health consequences for the child.
type 1) or being diagnosed with gestational diabetes mellitus Although the underlying mechanisms linking mater-
(GDM) during pregnancy. A proportion of these women will nal obesity with adverse outcome remain unclear, the
have previously undiagnosed type 2 diabetes. Due to the lack Hyperglycemia and Adverse Pregnancy Outcomes study
of consensus in diagnostic and screening criteria for GDM demonstrated that fasting hyperglycemia was correlated
during pregnancy, accurate estimates of risk and prevalence with increased risk of adverse outcomes including cesar-
of GDM are difficult to define. However, a meta-analysis in ean section and fetal macrosomia. This relationship exists
2007 demonstrated that compared to normal-weight preg- even when blood glucose levels are within the normal range.
nant women, the risk of GDM was 3.56 (3.05–4.21)-fold higher Fasting blood glucose levels are higher throughout preg-
if the woman was obese and 8.56 (5.07–16.04)-fold higher if nancy in obese women than in normal-weight women,13
259
260 Pharmacological treatment for the obese gestational diabetes mellitus patient
so it is likely that relative hyperglycemia may in part explain glucose absorption and substrate availability for hepatic
some of the excess in morbidity observed in obese pregnan- gluconeogenesis.21 Metformin is actively transported to
cies. However, it is likely that other factors in addition to hepatocytes via organic cation transporter 1. It inhibits the
hyperglycemia and insulin resistance such as micronutrient mitochondrial respiratory complex I leading to reduced
deficiency, hyperlipidemia, and inflammation play a role in ATP synthesis and activation of AMP-activated kinase.
the pathogenesis of morbidities in the obese.14,15 This is sup- This results in a reduction of lipid synthesis, inhibition
ported by studies that demonstrate the effects of maternal of gluconeogenesis, and a decrease in circulating insulin
glycemia and obesity, which are independent and additive for and glucose. Metformin also stimulates insulin-mediated
outcomes including cesarean section and macrosomia.16,17 glucose uptake by the skeletal muscle and liver.21 This,
Thus, although there is a clear rationale for treating hyper- together with reduced gluconeogenesis in the liver, increases
glycemia and insulin resistance caused by GDM in the obese insulin sensitivity.
to improve outcome, a therapy or therapies that could also In obese nonpregnant women with type 2 diabetes, the
treat maternal obesity, for example, minimizing maternal apparent clearance and volume of distribution of metformin
weight gain and reducing inflammation, might have the best are influenced by the lean body weight. Markers of renal
chance of optimizing pregnancy outcome. function such as age and serum creatinine also influence
clearance and distribution.22 During pregnancy, metformin
clearance increases with gestation due to enhanced renal
Overview of management of GDM elimination. To maintain a therapeutic effect, the metformin
dose may therefore need to be increased by up to 20% in the
in the obese third trimester.23 Whether increased maternal BMI has an
Dietary modification with home blood glucose monitoring additional influence on clearance and volume of distribution
is the first-line management for GDM in obese pregnant of metformin during pregnancy is not known.
woman. This is covered in depth in Chapter 17. The key prin-
ciples are to maintain a healthy, balanced diet, for example, Safety of metformin in pregnancy Over the last 10 years in
using the “eatwell plate” as per dietary recommendations the United Kingdom, metformin has increasingly replaced
for nonpregnant individuals with newly diagnosed diabetes, insulin as the first-line pharmacological agent for the
with particular emphasis on avoiding refined carbohydrates, treatment of GDM if women fail to achieve glycemic control
sugars, and fats. For most individuals, adherence to these after dietary modification. This change in practice is in
dietary recommendations will result in a calorie deficit. In large part due to at least 10 clinical studies (5 randomized
pregnancy, there are no guidelines to suggest that weight controlled trials, 2 prospective nonrandomized trials, and
loss is recommended/advisable during pregnancy.18 The best 3 retrospective studies) that have all demonstrated that
available advice for gestational weight gain in pregnancy is metformin is safe in pregnancy and achieves comparable
from the Institute of Medicine Guidelines, which advises a (or better) glycemic control compared to insulin or the
sulfonylurea glibenclamide.24–32
weight gain of 5–9 kg during pregnancy for all women with
BMI > 30 kg/m2.19 This advice has not been looked at in Metformin crosses the placenta, with umbilical cord
the context of dietary advice for GDM, but for most obese concentrations at the time of delivery being at least half
women it would seem that advice to maintain weight during the maternal concentrations.23,33 In some cases, they exceed
pregnancy is not unsafe. maternal concentrations. It is therefore important to follow
In many women, dietary modification achieves accept- up babies exposed to metformin in utero.
able glycemic control and no further treatment is required.
However, in women with GDM who remain hyperglycemic Short-term outcomes
despite dietary modification, pharmacological therapy is the Maternal outcomes Metformin is generally acceptable
next stage in management. Subcutaneous insulin therapy has to women and well tolerated during pregnancy, with only
traditionally been regarded as the next stage in management, 2%–7% of women being unable to take it due to its mainly
but with new evidence of safety and efficacy of oral hypogly- gastrointestinal side effects. Other reported side effects are
cemic agents in pregnancy, these are increasingly being used either very rare (lactic acidosis) or relatively minor (erythema
as second-line treatment after failure of dietary modification. in hypersensitive individuals). There is no evidence that
Insulin is then added as third-line therapy if oral hypoglyce- maternal BMI affects the side effect profile of metformin
mic agents do not meet glycemic targets or are not tolerated. during pregnancy.
Between 10% and 46% of women with GDM treated
with metformin require insulin in addition to metformin
Pharmacological treatment to maintain euglycemia as pregnancy progresses.24,26,28,29,31
Higher maternal BMI and fasting blood glucose at diagnosis
Oral agents and early gestation at diagnosis of GDM are associated with
Metformin increased need for supplemental insulin at an earlier gesta-
Mechanism of action Metformin is a synthetically derived tion.27,28 Difference in the racial groups, time of blood glu-
biguanide.20 After oral ingestion, it is rapidly absorbed and cose monitoring (pre- or postprandial), and the threshold
transported via the portal vein to the liver where it inhibits for starting supplementary therapy (1 vs. 2 or 3 high home
Pharmacological treatment 261
blood glucose monitoring readings) may also explain the Long-term outcomes
wide range in need for supplemental insulin between differ- Maternal outcomes The impact of metformin treatment
ent trials. However, in those women who do require supple- on maternal weight postpartum is less certain. A follow-up
mental insulin, the dose of insulin required is consistently study of women with PCOS who were randomized to
lower in women taking metformin and insulin as opposed metformin (2000 mg daily) or placebo from first trimester
to metformin alone. For example, in the Metformin versus to delivery reported maternal and offspring follow-up
Insulin for the Treatment of Gestational Diabetes (MiG) at 1 year. Despite the use of metformin being associated
trial, participants on metformin required a significantly with less weight gain during pregnancy (3.2 ± 2.0 vs. 4.2 ±
lower median dose of 42 U insulin, compared to the patients 4.3 kg, metformin compared to placebo; p < 0.03), women
who received insulin therapy alone required a median dose randomized to metformin lost less weight from the end of
of 50 U.28 pregnancy to 1 year postpartum (−2.1 ± 3.6 vs. −4.1 ± 4.9 kg;
Excessive weight gain during pregnancy in obese p < 003) and had higher BMIs (30.6 ± 8.1 vs. 27.6 ± 6.1 kg/m2;
women is associated with increased risk of complica- p < 0.004) compared to those in the placebo group.35 This is
tions including preterm birth, cesarean delivery, macro- despite there being no difference in maternal BMI between
somia, and childhood obesity. 34 Trials by Balani et al., 24 metformin and placebo groups during the first trimester at
Niromanesh et al., 31 and Rowan et al. 28 reported less ges- randomization (32.7 ± 6.9 vs. 32.0 ± 7.3 kg/m2, metformin
tational weight gain in women on metformin compared compared to placebo; p = 0.51). To date, there are no data
to other pharmacological agents. The underlying mecha- on postpartum weight and BMI in the obese GDM managed
nism for this is not clear. In management of people with with metformin during pregnancy.
type 2 diabetes (who are not pregnant), metformin is
regarded as being weight neutral and is not prescribed as Offspring outcomes Maternal obesity is associated with
a “weight loss” drug. In randomized controlled trials in increased risk of offspring obesity due in part to fat
women with polycystic ovarian syndrome (PCOS), with preferentially being stored in visceral as opposed to
associated insulin resistance but not overt type 2 diabetes, subcutaneous depots.36
treatment with metformin has not been associated with The MiG trial has followed up children at 2 years post-
significant weight loss. 20 However, limiting gestational delivery.28,37 Compared to the insulin group, the children
weight gain has potential health benefits for mother and exposed to metformin in utero had larger subscapular and
offspring, particularly in the obese. biceps skinfolds and greater upper arm circumferences.
There was no difference in total percentage body fat, fat
Fetal outcomes There is no evidence that metformin mass, or central fat measures between the two groups. The
treatment is associated with increased risk of congenital investigators concluded that in utero exposure to metfor-
anomalies. In the majority of trials, no difference was min may predispose children to laying down fat stores in
identified in frequency of large for gestational age, small for healthier subcutaneous fat depots, thus reducing the risk
gestational age, macrosomia, birth weight, and neonatal unit of developing insulin resistance and obesity in later life.
admission.26–29,31,32 Metformin is also associated with lower Others have subsequently cautioned against overinterpret-
risk of neonatal hypoglycemia compared to insulin. ing the results due to the crude indirect measures used to
The relationship between metformin use and preterm measure visceral fat depots, small numbers at follow-up,
birth rates is less clear. In the MiG trial, the rates of spon- and investigators not being blinded to the treatment group
taneous (5.0% vs. 3.5%) and iatrogenic (7.2% vs. 4.1%) pre- of the mother.38
term birth were higher in women treated with metformin If alterations in offspring body composition do exist fol-
(n = 363) compared to insulin (n = 370). 28 This reached lowing in utero exposure to metformin, whether these per-
significance when both spontaneous and iatrogenic pre- sist into childhood is not certain. Albeit in a different patient
term birth rates were combined (12.1% vs. 7.6%; metfor- cohort, Ro et al. reported follow-up of a randomized con-
min compared to insulin treated; p < 0.04). In contrast, trolled trial on women with PCOS randomized to metfor-
there was no difference in spontaneous preterm birth min or placebo in pregnancy.39 At 8 years, there were no
rates (p = 0.3) between metformin (n = 50 recruited, one differences in offspring body composition, weight, or height
spontaneous delivery at 36 weeks gestation) and insulin between those exposed to metformin and those exposed to
(n = 50, three spontaneous deliveries at 31, 33, and 36 weeks placebo. Children exposed to metformin in utero did how-
therapy) in the Ijas trial, 27 and there was an increase in ever have a higher fasting glucose level (4.93 vs. 4.60 mmol/L,
iatrogenic (15.0% vs. 7.6%) but not spontaneous (4.2% vs. p = 0.04) and a trend toward a lower LDL cholesterol level
4.9%) preterm birth rates in the insulin-treated (n = 399) (2.42 vs. 2.99 mmol/L, p = 0.07) and higher systolic blood
compared to metformin-treated (n = 471) group in the pressure (106 vs. 101 mmHg, p = 0.05).
Goh trial (p < 0.005). 26 Maternal obesity is associated with Although the follow-up studies are reassuring about the
increased risk of iatrogenic but not spontaneous preterm safety of metformin during pregnancy, caution should be
birth rates. 5 Differences in maternal BMI between trial exercised about concluding that its use in the obese GDM
participants may therefore in part explain inconsistencies will reduce the risk of the offspring in developing insulin
about whether metformin does or does not influence pre- resistance in later life due to redistribution of fat depots.
term birth rate. Further long-term follow-up studies are needed.
Glibenclamide maintain euglycemia as pregnancy progresses.50–54 Although

Mechanism of action Glibenclamide is a second-generation the presence of maternal obesity per se does not predict treat-
sulfonylurea, an insulin secretagogue. It binds to pancreatic ment failure, factors such as early diagnosis and treatment of
β-cell ATP calcium channel receptors and stimulates release gestational diabetes, baseline hyperglycemia, increased mater-
of insulin from the functional cell mass of the pancreas. nal age, and parity, which are more prevalent in the obese, do
It also increases insulin sensitivity in peripheral tissues.40,41 predict increased risk of failure.50–54 In nonpregnancy, glib-
Glibenclamide is therefore an effective treatment for the enclamide is associated with moderate weight gain, and ges-
inadequate insulin secretion and insulin resistance that tational weight gain in women treated with glibenclamide is
occur in the obese GDM. comparable to women with GDM treated with insulin.
Glibenclamide is well absorbed following oral admin- There are conflicting data about whether compared to
istration. It reaches peak concentration in approximately insulin therapy glibenclamide is associated with no change
3 hours, and following metabolism in the liver, it has a half- or increased risk of preeclampsia.58 In a retrospective study
life of 8 hours. Glibenclamide is cleared more rapidly in preg- by Jacobson et al.,58 outcomes were compared in women
nancy due to its being metabolized by the cytochrome P450 treated with insulin (n = 268) and with glibenclamide
enzymes, the activity of two (CYP2C9 and CYP3A) being (n = 236). Women treated with glibenclamide were more
increased in pregnancy. In nonpregnant women with type than twice as likely to develop preeclampsia (odds ratio [OR]
2 diabetes, studies demonstrate that obesity has no effect on 2.32; 95% confidence interval [CI] 1.17–4.63) compared to
glibenclamide pharmacokinetics and pharmacodynamics.42 those managed with insulin. In contrast, in a smaller study
There are no studies that have investigated whether obesity (n = 78 treated with insulin; n = 44 treated with gliben-
affects glibenclamide pharmacokinetics and pharmacody- clamide), there was no significant differences in the risk of
namics during pregnancy. preeclampsia between treatment groups.59 Maternal obesity
is associated with increased risk of developing preeclamp-
Safety of glibenclamide in pregnancy Glibenclamide was the sia5; thus until there is more conclusive trial evidence, it may
first oral agent to be tested in clinical trials to manage GDM be sensible to use alternative pharmacological agents in the
in pregnancy.43 It has subsequently been trialed in a large obese mother with GDM with other risk factors for develop-
number of randomized controlled trials, nonrandomized ing preeclampsia such as essential hypertension.
trials, and retrospective studies.44–54 To date, most trials
demonstrate that glibenclamide is effective in most women Fetal outcomes There is no evidence that glibenclamide
in optimizing glycemic control. is associated with increased risk of fetal anomaly. In the
Initial clinical trials43 suggested that fetal exposure to randomized controlled trial by Langer et al., in which women
glibenclamide was minimal because levels were undetectable were randomized to receive glibenclamide or metformin, the
in cord blood.55 However, more recent studies demonstrate trial was powered on the primary outcome of “desired level of
that glibenclamide crosses the placental barrier with a cord- glycemic control” and not less common neonatal outcomes.43
to-maternal plasma concentration ratio of 0.7.56 Although The largest retrospective observational study compared
glibenclamide is metabolized by the fetus and the placenta outcomes in women with GDM treated with insulin (n =
actively pumping glibenclamide from the fetus to the mother 8609, 80.6%) with oral hypoglycemic agents (n = 2073,
via the ABC transporters, fetal drug concentrations reflect 19.4% or which 99% were receiving glibenclamide).49 After
maternal serum levels.57 controlling for confounders including maternal weight,
ethnicity, parity, and gestation at diagnosis, oral therapy was
Short-term outcomes associated with increased risk of admission to the neonatal
Maternal outcomes Glibenclamide is generally well tolerated unit (OR [95% CI], 1.46 [1.07–2.00]), macrosomia (OR [95%
by women during pregnancy. Hypoglycemia and weight gain CI], 1.29 [1.03–1.64]), and, if GDM was diagnosed early
are the most common side effects reported by individuals in pregnancy, stillbirth (OR [95% CI], 4.68 [1.02–21.5]). A
taking glibenclamide who are not pregnant. Therefore, small randomized controlled trial in women with GDM
women who are started on glibenclamide treatment will randomized to insulin (n = 50) and glibenclamide (n = 49) has
need to be advised about symptoms and management of demonstrated a 4.9% increased rate of intrauterine growth
hypoglycemia and to take the medication with food. Weight restriction (p < 0.01). Maternal obesity is independently
gain is a well-recognized side effect of glibenclamide due to associated with increased risk of neonatal unit admission,
stimulation of insulin secretion with consequent enhanced macrosomia, intrauterine growth restriction, and stillbirth.5
glucose absorption and storage in adipose tissue, but there Whether maternal obesity contributes to increased risk
are no data on whether the use of glibenclamide in GDM has associated with glibenclamide needs to be evaluated in future
a significant impact on gestational weight gain. Although substantive trials.
side effects such as skin itching, rashes, mild nausea, and
heartburn are relatively common, these usually settle if Long-term outcomes
treatment persists. Other side effects such as elevated liver Maternal and offspring outcomes To date, there are no
enzymes are less common, and it only rarely causes jaundice. studies that report long-term maternal and offspring outcomes
Between 16% and 21% of women with GDM treated with following glibenclamide use for GDM in either women of
glibenclamide require insulin in addition to glibenclamide to normal weight or obese.
Pharmacological treatment 263
Insulin at the injection site and/or need to inject more than once
Subcutaneous insulin therapy is recognized as the stan- (most insulin pens have a maximum of 80 units), which may
dard for the treatment of GDM to achieve euglycemia after be cumbersome for the woman and lead to poor compli-
failure of dietary therapy in order to avoid the short- and ance, as observed in nonpregnancy.61 However, this would
long-term complications of poorly controlled glucose during be unusual, even in an obese woman with GDM.
pregnancy. A recent systematic review and meta-analysis of Increased prepregnancy BMI is a recognized risk factor
randomized controlled trials comparing therapeutic inter- for the need for insulin treatment of GDM. For example,
vention for GDM with standard antenatal care included data in one study, prepregnancy BMI > 30 kg/m2 was indepen-
from 10 studies including 3881 patients.60 In these studies, dently associated with the need for insulin treatment with
interventions used in the experimental group were dietary an odd ratio of 2.35 (1.33–4.13).62 In this study, women with
modification and glucose monitoring, and insulin was GDM were not offered oral hypoglycemic agents after fail-
administered if glycemic targets were not met with dietary ure of dietary therapy alone, and future studies will inform
modification. Treatment for GDM significantly reduced whether use of these agents is associated with decreased need
the risk for macrosomia (relative risk [RR], 0.47; 95% CI, for insulin in the obese mother with GDM.
0.38–0.57), large-for-gestational-age births (RR, 0.55;
95% CI, 0.45–0.67), shoulder dystocia (RR, 0.42; 95% CI,
What are the disadvantages for women of insulin
0.23–0.77), and gestational hypertension (RR, 0.68; 95% CI,
therapy? Given that insulin therapy has been regarded
0.53–0.87) without causing any significant increase in the
as the standard therapy and there is a wealth of data and
risk for small-for-gestational-age babies. No significant dif- experience of its use in women with type 1 and type 2
ference was observed between the two groups regarding diabetes, why would women with GDM not want to be
perinatal/neonatal mortality, neonatal hypoglycemia, birth treated with insulin during pregnancy? Insulin has several
trauma, preterm births, preeclampsia, cesarean section, and disadvantages for women, often requiring multiple daily
labor induction. The heterogeneity of studies, for example, injections and with the recognized risks of hypoglycemia
differences in study population, glycemic targets, and insu- and weight gain. Many women fear injecting themselves
lin regimens, does make comparison of studies challeng- with insulin, and in comparative trials, women prefer taking
ing. Nevertheless, it is accepted that subcutaneous insulin oral agents to insulin therapy.28 Having a diagnosis of GDM
therapy is effective at lowering blood glucose in women with has been shown to increase a woman’s anxiety and result in
GDM and that this is associated with improved outcomes for a less positive pregnancy experience than in women without
mother and baby. GDM.63 Treatment with insulin appears to have further
With changes in clinical practice over the last decade detrimental effects on maternal well-being in pregnancy.
in the United Kingdom, insulin is increasingly used as a A recent survey of 1372 Australian women including 393 who
third-line agent to supplement metformin therapy when reported their experiences of having GDM found that women
metformin alone fails to control glucose levels or if metfor- taking insulin were significantly more likely to experience
min is not tolerated. Indeed in a survey of maternity units shock, fear, or anxiety.64 This study highlights the need for
in Scotland in 2013, there was almost a universal use of healthcare professionals to provide additional support to these
insulin as a third-line agent (Stirrat, personal communi- women when their treatment for GDM is intensified.
cation). Details of types of insulin preparation and typical The increased insulin resistance associated with obesity in
insulin regimens for managing diabetes in pregnancy are pregnancy may mean that insulin requirements are greater
described in detail elsewhere (Chapter 22). As in standard for obese women with GDM than for women of normal
diabetes management, diabetologists will tailor an individ- weight. These larger doses of insulin may increase the risk of
ual’s insulin regimen and preparation after consideration of excess gestational weight gain due to the anabolic actions of
the woman’s daily blood glucose profile and lifestyle. Thus insulin described earlier. Women with GDM who are com-
women may require prandial quick-acting insulin and/ menced on insulin treatment should therefore have addi-
or basal (long-acting) insulin or a basal–bolus regimen tional dietetic input to minimize insulin-associated weight
depending on the situation, and insulin requirements are gain. Likewise, it is also important that obese women with
likely to change during pregnancy. The 2013 Scottish sur- GDM are advised to avoid excessive gestational weight gain
vey highlighted that a range of different insulin preparations as weight gain in the nonpregnant obese individual is associ-
and regimens was used, including the use of fixed mixtures ated with further increased insulin requirements.65 If these
of insulin (Stirrat, personal communication). Women with observations are extrapolated to pregnancy, excessive gesta-
GDM who need to start insulin therapy will require more tional weight gain could lead to increased insulin require-
intensive management during pregnancy, seeing a diabetes ments and a consequent vicious cycle of increasing insulin
specialist nurse and diabetes consultant for education about doses and increasing weight gain. Use of metformin in com-
insulin administration and dose titration as well as advice bination with insulin is associated with a lower total daily
about management of hypoglycemia. There is no evidence requirement of insulin,28 and the combination of the weight-
that obese women with GDM should use a different injection neutral effect of metformin and the lower insulin require-
site for insulin administration than is advised in a nonpreg- ments may be beneficial in limiting gestational weight gain
nant individual. If very large doses of insulin are required in the obese woman with GDM. However, further studies are
to control blood glucose, this may be associated with pain needed.
Summary that alternative methods of delivering effective care will be

needed. Current developments include evaluating the use of
In summary, it is highly likely that levels of maternal obesity telemetry as an alternative method of monitoring glycemic
and GDM will continue to rise in antenatal populations. As control in GDM reducing the need for hospital attendance
treatment escalates from dietary to oral agents through to and early-stage trials exploring combination therapy with
polypharmacy with insulin, the level of intervention and the metformin and glibenclamide as opposed to adding insulin
frequency of monitoring increase. At a national and interna- as third-line management. The results of these studies and
tional level, this is having a huge impact on the design and others are likely to fundamentally alter the management of
delivery of antenatal services with current models of antena- GDM. This will have a particular impact on the obese due
tal care increasingly being unsustainable. It is therefore likely to the high prevalence of GDM in this antenatal population.
REFERENCES
1. Ng M et al. Global, regional, and national prevalence of over- 19. Rasmussen KM, Yaktine AL, eds. Weight Gain during Pregnancy:
weight and obesity in children and adults during 1980–2013: A Reexamining the Guidelines. National Academies Press:
systematic analysis for the Global Burden of Disease Study 2013. Washington, DC, 2009.
Lancet 2014; 384(9945): 766–781. 20. Sivalingam VN, Myers J, Nicholas S, Balen AH, Crosbie EJ.
2. Heslehurst N et al. Trends in maternal obesity incidence rates, Metformin in reproductive health, pregnancy and gynaecologi-
demographic predictors, and health inequalities in 36,821 cal cancer: Established and emerging indications. Hum Reprod
women over a 15-year period. BJOG 2007; 114(2): 187–194. Update 2014; 20(6): 853–868.
3. Chu SY et al. Maternal obesity and risk of gestational diabetes 21. Pernicova I, Korbonits M. Metformin—Mode of action and clini-
mellitus. Diabetes Care 2007; 30(8): 2070–2076. cal implications for diabetes and cancer. Nat Rev Endocrinol
4. Chu SY, Kim SY, Lau J. Re: Prepregnancy BMI and the risk of 2014; 10(3): 143–156.
gestational diabetes: A systematic review of the literature with 22. Bardin C et al. Population pharmacokinetics of metformin in
meta-analysis. Obes Rev 2009; 10(4): 487–488; author reply obese and non-obese patients with type 2 diabetes mellitus. Eur
489–490. J Clin Pharmacol 2012; 68(6): 961–968.
5. Denison FC et al. Association between maternal body mass 23. Hughes RC et al. Effect of pregnancy on the pharmacokinetics of
index during pregnancy, short-term morbidity, and increased metformin. Diabet Med 2006; 23(3): 323–326.
health service costs: A population-based study. BJOG 2014; 24. Balani J et al. Pregnancy outcomes in women with gestational
121(1): 72–81; discussion 82. diabetes treated with metformin or insulin: A case-control study.
6. Catalano PM et al. Longitudinal changes in insulin release and Diabet Med 2009; 26(8): 798–802.
insulin resistance in nonobese pregnant women. Am J Obstet 25. Ekpebegh CO et al. A 10-year retrospective analysis of pregnancy
Gynecol 1991; 165(6 Pt. 1): 1667–1672. outcome in pregestational type 2 diabetes: Comparison of insulin
7. Kuhl C. Etiology and pathogenesis of gestational diabetes. and oral glucose-lowering agents. Diabet Med 2007; 24(3): 253–258.
Diabetes Care 1998; 21(Suppl. 2): B19–B26. 26. Goh JE, Sadler L, Rowan J. Metformin for gestational diabetes in
8. Buchanan TA et al. Insulin sensitivity and B-cell responsiveness routine clinical practice. Diabet Med 2011; 28(9): 1082–1087.
to glucose during late pregnancy in lean and moderately obese 27. Ijas H et al. Metformin should be considered in the treatment
women with normal glucose tolerance or mild gestational diabe- of gestational diabetes: A prospective randomised study. BJOG
tes. Am J Obstet Gynecol 1990; 162(4): 1008–1014. 2011; 118(7): 880–885.
9. Catalano PM et al. Gestational diabetes and insulin resistance: 28. Rowan JA et al. Metformin versus insulin for the treatment of ges-
Role in short- and long-term implications for mother and fetus. tational diabetes. N Engl J Med 2008; 358(19): 2003–2015.
J Nutr 2003; 133(5 Suppl. 2): 1674S–1683S. 29. Tertti K et al. Metformin vs. insulin in gestational diabetes. A ran-
10. Denison FC, Norrie G, Graham B, Lynch J, Harper N, Reynolds domized study characterizing metformin patients needing addi-
RM. Increased maternal BMI is associated with an increased risk tional insulin. Diabetes Obes Metab 2013; 15(3): 246–251.
of minor complications during pregnancy with consequent cost 30. Gandhi P et al. Introduction of metformin for gestational diabetes
implications. BJOG 2009; 116(11): 1467–1472. mellitus in clinical practice: Has it had an impact? Eur J Obstet
11. Denison FC et al. Maternal obesity, length of gestation, risk of Gynecol Reprod Biol 2012; 160(2): 147–150.
postdates pregnancy and spontaneous onset of labour at term. 31. Niromanesh S et al. Metformin compared with insulin in the
BJOG 2008; 115(6): 720–725. management of gestational diabetes mellitus: A randomized clini-
12. Sebire NJ et al. Maternal obesity and pregnancy outcome: A cal trial. Diabetes Res Clin Pract 2012; 98(3): 422–429.
study of 287,213 pregnancies in London. Int J Obes Relat Metab 32. Spaulonci CP et al. Randomized trial of metformin vs insulin in
Disord 2001; 25(8): 1175–1182. the management of gestational diabetes. Am J Obstet Gynecol
13. Scholl TO et al. Maternal glucose concentration influences fetal 2013; 209(1): 34 e1–34 e7.
growth, gestation, and pregnancy complications. Am J Epidemiol 33. Eyal S et al. Pharmacokinetics of metformin during pregnancy.
2001; 154(6): 514–520. Drug Metab Dispos 2010; 38(5): 833–840.
14. Ramsay JE et al. Maternal obesity is associated with dysregula- 34. Catalano PM. Management of obesity in pregnancy. Obstet
tion of metabolic, vascular, and inflammatory pathways. J Clin Gynecol 2007; 109(2 Pt. 1): 419–433.
Endocrinol Metab 2002; 87(9): 4231–4237. 35. Carlsen SM, Martinussen MP, Vanky E. Metformin’s effect on
15. Challier JC et al. Obesity in pregnancy stimulates macrophage first-year weight gain: A follow-up study. Pediatrics 2012; 130(5):
accumulation and inflammation in the placenta. Placenta 2008; e1222–e1226.
29(3): 274–281. 36. Norman JE, Reynolds RM. The consequences of obesity and
16. Alberico S et al. The role of gestational diabetes, pre-pregnancy excess weight gain in pregnancy. Proc Nutr Soc 2011; 70(4):
body mass index and gestational weight gain on the risk of new- 450–456.
born macrosomia: Results from a prospective multicentre study. 37. Rowan JA et al. Metformin in gestational diabetes: The offspring
BMC Pregnancy Childbirth 2014; 14: 23. follow-up (MiG TOFU): Body composition at 2 years of age.
17. Ovesen P et al. Maternal and neonatal outcomes in pregnan- Diabetes Care 2011; 34(10): 2279–2284.
cies complicated by gestational diabetes. A nation-wide study. 38. Barbour LA et al. Comment on: Rowan et al. Metformin in gesta-
J Matern Fetal Neonatal Med 2015; 28(14): 1720–1724. tional diabetes: The offspring follow-up (MiG TOFU): Body com-
18. Poston L, Chappell LC. How should women be advised on weight position at 2 years of age. Diabetes Care 2011; 34: 2279–2284.
management in pregnancy? BMJ 2012; 344: e2774. Diabetes Care 2012; 35(3): e28; author reply e30.
References 265
39. Ro TB et al. Growth, body composition and metabolic profile of 52. Kahn BF et al. Predictors of glyburide failure in the treatment of
8-year-old children exposed to metformin in utero. Scand J Clin gestational diabetes. Obstet Gynecol 2006; 107(6): 1303–1309.
Lab Invest 2012; 72(7): 570–575. 53. Kremer CJ, Duff P. Glyburide for the treatment of gestational dia-
40. Groop L et al. Different effects of glyburide and glipizide on betes. Am J Obstet Gynecol 2004; 190(5): 1438–1439.
insulin secretion and hepatic glucose production in normal and 54. Rochon M et al. Glyburide for the management of gesta-
NIDDM subjects. Diabetes 1987; 36(11): 1320–1328. tional diabetes: Risk factors predictive of failure and associ-
41. Groop LC et al. Dose-dependent effects of glyburide on insulin ated pregnancy outcomes. Am J Obstet Gynecol 2006; 195(4):
secretion and glucose uptake in humans. Diabetes Care 1991; 1090–1094.
14(8): 724–727. 55. Elliott BD et al. Insignificant transfer of glyburide occurs across
42. Jaber LA, Ducharme MP, Halapy H. The effects of obesity on the the human placenta. Am J Obstet Gynecol 1991; 165(4 Pt. 1):
pharmacokinetics and pharmacodynamics of glipizide in patients 807–812.
with non-insulin-dependent diabetes mellitus. Ther Drug Monit 56. Caritis SN, Hebert MF. A pharmacologic approach to the use of
1996; 18(1): 6–13. glyburide in pregnancy. Obstet Gynecol 2013; 121(6): 1309–1312.
43. Langer O et al. A comparison of glyburide and insulin in women 57. Hebert MF et al. Are we optimizing gestational diabetes treat-
with gestational diabetes mellitus. N Engl J Med 2000; 343(16): ment with glyburide? The pharmacologic basis for better clinical
1134–1138. practice. Clin Pharmacol Ther 2009; 85(6): 607–614.
44. Silva JC et al. Glibenclamide in the treatment for gestational dia- 58. Jacobson GF et al. Comparison of glyburide and insulin for the
betes mellitus in a compared study to insulin. Arq Bras Endocrinol management of gestational diabetes in a large managed care
Metabol 2007; 51(4): 541–546. organization. Am J Obstet Gynecol 2005; 193(1): 118–124.
45. Bertini AM et al. Perinatal outcomes and the use of oral hypogly- 59. Ramos GA et al. Comparison of glyburide and insulin for the
cemic agents. J Perinat Med 2005; 33(6): 519–523. management of gestational diabetics with markedly elevated
46. Anjalakshi C et al. A prospective study comparing insulin and oral glucose challenge test and fasting hyperglycemia. J Perinatol
glibenclamide in gestational diabetes mellitus in Asian Indian 2007; 27(5): 262–267.
women. Diabetes Res Clin Pract 2007; 76(3): 474–475. 60. Poolsup N, Suksomboon N, Amin M. Effect of treatment of ges-
47. Ogunyemi D, Jesse M, Davidson, M. Comparison of glyburide tational diabetes mellitus: A systematic review and meta-analysis.
versus insulin in management of gestational diabetes mellitus. PLOS ONE 2014; 9(3): e92485.
Endocr Pract 2007; 13(4): 427–428. 61. Crasto W et al. Insulin U-500 in severe insulin resistance in type 2
48. Lain KY et al. Neonatal adiposity following maternal treatment of diabetes mellitus. Postgrad Med J 2009; 85(1002): 219–222.
gestational diabetes with glyburide compared with insulin. Am J 62. Sapienza AD et al. Factors predicting the need for insulin therapy
Obstet Gynecol 2009; 200(5): 501 e1–501 e6. in patients with gestational diabetes mellitus. Diabetes Res Clin
49. Cheng YW et al. Treatment of gestational diabetes mellitus: Pract 2010; 88(1): 81–86.
Glyburide compared to subcutaneous insulin therapy and asso- 63. Kim C, Vahratian A. Self-rated health and health care use among
ciated perinatal outcomes. J Matern Fetal Neonatal Med 2012; women with histories of gestational diabetes mellitus. Diabetes
25(4): 379–384. Care 2010; 33(1): 41–42.
50. Chmait R, Dinise T, Moore T. Prospective observational study to 64. Morrison MK, Lowe JM, Collins CE. Australian women’s expe-
establish predictors of glyburide success in women with gesta- riences of living with gestational diabetes. Women Birth 2014;
tional diabetes mellitus. J Perinatol 2004; 24(10): 617–622. 27(1): 52–57.
51. Conway DL, Gonzales O, Skiver D. Use of glyburide for the 65. Yki-Jarvinen H et al. Effect of obesity on the response to insu-
treatment of gestational diabetes: The San Antonio experience. lin therapy in noninsulin-dependent diabetes mellitus. J Clin
J Matern Fetal Neonatal Med 2004; 15(1): 51–55. Endocrinol Metab 1997; 82(12): 4037–4043.
33 Role of exercise in reducing
the risks of gestational diabetes
mellitus in obese women
Raul Artal
decades, this concern has significantly diminished; however,

Introduction prescription of lifestyle interventions that include judicious
The World Health Organization has identified a worldwide gestation weight gain and physical activity for pregnant
epidemic of obesity and diabetes and designated them as women are not yet widely accepted or prescribed.
leading causes for premature mortality.1 These conditions The 2009 Institute of Medicine (IOM) recommenda-
stem primarily from the sedentary lifestyle adopted by the tions for gestational weight gain13 for overweight and obese
contemporary society. Obesity is recognized as a chronic women not only are counterintuitive14,15 but could also fur-
relapsing disease and is also a major cause for the diabetes ther amplify the risk of developing gestational diabetes, since
epidemics. The benefits of adopting a healthy lifestyle in pregnancy is a diabetogenic state characterized by progres-
nonpregnant women, which includes weight reduction and sive increase in insulin resistance, which worsens with addi-
physical activity, to prevent or manage diabetes has been well tional weight gain. Recent studies that aimed at evaluating
established in large clinical trials in the United States, China, gestational weight gain below the IOM recommendations
and Finland.2–4 have reported no adverse effects, no restricted fetal growth
Pregnancy and gestational weight gain have been a major (one of the IOM concerns) and better outcomes.15–17
contributor to the obesity epidemic in part because of the A strong association between higher BMIs, maternal
past practice and reluctance to prescribe lifestyle modifica- plasma glucose levels, and pregnancy complications has
tion regimens to pregnant women. After pregnancy, 16.1% been demonstrated by the Hyperglycemia and Pregnancy
of nulliparous and 19.3% of parous women become obese.5 Adverse Outcome study.11
Many countries, among them the United States, have expe- The American Diabetes Association (ADA) in its
rienced a sharp rise in the incidence of gestational diabetes, Standards of Medical Care in Diabetes—2014 recommends
of whom 60% will develop type 2 diabetes mellitus within that among nonpregnant individuals at high risk for
4 years.6 Women who are active prior to pregnancy and developing diabetes, structured programs that emphasize
throughout pregnancy have the lowest prevalence of gesta- lifestyle changes that include moderate weight loss (7% of
tional diabetes.7–9 body weight) and regular physical activity (150 minutes/
Women with a BMI of 30–35 have an adjusted odds ratio week) with dietary strategies including reduced calories and
(OR) of 2.6, and women with BMI greater than 35 have an reduced intake of dietary fat should be prescribed.18 Over
adjusted OR of 4.0 for gestational diabetes compared with the past years, information has been accumulated to safely
women with a BMI of less than 30.10 prescribe similar regimens of judicious weight gain and
A multitude of maternal and fetal comorbidities have physical activities to pregnant women at risk for diabetes or
been linked to maternal obesity, gestational diabetes, and those that are affected by either gestational diabetes or type
type 2 diabetes.11 2 diabetes.16,17,19–23
Historically, pregnant diabetic women were precluded The American Congress of Obstetricians and Gynecol
from engaging in physical activities; to the contrary, they ogists (ACOG) in its guidelines for exercise during pregnancy
were prescribed prolonged periods of hospitalization or and postpartum recommends to obese pregnant women to
home bed rest.12 The reluctance of prescribing lifestyle inter- engage in healthy lifestyle modification that includes physi-
ventions to pregnant women stemmed primarily from the cal activities and judicious diets. ACOG supports the rec-
fear of potentially causing harm to the fetus. With the advent ommendation that a moderate exercise program should be
of sophisticated imaging and fetal monitoring over the past part of the prevention and/or treatment plan for women with
266
Pathophysiology 267
gestational diabetes mellitus.24,30 ACOG also recommends The glycogen repletion from the muscles and liver increases
that “for an obese pregnant woman who is gaining less the insulin action that persists after exercise. Adiponectin
weight than recommended but has an appropriately growing regulates insulin sensitivity and glucose homeostasis, and
fetus, no evidence exists that encouraging increased weight low levels of adiponectin are associated with beta cell dys-
gain to conform with the 2009 IOM guidelines will improve function.39 Muscle binding of adiponectin translocates a
maternal or fetal outcomes.”25 glucose transporter type 4 (GLUT4)—which is an insulin-
In 1985, the concept of lifestyle modification was intro- regulated protein found in adipose and striated muscle tis-
duced for the management of pregnant diabetics with a study sue and is responsible for the glucose transport into cells.
comparing hormonal responses to light exercise of diabetic Exercise increases glucose uptake by GLUT4 translocation,
and nondiabetic pregnant women, and it was suggested that increasing the amount of aerobic enzymes in the skeletal
physical activity could be considered a safe intervention in muscles by as much as 20-fold upregulating lipid oxidation
pregnancy as well to improve and control glycemia.26 and mitochondrial biogenesis.
Over the subsequent years, different investigators have Exercise thus improves glycemia primarily through
confirmed and proposed various lifestyle intervention man- two mechanisms: (1) insulin-stimulated muscle glucose
agement schemes to prevent or manage gestational diabetes uptake and (2) insulin-independent glucose transporta-
and type 2 diabetes prior to and during pregnancy.27,28 The tion (GLUT4). As such exercise reduces the hepatic glucose
ADA in its Executive Summary: Standards of Medical Care in production and glycogen metabolism and also lowers total
Diabetes—201418 states that “Women with a history of GDM cholesterol and triglyceride while raising high-density lipo-
found to have prediabetes should receive lifestyle inter- protein levels.
vention or metformin to prevent diabetes.” The American Leptin has a key role in regulating energy intake and
College of Sports Medicine (ACSM) and the ADA issued energy expenditure; during pregnancy, its level increases by
a joint statement in 2010: “Epidemiological studies sug- 150%–200% in the second and third trimester over levels in
gest that higher levels of physical activity may reduce risk the first trimester and twofold to threefold above nonpreg-
of developing gestational diabetes mellitus prior and during nant levels.40 In leptin-deficient subjects with lipodystrophy,
pregnancy.” Randomized clinical trials suggest that moder- decreased production of leptin by adipose tissue has been
ate exercise may lower maternal blood glucose levels in gesta- considered to be a cause for insulin resistance41; conversely,
tional diabetes mellitus.29 The Royal College of Obstetricians exercise-induced reductions in leptins have been ascribed to
and Gynaecologists and the Canadian Diabetes Association changes in energy balance, improvements in insulin sensi-
(CDA) consider exercise an adjunctive therapy for women tivity, and lipid metabolism.42
with gestational diabetes.31,32 Insulin sensitivity progressively decreases in pregnancy,
The CDA guidelines also specify that the frequency, and to counteract this process and maintain euglycemia
intensity, type, and duration of activity should be based on during the third trimester in healthy pregnant women, insu-
each individual’s obstetric risk.32 lin secretion increases two to four times.43
The acquired decrease in insulin sensitivity particu-
larly in obese women in pregnancy is the physiological
Pathophysiology background and rationale for recommending exercise as a
logical intervention to either prevent or manage gestational
The physiological and endocrinological changes in preg- diabetes and prevent further progression to diabetes during
nancy are significant; pregnancy is considered to be a dia- or after pregnancy.
betogenic state and characterized by progressive insulin While prescribing exercise in pregnancy, the anatomical
resistance, similar to the insulin resistance observed in and physiological changes associated with pregnancy have to
patients with type 2 diabetes.24,33,34 be considered.34
In overweight and obese patients, insulin resistance is by
and large a result of the progressive increase in adiposity and
hyperglycemia; increased fat storage will further increase Exercise prescription for sedentary overweight or obese
the insulin resistance. During pregnancy, a major contribu- women to prevent or manage gestational diabetes
tor to insulin resistance is the accumulating adipose tissue, The basic components of exercise prescription include fre-
which acts as an active organ, and placental and other hor- quency, intensity, duration, and type.
mones such as human placental lactogen, cortisol, proges- Energy expenditure and thus insulin sensitivity depend
terone, estrogen, tumor necrosis factor alpha, adiponectin, on the frequency and amount of work (exercise) intensity and
interleukin-6, resistin, and plasminogen activator inhibitor. duration. As to type, aerobic exercise is most frequently pre-
Excessive weight gain in pregnancy and the increased adi- scribed to pregnant women, while experience with resistance
posity lead to further increased release of adipokines and exercise in pregnancy is more limited. Some of the studies
leptin that increase the oxidative stress that contributes to conducted in pregnancy to prevent or manage gestational
the insulin resistance.35,36 Exercise can reverse or prevent diabetes mellitus meet the requirements set for lifestyle
this process by stimulating muscle and liver glucose uptake modification by the ADA for nonpregnant, while other
and nonoxidatively metabolizing glucose into glycogen.37,38 studies appear to not have met the ADA recommendation
268 Role of exercise in reducing the risks of gestational diabetes mellitus in obese women
or quantified the interventions with objective physiological 40

measures to assess the efficacy of the interventions, or the
training effect for energy expenditure, or measured the 20
aerobic capacity at regular intervals to evaluate the level of
Change from rest (%)

training and efficacy, of the intervention (Table 33.2). 0
It is important to emphasize that bouts of exercise also
play an important role in upregulating insulin sensitivity.44 –20
In our experience, 15 minutes of low-intensity frequent
bouts of exercise do not result in significant changes in blood –40
glucose levels in normal nondiabetic pregnant subjects45 but
do make a difference and improve glycemia in sedentary –60
patients with gestational diabetes, particularly obese sub-
jects.46 It is important to note that in healthy normal preg- –80
nant subjects, continuous prolonged exercise in pregnancy 0 1.26 2.51 3.77 5.02 6.28
at about 55% VO2 or higher could significantly decrease glu- (300) (600) (900) (1200) (1500)
cose levels and result in hypoglycemia after 45–60 minutes Total energy expenditure during exercise, MJ (kcal)
of continuous exercise. Comparison has been studied for the
same subjects during and after pregnancy47 (Figure 33.1). Figure 33.2 Energy expenditure during exercise. (From
Thus, continuous strenuous exercise beyond 45–60 minutes Magkos, F. et al., Clin. Sci., 114, 59, 2008. With permission.)
in pregnancy requires some form of supervision.
In the prospective nurses’ health study, it has been dem- results can be obtained in pregnancy through several bouts
onstrated that prior to pregnancy, it is necessary to expend of exercise per week as has been illustrated in our own
at least 1700 kcal/week (the equivalent of 16 hours/week) studies27 (Table 33.1).
in light exercise or at least 840 kcal/week (the equivalent In terms of intensity, in a joint ACSM and American
of 8 hours/week) in vigorous exercise to reduce the risk of Heart Association exercise recommendations, light physical
developing gestational diabetes (Figure 33.2).48 activity is defined as requiring less than 3 METs, moderate
The nurses’ health study has determined that the relative activities 3–6 METS, and strenuous activities greater than
risk of gestational diabetes mellitus (GDM) was reduced by 6 METs.55
17% with prepregnancy light exercise and 30% with more METs or metabolic equivalents are used to quantify the
vigorous exercise.48 In another study, it has been demon- metabolic cost of exercise in terms of oxygen consumption.
strated that the exercise-induced changes with a homeostasis The equivalent of 3.5 mL O2/kg/minute is 1 MET and is
model assessment (HOMA) of insulin resistance, measured obtained at rest. Walking at a moderate pace for 1.5 miles/
the following day postexercise, were negatively and curvilin- day will expend approximately 6 METs. Walking is the type
early correlated to energy expenditure (Figure 33.2). of activity that can be readily prescribed and accepted by
The HOMA score is a measure of insulin resistance sedentary, overweight, or obese subjects.
decreased by 30% in subjects who expended more than Relevant to exercise prescription for pregnant diabetic
900 kcal (strenuous) during one bout of exercise.49 Similar patients and based on the aforementioned studies, we have
established that pregnant women could safely complete at
least one moderate bout of exercise of 30–45 minutes/day
5.5
Pregnant and if possible one bout of exercise after each meal to expend
Nonpregnant at least 200 kcal or more per day or approximately 6 METs/
5 ±1SEM day and 1400 kcal/week, the necessary energy expenditure
* to obtain or maintain euglycemia in pregnancy. Less than
Glucose (mmol/L)
* 16 MET hour/week would not suffice to either prevent GDM

*
4.5
or manage hyperglycemia of patients who have developed
GDM A2 or have preexisting diabetes. In our experience,
4 within 10 days of initiating an exercise program in GDM
patients, 60% will attain euglycemia.21,27,46 Physical activity
in early and throughout pregnancy can prevent or reduce
3.5
GDM risk.
Exercise Rest Meeting the ADA criteria for the prevention or man-
3 agement of type 2 diabetes in nonpregnant subjects with
0 15 30 45 60 75 impaired glucose tolerance, impaired fasting glucose, or
Time (minutes) A1C 5.7%–6.4% in our and other investigators’ experi-
ence has been validated in pregnancy as well (Table 33.2).
Figure 33.1 Glucose concentrations during prolonged
exercise: Pregnant versus nonpregnant. *p < 0.05. (From
The combination of nutritional control and exercise has
Soultanakis, H.N. et al., Semin. Perinatol., 20(4), 315, 1996. been demonstrated to better control glycemia also in
With permission.) pregnancy.27,28 It also appears that meeting the criteria
Practice points 269
Table 33.1 Lifestyle intervention of diet and exercise for obese GDM
patients study 1 (2007)
Subjects who lost weight or no Subjects who gained weight

weight change (n = 30) (n = 66)
Characteristic Mean (SD) n % Mean (SD) n %
Birth weight (g) 3286 (399) 3340 (612)
Normal 23 96 43 77
LGA (>4000 g) 1 4 10* 18*
SGA (<2500 g) 0 0.0 3 5
GA (week) 39 (1) 38 (2)
C-section 10 42 28 49
Vaginal 14 58 29 51
Exercise and diet 18 60 21 32*
Diet only 12 40 45 68*
Source: Reproduced from Artal, R. et al., Appl. Physiol. Nutr. Metab., 32(3), 596, 2007. With
permission.
*p < 0.05 compared to lost weight/no weight change.
established by the ADA for nonpregnant individuals to Exercise prescription for sedentary, overweight and
reduce the risk for or manage diabetes may be valid for obese subjects
pregnant subjects as well. This is illustrated in some but not Over the years, we have developed exercise prescription
all of the published trials in pregnant patients (Table 33.2). regimens for sedentary, overweight, or obese women who
Clearly in many of these intervention trials in pregnancy, are at risk for or acquired gestational diabetes or affected
the subjects did not engage in activities sufficient to elicit by type 2 diabetes.27 Table 33.3 illustrates our current
normoglycemia.22 It also appears that in many of the inter- recommendations.
ventional studies, the subjects’ BMIs were by and large at or
below 25–30. Significant is the fact that among subjects that
engaged in even a minimum amount of exercise, certain
comorbidities were precluded among them, fetal macroso-
Summary and conclusions
mia and reduced cesarean sections. Pregnancy poses a risk for pregnant women particularly
A potential cause for intervention failure in many of overweight, obese, or sedentary women to develop gesta-
the trials could be related to additional gestational weight tional diabetes. Women that engage in any recreational
gain. In our recent experience in the care of obese pregnant physical activity during the first 20 weeks of pregnancy
women with a BMI exceeding 30, a gestational weight gain experience a 48% reduction in developing gestational diabe-
of less than 5 kg for the duration of pregnancy has favorable tes when compared to sedentary women.8,57
outcomes.15 Several organizations based their recommendations on
In a population-based birth cohort study conducted in reviews and meta-analyses of randomized and nonrandom-
Central New York, we compared GDM prevalence among ized clinical trials58 have reached the conclusion that ante-
pregnant subjects who engaged in physical activities to preg- natal lifestyle interventions of judicious gestational weight
nant subjects who did not. Physical activity was associated gain and exercise could reduce or prevent gestational diabetes
with a 50% reduction in the rates of GDM among women in the overweight and obese pregnant women and could be
with a BMI greater than 33 kg/m2 (OR = 1.9, 95%, confidence safely prescribed to pregnant patients who have acquired
interval 1.2–3.1) (see Figure 33.3). gestational diabetes or type 2 diabetes.
This study suggests that exercise has a stronger and Pregnancy provides a unique opportunity to adopt
more immediate protective effect in women with higher lifestyle modification, since pregnant women have better
BMI and that lifestyle interventions may be more effec- access to medical care and supervision than any other time.
tive and reduce insulin resistance and improve insulin
sensitivity in subjects who are obese and insulin resistant
than in those who are obese but metabolically healthy; this Practice points
distinction remains to be studied. Since the data obtained
for this study were obtained from birth certificates, we ●● Obesity is currently the leading cause for gestational
could not determine the timing, intensity, and duration of diabetes.
exercise and if the activities were conducted throughout ●● Additional gestation weight gain in obese pregnant
pregnancy.8 women increases the risk for gestational diabetes.
Table 33.2 Lifestyle interventions for overweight women
ADA
Author Study n BMI GWG (kg)b criteria LGAa Comment
1. Artal Intervention 39 35.2 ± 7.2 21.5 Yes 11.8% GDM controlled %
(2007)21 Diet 57 33.5 ± 9.2 29.8 No 15.2% macrosomia in subjects
gaining weight—17.9%
Vs. losing/no
change—4.2%
2. Barakat Intervention 40 22.7 ± 2.8 12.5 ± 3.2 No No No GDM
(2012)50 Control 43 23.0 ± 2.9 13.8 ± 3.1 No Three subjects with GDM
3. de Barros Intervention 32 29.6 ± 4.1 30.4 ± 4.3 No 3.1 Resistance exercise
(2010)53 Control 32 29.4 ± 4.0 29.9 ± 4.1 9.4
4. Bo S Intervention 101 27.6 ± 4.1 NA Yes 9.9% Exercise but not behavior
(2014)54 Behavior 99 27.5 ± 4.2 NA No improved post prenatal
glucose
5. Bung Intervention 17 35.2 ± 4.8 NA No 10.1% Exercise and diet as efficient
et al. Insulin 17 35.9 ± 3.9 NA — 11.7% as insulin for GDM
(1991)46 management
6. Dempsey Intervention Continuous, NR Yes NR 74% lower GDM risk
(2004)57 ≥6 hours/ 326 categorical, 56% lower GDM risk
week and
289
≤6 hours/ multivariable
week models
None 294
7. Hopkins Intervention 47 26.7 NR Yes 4.3% No effect on insulin
(2009)59 Control 37 25.5 No 5.4% sensitivity in either group
Self-reported compliance
Lower cord IGF—1 in
exercise subjects
8. Asbee Intervention 57 25.5 25.1 No NR C/S 25%
(2009)60 Control 43 25.6 40.1 No NR C/S 58.3%
No difference in GDM
9. Kinnunen Intervention 128 >25 46.8% (above Yes 2.7% GDM—15.8%
(2012)61 Control 110 >25 recommendation) No 2.8% GDM—12.4%
54.4% (above LGA—reduced in both
recommendation) groups
Physical activity not
quantified
a LGA, large for gestational age.
b GWG, gestational weight gain.
14
n = 12, 290
12
10
No exercise
% With GDM
Exercise
8
0
<18 18–20 21–23 24–26 27–29 30–32 33–35 36–38 39–41 42+
Body mass index (BMI)
Figure 33.3 Prevalence of GDM by BMI and Physical Activity Status 1995–1996 (Central New York). (From Dye, T.D. et al., Am.
J. Epidemiol., 146(11), 961, 1997. With permission.)
References 271
●● Exercise or limited gestational weight gain or in combina-

Table 33.3 Exercise prescription for sedentary,
tion can reduce the risk of gestational diabetes.
overweight, or obese women with GDM who are
●● A daily exercise routine of at least 30 minutes/day can
unaccustomed to exercise
prevent or improve hyperglycemia.
●● Pregnancy is an ideal time for behavioral modifications,
Previously sedentary Target exercise
since patients are under frequent and close supervision.
and/or overweight/ energy expenditure
obese pregnant women RPEa (MET hours/week)
Weeks 1–3 of training 12–14 ≥16 Research agenda
Weeks 3–6 of training 13–15 28
Weeks 6–9 GAb of 15–16 28 ●● Cost-effectiveness in obesity and diabetes prevention.
training ●● When are interventions likely best prevent gestational
a RPE, rate of perceived exertion 6 equals no exertion and 20 diabetes?
equals maximal exertion. ●● The role of educators and healthcare providers in instill-
b GA, gestational age. ing lifelong healthy behavior to prevent obesity and
comorbidities.
REFERENCES
1. World Health Organizations. Global Recommendations on 17. Wolff S, Legarth J, Yangsgaard K et al. A randomized trial of the
Physical Activity for Health, WHO: Geneva, Switzerland, 2010. effects of dietary counseling on gestational weight gain and glu-
2. Knoler WC, Barrett-Connor E, Fowler SE, Hammon RF, Lachin cose metabolism in obese pregnant women. Int J Obes 2008; 32:
JM, Walker EA, Nathan DM. Reduction in incidence of type 2 495–501.
diabetes with lifestyle interventions or metformin. N Engl J Med 18. Standards of medical care in diabetes—2014. Diabetes Care
2002; 346: 393–403. 2014; 37(1): S29.
3. Paul XR, Lig W, Wang JX et al. Effects of diet and exercise in 19. Artal R, Lockwood CJ, Brown HL. Weight gain recommendations
preventing NIDDM in people with impaired glucose tolerance. in pregnancy and the obesity epidemic. Obstet Gynecol 2010;
The DaQing IGT and Diabetes Study. Diabetes Care 1997; 20: 115(1): 152–155.
537–544. 20. Mottola MF, Giroux I, Gratton R et al. Nutrition and exercise pre-
4. Tuomilehto J, LindstrÖm J, Eriksson JG et al. Prevention of vent excess weight gain in overweight pregnant women. Med Sci
type 2 diabetes mellitus by changes in lifestyle among subjects Sports Exerc 2010; 42(2): 265–272.
with impaired glucose tolerance. N Engl J Med 2001; 344(18): 21. Artal R, Catanzaro RB, Gavard JA. A lifestyle intervention of
1343–1350. weight-gain restriction: Diet and exercise in obese women with
5. Robinson WR, Cheng MM, Hoggatt KY et al. Childbearing is not gestational diabetes mellitus. Appl Physiol Nutr Metab 2007;
associated with young women’s long-term obesity risk. Obesity 32(3): 596–601.
2014; 22: 1126–1132. 22. Oten-Ntim E, Varma R, Croker H et al. Lifestyle interventions for
6. Kelly C, Booth G. Diabetes in Canadian Women. BMC Womens overweight and obese pregnant women to improve pregnancy
Health 2004; 4(Suppl. 1): S16–S24. outcome: Systematic review and meta-analysis. BMC Med 2012;
7. Dempsey JC, Sorenson TK, Williams MA et al. Prospective study 14: 47.
of gestational diabetes risk in relation to maternal recreational 23. Stafne SN, Salvesen KA, Romundstad PR et al. Regular exer-
physical activity before and during pregnancy. Am J Epidemiol cise during pregnancy to prevent gestational diabetes. Obstet
2004; 159(7): 663–670. Gynecol 2012; 119: 29–36.
8. Dye TD, Knox KL, Artal R et al. Physical activity, obesity and 24. ACOG Practice Bulletin Number 137, 2013. Gestational diabetes
diabetes in pregnancy. Am J Epidemiol 1997; 146(11): 961–965. mellitus.
9. Chu SY, Callaghan WM, Kim SY et al. Maternal obesity and 25. ACOG Committee Opinion Number 548, 2013. Weight gain dur-
risk of gestational diabetes mellitus. Diabetes Care 2007; 30: ing pregnancy.
2070–2076. 26. Artal R, Wiswell R, Romem Y. Hormonal responses in exercise
10. Weiss JL, Molone FD, Emig D et al. Obesity, obstetric complica- in diabetic and non-diabetic pregnant patients. Diabetes 1985;
tions and caesarian delivery rate: A population-based screening 34(2): 78–80.
study. Am J Obstet Gynecol 2004; 190: 1091–1097. 27. Artal R, Zavorsky GS, Catanzaro RB. Exercise recommendations
11. HAPO Study Cooperative Research Group. Hyperglycemia and in women with gestational diabetes mellitus. In: Kim C, Ferrara A,
Adverse Pregnancy Outcome (HAPO) study: Associations with eds., Gestational Diabetes during and after Pregnancy, Springer:
maternal body mass index. Int J Obstet Gynecol 2008; 198(5): London, U.K., 2010, pp. 243–257.
525 e1–525 e5. 28. Mottola MF, Hammond J, McManus R et al. Effects of a controlled
12. Pedersen J. The Pregnant Diabetic and Her Newborn, 2nd ed., nutrition and mild exercise program on glucose metabolism in
Munksgaard: Copenhagen, Denmark, 1977, pp. 221–232. healthy pregnant women. Can J Appl Physiol 1999; 24: 468.
13. Rasmussen KM, Yaktine AL, eds. Weight Gain during Pregnancy: 29. The American College of Sports Medicine and The American
Re-Examining the Recommendations, The National Academic Diabetes Association: Joint position statement. Exercise and type
Press: Washington, DC, 2009. 2 diabetes. Diabetes Care 2010; 33: e147–e167.
14. Artal R, Lockwood CJ, Browh HL. Weight gain recommendations 30. ACOG. Physical activity and exercise during pregnancy and the
for pregnancy and the obesity epidemic. Obstet Gynecol 2010; postpartum period. Obstet Gynecol 2015; 126: 1326–1327.
115(1): 152–155. 31. Royal College of Obstetricians and Gynaecologists Exercise in
15. Kiel DW, Dodson EA, Artal R et al. Gestational weight gain and Pregnancy (RCOG Statement 4), 2006.
pregnancy outcomes in obese women: How much is enough? 32. Canadian Diabetes Association Clinical Practice Guidelines
Obstet Gynecol 2007; 110: 752–758. Expert Committee. Canadian Diabetes Association clinical prac-
16. Vinter CA, Jensen DM, Ovesen P et al. The LIP (Lifestyle in tice guidelines for the prevention and management of diabetes in
Pregnancy) study: A randomized controlled trial of lifestyle inter- Canada. Can J Diabetics 2003; 27: S99–S105.
vention in 360 obese pregnant women. Diabetes Care 2011; 34: 33. Buchanan TA, Xiang AH. Gestational diabetes mellitus. J Clin
2502–2507. Invest 2005; 115(3): 485–491.
34. Artal R, O’Toole M. Guidelines of the American College of 48. Zhang C, Solomon CG, Manon JE, Hu FB. A prospective study of
Obstetricians and Gynecologists for exercise during preg- pregravid physical activity and sedentary behaviors in relation to
nancy and the postpartum period. Br J Sports Med 2003; 37: the risk for gestational diabetes mellitus. Arch Intern Med 2006;
6–12. 166: 543–548.
35. Miehle K, Stepan H, Fasshauer M. Leptin adiponectin and other 49. Magkos F, Tsekouras Y, Kavouras SA et al. Improved insulin sen-
adipokines in gestational diabetes mellitus and pre-eclampsia. sitivity after a simple bout of exercise is curvilinearly related to
Clin Endocrinol 2012; 76(1): 2–11. exercise energy expenditure. Clin Sci 2008; 114: 59–64.
36. Golbidi S, Laher I. Potential mechanisms of exercise in gesta- 50. Barakat R, Cordero Y, Coteron J et al. Exercise during pregnancy
tional diabetes. J Nutr Metab 2013; 285948. improves maternal glucose screen at 24–28 weeks: A random-
37. Hespel P, Verganwen L, Vandenberghe K, Richter EA. Important ized controlled trial. Br J Sports Med 2012; 46: 656–661.
role of insulin and flow in stimulating glucose intake in contract- 51. ACOG Committee Opinion, No. 267, 2002 (reaffirmed 2010).
ing skeletal muscle. Diabetes 1995; 44: 210–215. Exercise during pregnancy and the postpartum period.
38. Pencek RR, James FD, Lacy DB, Jabbour K, Williams PE, Fueger 52. Colberg SR, Sigal RJ, Fernhall B et al. Exercise and type 2 diabe-
PT, Wasserman DH. Interaction of insulin and prior exercise tes. Diabetes Care 2010; 33: e147–e167.
on the partitioning of an intestinal glucose load between 53. de Barros MC, Lopes MAB, Francisco RPV et al. Resistance exer-
splanchnic bed and skeletal muscle. J Clin Invest 1996; 98: cise and glycemic control in women with gestational diabetes
125–135. mellitus. Am J Obstet Gynecol 2010; 203: 556e1–556e6.
39. Retnokaran R, Hanley AJG, Raif N et al. Adiponectin and beta 54. Bo S, Rosato R, Ciccone G et al. Simple lifestyle recommenda-
cell dysfunction in gestational diabetes: Pathophysiological tions and the outcomes of gestational diabetes. Diabetes Obes
implications. Diabetologia 2005; 48(5): 993–1001. Metab 2014; 16(10): 1032–1035. doi:1111/dom.12289.
40. Laivuori H, Kaaja K, Koistinen H et al. Leptin during and after 55. Haskel WH, Lee IM, Pate RR et al. Physical activity and public
preeclamptic or normal pregnancy; its relation to serum insulin health: Updated recommendations for adults from the American
and insulin sensitivity. Metabolism 2000; 49: 259–263. College of Sports Medicine and the American Heart Association.
41. Oral EA, Simha V, Ruiz E et al. Leptin-replacement therapy for Circulation 2007; 116(9): 1081–1093.
lipodystrophy. N Engl J Med 2002; 346: 570. 56. Borg GA. Psychophysical basis of perceived exertion. Med Sci
42. Kraemer RR, Chu H, Costracane VD. Leptin and exercise. Exp Sports Exerc 1982; 14(5): 377–381.
Biol Med 2002; 227(9): 701–708. 57. Dempsey JC, Butler CL, Sorenson TK et al. A case-control study
43. Catalano PM, Tyzbir ED, Roman NM et al. Longitudinal changes of maternal recreational physical activity and risk of gestational
in insulin release and insulin resistance in nonobese pregnant diabetes mellitus. Diabetes Res Clin Pract 2004; 66: 203–215.
women. Am J Obstet Gynecol 1999; 165: 1667–1672. 58. Thangaratinam S, Rogozinska E, Jolly K et al. Effects of interven-
44. Heath GW, Gavin JR III, Hinderliter JM et al. Effects of exercise tion in pregnancy on maternal weight and obstetric outcomes:
on glucose and insulin sensitivity. J Appl Physiol 1983; 55(2): Meta-analysis of randomized evidence. BMJ 2012; 344: e2088.
512–517. 59. Hopkins SA, Baldi JC, Cutfield WS et al. Exercise training in preg-
45. Artal RM, Platt LD, Sperling M et al. Exercise in pregnancy: I. nancy reduces offspring size without changes in maternal insulin
Maternal cardiovascular and metabolic responses in normal sensitivity. J Clin Endocrinol Metab 2010; 95: 2080–2088.
pregnancy. Am J Obstet Gynecol 1981; 140: 123–127. 60. Albee SM, Jenkins TR, Butler JR et al. Preventing excessive weight
46. Bung P, Artal R, Khodiguian N, Kjos S. Exercise in gestational dia- gain during pregnancy through dietary and lifestyle counseling: A
betes. An optional therapeutic approach. Diabetes 1991; 40(2): randomized controlled trial. Obstet Gynecol 2009; 113: 305–312.
182–185. 61. Kinnunen TI, Raitanen J, Aittasalo M, Luoto R. Preventing exces-
47. Soultanakis HN, Artal R, Wiswell RA. Prolonged exercise in sive gestational weight gain–a secondary analysis of a c luster–ran-
pregnancy: Glucose homeostasis, ventilatory and cardiovascular domized controlled trail. Eur J Clin Nutr 2012; 66(12): 1344–1350.
responses. Semin Perinatol 1996; 20(4): 315–327.
34 Role of bariatric surgery in obese
women planning pregnancy
Ron Charach and Eyal Sheiner
of postbariatric surgery parturients. This chapter addresses

Introduction the safety of bariatric procedures in pregnancy as compared
Obesity is an epidemic affecting both developed and devel- to the general population and whether it is safer than preg-
oping countries and is defined as prepregnancy body mass nancy in the obese. General clinical guidelines and recom-
index (BMI) ≥30 kg/m2.1 Excessive weight gain and obesity mendations are provided throughout the text.
as a chronic disease are considered to be two of the most
pressing concerns in the developed world due to their grow-
ing threat to health in countries all over the world. The prev-
alence of overweight and obesity has increased dramatically Obesity impacts on pregnancy and
with alarming rates over the recent decades and has nearly fertility (Table 34.1)
doubled since 1980.2–4 In 2008, more than 1.4 billion adults
20 years and older were overweight. Of these, nearly 300 mil- Fertility and infertility treatments
lion women were obese.4 In the United States, nearly 35% of Obesity appears to be associated with reduced fecundity
adults were obese and over one-third of women 20 years and rates and increased time to pregnancy. Van der Steeg et al.11
older were obese in 2011–2012.5 Excessive weight gain and reported that obese women, even with a regular cycle, with
obesity is the fifth leading risk for global deaths, and accord- a BMI over 29 kg/m2 have lower pregnancy rates compared
ing to the latest WHO fact sheet, at least 2.8 million adults with those with normal weight. Other studies also found
die each year as a result of being overweight or obese. Obesity an association between obesity and reduced fecundity and
and excessive weight gain may also contribute substantially suggested that weight loss could increase fecundity for over-
to the burden of chronic health conditions such as diabetes, weight and obese women.12,13 Not all studies are consistent
ischemic heart disease, stroke, respiratory problems, cer- regarding infertility treatments in obese women. Several
tain cancers (endometrial, breast, and colon), and osteoar- studies have shown that higher BMI was inversely related to
thritis.4,6 In concordance to this worldwide increase in the follicular growth in response to gonadotropins and is asso-
prevalence of overweight and obesity, close to one-third of ciated with reduced probability of achieving pregnancy in
women of childbearing age (20–39 years) are classified as women receiving assisted reproduction.14–16 Other studies
obese, and an additional quarter of women in this age group argued that although higher threshold dose of gonadotro-
are overweight.3,7,8 Those disturbing findings can eventu- pins and more days of stimulation were required, there is
ally lead to adverse consequences for their future reproduc- no difference in the overall outcome and success of treat-
tive health, pregnancy, and long-term health.9 Significant ment cycles in comparison to normal-weight women.17,18
reduction in obesity-related reproductive complications may In a systemic review of 33 studies including almost 48,000
be achieved by weight loss. Surgical treatment has a great treatment cycles of in vitro fertilization or intracytoplasmic
potential to treat obesity in women of reproductive age, as sperm injection treatments in women who were overweight
Sjostrom reported bariatric surgery to be the most efficient or obese (BMI > 25), significantly higher miscarriage rate
means of weight loss in severely obese patients.10 (relative risk, RR = 1.31, p < 0.0001) as well as lower clini-
This chapter attempts to review the most recent avail- cal pregnancy (RR = 0.90, p < 0.0001), and live birth rates
able data on the role of bariatric surgery in the obese preg- (RR = 0.84, p = 0.0002) were found compared to women with
nant woman and discusses the management and outcomes normal weight (BMI < 25).19
273
274 Role of bariatric surgery in obese women planning pregnancy
emergency cesarean delivery (CD), and this risk increases with

Table 34.1 Adverse fertility and pregnancy
increasing obesity.22,35,36 High rates of associated anesthesia
outcomes found to be associated with obesity
complications, wound disruption, infection, and deep venous
thromboembolism were also found in obese parturient.21,37–39
Fertility
Whether obese woman are at increased risk for
●● Reduced fecundity
●● Higher threshold for successful gonadotropin postpartum hemorrhage is a matter of controversy. While
treatments several small studies did not find an association,25,40,41 Sebire
●● Lower clinical pregnancy rate et al., in a population-based cohort study, reported a 30%
Antepartum increased risk for a major postpartum hemorrhage (PPH)
●● Spontaneous abortion in gravidas with moderately raised BMI and about 70%
●● Gestational diabetes mellitus increased risk for women with very raised BMI compared
●● Pregnancy-induced hypertension
with women of normal BMI.42
●● Exacerbation of chronic maternal diseases
Intrapartum and postpartum

●● Longer first stage of labor Perinatal complications
●● Increased risk of cesarean delivery
●● Anesthesia complications
Adverse perinatal outcomes may be associated with overweight
●● Wound disruption and obesity. Maternal obesity appears to be associated with an
●● Infection increase in the rates of some congenital malformations (espe-
●● Deep venous thromboembolism
cially neural tube defects [NTDs]), and the risk may increase
●● Postpartum hemorrhage
with increasing maternal weight.43–45 In a systematic review
Perinatal (39 studies) and meta-analysis (18 studies), Stothard et al.44
●● Congenital malformations (especially neural tube reported that maternal obesity is likely to be associated with a
defects) small increased risk of a range of structural anomalies such as
●● Fetal and early neonatal death
●● Large for gestational age

spina bifida, NTD, cardiovascular anomalies, septal anomalies,
●● Shoulder dystocia cleft palate, cleft lip and palate, anorectal atresia, hydrocephaly,
●● Future metabolic syndrome and limb reduction anomalies. It should be emphasized that
these findings may be partially explained by technical prob-
lems with ultrasound detection (reducing rates of early termi-
Antepartum complications nations) and lower folic acid levels (body distribution).46,47
Maternal obesity has consistently been associated with
Pregestational maternal obesity is known to be a signifi-
increased rates of fetal and early neonatal death. Chu et al.,48
cant risk factor for adverse outcomes during pregnancy.20–22
in a meta-analysis of nine controlled studies, reported
During early pregnancy, there is an increased risk for spon-
that maternal obesity is associated with an increased risk
taneous abortion.23 In a systematic review of 16 studies,
of stillbirth in comparison to normal-weight women. The
patients with a BMI of ≥25 kg/m2 had significantly higher
mechanism of fetal compromise has not been fully deter-
risk of abortion (odds ratio [OR], 1.67).24 Later during preg-
mined and is likely to be multifactorial. Neonatal death may
nancy, maternal metabolic manifestations of obesity and
be caused largely due to pregnancy complications leading to
metabolic syndrome may be apparent.
preterm birth or other disorders.49
A higher prevalence of gestational diabetes mellitus
Several studies have shown that pregestational mater-
(GDM) and pregnancy-induced hypertension (PIH) was
nal obesity and excessive weight gain during pregnancy
found in obese women.20–23,25–27 A dose-dependent effect has
are independent risk factors50 for macrosomic (>4000 g)
also been demonstrated between increasing maternal BMI
or large-for-gestational-age (LGA) neonates.22,25,37,40,51,52
and risk for GDM and PIH.27,28
Increasing prepregnancy weight and maternal BMI has a
In obese gravidas, physiologic changes during pregnancy
linear relationship with birth weight and macrosomia.52–54
may exacerbate maternal chronic problems such as sleep
One potential sequela of fetal LGA might be shoulder dys-
apnea, nonalcoholic fatty liver disease, and cardiac or renal
tocia (SD). Indeed, several studies found that the risk for SD
dysfunction.29
increased with increased maternal BMI.20,55
Intrapartum and postpartum complications

During intrapartum, it appears that in overweight and obese Long-term complication and metabolic programming
women, first-stage labor progression is significantly slower than Recently, there is growing evidence in epidemiologi-
that of normal-weight women.30–34 Moreover, labor proceeds cal studies and animal models that embryonic, fetal, and
more slowly as BMI increases.30–34 Vahratian et al., in a cohort early postnatal subtle environment effects might appear as
study of 612 nulliparous women, found that the median dura- health issues later in life (i.e., obesity, type 2 diabetes mel-
tion of labor from 4 to 10 cm was significantly longer for litus). These effects are known to be designated as metabolic
both overweight and obese women, compared to normal- programing.56,57 Laitinen et al.58 reported that offspring of
weight women (7.5, 7.9, and 6.2 hours, respectively).34 The overweight or obese mothers had higher mean BMIs at each
obese patient is also at an increased risk of both elective and age point tested than did children born to underweight or
Types of bariatric procedures 275
normal-weight mothers. Children exposure to maternal Types of bariatric procedures

obesity posed a strong predictor for significant risk of future
metabolic syndrome.59 Hochner et al.60 found that maternal To date, there are three fundamental types of bariatric
size both before and during pregnancy was also associated surgical procedures: (1) malabsorptive, (2) restrictive, and
with cardiometabolic risk factors in young adult offspring. (3) procedures with both a restrictive and malabsorptive
component (Table 34.2).
Treating obesity: What are the Malabsorptive procedures

options? Pure malabsorptive procedures, such as biliopancreatic diver-
sion (BPD) and jejunoileal bypass, cause weight loss primar-
It is strongly advised to encourage obese women to undertake ily by decreasing the absorption of nutrients either through
a preconception assessment, counseling, and weight reduction diversion of the biliopancreatic secretions that are essential
program.1 Preconception weight reduction has been proven for the absorption of micronutrients or bypassing a significant
to be the best way to decrease medical and obstetric compli- portion of small bowel surface area. This type of procedures
cations in obese women planning pregnancy.61,62 Currently, can bring about superior weight loss results, but they may also
there are two main options for obesity treatment: nonsurgical bring about serious metabolic complications and micronutri-
treatment and bariatric surgery. Nonsurgical treatment is usu- ent deficiencies. Due to their complexity and the linkage to
ally multicomponent and includes dietary changes, exercise, significant safety concerns, they are now rarely performed.65
behavior modification, and various pharmacotherapies. Until
recently, observational studies and randomized controlled tri-
Restrictive procedures
als were not powered enough to detect differences between
those two approaches on multiple health outcomes. In 2007, Restrictive procedures, such as laparoscopic adjustable gas-
Sjöström10 conducted the Swedish Obese Subjects study, which tric banding (LAGB), VBG, and probably also sleeve gas-
is currently one of the best sources of evidence regarding long- trectomy (with controversy), are frequently performed in
term consequences of bariatric surgery. This prospective, non- the United States and Europe.66 The least invasive procedure
randomized trial included 4047 obese persons, of which 2010 is LABG, which is completely reversible and has the lowest
individuals underwent bariatric surgery and 2037 matched mortality rates.65 In this procedure, a fluid-filled band is
controls were given conventional treatment. Maximal weight placed around the stomach near the fundus, reducing gastric
loss after bariatric surgery was achieved after 1–2 years and volume, reservoir capacity, and eventually solid food intake.
was in average 32% for gastric bypass, 25% for vertical-banded Although these procedures are simpler and safer in compari-
gastroplasty (VBG), and 20% for gastric banding. These find- son to malabsorptive procedures, they tend to produce more
ings were opposed to body weight change of only 1%–2% for gradual weight loss and have the disadvantages of frequent
the conventional treatment group.10 After 10 years, the weight office visits for band adjustment and that some high-energy
losses were, respectively, 25%, 16%, and 14% in average and foods (soft solids and liquid) may bypass the restriction.
were maintained also in the 15 years’ follow-up. The authors
concluded that surgery is the only effective treatment for obe- Mixed procedures: Combination of restriction and
sity with a dramatic positive effects on most cardiovascular risk malabsorption
factors and type 2 diabetes mellitus and with significant reduc- The most commonly performed procedure in the world is
tion of mortality (adjusted hazards ratio of 0.71, p = 0.01).10 currently of the combined type with the Roux-en-Y gastric
Likewise, a recent meta-analysis that included 11 studies with bypass (RYGB) accounting for 46.6% of all bariatric pro-
796 individuals (range of mean BMI at baseline 30–52) demon- cedures.66 In the RYGB, a small gastric pouch restricts oral
strated that bariatric surgery is more efficient than nonsurgi- intake, and small bowel reconfiguration provides additional
cal treatment for obesity for up to 2 years of follow-up. Higher mild malabsorption and neuroendocrine changes favoring
rates of body weight reduction and remission of type 2 diabetes weight loss.65
and metabolic syndrome were found in the bariatric surgery
group.63 Nevertheless, it should be emphasized that bariatric
surgery could also lead to some postoperative complications Table 34.2 Types of bariatric procedures
such as iron-deficiency anemia (mainly in malabsorptive bar-
iatric surgery) and reoperations.63 Malabsorptive
●● Biliopancreatic diversion
Given the limited long-term success of lifestyle modifica- ●● Jejunoileal bypass
tions and medical treatments, rates of surgical weight loss pro- ●● Biliopancreatic diversion with duodenal switch
cedures have increased dramatically over the recent decade.
Restrictive
Between the years 1998 and 2005, the incidence of bariat- ●● Laparoscopic adjustable gastric band
ric surgery in the United States has increased by 800%.64 In ●● Vertical banded gastroplasty
women within the 18 to 45 years age group, it has accounted ●● Sleeve gastrectomy (may also have malabsorptive effect)
for 83% of procedures.64 Those findings stress the need to

Mixed procedures
address this new obstetric group with unique risks and pos- ●●Roux-en-Y gastric bypass
sible fetal and maternal outcomes.
General outcomes oral iron treatments along with multivitamin, folate, and
Inconsistency exists regarding the long-term outcomes B12 supplementation are imperative. Gurewitsch et al.74 even
of different types of bariatric surgery. A recent systematic proposed that female patients who had RYGB in the child-
review of all bariatric procedures with a follow-up of 10 or bearing age should undergo preconception treatment with
more years showed greater than 50% excess weight loss for parenteral iron to avoid the risks of transfusion therapy dur-
all current procedures. The weight losses for LAGB and other ing pregnancy.
bariatric procedures were substantial and with similar excess Marceau et al.75 reported a significant decrease in mean
weight loss in the long term.67 Angrisani et al., however, in a serum albumin concentration among women after BPD,
recent study that compared outcomes of patients randomly with 1.6% of them requiring parenteral nutrition. Eerdekens
assigned to undergo LAGB or RYGB at 10 years, reported et al.76 described five case reports of postoperative excessive
that RYGB had better results than LAGB in terms of excess vomiting or fat malabsorption that eventually resulted in
weight loss results (76.2% vs. 46.2%) at 10 years’ follow-up.68 severe vitamin K deficiency and fetal cerebral hemorrhage.
Nevertheless, LAGB was found to be safer in terms of early Other case reports have described NTD possibly related to
and long-term surgical complication.68 folate deficiency,77 fetomaternal electrolyte imbalances,78
and microphthalmia attributed to vitamin A deficiency.79
In light of these findings, regardless of the type of bariatric
procedure, routine nutritional screening, recommendations
Postbariatric surgery: Pregnancy for appropriate supplements, and monitoring compliance of
complications and management this unique population are obligatory.73 Concerns about low
birth weight caused by malnourishment of mother and fetus
Since concerns arose regarding the potential impact of are discussed later in this chapter.
bariatric surgery on future pregnancies, a multidisciplinary
team approach of a bariatric surgeon and a maternal–fetal
medicine specialist is probably the best way to manage Screening for gestational diabetes mellitus
subsequent pregnancies. Special concern should be given Glucose screening in several postoperative patients should
to screening for gestational diabetes, nutritional guidelines be managed differently as compared to the general popula-
and supplementation, gastrointestinal complaints, and tion. Glucose challenge test (GCT) or glucose tolerance test
gravidas who still remain obese during pregnancy.69 For are not well tolerated in those patients who have undergone
the appropriate management of postbariatric pregnancy, a procedure with malabsorptive component such as RYGB,
see Table 34.3. due to the “dumping syndrome” that occurs in approxi-
mately 50% of these patients.80 “Dumping syndrome” is
caused by ingestion of high amounts of refined sugars that
Nutritional deficiencies
leads to a hyperosmolar environment and fluid shift from
Several case studies have demonstrated nutritional deficien- the intravascular compartment to the small bowel lumen,
cies in postoperative pregnancies.70–72 Procedures involving causing distension, cramping, nausea, vomiting, and diar-
malabsorption, such as RYGB, appear more likely to cause rhea. Reasonable alternatives for screening are measuring
nutritional deficiencies due to hormonal and nutritional glycated hemoglobin (A1C) and assuming overt diabetes is
changes as compared to purely restrictive procedures such as present if it is elevated (≥6.5%) and measuring fasting and
LAGB. Although only one case of anemia was demonstrated postprandial blood sugars for 1–2 weeks in the second tri-
in a study of 79 consecutively enrolled, previously banded mester.81,82 Restrictive-type bariatric procedures such as
pregnant patients,70 all types of bariatric surgery may be LAGB are not reported to cause dumping syndrome; there-
complicated by iron deficiency. Malabsorptive procedures fore, women after LAGB procedure can undergo standard
may be more refractory to treatment and are also prone to be testing for GDM.
associated with an increased risk of calcium, vitamin D, and
vitamin B12 deficiencies.73 Therefore, prenatal and perinatal
Gastrointestinal complications
Life-threatening surgical complications of both malabsorp-
Table 34.3 Appropriate management of
tive and restrictive procedures might be nonspecific, resemble
postbariatric pregnancy
a number of other conditions, and could be difficult for obste-
tricians to recognize. For this reason, all pregnant women
●● Routine screening for nutritional deficiencies with a history of bariatric surgery with severe abdominal pain
●● Usage of appropriate supplements and monitoring
compliance should also be examined by a general surgeon.
●● Alternative screening ways for GDM (after malabsorp- Bowel obstruction during pregnancy in women who had
tive procedure) previously undergone bariatric surgery has been reported in
●● Careful evaluation by a general surgeon in cases of several case reports, and few of them reported fetal and mater-
severe abdominal pain
nal deaths.83,84 In those patients, common obstetric complaints
●● Avoidance of conception for 12 months following bar-
iatric surgery such as hyperemesis, esophageal reflux, and contractions
might mimic early bowel obstruction symptoms and should be
Postbariatric surgery: Pregnancy complications and management 277
evaluated carefully. Most cases are attributed to bowel ischemia high number of unintended pregnancies.95,96 Theoretically,
and internal hernia formation after RYGB.84–87 The estimated conception during this time period provokes a risk for sur-
prevalence of this complication is up to 2%, and it is contributed gical complications, nutritional-related complications, and
by elevated abdominal pressure and cephalad intestinal dis- malformations.97 It seems reasonable that additional nutri-
placement caused by the enlarging gravid uterus.84 Computed tional requirements of the developing fetus and the dramatic
tomography scan and even exploratory surgery may be neces- decline in maternal intake of nutrition might lead to intra-
sary in highly suspicious clinical circumstance.81,88 uterine growth restriction (IUGR).98 When considering the
Among women who have undergone LAGB, band slip- consequences of early postbariatric pregnancies, the follow-
page has been reported in the literature89,90; nevertheless, ing issues should be addressed:
as other postbariatric surgical complications, systematic
studies have not reported high rates of this complication. 1. Pregnancy outcome and safety: In order to optimize
Haward et al.91 recently demonstrated that pregnancy does weight loss and reduce the potentially adverse effects of
not increase the risk of primary or overall band revision nutrient fluctuations and deficiencies after surgery,75,81,88
after LAGB. However, the authors mention that shorter time women are currently advised to avoid conception for
between LAGB operation and pregnancy may predispose to 12–18 and sometimes 24 months following bariatric
band revisions (p = 0.03). surgery.64,82,88,94,99,100 The 2009 ACOG Practice Bulletin
One of the major benefits of LAGB is the band adjustabil- on pregnancy after bariatric surgery advised that “some
ity that allows varying degrees of weight reduction or gain authorities have recommended waiting 12–24 months
and control of gastrointestinal symptoms such as nausea after bariatric surgery before conceiving so that the fetus
and vomiting. Some surgeons prophylactically adjust the is not exposed to a rapid weight loss environment and
band: Ducarme et al.92 demonstrated a 60% band adjustment so that the patient can achieve full weight loss goals.”82
rate for the prevention of nausea and vomiting in the first This undetermined advice reflects the growing num-
trimester of pregnancy. However, Martin et al.93 reported bers of observational studies that do not demonstrate a
that although some bands were adjusted to alleviate nausea difference in pregnancy outcomes in early postbariat-
and vomiting, the gastric bands were not routinely deflated ric surgery gestations as compared to later ones.70,101–103
during pregnancy. Dixon et al.70 found that patients often Our group100 compared the pregnancy outcome of 104
required close monitoring and band adjustment in order to patients who conceived during the first postoperative
meet the recommended gestational weight gain and to relieve year to 385 patients who conceived after the first postop-
excessive nausea and vomiting. They also reported that some erative year. No significant difference was found between
fluid was removed at 36 weeks in order to reduce the risk of the two groups nor was a shorter time interval to con-
gastric restriction during delivery. Currently, there is no gen- ception associated with an adverse perinatal outcome or
eral practical consensus as to whether or not bands should be pregnancy complications. As a possible explanation, Kral
deflated once pregnancy is detected. et al.104 suggested that in oppose to starvation state, the
lack of abnormal starvation stress after bariatric surgery
does not expose the fetus to an excess of glucocorticoids,
Gestational weight gain
therefore being less likely to induce IUGR in the neonate.
Currently, there are no clear guidelines for optimal weight In light of the theoretical risks raised, avoiding preg-
gain during pregnancy in women who have undergone nancy for at least 1 year is still advised, but if conception
bariatric surgery. Nutritionist consultation after concep- does occur less than a year after surgery, advising termi-
tion is important for the planning of dietary regimens nation of the pregnancy seems to be unnecessary due to
that are adjusted to the physiologic changes of pregnancy. the relative scarcity of complications uncovered to date.
Weight gain recommendations are based on Institute of Close observation alone may be sufficient to achieve a sat-
Medicine (IOM) guidelines, which are based on prepreg- isfactory outcome.
nancy BMI.82 The minimum weight gain standards sug- 2. Pregnancy weight gain and postpartum weight loss:
gested by the IOM are 230 g/week for obese women in the Currently, inconsistency exists regarding the effect of early
second and third trimester. Dixon et al. reported that if the pregnancy after surgery on gestational weight gain and
bariatric surgery team monitored the patient throughout postpartum weight loss. Dao et al.101 reported that women
the pregnancy, the weight gain was optimal and similar to who conceived within 12 months of surgery had less
the IOM recommendations.70 Due to concerns about fetal weight gain during pregnancy and less weight loss post-
growth, caloric restriction during pregnancy is not recom- partum than those who waited longer before conceiving
mended even in the price of continuing to be overweight (mean gestational weight gain 1.8 vs. 15.5 kg). However,
after bariatric surgery.82 Wax et al.103 compared 52 postgastric bypass patients with
surgery-to-conception intervals of ≤18 months to those
Surgery-to-conception interval of >18 months and found no statistically significant dif-
After bariatric surgery, the first 1–2 years is often referred ferences in maternal pregnancy weight gain. Only limited
to as the “rapid weight loss phase.”82,94 During this period data are available on the long-term effects of time to con-
of time, fertility may be increased and effectiveness of cer- ception after surgery on total weight loss with some stud-
tain contraceptives may be decreased, possibly leading to ies reporting no negative effect on total weight loss.101,105
Table 34.4 Summary of selected studies on gestational diabetes mellitus in pregnancy after bariatric surgery
Study (author, Type of

year, country) Cases Comparison group surgery Main findings
Lesko and 70 deliveries 140 deliveries matched BS (85.7% Significant decrease in rate of GDM in BS
Peaceman126 for presurgical RYGB) patients (0%) as compared with control
2012, USA BMI ± 6 groups (morbidly obese 16.4%, obese
9.3%).
Sheiner et al.111 298 Pregnancy cohort BS Higher rate of GDM in postsurgical patients
2004, Israel deliveries without BS, as compared to the general population.
n = 158,912 No association when confounders were
eliminated by logistic regression analysis.
Ducarme et al.92 13 deliveries 414 deliveries in obese LAGB Decreased rate of GDM post-LAGB as
2007, France women compared to obese controls (0% vs.
22.1%).
Aricha-Tamir 144 144 presurgical deliveries BS (3.4% Decreased risk for diabetes mellitus
et al.29 2012, deliveries in the same women RYGB) postsurgery (significant for the total
Israel procedures, mainly LAGB) (OR 0.2).
Dixon et al.70 79 deliveries 79 deliveries matched for LAGB Lower rate of GDM after LAGB as compared
2005, Australia BMI, age, and parity to obese controls (6.3% vs. 19%).
Burke et al.108 354 346 presurgical deliveries BS (87%/75% Women with delivery after BS had lower
2010, USA deliveries RYGB) incidences of GDM (8% vs. 27%, OR
0.23).
Weintraub et al.107 507 301 presurgical deliveries LAGB Decreased risk for diabetes mellitus after BS
2008, Israel deliveries compared to presurgical deliveries.
Skull et al.114 49 deliveries 31 presurgical deliveries LAGB Incidence of GDM was significantly
2004, Australia (44 in the same women reduced in the LAGB group (8% vs. 27%
women) p = 0.048).
Patel et al.102 26 deliveries 254 deliveries stratified RYGB NS
2008, USA by BMI
Wax et al.103 38 deliveries 76 deliveries matched for RYGB NS
2008, USA maternal age and prior
CS
Abbreviations: BS, bariatric surgery; CS, cesarean section; LAGB, laparoscopic adjustable gastric banding; NS, no significant difference; RYGB,
Roux-en-Y gastric bypass.
Obstetric outcomes 298 postoperative deliveries were compared to 158,912

deliveries of the general population by Sheiner et al., a
Gestational diabetes mellitus (Table 34.4) higher rate of GDM was found among women who under-
Obese women clearly have an increased risk for developing went bariatric surgery (9.4% vs. 5.0%, p = 0.001).111 It was
GDM106 making surgical weight loss procedures a poten- because the postoperative women were still more likely to
tially good way to prevent it. be obese, as compared with the general population. No sig-
Weintraub et al.107 in a retrospective study that com- nificantly increased risk for GDM in postoperative women
pared pregnancy outcomes of 301 pregnancies before bar- was found after confounders such as maternal BMI were
iatric surgery to 507 pregnancies after bariatric surgery, neutralized. Other studies have also found higher preva-
Weintraub et al.107 found that there was a significantly lence of GDM among p ostoperative women in comparison
lower prevalence of GDM in the postbariatric surgery to the general population.112–114 Clinicians can be reassured
group (17.3% vs. 11.0%, respectively). Similar results were that women with a history of bariatric surgery who eventu-
observed by Burke et al.108 in a recent retrospective study ally develop GDM during subsequent pregnancies appear
comparing 346 women who had deliveries before bariat- to have similar perinatal outcomes and glycemic control as
ric surgery to 354 women who had deliveries after bariat- other women with GDM.115
ric surgery. In their study, the incidences of GDM in the To date, despite the controversial magnitude of its effect,
postbariatric group was even equivalent to the general it seems that bariatric surgery does, in fact, lower the risk
population (8% vs. 27%, OR 0.23, 95% confidence interval of GDM in subsequent pregnancies, although this risk does
[CI] 0.15–0.36).109 Other observational studies consistently not completely normalize to general population levels, per-
report a lower prevalence of GDM among women who have haps due to residual obesity among postoperative women.
had bariatric surgery than among obese women who have Whether the type of bariatric procedure impacts the course
not undergone this surgery110 (Figure 34.1). However, when and outcome of GDM is unclear.
Fetal outcomes 279
30 Mode of delivery (Table 34.6)

Obese patients are at increased risk of both elective and
25 emergency CD, and this risk increases with increasing obe-
sity.22,36,122 When compared to the general obstetric popu-
20 lation, observational studies reported higher CD rates in
postbariatric surgery pregnancies.102,111 This is not surpris-
ing and could be explained by the fact that postoperative
15
patients are typically older and more likely to be obese than
women in the general obstetrical population.103,115 Sheiner
10 et al.22 also suggested that this higher prevalence of cesarean
section seen in postoperative patients may be explained by
5 caregiver bias, mentioning that clinicians should have a clear
indication before operating these patients.
0
Postoperative patients tend to have either similar or
Skull et al. Aricha- Burke et al. Weintraub lower rate of CD compared to obese women who have not
Tamir et al. et al. undergone bariatric surgery.102,123
% GDM postbariatric % GDM prebariatric
Figure 34.1 Selected studies presenting rates of gestational

diabetes mellitus before and after bariatric surgery.
Fetal outcomes
Birth weight (Table 34.7)
In postoperative women with type 2 diabetes who reduced Inappropriate maternal weight gain (especially during the
body weight and normalized serum glucose before preg- rapid weight loss phase of the first 1–2 years postsurgery)
nancy, euglycemia may remain during pregnancy despite along with the elevated risk of nutritional deficiencies
pregnancy-related insulin resistance.116 stresses the need for careful monitoring for fetal growth in
postbariatric surgery patients.
Compared with obese women who have not undergone
Hypertensive disorders (Table 34.5) a bariatric procedure, observational studies have generally
Obesity is a known risk factor for the development of hyper- reported larger proportion of appropriate-for-gestational-age
tensive disorders of pregnancy, and it has been consistently infants among postbariatric surgery pregnancies.75,102,107
reported that maternal weight and BMI are independent
risk factors for preeclampsia, as well as other hypertensive LGA: Maternal obesity and excessive weight gain during
disorders.21,106,117–119 pregnancy are associated with macrosomic (>4000 g) or
Studies among postbariatric surgery women as com- LGA neonates.124 Most studies comparing postbariat-
pared to obese controls that examined this spectrum of ric surgery pregnancies to pregnancies in obese women
outcomes have generally demonstrated them to be lower have found decreased rates of LGA or macrosomic
in postoperative women.70,107,115,120 Aricha-Tamir et al., 29 infants.64,75,92,123,125 Weintraub et al.107 reported a reduc-
in a recent study of 288 women (144 preoperative preg- tion of more than 50% in the prevalence of macrosomia
nancies that were matched to 144 postoperative pregnan- (OR 0.45; 95% CI 0.21–0.94). When comparing postop-
cies of the same women), found a significant reduction erative pregnancies to the general population, there was
in hypertensive disorders in postoperative pregnan- no consistent reduction in the rates of LGA and macroso-
cies (adjusted OR 0.4, 95% CI 0.2–0.8). Bennett et al.119 mic infants.103,111 This could be explained perhaps by the
reported that women who delivered after surgery had sub- fact that postoperative women tend to be more obese than
stantially lower rates of preeclampsia and eclampsia (OR women in the general population, thereby still maintain-
0.20, 95% CI 0.09–0.44), chronic hypertension complicating a higher risk for macrosomia. These data suggest that
ing pregnancy (0.39, 0.20–0.74), and gestational hyper- rates of fetal macrosomia in subsequent pregnancies may
tension (0.16, 0.07–0.37). These lower rates remained even be reduced by weight loss after bariatric surgery.
after adjustment for age at delivery, surgical procedure, Growth restriction: Although inconsistent, some obser-
and preexisting diabetes (Figure 34.2). vational studies have demonstrated increased rates of
Sheiner et al. revealed no difference in the risk for hyper- IUGR and small-for-gestational-age (SGA) infants in
tensive disorders among pregnancies subsequent to restric- postbariatric surgery patients.75,111,121,123,125,126 A recent
tive versus malabsorptive-type procedures.99 population-based study from Denmark,123 conducted
Finally, studies comparing the risk for hypertensive among infants born after maternal bariatric surgery
disorders in postoperative women and general popula- (84% gastric bypass), revealed 2.3 times higher risk of
tion controls found that the incidence of preeclampsia SGA than infants born by a matched group of women
may drop down to be the same as the general obstetrical (prepregnancy BMI, age, parity, date of delivery, and
community.103,111,121 smoking) without bariatric surgery. Substantial maternal
Table 34.5 Summary of selected studies on hypertensive disorders in pregnancy after bariatric surgery
Study (author, Type of

year, country) Cases Comparison group surgery Main findings
Sheiner et al.111 298 Pregnancy cohort without BS • No difference in rate of mild preeclampsia and
2004, Israel deliveries BS, n = 158,912 severe preeclampsia among postsurgical
women and normal controls.
• Increased risk for chronic hypertension in
postsurgical women.
• No increased risk of total hypertensive disorders
based on multivariate logistic regression.
Ducarme 13 414 deliveries in obese LAGB • Decreased rate of preeclampsia in post-LAGB
et al.92 2007, deliveries women as compared to obese controls (0% vs. 3.1%).
France • NS in PIH.
Aricha-Tamir 144 144 presurgical deliveries BS (3.4% Decreased risk for hypertensive disorders among
et al.29 2011, deliveries in the same women RYGB) postsurgical (mainly LAGB) deliveries as
Israel compared to presurgical deliveries of the same
women (OR 0.4).
Dixon et al.70 79 1. 40 presurgical LAGB • Lower rate of PIH and preeclampsia among
2005, deliveries deliveries in the same post-LAGB women as compared to obese
Australia women controls (10% vs. 38% and 5% vs. 25%,
2. 79 deliveries matched respectively) and preoperative pregnancies
for BMI, age, and parity (10% vs. 45% and 5% vs. 28%, respectively).
3. Community • NS in PIH as compared to community.
Bennett et al.119 316 269 presurgical deliveries BS (84% Women who delivered after surgery had
2010, USA deliveries RYGB) substantially lower rates of preeclampsia and
eclampsia (OR 0.20), chronic hypertension
complicating pregnancy (OR 0.39), and
gestational hypertension (OR 0.16) even after
adjustment for age at delivery, multiple
pregnancy, surgical procedure, preexisting
diabetes, and insurance plan.
Weintraub 507 301 presurgical deliveries LAGB Decreased risk for total hypertensive disorders and
et al.107 deliveries severe preeclampsia in postsurgical as compared
2008, Israel to presurgical deliveries (OR 0.4).
Patel et al.102 26 254 deliveries stratified RYGB NS
2008, USA deliveries by BMI
Wax et al.103 38 76 deliveries matched for RYGB Increased rate of hypertension among post-RYGB
2008, USA deliveries maternal age and prior women and age-/prior cesarean section–matched
CS controls; however, the increased risk was
eliminated once BMI was controlled for.
Abbreviations: BS, bariatric surgery; CS, cesarean section; LAGB, laparoscopic adjustable gastric banding; NS, no significant difference was
found; OR, odds ratio; PIH, pregnancy-induced hypertension; RYGB, Roux-en-Y gastric bypass.
weight loss and induced malabsorption were thought by an SGA infant (<2500 g) among post-LAGB patients versus
the authors as the reason for the increase in SGA births. obese controls. The authors hypothesized that this improved
When compared to the morbidly obese group, Lesko and outcome may be related to the reduction in the risk of pre-
Peaceman,126 in a recent study, found that neonates of eclampsia. Another study comparing 38 post-RYGB cases
postoperative women were significantly more likely to be with 76 general population controls found no difference in
SGA (5.0% vs. 17.4%, respectively). Similar results were mean birth weight or IUGR among the two groups.103
demonstrated by our group111 in a study among postop-
erative women (procedures of all types) and the general
Congenital anomalies (Table 34.8)
obstetric population (5% vs. 2%, respectively). However,
when variables such as obesity and hypertensive disor- Although deficiencies in a number of nutrients have been
ders were controlled, this relationship disappeared. This reported in case studies68–70 and concerns about malnour-
may highlight the importance of residual confounders ishment of mother and fetus were made, systematic studies
when interpreting the results of studies commenting on have repeatedly failed to link history of bariatric surgery
SGA or IUGR in postoperative pregnancies. with poor perinatal outcomes.75,92,107,111 Increased risk of con-
genital anomalies has been found in general in overweight
In contrast, in a small study including 13 post-LAGB and obese women, and postbariatric surgery patients must
patients, Ducarme et al. reported a decreased rate of delivering be balanced with this known risk.
Postbariatric surgery: Fertility 281
16 Reassuring findings were also reported by Sheiner et al. in a

large study that examined the risk for congenital malforma-
14
tions in 298 pregnancies after bariatric surgery and 158,912
12 deliveries in the general obstetrical population. In their
study, no significant difference was found between the two
10
groups.111 The study was powered enough to conclude that no
8 differences were found.
It is obligatory to be aware for any deficiencies and treat
6
them once detection is made. Although it is not known if
4 women require higher doses of folic acid after weight loss
surgery,82 patients should be encouraged to take adequate
2 folic acid supplementation (800 mcg daily, or even 5 mg) to
0 decrease the risk along with screening via maternal serum
PET/eclampsia Chronic Gestational PET/eclampsia alpha-fetoprotein and ultrasound.
superimposed HTN HTN
on preexisting
hypertension
% postbariatric % prebariatric
Postbariatric surgery: Fertility
It is well established in literature that obesity can be asso-
Figure 34.2 The risk for hypertensive disorders. HTN, hyper- ciated with oligoovulation/anovulation that can reduce
tension; PET, preeclampsia toxemia; PIH, pregnancy-induced fecundity rates. Studies have shown that weight loss (both
hypertension. (Adapted from Bennett, W.L. et al., Br. Med. J.,
340, c1662, 2010.) nonsurgical and surgical) could increase fertility for over-
weight and obese women.11–13 In one series, 15 of 32 women
that were unsuccessful at becoming pregnant prior to bariat-
A potentially major concern is the association between ric surgery became pregnant following surgery.75 In another
bariatric surgery and NTD. This association was reported in study by Teitelman et al.,129 70 out of 98 anovulatory women
a number of case reports127,128 and may be attributed to the regained normal menstrual cycles after surgery, and those
biologic plausibility of folate deficiency and hyperhomocys- who regained ovulation lost significantly more weight than
teinemia in postoperative woman.121,128 However, Weintraub those who remained anovulatory. In a recent retrospective
et al.107 did not report a higher rate of fetal malformations study of 110 obese infertile women who were unsuccess-
among 507 deliveries after bariatric surgery as compared to ful in their attempts to conceive prior to bariatric surgery,
301 deliveries before bariatric surgery after controlling for 69 became pregnant after surgical weight loss. The authors
maternal age and preterm delivery in a multivariate model. found only the amount of weight loss (OR 20.2, p = 0.001)
Table 34.6 Summary of selected studies on cesarean delivery rate in pregnancy after bariatric surgery
Study (author, Comparison Type of

year, country) Cases group surgery Main findings
Sheiner et al.111 298 Pregnancy BS Higher rate of CS even after controlling for
2004, Israel deliveries cohort confounders using logistic regression and
without BS, Mantel–Haenzel analyses (OR 2.4)
n = 158,912
Ducarme et al.92 13 deliveries 414 deliveries in LAGB Decreased rate CS among post-LAGB women as
2007, France obese women compared to obese controls (15.3% vs. 34.4%)
Burke et al.108 354 346 presurgical BS (87%/75% Women with delivery after BS had lower
2010, USA deliveries deliveries RYGB) incidences of CS (28% vs. 43%, OR 0.53)
Weintraub et al.107 507 301 presurgical LAGB Increased risk for CS among postsurgical deliveries
2008, Israel deliveries deliveries (OR −1.9)
Patel et al.102 2008, 26 deliveries 254 deliveries RYGB • Higher rate of CS among post-RYGB women as
USA stratified by compared to nonobese controls
BMI • NS when compared to obese or severely obese
controls
Wax et al.103 2008, 38 deliveries 76 deliveries RYGB NS
USA matched for
maternal age
and prior CS
found; OR, odds ratio; RYGB, Roux-en-Y gastric bypass.
Table 34.7 Summary of selected studies on birth weight in pregnancy after bariatric surgery
Study (author, year, Type of

country) Cases Comparison group surgery Main findings
Lesko et al. 126 70 deliveries 140 deliveries BS (85.7% • Neonates were significantly more likely to
2012, USA matched for RYGB) be SGA in the BS group (17.4%) than the
presurgical morbidly obese group (5.0%) (OR 3.94).
BMI ± 6 • The incidence of macrosomia (defined as
birth weight more than 4000 g) was lower
as compared with control group (OR 0.21).
Sheiner et al.111 298 deliveries Pregnancy cohort BS • Higher rate of IUGR (5% vs. 2%) and
2004, Israel without BS, macrosomia in postsurgical patients as
n = 158,912 compared to the general population.
• IUGR was not significant when
confounders were controlled for.
• Macrosomia remained significant after
control of confounders (OR 2.1).
Ducarme et al.92 13 deliveries 414 deliveries in LAGB Decreased rate of SGA (7.7% vs. 10.6%) and
2007, France obese women macrosomic infants (7.7% vs. 14.6%)
post-LAGB as compared to obese controls.
Dixon et al.70 2005, 79 deliveries 79 deliveries matched LAGB NS
Australia for BMI, age, and
parity
Weintraub et al.107 507 deliveries 301 presurgical LAGB • Decreased risk for macrosomia after BS as
2008, Israel deliveries compared to deliveries before surgery.
• NS in risk for SGA infants.
Josefsson et al.125 126 deliveries Pregnancy cohort BS Women surgically treated before their first child
2011, Sweden without BS more often delivered SGA infants compared
n = 188,500 with the children born to the reference
mothers (5.6% vs. 2.1%).
Patel et al.102 26 deliveries 254 deliveries RYGB • Decreased rate of macrosomia post-RYGB
2008, USA stratified by BMI (0% vs. 18.5%) as compared to severely
obese controls.
• NS in rate of IUGR.
Wax et al.103 2008, 38 deliveries 76 deliveries matched RYGB NS
USA for maternal age
and prior CS
Abbreviations: BS, bariatric surgery; CS, cesarean section; IUGR, intrauterine growth restriction; LAGB, laparoscopic adjustable gastric banding;
NS, no significant difference was found; OR, odds ratio; RYGB, Roux-en-Y gastric bypass; SGA, small for gestational age.
Table 34.8 Summary of selected studies on birth defects rate in pregnancy after bariatric surgery
Type of
Study (author, year, country) Cases Comparison group surgery Main findings
Sheiner et al.111 2004, Israel 298 deliveries Pregnancy cohort BS NS
without BS,
n = 158,912
Weintraub et al.107 2008, Israel 507 deliveries 301 presurgical LAGB Increased rate of fetal
deliveries malformation (OR 2.5) but NS
once maternal age and
preterm delivery were
controlled for
Patel et al.102 2008, USA 26 deliveries 254 deliveries RYGB NS
stratified by BMI
Wax et al.103 2008, USA 38 deliveries 76 deliveries matched RYGB NS
for maternal age
and prior CS
found; OR, odds ratio; RYGB, Roux-en-Y gastric bypass.
References 283
and the achieved BMI (p = 0.001) after surgery were good 4. Postbariatric surgery parturients are becoming a new
predictors for successful pregnancy regardless of the type of obstetric group with unique risks and possible fetal and
surgical procedure the woman underwent.130 maternal outcomes.
However, in spite of these encouraging data, when 5. Pregnancy after bariatric surgery appears to be generally
compared to nonobese controls, studies have demon- safer than pregnancy in obese women.
strated inconsistency regarding fertility outcomes as some 6. Routine nutritional screening, recommendations for
reports do not demonstrate any significant difference in appropriate supplements, and monitoring compliance
fertility between normal and postoperative women, while are obligatory for postoperative pregnant women.
others show slightly worse outcomes.109 Our group111 had 7. Life-threatening surgical complications might be
found elevated rates of infertility treatments in postbar- nonspecific and could be difficult for obstetricians to
iatric surgery patients as compared to the general popula- recognize. All pregnant women with a history of bariat-
tion even after controlling for potential confounders such ric surgery with severe abdominal pain should also be
as maternal obesity. Nevertheless, it should be emphasized examined by a general surgeon.
that infertility treatments may not be a perfect marker 8. There is no general practical consensus in women post-
for fertility status. Contrary to our prior findings, more LAGB to prophylactically deflate the band once preg-
recent studies comparing postbariatric patients to obese nancy is detected.
or preoperative parturients found no significant differ- 9. Weight gain recommendations are based on IOM guide-
ence in the rate of infertility treatment before or after lines. Caloric restriction during pregnancy is not recom-
surgery. 29,107 mended even at the price of continuing to be overweight.
In conclusion, generally, postsurgical women show 10. GCT is not well tolerated in patients who have undergone
increased fertility when compared to obese controls.109 a malabsorptive procedure; alternative screening ways
When compared to nonobese controls, studies have demon- for GDM should be implemented.
strated inconsistent results. Currently, it is agreed by most 11. Severely obese women should be offered bariatric sur-
experts that bariatric surgery is not the primary treatment gery with a sufficient interval before conception to allow
for infertility in severely obese women.112,131 weight reduction.
12. More powered studies are needed in order to conclude
about optimal surgical-to-conception interval; currently
Summary of key points women are advised to avoid conception for 12 months
following bariatric surgery.
1. Obesity is an epidemic and is considered to be one of the 13. Postoperative patients tend to have either a similar or
most pressing concerns in the developed world. lower rate of CD compared to obese women who have
2. Close to one-third of women of childbearing age (20– not undergone bariatric surgery.
39 years) are classified as obese in the western society, 14. The risk of obstetrical complications, such as GDM and
and an additional quarter of women in this age group hypertensive disorders as well as fetal macrosomia, is
are overweight, a fact that can eventually lead to adverse significantly reduced following bariatric surgery. There
consequences for their future reproductive health, preg- is a possible risk for SGA infants.
nancy, and long-term health. 15. Systematic studies have repeatedly failed to link the his-
3. Rates of surgical weight loss procedures have increased tory of bariatric surgery with poor perinatal outcome.
dramatically over the recent decade as surgery is prob- 16. Generally, postsurgical women show increased fertility
ably the best treatment for obesity. when compared to obese controls.
REFERENCES
1. American College of Obstetricians and Gynecologists. ACOG 7. O’Brien PE, Brown WA, Dixon JB. Obesity, weight loss and bar-
committee opinion no. 549: Obesity in pregnancy. Obstet iatric surgery. Med J Aust September 19, 2005; 183(6): 310–314.
Gynecol January 2013; 121(1): 213–217. 8. Sarwer DB, Allison KC, Gibbons LM, Markowitz JT, Nelson DB.
2. Cnattingius S, Bergstrom R, Lipworth L, Kramer MS. Prepregnancy Pregnancy and obesity: A review and agenda for future research.
weight and the risk of adverse pregnancy outcomes. N Engl J Med J Womens Health (Larchmt). July–August 2006; 15(6): 720–733.
January 15, 1998; 338(3): 147–152. 9. Chin JR, Murtaugh MA, Silver R. Obesity: Implications for wom-
3. Hedley AA, Ogden CL, Johnson CL, Carroll MD, Curtin LR, en’s reproductive health. Curr Epidemiol Rep 2014; 1: 17–26.
Flegal KM. Prevalence of overweight and obesity among US chil- 10. Sjöström L. Bariatric surgery and reduction in morbidity and
dren, adolescents, and adults, 1999–2002. JAMA June 16, 2004; mortality: Experiences from the SOS study. Int J Obes (Lond)
291(23): 2847–2850. December 2008; 32(Suppl. 7): S93–S97.
4. WHO. Obesity and overweight. Fact sheet. Updated March 11. van der Steeg JW, Steures P, Eijkemans MJ et al. Obesity affects
2013. Report No.: 311. spontaneous pregnancy chances in subfertile, ovulatory women.
5. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of obe- Hum Reprod February 2008; 23(2): 324–328.
sity among adults: United States, 2011–2012. Data Brief. NCHS. 12. Gesink Law DC, Maclehose RF, Longnecker MP. Obesity and
2013. Report No.: 131. time to pregnancy. Hum Reprod February 2007; 22(2): 414–420.
6. Must A, Spadano J, Coakley EH, Field AE, Colditz G, Dietz WH. 13. Ramlau-Hansen CH, Thulstrup AM, Nohr EA, Bonde JP, Sorensen
The disease burden associated with overweight and obesity. TI, Olsen J. Subfecundity in overweight and obese couples. Hum
JAMA October 27, 1999; 282(16): 1523–1529. Reprod June 2007; 22(6): 1634–1637.
14. Wang JX, Davies M, Norman RJ. Body mass and probability of 37. Hood DD, Dewan DM. Anesthetic and obstetric outcome in
pregnancy during assisted reproduction treatment: Retrospective morbidly obese parturients. Anesthesiology December 1993;
study. Br Med J November 25, 2000; 321(7272): 1320–1321. 79(6): 1210–1218.
15. Petersen GL, Schmidt L, Pinborg A, Kamper-Jorgensen M. The 38. Perlow JH, Morgan MA. Massive maternal obesity and periop-
influence of female and male body mass index on live births after erative cesarean morbidity. Am J Obstet Gynecol February 1994;
assisted reproductive technology treatment: A nationwide register- 170(2): 560–565.
based cohort study. Fertil Steril May 2013; 99(6): 1654–1662. 39. Myles TD, Gooch J, Santolaya J. Obesity as an independent risk
16. Crosignani PG, Ragni G, Parazzini F, Wyssling H, Lombroso G, factor for infectious morbidity in patients who undergo cesar-
Perotti L. Anthropometric indicators and response to gonado- ean delivery. Obstet Gynecol November 2002; 100(5 Pt. 1):
trophin for ovulation induction. Hum Reprod March 1994; 9(3): 959–964.
420–423. 40. Bianco AT, Smilen SW, Davis Y, Lopez S, Lapinski R, Lockwood
17. Balen AH, Platteau P, Andersen AN et al. The influence of body CJ. Pregnancy outcome and weight gain recommendations for
weight on response to ovulation induction with gonadotrophins the morbidly obese woman. Obstet Gynecol January 1998; 91(1):
in 335 women with world health organization group II anovula- 97–102.
tory infertility. BJOG October 2006; 113(10): 1195–1202. 41. Calandra C, Abell DA, Beischer NA. Maternal obesity in preg-
18. Souter I, Baltagi LM, Kuleta D, Meeker JD, Petrozza JC. Women, nancy. Obstet Gynecol January 1981; 57(1): 8–12.
weight, and fertility: The effect of body mass index on the out- 42. Sebire NJ, Jolly M, Harris JP et al. Maternal obesity and preg-
come of superovulation/intrauterine insemination cycles. Fertil nancy outcome: A study of 287,213 pregnancies in London. Int J
Steril March 1, 2011; 95(3): 1042–1047. Obes Relat Metab Disord August 2001; 25(8): 1175–1182.
19. Rittenberg V, Seshadri S, Sunkara SK, Sobaleva S, Oteng-Ntim 43. Rasmussen SA, Chu SY, Kim SY, Schmid CH, Lau J. Maternal obe-
E, El-Toukhy T. Effect of body mass index on IVF treatment out- sity and risk of neural tube defects: A metaanalysis. Am J Obstet
come: An updated systematic review and meta-analysis. Reprod Gynecol June 2008; 198(6): 611–619.
Biomed Online. October 2011; 23(4): 421–439. 44. Stothard KJ, Tennant PW, Bell R, Rankin J. Maternal overweight
20. Cedergren MI. Maternal morbid obesity and the risk of adverse preg- and obesity and the risk of congenital anomalies: A systematic
nancy outcome. Obstet Gynecol February 2004; 103(2): 219–224. review and meta-analysis. JAMA February 11, 2009; 301(6):
21. Robinson HE, O’Connell CM, Joseph KS, McLeod NL. Maternal 636–650.
outcomes in pregnancies complicated by obesity. Obstet Gynecol 45. Gilboa SM, Correa A, Botto LD et al. Association between pre-
December 2005; 106(6): 1357–1364. pregnancy body mass index and congenital heart defects. Am J
22. Sheiner E, Levy A, Menes TS, Silverberg D, Katz M, Mazor M. Obstet Gynecol January 2010; 202(1): 51.e1–51.e10.
Maternal obesity as an independent risk factor for caesarean 46. Hendler I, Blackwell SC, Treadwell MC, Bujold E, Sokol RJ,
delivery. Paediatr Perinat Epidemiol May 2004; 18(3): 196–201. Sorokin Y. Does advanced ultrasound equipment improve the
23. Catalano PM. Management of obesity in pregnancy. Obstet adequacy of ultrasound visualization of fetal cardiac structures
Gynecol February 2007; 109(2 Pt. 1): 419–433. in the obese gravid woman? Am J Obstet Gynecol June 2004;
24. Metwally M, Ong KJ, Ledger WL, Li TC. Does high body mass 190(6): 1616–1619; discussion 1619–1620.
index increase the risk of miscarriage after spontaneous and 47. Hendler I, Blackwell SC, Bujold E et al. The impact of maternal
assisted conception? A meta-analysis of the evidence. Fertil Steril obesity on midtrimester sonographic visualization of fetal car-
September 2008; 90(3): 714–726. diac and craniospinal structures. Int J Obes Relat Metab Disord
25. Gross T, Sokol RJ, King KC. Obesity in pregnancy: Risks and out- December 2004; 28(12): 1607–1611.
come. Obstet Gynecol October 1980; 56(4): 446–450. 48. Chu SY, Kim SY, Lau J et al. Maternal obesity and risk of stillbirth:
26. Gaillard R, Steegers EA, Hofman A, Jaddoe VW. Associations of A metaanalysis. Am J Obstet Gynecol September 2007; 197(3):
maternal obesity with blood pressure and the risks of gestational 223–228.
hypertensive disorders. The generation R study. J Hypertens May 49. Chen A, Feresu SA, Fernandez C, Rogan WJ. Maternal obesity
2011; 29(5): 937–944. and the risk of infant death in the United States. Epidemiology
27. O’Brien TE, Ray JG, Chan WS. Maternal body mass index and the January 2009; 20(1): 74–81.
risk of preeclampsia: A systematic overview. Epidemiology May 50. Bhattacharya S, Campbell DM, Liston WA, Bhattacharya S.
2003; 14(3): 368–374. Effect of body mass index on pregnancy outcomes in nulliparous
28. Torloni MR, Betran AP, Horta BL et al. Prepregnancy BMI and the women delivering singleton babies. BMC Public Health July 24,
risk of gestational diabetes: A systematic review of the literature 2007; 7: 168.
with meta-analysis. Obes Rev March 2009; 10(2): 194–203. 51. Briese V, Voigt M, Wisser J, Borchardt U, Straube S. Risks of preg-
29. Aricha-Tamir B, Weintraub AY, Levi I, Sheiner E. Downsizing nancy and birth in obese primiparous women: An analysis of
pregnancy complications: A study of paired pregnancy outcomes German perinatal statistics. Arch Gynecol Obstet February 2011;
before and after bariatric surgery. Surg Obes Relat Dis July– 283(2): 249–253.
August 2012; 8(4): 434–439. 52. Johnson JW, Longmate JA, Frentzen B. Excessive maternal weight
30. Kominiarek MA, Zhang J, Vanveldhuisen P, Troendle J, Beaver J, and pregnancy outcome. Am J Obstet Gynecol August 1992;
Hibbard JU. Contemporary labor patterns: The impact of mater- 167(2): 353–370; discussion 370–372.
nal body mass index. Am J Obstet Gynecol September 2011; 53. Abrams BF, Laros RK, Jr. Prepregnancy weight, weight gain, and
205(3): 244.e1–244.e8. birth weight. Am J Obstet Gynecol March 1986; 154(3): 503–509.
31. Norman SM, Tuuli MG, Odibo AO, Caughey AB, Roehl KA, 54. Frentzen BH, Dimperio DL, Cruz AC. Maternal weight gain:
Cahill AG. The effects of obesity on the first stage of labor. Obstet Effect on infant birth weight among overweight and average-
Gynecol July 2012; 120(1): 130–135. weight low-income women. Am J Obstet Gynecol November
32. Zhang J, Bricker L, Wray S, Quenby S. Poor uterine contractility 1988; 159(5): 1114–1117.
in obese women. BJOG March 2007; 114(3): 343–348. 55. Mazouni C, Porcu G, Cohen-Solal E et al. Maternal and anthropo-
33. Chin JR, Henry E, Holmgren CM, Varner MW, Branch DW. morphic risk factors for shoulder dystocia. Acta Obstet Gynecol
Maternal obesity and contraction strength in the first stage of Scand 2006; 85(5): 567–570.
labor. Am J Obstet Gynecol August 2012; 207(2): 129.e1–129.e6. 56. Waterland RA, Garza C. Potential mechanisms of metabolic
34. Vahratian A, Zhang J, Troendle JF, Savitz DA, Siega-Riz AM. imprinting that lead to chronic disease. Am J Clin Nutr February
Maternal prepregnancy overweight and obesity and the pattern 1999; 69(2): 179–197.
of labor progression in term nulliparous women. Obstet Gynecol 57. Taylor PD, Poston L. Developmental programming of obesity in
November 2004; 104(5 Pt. 1): 943–951. mammals. Exp Physiol March 2007; 92(2): 287–298.
35. Gunatilake RP, Smrtka MP, Harris B et al. Predictors of failed trial 58. Laitinen J, Power C, Jarvelin MR. Family social class, maternal
of labor among women with an extremely obese body mass index. body mass index, childhood body mass index, and age at men-
Am J Obstet Gynecol December 2013; 209(6): 562.e1–562.e5. arche as predictors of adult obesity. Am J Clin Nutr September
36. Poobalan AS, Aucott LS, Gurung T, Smith WC, Bhattacharya S. 2001; 74(3): 287–294.
Obesity as an independent risk factor for elective and emergency 59. Boney CM, Verma A, Tucker R, Vohr BR. Metabolic syndrome
caesarean delivery in nulliparous women—Systematic review in childhood: Association with birth weight, maternal obesity,
and meta-analysis of cohort studies. Obes Rev January 2009; and gestational diabetes mellitus. Pediatrics March 2005; 115(3):
10(1): 28–35. e290–e296.
References 285
60. Hochner H, Friedlander Y, Calderon-Margalit R et al. Associations 83. Loar PV, 3rd, Sanchez-Ramos L, Kaunitz AM, Kerwin AJ, Diaz J.
of maternal prepregnancy body mass index and gestational Maternal death caused by midgut volvulus after bariatric surgery.
weight gain with adult offspring cardiometabolic risk factors: The Am J Obstet Gynecol November 2005; 193(5): 1748–1749.
Jerusalem perinatal family follow-up study. Circulation March 20, 84. Moore KA, Ouyang DW, Whang EE. Maternal and fetal deaths
2012; 125(11): 1381–1389. after gastric bypass surgery for morbid obesity. N Engl J Med
61. Getahun D, Ananth CV, Peltier MR, Salihu HM, Scorza WE. August 12, 2004; 351(7): 721–722.
Changes in prepregnancy body mass index between the first and 85. Efthimiou E, Stein L, Court O, Christou N. Internal hernia after
second pregnancies and risk of large-for-gestational-age birth. gastric bypass surgery during middle trimester pregnancy result-
Am J Obstet Gynecol June 2007; 196(6): 530.e1–530.e8. ing in fetal loss: Risk of internal hernia never ends. Surg Obes
62. Getahun D, Kaminsky LM, Elsasser DA, Kirby RS, Ananth CV, Relat Dis May–June 2009; 5(3): 378–380.
Vintzileos AM. Changes in prepregnancy body mass index 86. Kakarla N, Dailey C, Marino T, Shikora SA, Chelmow D.
between pregnancies and risk of primary cesarean delivery. Am J Pregnancy after gastric bypass surgery and internal hernia forma-
Obstet Gynecol October 2007; 197(4): 376.e1–376.e7. tion. Obstet Gynecol May 2005; 105(5 Pt. 2): 1195–1198.
63. Gloy VL, Briel M, Bhatt DL et al. Bariatric surgery versus non- 87. Charles A, Domingo S, Goldfadden A, Fader J, Lampmann R,
surgical treatment for obesity: A systematic review and meta- Mazzeo R. Small bowel ischemia after Roux-en-Y gastric bypass
analysis of randomised controlled trials. Br Med J October 22, complicated by pregnancy: A case report. Am Surg March 2005;
2013; 347: f5934. 71(3): 231–234.
64. Maggard MA, Yermilov I, Li Z et al. Pregnancy and fertility fol- 88. Beard JH, Bell RL, Duffy AJ. Reproductive considerations and
lowing bariatric surgery: A systematic review. JAMA November pregnancy after bariatric surgery: Current evidence and recom-
19, 2008; 300(19): 2286–2296. mendations. Obes Surg August 2008; 18(8): 1023–1027.
65. Hng KN, Ang YS. Overview of bariatric surgery for the physician. 89. Van Mieghem T, Van Schoubroeck D, Depiere M, Debeer A,
Clin Med October 2012; 12(5): 435–440. Hanssens M. Fetal cerebral hemorrhage caused by vitamin K
66. Buchwald H, Oien DM. Metabolic/bariatric surgery worldwide deficiency after complicated bariatric surgery. Obstet Gynecol
2011. Obes Surg April 2013; 23(4): 427–436. August 2008; 112(2 Pt. 2): 434–436.
67. O’Brien PE, MacDonald L, Anderson M, Brennan L, Brown 90. Kirshtein B, Lantsberg L, Mizrahi S, Avinoach E. Bariatric emer-
WA. Long-term outcomes after bariatric surgery: Fifteen-year gencies for non-bariatric surgeons: Complications of laparo-
follow-up of adjustable gastric banding and a systematic review scopic gastric banding. Obes Surg November 2010; 20(11):
of the bariatric surgical literature. Ann Surg January 2013; 1468–1478.
257(1): 87–94. 91. Haward RN, Brown WA, O’Brien PE. Does pregnancy increase
68. Angrisani L, Cutolo PP, Formisano G, Nosso G, Vitolo G. the need for revisional surgery after laparoscopic adjustable gas-
Laparoscopic adjustable gastric banding versus Roux-en-Y gastric tric banding? Obes Surg September 2011; 21(9): 1362–1369.
bypass: 10-year results of a prospective, randomized trial. Surg 92. Ducarme G, Revaux A, Rodrigues A, Aissaoui F, Pharisien I,
Obes Relat Dis May–June 2013; 9(3): 405–413. Uzan M. Obstetric outcome following laparoscopic adjustable
69. Kominiarek MA. Preparing for and managing a pregnancy after bar- gastric banding. Int J Gynaecol Obstet September 2007; 98(3):
iatric surgery. Semin Perinatol December 2011; 35(6): 356–361. 244–247.
70. Dixon JB, Dixon ME, O’Brien PE. Birth outcomes in obese 93. Martin LF, Finigan KM, Nolan TE. Pregnancy after adjustable gas-
women after laparoscopic adjustable gastric banding. Obstet tric banding. Obstet Gynecol June 2000; 95(6 Pt. 1): 927–930.
Gynecol November 2005; 106(5 Pt. 1): 965–972. 94. Karmon A, Sheiner E. Timing of gestation after bariatric surgery:
71. Guelinckx I, Devlieger R, Vansant G. Reproductive outcome after Should women delay pregnancy for at least 1 postoperative year?
bariatric surgery: A critical review. Hum Reprod Update March– Am J Perinatol June 2008; 25(6): 331–333.
April 2009; 15(2): 189–201. 95. Merhi ZO. Challenging oral contraception after weight loss by
72. Dixon JB, Dixon ME, O’Brien PE. Elevated homocysteine levels bariatric surgery. Gynecol Obstet Invest 2007; 64(2): 100–102.
with weight loss after lap-band surgery: Higher folate and vitamin 96. Victor A, Odlind V, Kral JG. Oral contraceptive absorption and
B12 levels required to maintain homocysteine level. Int J Obes sex hormone binding globulins in obese women: Effects of jeju-
Relat Metab Disord February 2001; 25(2): 219–227. noileal bypass. Gastroenterol Clin North Am September 1987;
73. Ziegler O, Sirveaux MA, Brunaud L, Reibel N, Quilliot D. Medical 16(3): 483–491.
follow up after bariatric surgery: Nutritional and drug issues. 97. Wittgrove AC, Jester L, Wittgrove P, Clark GW. Pregnancy follow-
General recommendations for the prevention and treatment of ing gastric bypass for morbid obesity. Obes Surg August 1998;
nutritional deficiencies. Diabetes Metab December 2009; 35(6 8(4): 461–464; discussion 465–466.
Pt. 2): 544–557. 98. Iavazzo C, Ntziora F, Rousos I, Paschalinopoulos D. Complications
74. Gurewitsch ED, Smith-Levitin M, Mack J. Pregnancy follow- in pregnancy after bariatric surgery. Arch Gynecol Obstet August
ing gastric bypass surgery for morbid obesity. Obstet Gynecol 2010; 282(2): 225–227.
October 1996; 88(4 Pt. 2): 658–661. 99. Sheiner E, Balaban E, Dreiher J, Levi I, Levy A. Pregnancy out-
75. Marceau P, Kaufman D, Biron S et al. Outcome of pregnancies come in patients following different types of bariatric surgeries.
after biliopancreatic diversion. Obes Surg March 2004; 14(3): Obes Surg September 2009; 19(9): 1286–1292.
318–324. 100. Sheiner E, Edri A, Balaban E, Levi I, Aricha-Tamir B. Pregnancy
76. Eerdekens A, Debeer A, Van Hoey G et al. Maternal bariatric outcome of patients who conceive during or after the first year
surgery: Adverse outcomes in neonates. Eur J Pediatr February following bariatric surgery. Am J Obstet Gynecol January 2011;
2010; 169(2): 191–196. 204(1): 50.e1–50.e6.
77. Haddow JE, Hill LE, Kloza EM, Thanhauser D. Neural tube 101. Dao T, Kuhn J, Ehmer D, Fisher T, McCarty T. Pregnancy out-
defects after gastric bypass. Lancet June 7, 1986; 1(8493): 1330. comes after gastric-bypass surgery. Am J Surg December 2006;
78. Weissman A, Hagay Z, Schachter M, Dreazen E. Severe maternal 192(6): 762–766.
and fetal electrolyte imbalance in pregnancy after gastric surgery 102. Patel JA, Patel NA, Thomas RL, Nelms JK, Colella JJ. Pregnancy
for morbid obesity. A case report. J Reprod Med November 1995; outcomes after laparoscopic Roux-en-Y gastric bypass. Surg
40(11): 813–816. Obes Relat Dis January–February 2008; 4(1): 39–45.
79. Smets KJ, Barlow T, Vanhaesebrouck P. Maternal vitamin A defi- 103. Wax JR, Cartin A, Wolff R, Lepich S, Pinette MG, Blackstone J.
ciency and neonatal microphthalmia: Complications of biliopan- Pregnancy following gastric bypass for morbid obesity: Effect of
creatic diversion? Eur J Pediatr July 2006; 165(7): 502–504. surgery-to-conception interval on maternal and neonatal out-
80. Maggard MA, Shugarman LR, Suttorp M et al. Meta-analysis: comes. Obes Surg December 2008; 18(12): 1517–1521.
Surgical treatment of obesity. Ann Intern Med April 5, 2005; 104. Kral JG, Kava RA, Catalano PM, Moore BJ. Severe obesity: The
142(7): 547–559. neglected epidemic. Obes Facts April 25, 2012; 5(2): 254–269.
81. Wax JR, Pinette MG, Cartin A, Blackstone J. Female reproduc- 105. Alatishe A, Ammori BJ, New JP, Syed AA. Bariatric surgery
tive issues following bariatric surgery. Obstet Gynecol Surv in women of childbearing age. QJM August 2013; 106(8):
September 2007; 62(9): 595–604. 717–720.
82. American College of Obstetricians and Gynecologists. ACOG 106. Burstein E, Levy A, Mazor M, Wiznitzer A, Sheiner E. Pregnancy
practice bulletin no. 105: Bariatric surgery and pregnancy. Obstet outcome among obese women: A prospective study. Am J
Gynecol June 2009; 113(6): 1405–1413. Perinatol October 2008; 25(9): 561–566.
107. Weintraub AY, Levy A, Levi I, Mazor M, Wiznitzer A, Sheiner 119. Bennett WL, Gilson MM, Jamshidi R et al. Impact of bariatric sur-
E. Effect of bariatric surgery on pregnancy outcome. Int J gery on hypertensive disorders in pregnancy: Retrospective analysis
Gynaecol Obstet December 2008; 103(3): 246–251. of insurance claims data. Br Med J April 13, 2010; 340: c1662.
108. Burke AE, Bennett WL, Jamshidi RM et al. Reduced incidence 120. Sjostrom L, Narbro K, Sjostrom CD et al. Effects of bariatric surgery
of gestational diabetes with bariatric surgery. J Am Coll Surg on mortality in Swedish obese subjects. N Engl J Med August 23,
August 2010; 211(2): 169–175. 2007; 357(8): 741–752.
109. Willis K, Sheiner E. Bariatric surgery and pregnancy: The magi- 121. Dixon JB, Dixon ME, O’Brien PE. Pregnancy after lap-band surgery:
cal solution? J Perinat Med March 2013; 41(2): 133–140. Management of the band to achieve healthy weight outcomes.
110. Kjaer MM, Nilas L. Pregnancy after bariatric surgery—A review Obes Surg February 2001; 11(1): 59–65.
of benefits and risks. Acta Obstet Gynecol Scand March 2013; 122. Gilead R, Yaniv Salem S, Sergienko R, Sheiner E. Maternal “iso-
92(3): 264–271. lated” obesity and obstetric complications. J Matern Fetal Neonatal
111. Sheiner E, Levy A, Silverberg D et al. Pregnancy after bariatric Med December 2012; 25(12): 2579–2582.
surgery is not associated with adverse perinatal outcome. Am J 123. Kjaer MM, Lauenborg J, Breum BM, Nilas L. The risk of adverse
Obstet Gynecol May 2004; 190(5): 1335–1340. pregnancy outcome after bariatric surgery: A nationwide register-
112. Karmon A, Sheiner E. Pregnancy after bariatric surgery: A com- based matched cohort study. Am J Obstet Gynecol June 2013;
prehensive review. Arch Gynecol Obstet May 2008; 277(5): 208(6): 464.e1–464.e5.
381–388. 124. Ehrenberg HM, Mercer BM, Catalano PM. The influence of obe-
113. Bar-Zohar D, Azem F, Klausner J, Abu-Abeid S. Pregnancy sity and diabetes on the prevalence of macrosomia. Am J Obstet
after laparoscopic adjustable gastric banding: Perinatal out- Gynecol September 2004; 191(3): 964–968.
come is favorable also for women with relatively high ges- 125. Josefsson A, Blomberg M, Bladh M, Frederiksen SG, Sydsjo G.
tational weight gain. Surg Endosc October 2006; 20(10): Bariatric surgery in a national cohort of women: Sociodemographics
1580–1583. and obstetric outcomes. Am J Obstet Gynecol September 2011;
114. Skull AJ, Slater GH, Duncombe JE, Fielding GA. Laparoscopic 205(3): 206.e1–206.e8.
adjustable banding in pregnancy: Safety, patient tolerance 126. Lesko J, Peaceman A. Pregnancy outcomes in women after bariatric
and effect on obesity-related pregnancy outcomes. Obes Surg surgery compared with obese and morbidly obese controls. Obstet
February 2004; 14(2): 230–235. Gynecol March 2012; 119(3): 547–554.
115. Sheiner E, Menes TS, Silverberg D et al. Pregnancy outcome 127. Pelizzo G, Nakib G, Alfei A et al. Fetal neural tube defects in
of patients with gestational diabetes mellitus following bar- pregnant women previously submitted to bariatric surgery: More
iatric surgery. Am J Obstet Gynecol February 2006; 194(2): attention to a new emerging entity. Prenat Diagn February 2013;
431–435. 33(2): 196–197.
116. Adami GF, Murelli F, Briatore L, Scopinaro N. Pregnancy in 128. Moliterno JA, DiLuna ML, Sood S, Roberts KE, Duncan CC.
formerly type 2 diabetes obese women following biliopancre- Gastric bypass: A risk factor for neural tube defects? Case report.
atic diversion for obesity. Obes Surg September 2008; 18(9): J Neurosurg Pediatr May 2008; 1(5): 406–409.
1109–1111. 129. Teitelman M, Grotegut CA, Williams NN, Lewis JD. The impact
117. Ehrenthal DB, Jurkovitz C, Hoffman M, Jiang X, Weintraub WS. of bariatric surgery on menstrual patterns. Obes Surg November
Prepregnancy body mass index as an independent risk factor for 2006; 16(11): 1457–1463.
pregnancy-induced hypertension. J Womens Health (Larchmt) 130. Musella M, Milone M, Bellini M, Sosa Fernandez LM, Leongito M,
January 2011; 20(1): 67–72. Milone F. Effect of bariatric surgery on obesity-related infertility.
118. Sibai BM, Ewell M, Levine RJ et al. Risk factors associated with Surg Obes Relat Dis July–August 2012; 8(4): 445–449.
preeclampsia in healthy nulliparous women. The calcium for pre- 131. Merhi ZO. Weight loss by bariatric surgery and subsequent fertility.
eclampsia prevention (CPEP) study group. Am J Obstet Gynecol Fertil Steril February 2007; 87(2): 430–432.
November 1997; 177(5): 1003–1010.
35 Fetal lung maturity
Gian Carlo Di Renzo, Giulia Babucci, and Graziano Clerici
In a diabetic pregnancy, however, it is not advisable to

Introduction assume that the risk of neonatal RDS is cancelled after 37 weeks.
The functional immaturity of lung tissue leads to an acute In fact, full-term infants who were admitted to the intensive
progressive breathing failure, the so-called respiratory distress care unit with respiratory problems were most probably born
syndrome (RDS). RDS includes some common clinical signs from mothers with diabetes. This confirms that neonates who
such as tachypnea, expiratory grunting, nasal flaring, and cos- are born from mothers who have pregnancy complicated by
tal retractions. It is caused by deficient surfactant production medical pathologies, especially diabetes, have frequently an
causing decreased lung compliance resulting in hypoxia.1–4 outcome complicated by respiratory diseases/morbidities.13
Diabetic pregnancy, and particularly poorly controlled Clinical studies were performed to investigate the inci-
maternal diabetes, represents one of the most important dence of respiratory morbidity in term infants hospitalized in
risks for RDS. NICU to identify risk factors for such admissions. By those
Though the perinatal mortality rate in pregnancies studies, results have shown that diabetes and cesarean deliv-
complicated by diabetes has declined, conditions such as ery, frequently related, are the most important risk factors for
congenital malformation and perinatal morbidity (prema- NICU admission for RDS and other respiratory morbidities.13
turity, hypoglycemia, macrosomia, asphyxia, hypocalcemia, Infants delivered before 39 weeks were more likely to be
hyperbilirubinemia, polycythemia, hyperviscosity, hyper- admitted to the NICU for respiratory morbidity than those
trophic cardiomyopathy, and respiratory distress) are still delivered after that limit. This is complicated by the fact that
common problems in newborns as demonstrated by a recent infants born from mothers with diabetes or hypertension
Australian study (Table 35.1).5–7 were more likely to be delivered prior to 39 weeks and fre-
Despite the significant improvement in neonatal care, the quently by cesarean section that, per se, increases the risk
morbidity for respiratory complications such as RDS in the of neonatal respiratory complications compared to infants
infants born from diabetic mothers is remarkable, as well as born from vaginal delivery.13
the cost of the resulting care.8 Thus, a correct management of diabetes during preg-
In this kind of high-risk pregnancy, metabolic (glycemia) nancy particularly to reach a good glycemic control in order
control and evaluation of fetal well-being with serial ultra- to reduce the need to intervene at an early stage is critical.
sound scans are crucial in the maternal clinic care.
Moreover, all these aspects are important in planning
the site, time, and mode of delivery in order to improve Pathophysiology of fetal lung
offspring outcome.9 Particularly, the time of delivery plays maturation in diabetic pregnancies
a key role in the improvement of neonatal outcome, and it is
related to maternal metabolic control and the evaluation of Neonatal pulmonary function of the infants of diabetic
fetal well-being and fetal lung maturity that must be clearly mothers is suboptimal compared with infants of nondiabetic
documented.10 women matched for gestational age.14
Until recently, RDS was one of the most common and The mechanism by which maternal diabetes influences
serious diseases in infants of diabetic mothers. In the 1970s, lung development remains unknown. Although the patho-
improved management of diabetic pregnancies resulted in a physiologic functions underlying the occurrence of this
decline in its incidence from 31% to 3%.11 problem remain unclear, fetal hyperglycemia and/or hyper-
The observation of Kulovich indicates that nondiabetic insulinemia, reply to variations in maternal blood, are
fetus achieves pulmonary maturity at a mean gestational involved in delayed pulmonary maturation influencing pul-
age of 34–35 weeks. Interestingly, by 37 completed weeks, monary surfactant biosynthesis.14,15
86% of the lecithin/sphingomyelin (L/S) ratio and 68% of the The loss of function may be due to inadequate production of
phosphatidylglycerol (PG) values were in the normal range alveolar surfactant or abnormal lung maturation and function.
in diabetic pregnant women, and no significant differences The reason for the delay in lung maturation in infants of
were found from the diabetic and nondiabetic groups.4,12 diabetic mothers has been the subject of several theories.
287
Table 35.1 SCN/NICU admission of live born babies, by maternal diabetes in pregnancy
status, 2005–2007
Gestational diabetes
Preexisting diabetes mellitus No diabetes
Number Percent Number Percent Number Percent
Admitted 2,843 57.5 12,734 31.9 110,546 13.9
Not admitted 2,104 42.5 27,237 68.1 685,907 86.1
Total 4,947 100.0 39,976 100.0 796,628 100.0
Age-standardized rate (percent) (95% confidence interval)
Admitted 58.2 (57.7–58.7) 31.8 (31.2–32.3) 13.9 (13.5–14.2)
Not admitted 41.8 (41.4–42.2) 68.2 (67.4–69.0) 86.1 (85.2–87.0)
Source: AIHW analysis of NPDC data, Diabetes in pregnancy: Its impact on Australian women and their babies.
Copyright Australian Institute of Health and Welfare 2010. This publication is part of the Australian Institute
of Health and Welfare’s Diabetes Series. A complete list of the Institute’s publications is available from the
Institute’s website: www.aihw.gov.au.
Some authors have reported irregular production in phos- in pregnancies with poor glycemic control regardless of
pholipids as shown by delayed appearance of PG and the the class of diabetes. 24–28
early appearance of phosphatidylinositol (PI). Others have Bourbon et al. proposed that elevated maternal plasma
suggested that insulin interferes with the regular timing of level of myoinositol in diabetic women may inhibit or delay
lung maturation induced by glucocorticoids, while some oth- the production of PG in their fetuses.29
ers have suggested that the delay in maturation is the result In a recent study of lung histogenesis during intrauterine
of poor glycemic control regardless of class of diabetes.4,16–18 development of diabetic Sprague Dawley rats at 18, 19, and
In vitro studies have largely documented that insu- 21 days of gestation, a delay in alveolarization in the diabetic
lin can interfere with substrate availability for surfactant group was observed, which was confirmed by transmission
biosynthesis. Smith et al.19 demonstrated that, when insulin electron microscopy. In the diabetic group, a small quantity of
was added in the presence of cortisol to fetal lung cell cul- protein D (associated with surfactant) was detected. The same
tures, steroid-enhanced lecithin synthesis was abolished. fetuses, of the diabetic group, presented a delay in lung his-
Engle et al.14 found that higher levels of insulin resulted togenesis and in the differentiation of type II pneumocytes.30
in declined glucose and choline uptake by fetal rat type II It is suspected that neonatal respiratory problems in these
alveolar cells. infants have also a histologic basis in addition to biochemi-
However, from other animal studies, it appears that cal origin. Previously, Pinter31 demonstrated a decreased
altered expression factors or growth factors may have an fluid clearance and lack of thinning of the lung’s connective
important regulatory influence. Particularly, the role of the tissue compared with controls in the fetal lung of diabetic
epidermal growth factor (EGF) in the development of fetal rat rats. Bhavnan32 reported higher lung glycogen levels and
lung in the case of maternal diabetes was examined. In order reduced pulmonary compliance in offsprings of diabetic
to evaluate the possible role of glucose for the expression of rabbits compared with controls.
EGF and the growth of the pulmonary tissue, they performed It is well known that glucose balance has an effect on the
in vitro studies with organotypic cultures of fetal alveolar incidence of the hyaline membrane disease (Figure 35.1).
cells derived from control rats. The results also indicate that Several studies have attempted to explain the mechanism of
glucose has a role in the expression of EGF and that cells hyaline membrane disease. Hawdon and Aynsley-Green,33
exposed to high glucose concentrations are less responsive to in an investigation of type II pneumocytes in rats and rab-
EGF probably due to decreased thymidine incorporation.20 bits, have shown that insulin inhibits the cortisol-dependent
Carlson et al. have shown that insulin blocks cortisol production of phosphatidylcholine, apparently as a conse-
action at the level of the fibroblasts by reducing the produc- quence of the inhibited production of one of the precursors
tion of the fibroblast-pneumocyte factor.21 of phosphatidylcholine, fibroblast-pneumocyte factor. In
Some authors22 reported abnormal timing of phospho- rats, high glucose levels block the transformation of choline
lipid production in diabetic pregnancy, as indicated by a to phosphatidylcholine, and butyrate blocks the translation
delay in the appearance of PG in the amniotic fluid only of mRNA into surfactant proteins.
in class A (white) of diabetic pregnancies. Smith23 pos- In a demonstration of the central role of glucose homeo-
tulated that insulin interferes with the normal timing of stasis and decreased insulin resistance in the prevention of
glucocorticoid-induced pulmonary maturation in the fetus. gestational diabetes and/or improvement of its outcome,
Some investigators have disagreed with these find- recent studies have used an insulin sensitizer to be admin-
ings, reporting that fetal lung maturation occurred later istered to pregnant women. In these recent studies, 2 mg of
Phospholipids 289
80 5.9 mmol/L and 1 hour postprandial below 7.8 mmol/L), serial

ultrasound scans to monitor fetal growth (to prevent fetal
macrosomia), and better techniques of antepartum surveil-
lance, the need to perform preterm delivery has decreased.
60
In fact, pregnant women with diabetes who have normal fetal
Perinatal mortality (%)
growth and good glycemic control should be waiting for spon-

taneous labor; on the contrary, in cases of poor metabolic con-
40 trol and/or with evident fetal macrosomia, they can be offered
elective birth preferably not before 38 completed weeks.3
Considering the risks of delayed lung maturity in fetuses
of diabetic mothers, the assessment of fetal lung maturity
20 is mandatory in the case of very poor metabolic control in
diabetes and pregnancy with uncertain date.37–40
Poorly controlled diabetes is one of the conditions that
recommends performing amniocentesis, during the third
0
trimester of pregnancy, in order to analyze amniotic fluid
19 922
46 5
56 5
66 5
70 9
73 2
76 5
81 0
89 8
1
composition that reflects the status of the fetal lung.
19 194
19 195
19 196
19 196
19 197
19 197
19 198
19 198
99
–1
–1
–
–
A number of diagnostic methods with a high degree of
90
26
18
accuracy have been developed and are now available.

Figure 35.1 Perinatal mortality after introduction of
insulin therapy and amniotic fluid analysis. (Modified from
Moore, T.R., Diabetes in pregnancy, in: Creasy, R.K. and
Resnik, R., eds., Maternal Fetal Medicine, 4th ed., Saunders,
Phospholipids
Philadelphia, PA, 1999, pp. 964–995; Croze, M.L. and Soulage,
C.O., Biochimie, 95(10), 1811, 2013.) The L/S ratio was introduced by Gluck and colleagues in
1971.41 The test depends on the flow of fetal lung fluid into
the amniotic fluid changing its phospholipid composition.
myoinositol is administered to pregnant women in conjunc-
The result is expressed as the ratio of lecithin (phosphatidyl-
tion with the classic 200 µg of folic acid and only 400 µg of
choline) to sphingomyelin.
folic acid in the control group.
Sphingomyelin is a general membrane lipid and is not
The results showed a significant reduction in the inci-
related to lung maturational events. The sphingomyelin con-
dence of gestational diabetes in the group to which myoino-
tent in the amniotic fluid tends to fall from about 32 weeks
sitol was administered or at least a more adequate glycemic
gestational age to term, whereas the more saturated lecithin
control in the case of pregnant women who develop gesta-
concentration (a large part from the fetal lung) increases.
tional diabetes. The available data provide enough evidence
The L/S ratio for normal pregnancies is less than 0.5 at
for implementing a large-scale, high-quality randomized
20 weeks, gradually increasing to 1.0 at 32 weeks and around
controlled trial (RCT) looking at the efficacy of inositol in
35 weeks achieving a value of 2.0 correlating it with fetal lung
the prevention of GDM in high-risk women.34–37
maturity; empirically RDS is unlikely if the ratio is more
than 2.0 (Figures 35.2 and 35.3).
Evaluation of fetal lung maturity 12

Fetal lung maturity assessment has become a very impor-
Phospholipids in amniotic fluid
10 35 weeks 38.5 weeks

tant tool in the management of high-risk pregnancies, espe-
cially the diabetic ones. The evaluation of the ratio between 8
(mg/100 mL)
lecithin and sphingomyelin and the amount of the PG by

amniocentesis in the amniotic fluid is the method most 6
widely used to identify fetal lung maturity. Particularly, the
4
evidence of L/S above 2.5 and the presence of PG should be
used to predict fetal lung maturity.35 2
Moreover, amniocentesis in the third trimester is not
associated with increased risk of procedure-related compli- 0
cations.36 About 10 mL of amniotic fluid should be collected 12 16 20 24 28 32 36 40
and analyzed. Weeks gestation
In the past, elective preterm delivery for diabetic pregnan-
Lecithin Sphingomyelin
cies was common in order to avoid unexpected intrauterine
death, which is one of the most important complications of
Figure 35.2 Changes in the concentrations of lecithin and
gestational diabetes. This practice often resulted in high inci- sphingomyelin in the amniotic fluid. The vertical lines indicate
dence of neonatal morbidity and mortality. With the improve- achieved pulmonary maturity in nondiabetic (black) and
ment of glycemic control (target values blood glucose 3.5 and diabetic (red) pregnancy.
8 The appearance of the acidic phospholipid PG may be

delayed by fetal hyperinsulinemia, and this delay is associ-
ated with an increased incidence of RDS. In fact, the appear-
6 ance of PG has been reported to be delayed in pregnancies
of diabetic women and, for that, remains a reliable predic-
tor of pulmonary maturity. However, the absence of PG in
L/S ratio
4 diabetic pregnancy does not assume the diagnosis of pulmo-

nary immaturity, since PG fails to appear in 10% of amniotic
fluid samples by 40 weeks gestation and the presence of PG
2 is reported in only 47% of samples studied for both diabetic
and nondiabetic patients with mature L/S ratios and gesta-
tion of 34 weeks or beyond.49
0 Prior to lung maturation, there is a progressive increase of
18 22 26 30 34 38 42 PI in the amniotic fluid particularly from 26 to 33–35 weeks.
Weeks gestational age PG appears in the amniotic fluid as the percentage of PI
L/S PG PI begins its decline.43 The “lung profile” is a test that combines
the L/S ratio with measurements of the percentage of disatu-
Figure 35.3 Lecithin/sphingomyelin ratio, phosphatidylglycerol, rated (acetone precipitable) lecithin (phosphatidylcholine),
and phosphatidylinositol in the amniotic fluid from normal PI, and PG in the amniotic fluid.22 The information provided
pregnancies. (Data from Gluck, L. et al., Am. J. Obstet. Gynecol., by this profile enhances the accuracy of diagnosing fetal lung
120, 142, 1974; De Luca, A.K. et al., Acta Obstet. Gynecol., 88, maturity and gives further information on lung develop-
1036, 2009; Hallman, M. et al., Am. J. Obstet. Gynecol., 125,
613, 1976; Piper, J.M., Semin. Perinatol., 26, 206, 2002.) ment. In a small group of cases, the specificity increased from
69% to 93% by substituting the lung profile for the L/S ratio.11
The evaluation of the L/S ratio by chromatography is the The introduction of the immunological (Amniostat-
standard and reliable method with sensitivity of 83%–97% FLM) evaluation of PG provides rapid result with minimal
and specificity of 98%.44,45 Nevertheless, chromatography is a equipment necessary. This method has a false-positive rate
challenging test for routine activity because it is complicated >50%. Nevertheless, literature data confirm the utility of this
to perform and requires special laboratory tools. approach as a screening test for their rapidity and simplicity
Many factors can influence the L/S ratio. Lecithin is found of execution, particularly for specificity in the case of con-
in many body fluids including blood, vaginal secretions, and tamination and/or diabetic patients.
gastrointestinal fluid. The results of PG assay by thin-layer chromatography
The value of L/S ratio has been questioned in diabetic and Amniostat-FLM were reported to be concordant in
pregnancies where there is an increase of false-negative rate 90%–95% of the cases.49
that could range from 3% to 27%. Most studies, however,
report a low incidence of RDS with an L/S ratio >2.46
Some authors show that L/S ratio may not be a reliable indi- Proteins
cator of pulmonary maturity in diabetic pregnancies. Those
authors affirm that even a ratio above 2.0 does not guarantee The predominant protein involved in surfactant metabolism
lung maturation.47,48 Diabetes may affect the secretion of the is a 35 kDa protein, called surfactant-associated protein 35
fetal lung fluid, resulting in a higher removal of phospholipids (SAP-35). This protein originates from the type II alveolar
from the alveolar lining and in an increase false-negative rate.48 cells. An increase in the level of this protein in the amni-
Other authors noted no difference between diabetic otic fluid occurs near term, and a significant correlation with
patients and controls.49 When maternal diabetes is well con- pulmonary maturity has been described.51,52 The measure-
trolled during pregnancy, the L/S ratio can be used to estab- ment of this protein is simply available using ELISA method
lish the risk of neonatal RDS.49 and monoclonal antibodies. For Hallman et al., the test can
Surfactant of mature lung contains PG, which is absent predict RDS with an accuracy similar to that of L/S ratio.52
early in gestation and only appears at about 35 weeks of In high-risk pregnancies such as diabetes or hypertension,
pregnancy.43 the levels of SAP-35 have lower correlation with fetal lung
PG is virtually present only in the lung tissue and surfac- maturity and are probably not reliable in these situations.52
tant. Thus, amniotic fluid contaminated with blood or meco-
nium can be analyzed for this substance.50
When PG is present, RDS does not occur except possibly Lipids
in the case of intrapartum acidemia and hypoxemia. With
trace amounts of PG in amniotic fluid, an incidence of RDS Cholesterol palmitate is present in the amniotic fluid at term.
<1% has been reported.49 Its role is unknown, but it could be involved in the transport
Many authors confirmed that the added evaluation of of palimitic acid that is used in the synthesis of saturated
the PG decreases the rate of false positives significantly and phosphatidylcholine. Thin-layer chromatography and den-
improves the specificity of the L/S ratio.49 sitometry are reported as simple methods for determining
Summary 291
this substance in the amniotic fluid.53 It was observed that the mirror of a bad maternal glycemic control, and/or an
levels of cholesterol palmitate were correlated with fetal uncertain dating of pregnancy.
lung maturity; on the contrary, a similar correlation was not Caution should be used in planning the time of delivery
demonstrated in diabetic pregnancies.54 of patients with a mature L/S ratio and absent PG. The L/S
ratio can be used to assess fetal lung maturity when glucose
levels have been well controlled in a diabetic pregnancy.
Fluorescence polarization When control has been unreliable or is difficult to assess,
positive PG or higher L/S ratio (>2.5) should be used to
There are two main systems used to measure fluorescent predict fetal lung maturity.34
polarization: the Fetal lung maturity analizer (FELMA) Ghidini et al.68 suggest that an L/S ratio of ≥3.0 repre-
microviscometer and the Tdx system. The fluorescent polar- sented the optimal trade-off between sensitivity (68%) and
ization of amniotic fluid is in large part determined by false-positive rate (6%) in the prediction of the presence
binding of the probe to phospholipid structures and to the of PG. Similarly, a significant relationship exists between
albumin as a predominant protein.55 Fluorescence polar- LB count values and the presence of PG. An LB count of
ization of the amniotic fluid is inversely related to the L/S ≥50,000 represents the optimal trade-off between sensitivity
ratio.56,57 The specificity of this method to predict RDS ranges (92%) and false-positive rate (0%) in the prediction of the
from 50% to 70%.58,59 The technique is not reliable when the presence of PG.
amniotic fluid is contaminated with blood or meconium. The presence of PG or an L/S ratio of ≥2.5 indicates that
Maternal diabetes may have variable effects on the fluores- the fetal lungs are mature. The confirmation of fetal lung
cent polarization values, and in this condition, this method is maturity should be obtained and used to plan the time of
an unreliable indicator of fetal lung maturity.60 delivery.
However, it is still unclear if there is a delay in fetal lung
maturation between diabetic and nondiabetic pregnant
Comparing current techniques women at the same period of gestation. Some studies have
confirmed a delay in pregnant women with diabetes, while
Several recent studies have examined the capacity of the vari-
other studies have not demonstrated a significant differ-
ous tests in today’s use for predicting neonatal lung maturity.9,61
ence in fetal lung maturity in pregnant women with diabetes
Lamellar bodies (LB)—1 to 5 μm lamellated structures synthe-
compared with the nondiabetic.4
sized by type 2 granular pneumocytes—can be counted using
Furthermore, it is still a matter of debate if preexisting
commercial blood cell analyzers.62 These storage granules con-
diabetes has a role in the delay of fetal lung maturation4
tain phospholipids, cholesterol, and several surfactant-specific
(Figure 35.4). Also, Di Renzo et al. in a study published in
proteins that become pulmonary surfactant. The first appear-
2011 investigated for the assessment of lung maturity. In the
ance of lamellar bodies in the cytoplasm of fetal pneumocytes
study, conducted between 2008 and 2010, 105 cases of gesta-
is between 20 and 24 weeks of gestation63; the lamellar bodies
tional diabetes were found that were assessed for fetal lung
become more numerous and larger and are secreted into fetal
maturity by evaluating the L/S ratio and the presence of PG
alveoli; and fetal breathing movements and exudation of lung
considering an L/S ratio of >4 and PG present at gestational
fluid carry these lamellar bodies into the amniotic fluid. The
age between 36 and 40 weeks. The vast majority was found
phospholipid content and the laminated structure change when
the fetal lung is mature. The measure of the surfactant/albumin
ratio in the amniotic fluid represents the fetal lung maturity
100
test. Other means of predicting the maturity include calculat-
90
ing the L/S ratio, measuring the amniotic fluid level of PG,64
performing the foam stability test65 or a fluorescence polar- 80
ization assay,66 and measuring amniotic fluid optical density 70
at 650 nm.67 All these tests measure some aspects of the pul- 60
monary surfactant contained in the amniotic fluid, and none
%
50
is perfect for the prediction of lung maturity. The ideal test 40
must be readily available and cheap with good predictive value. 30
Abd El Aal et al. with an important study found that a lamellar
20
body count is a good screening test for predicting neonatal lung
10
maturity. It is comparable to the fetal lung maturity assay by
fluorescence polarization and better in others. 0
<34 34–36 37 38 >39
Gestational age (weeks)
Summary Nondiabetic Pregestational DM Gestational DM
To date, the method most widely used to assess the fetal Figure 35.4 The controversy surrounding fetal lung maturity
lung maturity is the evaluation of the L/S ratio and PG. in diabetic pregnancy: a reevaluation. (From Langer, O., J.
Particular attention should be given to macrosomia, as Matern. Neonatal. Med., 12, 428, 2002. With permission.)
Table 35.2 Assessment of lung maturity: Results Table 35.3 Actions of glucocorticosteroids
77 mature Prenatal administration significantly increases content of

28 not mature (14 of them >38 weeks) elastin in the fetal lung.
10 retreatment with corticosteroids As transcriptional activators for the synthesis of surfactant-
associated proteins (SP-A and SP-B).
13 repeated after 3–5 days
On lung fibroblasts that elaborate fibroblast pneumocyte
11 mature
factor that subsequently acts on type II cells and
2 RDS stimulates surfactant synthesis.
Source: Di Renzo et al. (2011), unpublished data. Prenatal administration makes the immature lung more
responsive to exogenous surfactant.
Stimulate upregulation of corticotropin releasing hormone
to have reached the fetal lung maturity, but a substantial (CRH) output by the placenta.
minority (28 cases) was immature, half of which were found
to be at an age higher than 38 weeks of gestation. In the
immature cases, readministering corticosteroid has been ●● The Royal College of Obstetricians and Gynaecologists
necessary in giving a final result, however, of two cases of (RCOG)78,79 published the fourth edition of their guide-
RDS at birth (Table 35.2).69 line in October 2010.
In the last few years, new important information about

Induction of fetal lung maturity ACS to consider different unresolved topic issues has been
published: fetal, neonatal, and maternal effects of ACS,
When fetal lungs are immature, the infant will develop RDS. single versus multiple courses of ACS therapy, charac-
About 25% of untreated ones die within 28 days of birth, and teristics optimal steroids, the ideal dose, and the route of
another 25% will develop chronic lung disease. administration.71
Since 1972, when it was reported a decreased incidence Concerning the long-term effects on the use of multi-
of RDS in newborn from a mother who received a prena- ple steroid treatments in the last Cochrane Review (2006),
tal administration of corticosteroids,69 several randomized Roberts and Dalziel have demonstrated that ACS does not
trials about the role of steroids (and other drugs) to induce improve the risk, to the mother, of death, chorioamnionitis,
and/or improve pulmonary maturity have been conducted. or puerperal sepsis.72
Many reports have confirmed the original findings that Asztalos suggested that ACS reduces birth weight that is
antenatal administration of glucocorticoids to the mother associated with adverse adult outcomes like stroke, hyper-
is associated with a statistically significant reduction in the tension, coronary heart disease, and type 2 diabetes. In
incidence of RDS.70,71 animal studies, the long-term effects of ACS exposure on
Almost all have demonstrated the efficacy of corticoste- hypothalamic–pituitary–adrenal (HPA) axis activity and
roid treatment in reducing perinatal morbidity and mor- neurological function are largely documented. From that,
tality, as confirmed by an RCT that has compared a single it appears that ACS, regardless of dose exposure, can affect
course of corticosteroids with either placebo or no interven- the fetal development of the HPA axis causing permanent
tion, confirming the efficacy of this treatment.72 changes in this axis that persists through life and manifested
During the last decade, the strategy for the prevention of itself by chronic illness and behavioral changes.80–83
RDS has been directed toward the acceleration of fetal lung The possible long-term consequences of AGC exposure
maturity by administering various hormones to the mother, in HPA axis function and subtler neurodevelopmental out-
but at the present time, glucocorticoids remain the most comes, including adaptation to stress, cognition, behavior,
widely used agents. and the cardiovascular and immune responses, are confirmed
Glucocorticoids act in different ways on the matura- by another study of Waffarn and Davis.84 Further information
tion process,73 known already for many years, as listed in is required concerning the long-term effects into adulthood.
Table 35.3. In consideration of the effectiveness of ACS to reduce
The main institutions have been producing and updat- RDS and other complications, the guideline recommended
ing recommendations and guidelines on the use of antenatal the use of ACS in all women with a high risk of preterm
corticosteroids (ACSs): delivery (24–34 weeks) and after 34 weeks gestation only if
there is an evidence of pulmonary immaturity, particularly
●● The National Institutes of Health (NIH) published a in patients with conditions that cause a delay in the fetal lung
Consensus Development Conference Statement in 1994 maturation as diabetes.71
on the use of ACSs74 and in 2000 a revision to the use of Considering the choice of the type, dose, and mode of
repeated courses of ACS.75 corticosteroid administration, today, betamethasone, as the
●● In 2011, the American College of Obstetricians and steroid of choice, is recommended to be given in a course of
Gynecologists’ Committee on Obstetric Practice76,77 two doses of 12 mg administered intramuscularly 24 hours
whose conclusions were in agreement with NIH. apart79 (Table 35.4). Recent evidences appear to support the
Induction of fetal lung maturity 293
Table 35.4 Guidelines of the Royal College of Obstetricians and Gynaecologists: Antenatal Corticosteroids to
Reduce Neonatal Morbidity and Mortality
This is the fourth edition of this guideline, which was previously published in April 1996, December 1999
and February 2004. The previous guideline was entitled Antenatal Corticosteroids to Prevent Respiratory Distress Syndrome.
1. Purpose and scope
The aim of this guideline is to provide up-to-date information on the appropriate use of antenatal corticosteroid therapy
in women whose babies are at risk of complications owing to either preterm birth or elective cesarean section at term.
This guideline does not assess the effectiveness of tests in the prediction of preterm delivery (e.g., ultrasound scanning for
cervical length, cervical fibronectin measurement) or other interventions that may prevent preterm labour (e.g., tocolysis).
2. Background
There is evidence to suggest that antenatal corticosteroids are effective not only in reducing respiratory distress syndrome
(RDS) but also in reducing other complications of prematurity such as intraventricular haemorrhage (IVH). The title of this
guideline has been changed to Antenatal Corticosteroids to Prevent Neonatal Morbidity and Mortality to include all groups
of women and all outcomes.
3. Identification and assessment of evidence
This RCOG guideline was developed in accordance with standard methodology for producing RCOG Greentop
Guidelines. The Cochrane Library (including the Cochrane Database of Systematic Reviews), DARE, Embase, TRIP, Medline
and PubMed (electronic databases) were searched for relevant randomised controlled trials, systematic reviews and
meta-analyses and cohort studies. The search was restricted to articles published between 2002 to July 2008. The databases
were searched using the relevant MeSH terms, including all subheadings, and this was combined with a keyword search.
Search words included ‘steroids’, ‘premature labour’, ‘premature fetus’ and ‘membrane rupture’, and the search was limited
to humans and to the English language. The National Library for Health and the National Guidelines Clearing House were
also searched for relevant guidelines and reviews.
4. What are the benefits of antenatal corticosteroids?
Antenatal steroids are associated with a significant reduction in rates of neonatal death, RDS and intraventricular
haemorrhage and are safe for the mother.
Antenatal corticosteroids have no known benefits for the mother.
A Cochrane review of 21 studies (3885 women and 4269 infants) showed that treatment of women at risk of preterm
birth with a single course of antenatal corticosteroids reduced the risk of neonatal death by 31% (95% CI 19–42%), RDS by
44% (95% CI 31–57%) and intraventricular haemorrhage by 46% (95% CI 31%–67%).1 Antenatal corticosteroid use is also
associated with a reduction in necrotising enterocolitis, respiratory support, intensive care admissions and systemic
infections in the first 48 hours of life compared with no treatment or treatment with placebo.
5. At what gestation should antenatal steroids be used?
Clinicians should offer a single course of antenatal corticosteroids to women between 24+0 and 34+6 weeks of gestation
who are at risk of preterm birth.
Antenatal corticosteroids can be considered for women between 23+0 and 23+6 weeks of gestation who are at risk of
preterm birth.
The decision to administer corticosteroids at gestations less than 24+0 weeks should be made at a senior level taking all
clinical aspects into consideration.
The data are strongest for gestations between 26+0 and 34+6 weeks. The data for pregnancies between 24+0 and 26+0
weeks of gestation are scarce, with only one trial (49 infants) contributing data to the Cochrane review.1
The conclusion of the authors of the Cochrane review1 and the American Congress of Obstetricians and Gynecologists
(ACOG) Committee opinion (2008)2 is that, despite the paucity of data at earlier gestations, the reduction in outcomes
other than RDS at 26+0 weeks of gestation would suggest that there is some benefit in corticosteroid prophylaxis at earlier
gestations between 24+0 and 26+0 weeks.
In a prospective cohort of 4446 infants between 22+0 and 25+0 weeks of gestation, multivariable analyses showed that
those who received intensive care, were exposed to antenatal corticosteroids, were of female sex, were from singleton
pregnancies, and of higher birth weight (per each 100 g increment) had a reduced risk of death. Among survivors, the risk
of death, or profound or any neurodevelopmental impairment at 18–22 months corrected age, was reduced.3 These
reductions were similar to those associated with a 1-week increase in gestational age.
A retrospective cohort study on 181 infants born at 23 weeks of gestation revealed that those exposed to antenatal
corticosteroids had decreased odds of death (OR 0.32, 95% CI 0.12–0.84), with no significant differences in the occurrence
of necrotising enterocolitis among survivors (15.4% compared with 28.6%, P = 0.59) or severe intraventricular
haemorrhage (23.1% compared with 57.1%, P = 0.17). Only a complete course of corticosteroids was associated with a
decreased odds of death (OR 0.18, 95% CI 0.06–0.54).4 The study concluded that neonates at 23 weeks of gestation whose
mothers completed a course of antenatal corticosteroids had an associated 82% reduction in odds of death.
Evidence from the EPICure study, a prospective cohort study, showed that of 283 babies born at less than 26+0 weeks of
gestation assessed at 2.5 years and 241 assessed at 6 years, antenatal corticosteroids was associated with an increased
mental development index.5
(Continued)
Table 35.4 (Continued) Guidelines of the Royal College of Obstetricians and Gynaecologists: Antenatal
Corticosteroids to Reduce Neonatal Morbidity and Mortality
Trial data are scanty for pregnancies at the extreme of prematurity. Obstetricians currently have the discretion to
administer steroids before the 24th week of pregnancy, but the whole clinical picture needs to be taken into account with
respect to intact survival data as well as the chance of any survival based on antenatal assessment of viability at these
extremes (e.g., by estimation of fetal weight). In this context, we have advised caution and discussion at senior level prior to
considering antenatal corticosteroid administration at 23+0 to 23+6 weeks of gestation.
6. How long after administration is a course of antenatal corticosteroids most effective?
Antenatal corticosteroids are most effective in reducing RDS in pregnancies that deliver 24 hours after and up to 7 days
after administration of the second dose of antenatal corticosteroids.
Antenatal corticosteroid use reduces neonatal death within the first 24 hours and therefore should still be given even if
delivery is expected within this time.
Reduction in RDS is seen in infants born up to 7 days after the first dose (RR 0.46, 95% CI 0.35–0.60, nine studies, 1110
infants).1 No reduction in neonatal death, RDS or cerebro-ventricular haemorrhage is seen in infants delivered more than
7 days after treatment with antenatal corticosteroids.1
Antenatal corticosteroid use reduces neonatal death even when infants are born less than 24 hours after the first dose has
been given (RR 0.53, 95% CI 0.29–0.96, four studies, 295 infants).1
However, caution should be exercised in the interpretation of the data. The question as to whether the effects of antenatal
corticosteroids change with time to delivery cannot be answered by existing analyses, the main reason being that the time
periods chosen for subgroup analysis were arbitrary. All the babies born at term were in one subgroup. Analyses based on
subgroups defined by outcomes not known at randomisation are subject to considerable bias. This question would require
re-analysis of individual patient data to clarify whether the association is real.
7. How safe is the use of antenatal corticosteroids?
Women may be advised that the use of a single course of antenatal corticosteroids does not appear to be associated with
any significant short-term maternal or fetal adverse effects.
Evidence on the longer-term benefits and risks of a single course of antenatal corticosteroids shows no clear difference in
adverse neurological or cognitive effects. There is still insufficient evidence on the longer-term benefits and risks of multiple
courses of antenatal corticosteroids.
Studies in the sheep model have shown that injections with glucocorticoids enhance fetal lung maturation but are
associated with developmental and other functional alterations that are of concern. Weekly doses to the sheep mother are
associated with restricted fetal growth, delayed myelination of the central nervous system, altered blood pressure soon after
birth and increased insulin response to glucose challenge in early adulthood.
There have therefore been concerns about whether steroid administration may have adverse effects on the long-term
outcomes of children exposed in the antenatal period. Multivariate analyses in humans have shown that increasing the
number of glucocorticoid exposures, for the purpose of enhancing lung maturation prior to preterm birth, is associated with
reduced birth weight and behavioural disorders at 3 years of age.7
The Cochrane review and other studies have shown that a single course of corticosteroid therapy for preterm birth results
in benefit without causing significant adverse effects such as neonatal or maternal sepsis.1,8,9 The Cochrane review revealed
that no statistically significant differences were seen for chorioamnionitis (RR 0.91, 95% CI 0.70–1.18, 12 studies, 2485
women) or puerperal sepsis (RR 1.35, 95% CI 0.93–1.95, eight studies, 1003 women). There were no maternal deaths, but
the randomised controlled trials were underpowered to detect such a difference (RR 0.98, 95% CI 0.06–15.50, three
studies, 365 women).1
Long-term follow-up of survivors from randomised trials of antenatal corticosteroid therapy through childhood to
adulthood (up to 20 years of age) shows no clear adverse neurological or cognitive effects.10,11
A randomised controlled trial showed that children who had been exposed to repeat as compared with single courses of
antenatal corticosteroids did not differ significantly in physical or neurocognitive measures. However, there was a
nonsignificant higher risk of cerebral palsy among children who had been exposed to repeat doses of corticosteroids
(RR 5.7, 95% CI 0.7–46.7, P = 0.12). The number of children with cerebral palsy was small. This was, however, felt to be of
concern and warranting further study.12
Furthermore, studies on long-term cardiovascular risks, cognitive functioning, working memory and attention, psychiatric
morbidity, handedness or health-related quality of life on the survivors of the first and largest double-blind, placebo-
controlled, randomised trial of a single course of antenatal betamethasone for the prevention of neonatal RDS at age 31
found no differences between groups exposed to betamethasone and placebo.
8. Are there any contraindications to the use of antenatal corticosteroids?
Caution should be exercised when giving corticosteroid therapy to women with systemic infection including tuberculosis
or sepsis. Corticosteroids suppress the immune system, so there is a risk that their use may activate latent infections or
exacerbate fungal infections. In a woman with systemic infection, it may theoretically suppress the immune response to
infection. There is no evidence to suggest that a single course of corticosteroids would have a profound effect in women
with systemic infection, but caution should be exercised in its use.
Senior opinion should be sought when contemplating delaying delivery for steroid prophylaxis in cases of overt
chorioamnionitis.
(Continued)
Induction of fetal lung maturity 295
A large meta-analysis of observational studies reports that clinical chorioamnionitis is significantly associated with both
cystic periventricular leucomalacia (RR 2.6, 95% CI 1.7–3.9) and cerebral palsy (RR, 1.9, 95% CI 1.5–2.5).15,16 This would
suggest that with chorioamnionitis, a course of antenatal corticosteroids may be started, but should not delay delivery if
indicated by maternal or fetal condition.
9. Who should receive antenatal corticosteroids?
Antenatal corticosteroids should be given to all women at risk of iatrogenic or spontaneous preterm birth up to 34+6
weeks of gestation.
Antenatal corticosteroids should be given to all women for whom an elective cesarean section is planned prior to 38+6
weeks of gestation.
There is no evidence to support a practice of prophylactic steroids in women with a previous history of preterm delivery
or multiple pregnancy who show no signs of being at risk of iatrogenic or spontaneous preterm birth.
There is evidence of benefit in all major subgroups of preterm babies, such as women with premature rupture of
membranes and pregnancy-related hypertension syndromes as well as the subgroups discussed below.1 This benefit is
irrespective of race or gender. A single course of antenatal corticosteroids should be considered routine for preterm
delivery with few exceptions.
9.2 In women with diabetes mellitus
Diabetes mellitus is not a contraindication to antenatal corticosteroid treatment for fetal lung maturation.
Women with impaired glucose tolerance or diabetes who are receiving fetal steroids should have additional insulin
according to an agreed protocol and be closely monitored.
Women with either insulin-dependent diabetes or gestational diabetes were not entered into randomized controlled
trials of antenatal corticosteroid therapy. There is therefore no evidence from randomized controlled trials that antenatal
corticosteroid therapy is either safe or effective in these circumstances. Maternal hyperglycemia can adversely affect fetal
lung maturity. It is possible that any benefit of corticosteroids could be offset by corticosteroid-induced hyperglycemia.19
However, the National Institute of Health and Clinical Excellence (NICE) has published a clinical guideline for diabetes
in pregnancy that states that ‘diabetes should not be considered a contraindication to antenatal corticosteroids’.20 The
NICE guideline recommends that diabetic women receiving steroids should have additional insulin according to an agreed
protocol.
10. What is the best dose and route of administration for a course of antenatal corticosteroids?
Betamethasone 12 mg given intramuscularly in two doses or dexamethasone 6 mg given intramuscularly in four doses
are the steroids of choice to enhance lung maturation.
The most extensively studied regimens of corticosteroid treatment for the prevention of RDS are two doses of
betamethasone 12 mg given intramuscularly 24 hours apart or four doses of dexamethasone 6 mg given intramuscularly
12 hours apart. Evidence for other dosing regimens such as the commonly used two doses of betamethasone 12 mg given
12 hours apart is sparse (two trials, 92 women),29,30 but it would seem reasonable that as long as 24 mg of either drug is
given within a 24–48-hour period, any dosing regimen can be used.
A large nonrandomised retrospective study has suggested that infants exposed to antenatal betamethasone have less
neonatal cystic periventricular leukomalacia than infants exposed to antenatal dexamethasone.31 Another historical cohort
study used multivariate logistic regression analysis to compare the two steroid-treated groups with each other. It showed
the risk of neonatal death was lower with betamethasone than with dexamethasone (OR 0.44 for betamethasone versus
0.73 with dexamethasone, p < 0.05).32
The Cochrane review on antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm
birth suggests that betamethasone treatment causes a larger reduction in RDS than dexamethasone. Since then, in a
Cochrane review on different corticosteroid regimes (ten trials, 1089 women and 1161 infants), dexamethasone decreased
the incidence of intraventricular haemorrhage compared with betamethasone (RR 0.44, 95% CI 0.21–0.92, four trials, 549
infants).33
This review advised caution in suggesting benefit of dexamethasone over betamethasone. A randomised controlled trial
looking at long-term outcomes in children treated with dexamethasone would be required as this form of corticosteroid
has not been studied in the long term. For current practice, the use of either corticosteroid is acceptable.
Comparison of oral versus intramuscular administration of dexamethasone (one trial, 183 infants) suggests that oral
administration increased the incidence of neonatal sepsis (RR 8.48, 95% CI 1.11–64.93) in one trial of 183 infants.
No statistically significant differences were seen for other outcomes reported. The authors suggest that very few
conclusions about the optimum regimens can be made.
11. When should an antenatal course of corticosteroids be repeated?
Weekly repeat courses of antenatal corticosteroids reduce the occurrence and severity of neonatal respiratory disease,
but the short-term benefits are associated with a reduction in weight and head circumference. Weekly repeat courses are
not recommended.
A single rescue course may be considered with caution in pregnancies where the initial course was given at less than
26+0 weeks of gestation. Senior opinion should be sought if a rescue course is to be considered.
(Continued)
Animal studies and observational studies in humans have suggested that multiple courses of steroids may lead to
possible harmful effects including growth delay, brain developmental delay, lung development problems, necrotising
enterocolitis, maternal and neonatal sepsis, adrenal gland insufficiency and placental infarction.35–41 A systematic review
of 19 randomised controlled trials of repeat doses of antenatal corticosteroids in animals concluded that there might be
beneficial effects in terms of lung function but adverse effects on brain function and fetal growth.40
The effect of repeat courses of antenatal corticosteroids in humans is the subject of a Cochrane review by Crowther
et al.42 The review suggests that although repeat courses reduce the occurrence and severity of neonatal lung disease
(RR 0.60, 95% CI 0.48–0.75, three trials, 2139 infants) and the risk of serious health problems in the first few weeks of
life (RR 0.79, 95% CI 0.67–0.93, four trials, 2157 infants), they are associated with a reduction in some measures of
weight (Z-score [weighted mean difference] –0.13, 95% CI –26 to 0.00, one trial, 1144 infants) and with being small for
gestational age at birth (RR 1.63, 95% CI 1.12–2.37, two trials, 602 infants).
There is still insufficient evidence on the longer-term benefits and risks to be able to recommend the routine use of
repeat courses.
The Cochrane review did not include the results of several continuing trials, two of which have been completed since its
publication.43,44 The results of TEAMS (Trial of early and multiple steroids) are awaited.43 In the MACS (Multiple Courses of
Antenatal Corticosteroids for Preterm Birth) trial of 1858 women at 25–32 weeks of gestation, multiple courses of
antenatal corticosteroids (n = 937) or placebo (n = 921) every 14 days until week 33 or delivery (whichever came first)
found that although infants exposed to multiple courses of antenatal corticosteroids had similar morbidity and mortality to
those exposed to placebo (12.9% versus 12.5%), those receiving multiple doses of corticosteroids weighed less at birth
(2216 g versus 2330 g, p = 0.0026), were shorter (44.5 cm versus 45.4 cm, p < 0.001), and had a smaller head
circumference (31.1 cm versus 31.7 cm, p < 0.001). A regime of repeat doses should not be recommended.44
In 2000, a National Institutes of Health consensus statement concluded that ‘repeat courses of corticosteroids should
not be used routinely and should be reserved for women enrolled in randomised controlled trials’.45 A recent randomised
controlled trial recruited 437 patients with singleton or twin pregnancies at less than 33+0 weeks of gestation who had
completed a single course of corticosteroids before 30+0 weeks of gestation and randomised them to a single rescue
course of betamethasone (two 12 mg doses 24 hours apart) or placebo. There was a significant reduction in the primary
outcome of composite neonatal morbidity at less than 34 weeks of gestation in the rescue steroid group than in the
placebo group (43.9% versus 63.6%, OR 0.45, 95% CI 0.27–0.75, p = 0.002) and significantly decreased RDS, ventilator
support and surfactant use.
Perinatal mortality and other morbidities were similar in each group. No long-term outcomes were reported for this
study.46
Peltoniemi et al. looked at whether a single additional dose of 12 mg betamethasone given when preterm birth is
imminent before 34+0 weeks of gestation and at least a week after the full treatment course would lower the risk of RDS
and severe (grade 3 or 4) intraventricular haemorrhage compared with placebo. This study showed that more infants in the
steroid group required surfactant therapy. Post hoc analysis of data on 206 infants who were delivered within 1 to 24
hours indicated that the additional steroid dose tended to increase the risk of RDS and reduce the rate of intact survival.
No differences were found between the steroid and placebo groups in mortality rates or rates of severe intraventricular
haemorrhage. The results of this study would suggest that rescue doses may alter respiratory adaptation.47 A subanalysis
within the MACS study also showed no benefit from a rescue dose.
A rescue course of two doses of 12 mg betamethasone or four doses of 6 mg dexamethasone should only be considered
with caution in those pregnancies where the first course was given at less than 26+0 weeks of gestation and another
obstetric indication arises later in pregnancy.
In the absence of evidence, a rescue course of two doses of 12 mg betamethasone or four doses of 6 mg dexamethasone
should only be considered with caution in those pregnancies where the first course was given at less than 26+0 weeks of
gestation and another obstetric indication arises later in pregnancy. This would be justified by the paucity of data on the
efficacy of the current dosing regimens on babies less than 26+0 weeks of gestation.
Source: Anceschi, M.M. et al., Effects of corticosteroids on fetal lung maturation, in: The Surfactant System of the Lung, MacMillan Press,
London, U.K., 1991, pp. 23–27.
use of dexamethasone as the first-line treatment: four doses Conclusions

of 6 mg given every 12 hours. However, if dexamethasone is
not available, betamethasone may be used.85 If a pregnant woman has diabetes under poor control, the
Finally, despite the new evidence comparing single versus infant is at risk for RDS because of delayed lung matura-
multiple doses of steroids, it is not possible, at this moment, tion. The delayed lung maturation includes a delay in the
to modify the recommendations of the NIH Consensus appearance of surfactant and, probably, delayed lung struc-
Development Panel on the use of repeated courses of ACS tural maturation as a result of both high insulin and glucose
or to establish the number of doses or the interval between effects on the fetal lung. However, concerns about the reli-
them that are safer for the fetus.71 ability of the lung maturity test in diabetic pregnancies have
References 297
decreased as management of the diabetic pregnant women complications of diabetes like macrosomia. Furthermore,
has focused on the good control of blood glucose levels. when a preterm delivery is indicated such as in uncontrolled
It is largely accepted that a close blood glucose control diabetic patients, it is mandatory to use a combination of
should be maintained during pregnancy, in order to reduce tests for the prediction of fetal lung maturity.
REFERENCES
1. Al-Khalifah R, Al-Subaihin A, Al-Kharfi T, Al-Alaiyan S, AlFaleh 22. Kulovich MV, Gluck CKL. The lung profile: II. Complicated preg-
KM. Neonatal short-term outcomes of gestational diabetes nancy. Am J Obstet Gynecol 1979; 136: 64.
mellitus in Saudi mothers: A retrospective cohort study. J Clin 23. Smith BT. Pulmonary surfactant during fetal development and
Neonatol January–March 2012; 1(1): 29–33. the neonatal adaptation: Hormonal control. In: Robertson B, Van
2. Weindling MA. Offspring of diabetic pregnancy: Short-term out- Golde LMG, Batemburg JJ, eds., Pulmonary Surfactant. Elsevier:
comes. Semin Fetal Neonatal Med 2009; 14(Suppl. 2): 111–118. Amsterdam, the Netherlands, 1984, pp. 357–381.
3. Johns K, Olynik C, Mase R, Kreisman S, Tildesley H. Gestational 24. Ferroni KM, Gross TL, Sokol RJ. What affects fetal pulmonary
diabetes mellitus outcome in 394 patients. J Obstet Gynaecol maturation during diabetic pregnancy? Am J Obstet Gynecol
Can 2006; 28: 122–127. 1984; 150: 270.
4. Langer O. The controversy surrounding fetal lung maturity in dia- 25. Tyden O, Berne C, Erikkson UJ. Fetal maturation in strictly con-
betes pregnancy: A re-evaluation. J Matern Neonatal Med 2002; trolled diabetic pregnancy. Diabetes Res 1984; 1: 1314.
12: 428–432. 26. Landon MB, Gabbe SC, Piana R. Neonatal morbidity in preg-
5. Reece EA, Homko CJ. Infant of the diabetic mother. Semin nancy complicated by diabetes mellitus: Predictive value of
Perinatol 1994; 18: 459–469. maternal glycemic profiles. Am J Obstet Gynecol 1987; 156:
6. Hay WW Jr. Care of the infant of the diabetic mother. Curr Diab 1089–1093.
Rep February 2012; 12(1): 4–15. 27. Cunningam MD, Desai NS, Thomson SA. Amniotic fluid phos-
7. Inwin JA. The future role for a diabetes specialist midwife. Best phatidylglycerol in diabetic pregnancy. Am J Obstet Gynecol
Pract Res Clin Endocrinol Metab August 2010; 24(4): 653–662. 1978; 131: 712.
8. Watson D, Rowan J, Neale L, Battin MR, Aust NZ. Admissions to 28. Ylinen K. High maternal levels of hemoglobin A1c associated
neonatal intensive care unit following pregnancies complicated with delayed fetal lung maturation in insulin-dependent diabetic
by gestational or type 2 diabetes. J Obstet Gynaecol December pregnancies. Acta Obstet Gynecol Scand 1987; 66: 263–266.
2003; 43(6): 429–432. 29. Bourbon JR, Doucet E, Rieutort M. Role of myo-inositol in
9. De Luca AK, Nakazawa CY, Azevedo BC, Rudge MV, De Araújo impairment of fetal lung phosphatidylglycerol biosynthesis in the
Costa RA, Calderon IMI. Influence of glycemic control on fetal diabetic pregnancy: Physiological consequences of phosphati-
lung maturity in gestation affected by diabetes or mild hypergly- dylglycerol deficient surfactant in the newborn rat. Exp Lung Res
cemia. Acta Obstet Gynecol Scand 2009; 88: 1036–1040. 1986; 11: 195.
10. Kjos SL, Berkowitz KM, Kung B. Prospective delivery of reliably 30. Treviño-Alanís M, Ventura-Juárez J, Hernández-Piñero J, Nevárez-
dated term infants of mothers without determination of fetal Garza A, Quintanar-Stephano A, González-Piña A. Delayed
lung maturity: Comparison to historical control. J Matern Fetal lung maturation of foetus of diabetic mother rats develop with a
Neonatal Med 2002; 12: 433–437. diminish, but without changes in the proportion of type I and II
11. Kjos SL, Walther FJ, Montoro M. Prevalence and etiology if respi- pneumocytes, and decreased expression of protein D-associated
ratory distress in infants of diabetic mothers: Predictive value surfactant factor. Anat Histol Embryol June 2009; 38(3): 169–176.
of fetal lung maturation tests. Am J Obstet Gynecol 1990; 163: 31. Pinter E, Peyman JA, Snow K. Effects of maternal diabetes on fetal
898–903. rat lung ion transport: Contribution of alveolar and bronchiolar
12. Di Renzo GC, Anceschi MM, Guidetti R, Cosmi EV. Requirements epithelial cells to Na+,K+ ATPase expression. J Clin Invest 1991;
of perinatal prevention and treatment of respiratory distress syn- 87: 821–830.
drome. Eur Respir J 1989; 2(Suppl. 3): 68s–72s. 32. Bhavnani BR, Enhorning G, Ekelund L. Maternal diabetes and its
13. Horowitz K, Feldman D, Stuart B, Borgida A, Ming Y, Fang effect on biochemical and functional development of rabbit fetal
V, Herson V. Full-term neonatal intensive care unit admission lung. Biochem Cell Biol 1988; 66: 396–404.
in an urban community hospital: The role of respiratory mor- 33. Hawdon JM, Aynsley-Green A. Neonatal complications, includ-
bidity. J Matern Fetal Neonatal Med November 2011; 24(11): ing hypoglycemia. In: Dornhorst A, Hadden DR, eds., Diabetes
1407–1410. and Pregnancy: An International Approach to Diagnosis and
14. Engle M, Langan SM. The effects of insulin and hyperglycemia on Management. John Wiley & Sons: New York, 1996, pp. 303–318.
surfactant phospholipid biosynthesis in organotypic cultures on 34. Croze ML, Soulage CO. Potential role and therapeutic interests
type II pneumocytes. Biochim Biophys Acta 1983; 753: 6–13. of myo-inositol in metabolic diseases. Biochimie October 2013;
15. Grenache DG, Gronowski AM. Fetal lung maturity. Clin Biochem 95(10): 1811–1827.
2006; 39: 1–10. 35. D’Anna R, Scilipoti A, Giordano D, Caruso C, Cannata ML,
16. Moore TR. A comparison of amniotic fluid fetal pulmonary Interdonato ML, Corrado F, Di Benedetto A. myo-Inositol supple-
phospholipids in normal and diabetic pregnancy. Am J Obstet mentation and onset of gestational diabetes mellitus in pregnant
Gynecol 2002; 186: 641–650. women with a family history of type 2 diabetes: A prospective,
17. Piper JM, Samueloff A, Langer O. Outcome of amniotic fluid randomized, placebo-controlled study. Diabetes Care April 2013;
analysis and neonatal respiratory status in diabetic and nondia- 36(4): 854–857.
betic pregnancies. J Reprod Med 1995; 40: 780–784. 36. Matarrelli B, Vitacolonna E, D’Angelo M, Pavone G, Mattei
18. Piper JM, Xenakis EMJ, Langer O. Delayed appearance of pulmo- PA, Liberati M, Celentano C. Effect of dietary myo-inositol
nary maturation markers is associated with poor glucose control supplementation in pregnancy on the incidence of maternal
in diabetic pregnancies. J Matern Fetal Med 1998; 7: 148–153. gestational diabetes mellitus and fetal outcomes: A randomized
19. Smith BT, Giroud CJP. Insulin antagonism of cortisol action on controlled trial. J Matern Fetal Neonatal Med July 2013; 26(10):
lecithin synthesis by cultured fetal lung cells. J Pediatr 1975; 87: 967–972.
953–955. 37. Corrado F, D’Anna R, Di Vieste G, Giordano D, Pintaudi B,
20. Thulesen J, Poulsen SS, Nexø E, Raaberg L. Epidermal growth Santamaria A, Di Benedetto A. The effect of myoinositol supple-
factor and lung development in the offspring of the diabetic rat. mentation on insulin resistance in patients with gestational diabe-
Pediatr Pulmonol February 2000; 29(2): 103–112. tes. Diabet Med August 2011; 28(8): 972–975.
21. Carlson KS, Smith BT. Insulin acts on the fibroblast to inhibit glu- 38. Moore TR. Diabetes in pregnancy. In: Creasy RK, Resnik R, eds.,
cocorticoid stimulation of lung maturity. J Appl Physiol 1984; 57: Maternal Fetal Medicine, 4th ed., Saunders: Philadelphia, PA,
1577–1579. 1999, pp. 964–995.
39. Bartelsmeyer JA. Fetal lung maturity. In: Winn HN, Hobbins JCH, 65. Abd El Aal DEM, Elkhirshy AA, Atwa S, El-Kabsh MY. Lamellar
eds., Clinical Maternal–Fetal Medicine. Parthenon: London, U.K., body count as a predictor of neonatal lung maturity in high-risk
2000. pregnancies. Int J Gynaecol Obstet 2005; 89(1): 19–25.
40. Hordor JG, Poggi SH, Spong CY, Goodwin KM, Vink JS, Pezzullo 66. Wijnberge RL, Huisjes AJ, Voorbij AM. The accuracy of lamel-
JC, Ghidini A. Risk of third trimester amniocentesis: A case-con- lar body count and lecithin/sphingomyelin ratio in the prediction
trol study. Am J Perinatol 2006; 23(3): 177–180. of neonatal respiratory distress syndrome: A meta-analysis. Br J
41. NHS. Diabetes in pregnancy: Management of diabetes and its Obstet Gynecol 2001; 108: 583–588.
complications from pre-conception to the postnatal period. NICE 67. Ross GE, Bever FN, Uddin Z, Hockman EM, Herman BA.
clinical guideline 63, March 2008. Decreased laboratory testing for lecithin-to-sphingomyelin ratio
42. De Luca AK, Nakazawa CY, Azevedo BC, Rudge MV, De Araújo and phosphatidylglycerol after fetal lung maturity assessment
Costa RA, Calderon IM. Influence of glycemic control on fetal from lamellar body count in amniotic fluid. J Am Osteopath
lung maturity in gestation affected by diabetes or mild hypergly- Assoc 2002; 102: 423–428.
cemia. Acta Obstet Gynecol 2009; 88: 1036–1040. 68. Ghidini A, Spong CY, Goodwin K, Pezzullo JC. Optimal thresh-
43. Piper JM. Lung maturation in diabetes in pregnancy: If and when olds of the lecithin/sphingomyelin ratio and lamellar body
to test. Semin Perinatol 2002; 26: 206–209. count for the prediction of the presence of phosphatidyl glyc-
44. Grenache DG, Gronowski AM. Fetal lung maturity. Clin Biochem erol in diabetic women. J Matern Fetal Neonatal Med 2003; 12:
2006; 50: 108–112. 95–98.
45. Gluck L, Kulovich MV, Boerer Jr RC. Diagnosis of the respiratory 69. Sher G, Statland BE. Assessment of fetal pulmonary maturity by
distress syndrome by amniocentesis. Am J Obstet Gynecol 1971; the Lumadex foam stability index test. Obstet Gynecol 1983; 35:
109: 440. 444–449.
46. Gluck L, Kulovich MV, Borer RC. The interpretation and signifi- 70. Kesselman EJ, Figueroa R, Garry D, Maulik D. The usefulness of
cance of the lecithin/sphingomyelin ratio in amniotic fluid. Am J the TDx/TDxFLx fetal lung maturity II assay in the initial evalu-
Obstet Gynecol 1974; 120: 142. ation of fetal lung maturity. Am J Obstet Gynecol 2003; 188:
47. Hallman M, Kulovich M, Kirkpatrick E. Phosphatidylinositol and 1220–1222.
phosphatidylglycerol in amniotic fluid: Indices of lung maturity. 71. Sobczak M, Pertynska M, Wilczynski J. Evaluation of predictive
Am J Obstet Gynecol 1976; 125: 613. value of lung maturity in newborns based on effectiveness of
48. Kulovich MV, Hallman MB, Gluck L. The lung profile. I. Normal amniotic fluid study. Ginekol Pol 2000; 71: 340–346.
pregnancy. Am J Obstet Gynecol 1975; 135: 57. 72. Ghidini A, Spong CY, Goodwin K, Pezzullo JC. Optimal thresh-
49. Di Renzo GC, Cutuli A, De Graaf O et al. Problematiche diagnos- olds of the lecithin/sphingomyelin ratio and lamellar body
tiche nella valutazione della maturità fetale. In: AOGOI, Corso count for the prediction of the presence of phosphatidyl glyc-
di aggiornamento in Medicina Perinatale, Bormio, Italy, 1997, erol in diabetic women. J Matern Fetal Neonatal Med 2002; 12:
pp. 75–77. 95–98.
50. Robert MF, Neff RK. Association between maternal diabetes and 73. Liggins GC, Howie RN. A controlled trial of antepartum gluco-
the respiratory distress syndrome in the newborn. N Engl J Med corticoid treatment for prevention of the respiratory distress syn-
1976; 294: 357–360. drome in premature infants. Pediatrics 1972; 50: 515–525.
51. Gindes L, Chen R, Perri T et al. Perinatal morbidity in offspring of 74. Anceschi M, Maranghi L, Cosmi E. Prevention of fetal and neo-
diabetic mothers. Isr J Obstet Gynecol 2002; 12: 165–171. natal lung immaturity. In: Kurjak A, Chervenak F, eds., Textbook
52. Cutuli A, Clerici G, Di Renzo GC. Fetal lung maturity. In: of Perinatal Medicine, Information Health Care: Oxford, U.K.,
Textbook of Diabetes and Pregnancy. Hod M, Jovanovic L, Di 2006, pp. 1433–1445.
Renzo GC, De Leiva A, Langer O. Martin Dunitz Ed., New York, 75. Miracle X, Di Renzo GC, Stark A, Fanaroff A, Carbonell-Estrany,
2003; 276–288. Saling E (Coordinators of WAPM Prematurity working group).
53. Dubin SB. The laboratory assessment of fetal lung maturity. Am J Guideline for the use of antenatal corticosteroids for fetal matura-
Clin Pathol 1992; 97: 836. tion. J Perinat Med 2008; 36(3): 191–196.
54. Harvey D, Parkinson C, Campbell S. Risk of respiratory distress 76. Roberts D, Dalziel S. Antenatal corticosteroids for accelerating
syndrome. Lancet 1975; 1: 42. fetal lung maturation for woman at risk at preterm birth. Cochrane
55. McMahon MJ, Mimouni F, Miodovnik M. Surfactant associated database Syst Rev 2006; (3): CD004454.
protein (SAP-35) in amniotic fluid from diabetic and non diabetic 77. ACOG Committee on Obstetric Practice. ACOG Committee
pregnancies. Obstet Gynecol 1987; 70: 94–98. Opinion No. 475: Antenatal corticosteroid therapy for fetal matu-
56. Hallman M, Arjomaa P, Mizumoto M, Akino T. Surfactant pro- ration. Obstet Gynecol 2011; 117: 422.
teins in the diagnosis of fetal lung maturity. Am J Obstet Gynecol 78. Green-Top Guideline No. 7. RCOG. Antenatal corticosteroids to
1988; 158: 531–535. reduce neonatal morbidity and mortality. October 2010.
57. Ludimir J, Alvarez JG, Mennuti MT. Cholesterol palmitate as a pre- 79. Anceschi MM, Luzi G, Broccucci L, Di Renzo GC, Cosmi
dictor of fetal lung maturity. Am J Obstet Gynecol 1987; 157: 84–88. EV. Effects of corticosteroids on fetal lung maturation. In: The
58. Ludimir J, Alvarez JG, Landon MB et al. Amniotic fluid choles- Surfactant System of the Lung, The MacMillan Press: London,
terol palmitate in pregnancies complicated by diabetes mellitus. U.K., 1991, pp. 23–27.
Obstet Gynecol 1988; 72: 360–362. 80. Report on The Consensus Development Conference on the Effect
59. Tait JF, Franklin RW, Simpson JB, Ashwood ER. Improved fluo- of Corticosteroids for Fetal maturation on Perinatal Outcomes.
rescence polarization assay for use in evaluating fetal lung matu- U.S. Department of Health and Human Service, Public Health
rity. I. Development of the assay procedure. Clin Chem 1986; 32: Service, NIH Publication No. 95-3784, November 1994.
248–254. 81. Asztalos E. Antenatal corticosteroids: A risk factor for the devel-
60. Russell JC. A calibrated fluorescence polarization assay for assess- opment of chronic disease. J Nutr Metab 2012; 2012: 930591.
ment of fetal lung maturity. Clin Chem 1987; 33: 1177–1184. 82. Waffarn F, Davis EP. Effects of antenatal corticosteroids on the
61. Blumenfeld TA, Stark RI, James LS, George JD. Determination of hypothalamic-pituitary-adrenocortical axis of the fetus and new-
fetal lung maturity by fluorescence polarization of amniotic fluid. born: Experimental findings and clinical considerations. Am J
Am J Obstet Gynecol 1978; 130: 782–787. Obstet Gynecol December 2012; 207(6): 446–454.
62. Golde SH, Vogt JF, Gabbe SG, Cabal LA. Evaluation of the 83. Hofmeyr GJ. Antenatal corticosteroids for women at risk of
FELMA in microviscosimeter in predicting fetal lung maturity. preterm birth: RHL commentary. WHO: Geneva, Switzerland,
Obstet Gynecol 1979; 54: 639–642. February 2009.
63. Ashwood ER, Tait JF, Foerder CA. Improved fluorescence polar- 84. NIH Consensus Development Panel. Antenatal corticoste-
ization assay for use in evaluating fetal lung maturity. III. Clin roids revisited August 17–18, 2000. Obstet Gynecol 2001; 98:
Chem 1986; 32: 260. 144–150.
64. Barkau G, Hashiach S, Lanzer D. Determination of fetal lung 85. Committee on Obstetric Practice. COG committee opinion:
maturity from amniotic fluid microviscosity in high risk preg- Antenatal corticosteroid therapy for fetal maturation. Obstet
nancy. Obstet Gynecol 1982; 59: 615–623. Gynecol 2002; 99: 871–873.
36 Monitoring during the later stage
of pregnancy and during labor:
Glycemic considerations
Harold W. de Valk and Gerard H.A. Visser
In the same study, carbohydrate-to-insulin ratios (CIR) were

Introduction studied. CIR signifies the number of grams carbohydrate
Childbirth means a sudden change in glucose and insulin that can be managed by 1 unit of insulin. There was a large
homeostasis for both the mother and the newborn child. and steady decrease in CIR during pregnancy, most notably
An important threat for the neonate in the immediate at breakfast. This fall indicates that prandial insulin require-
postpartum period is neonatal hypoglycemia, related in ments (short-acting insulin) during pregnancy rise consid-
part to recent maternal hyperglycemia and fetal hyperin- erably, most notably at breakfast. This is a well-recognized
sulinism. Irrespective of the ongoing discussion on pre- phenomenon by patients and healthcare professionals.
cise cutoff levels in plasma glucose concentrations and In women with diabetes during pregnancy, endogenous
the potential long-term consequences, current endeavors insulin production is not sufficient to achieve normal glu-
are aimed at preventing neonatal hypoglycemia as well as cose levels. In women with type 1 diabetes, this is obvious as
ensuring rapid detection and treatment. In this chapter, it is in most women with type 2 diabetes. Exogenous insulin
glycemic physiology and pathophysiology during the is used to imitate the normal physiological situation, but at
later stages of normal and diabetic pregnancy are briefly present, we have still considerable distance to cover to reach
described followed by a description of current glucose that goal. It is wise to remind patients (and healthcare pro-
monitoring and insulin treatment management options as fessionals as well) that part of the inability to reach glyce-
well as some organizational issues. mic targets and part of the explanation of adverse events like
severe hypoglycemia are based on this inability to imitate
physiology. It can do a lot to alleviate patient feeling of guilt,
Physiology and pathophysiology fear, and personal inadequacy.
In women with gestational diabetes mellitus (GDM), the
The later stages in pregnancy are associated with an ever- ever-increasing insulin resistance, in some (obese) patients on
increasing degree of maternal insulin resistance, brought top of an already severe degree of insulin resistance, means
about by placental hormones and potentially exacerbated by that maintenance of normoglycemia can only be maintained
adipose tissue–derived interleukins, TNF-α, and other com- by increasing insulin production. When beta cells cannot
pounds. In normal conditions, the consequently increased meet this extra demand, hyperglycemia ensues. The timing
insulin requirement is met by increasing pancreatic insulin of this moment depends on the degree of insulin resistance
production to maintain normal plasma glucose levels. Normal and beta-cell secretory capacity and may vary from very early
maternal plasma glucose levels imply that there are probably in pregnancy (basically indicating a high risk of periconcep-
normal fetal glucose levels without excess stimulation of fetal tional abnormal glucose levels) to very late in the third tri-
insulin production. Generally speaking, insulin requirement mester. With a fixed time for screening for GDM, these late
decreases a little between 6 and 16 weeks of gestation and cases are easily missed (except in the case of fetal growth
sharply increases after week 20. This was exemplified ele- acceleration but that is diagnosis of GDM and not screening
gantly in a study by Mathiesen et al., using basal insulin infu- for GDM) with unrecognized hyperglycemia leading to neo-
sion rates in women with type 1 diabetes on pump therapy.1 natal hypoglycemia in women assumed to be normoglycemic.
299
300 Monitoring during the later stage of pregnancy and during labor
Glucose monitoring during later stages produce a profile that begins to approach informationwise
that of an offline continuous glucose monitor.3 Nevertheless,
of pregnancy and labor part of the patient population can be well regulated during
pregnancy without undue hyperglycemia or hypoglycemia.
Monitoring and treatment are aimed at maintaining mater-
These patients basically do not need offline or online con-
nal plasma glucose levels within a safe range to prevent fetal
tinuous glucose monitoring. In addition, some patients who
growth promotion, to prevent maternal hypo- and hypergly-
could theoretically benefit from more advanced monitoring
cemia, and to prevent neonatal hypoglycemia. Monitoring
find those methods too cumbersome and too distressing,
and treatment are closely interlinked. Monitoring basically
and linked to other negative emotions.
means nothing more than ensuring a steady flow of infor-
mation to the patient and the healthcare professional on
maternal plasma glucose levels combined with information
on food intake, stress, physical activity, as well other glu- Continuous glucose monitoring
cose level–modifying issues and information regarding fetal
The technique of continuous glucose monitoring has been
growth and amniotic fluid volume. These are all essential to
described in other chapters. Broadly speaking, two methods
provide the best opportunity to improve or maintain glyce-
exist: offline continuous glucose monitoring (oCGM) and
mic control. Monitoring is a means and not a goal in itself.
online (real time) continuous glucose monitoring (rtCGM).
The first question is at what time points do we need a
Both techniques measure interstitial and not plasma glucose
measurement? Usual glycemic profiles include fasting glu-
values, but these two generally correspond well. As a conse-
cose levels, preprandial glucose levels, and postprandial
quence of measuring in two different compartments, there
(1.5–2 hours) glucose levels. Fasting levels have not much to
may be a delay in glucose curves in the compartments with
do with the culinary events during the previous evening and
the change in the interstitial compartment following the
night. Fasting levels reflect the endogenous glucose produc-
change of the plasma compartment. This delay is variable but
tion and is used to determine the dose of long-acting insulin
may be as long as 20 minutes. This can pose a problem with
or the basal rate of the insulin pump during the night. Low
rtCGM but not so much with oCGM.
fasting glucose levels mean overdosing of insulin. Low pre-
oCGM was introduced first and provides generally up
prandial glucose levels, especially if they occur at more than
to 4–5 days of glucose values that are registered but not
one meal, can indicate that the insulin dose and plasma insu-
directly available to the patient or the healthcare profes-
lin levels are too high and suppress endogenous glucose pro-
sional. When we consider usefulness in the peripartum
duction, a key feature of the defense against hypoglycemia.
period, it is obvious that just an isolated oCGM can hardly
Prandial insulin doses, either by injection or by insu-
be expected to help. The landmark trial by Murphy et al.
lin pump, are determined by pre- and/or postprandial lev-
showed that oCGM in women with type 1 or type 2 diabe-
els observed in the previous days in combination with the
tes did result in a statistically significant decrease in macro-
size and content of the envisaged meal. The dose of prandial
somia, the primary end point.4 Neonatal hypoglycemia, as
insulin is dictated largely by the amount of carbohydrates,
an index of peripartum management, was similar in both
the content of the meal, and the degree of insulin resis-
groups (oCGM 8%, no oCGM 17%). The studied group was
tance. Insulin resistance is highest during the morning and
rather small (38 patients on oCGM and 33 patients not on
decreases during the day and is lowest in the evening and
oCGM), limiting the power to detect smaller differences. In
during the night, affecting insulin dosing and the CIR. It
conclusion, oCGM can reduce macrosomia with adequate
is clear that it takes clinical expertise and education from
implementation in an individualized multidisciplinary set-
patients, diabetes nurse educators, and endocrinologists to
ting, but there is no specific place for oCGM in peripartum
interpret and tease out glycemic profiles especially in preg-
control and during delivery in patients with type 1 or type
nancy and that the ability to do so is also highly dependent
2 diabetes. A recent Chinese study provided some evidence
on the information available. This information is related not
that oCGM may have an effect in women with GDM.5 In that
only to glucose levels but also to meals and, for example, to
study, patients with GDM were randomized to the interven-
physical activity. More details on insulin therapy during
tion group (oCGM and 7-point glucose profiles) and control
pregnancy are described elsewhere.2 Having established this
group (only glucose profiles). In this intensive study, the
framework, the second question is what the best method of
intervention group received oCGM weekly until satisfactory
glucose monitoring is.
control was reached followed by oCGM every 2–4 weeks.
They also performed glucose profiles similar to that in the
Finger-stick measurements control group. Nutritional counseling was based on glucose
data throughout the trial. oCGM was not employed around
Classic monitoring means glucose measurements in capil- childbirth. Insulin therapy was started significantly more in
lary blood obtained by finger stick. This can be painful and the intervention group compared to the control group (27.9%
is cumbersome as the patient needs to carry the (small) vs. 12.2%, p < 0.001). The incidence in the intervention group
device with her. It only provides a snapshot image of a con- of neonatal hypoglycemia (defined as a plasma glucose
tinuously changing plasma glucose level. Kerssen et al. have <2.2 mmol/L [40 mg/dL]) was significantly lower (5.5% vs.
shown that at least 10 measurements a day are required to 14%, p < 0.02). Interestingly, this difference was present in
Continuous glucose monitoring 301
diet-treated patients but not anymore in insulin-requiring defined as a plasma glucose level <2.5 mmol/L (<45 mg/dL),
patients. In the intervention group, there was also a lower the first measurement within 2 hours, and they employed
incidence of preeclampsia, primary Cesarean section, prophylactic early feeding starting 3 hours after delivery in
preterm delivery, macrosomia, hyperbilirubinemia, and all women. In the control group, glucose was measured every
respiratory distress syndrome. Again, these significant dif- hour during delivery. Such a close monitoring can mean that
ferences were present in diet-treated women, not in insulin- the difference with rtCGM may be quite limited. A sample
requiring women. The beneficial effect may be due to the strict size calculation was not performed, which is logical in a ret-
nutritional setup as well as to the ability to assess the post- rospective analysis. All of the women in the control group
prandial glucose curve in more detail with CGM. The lower were included and 45% of the women in the rtCGM group
incidence of neonatal hypoglycemia with CGM may be in (women had to have had rtCGM before delivery). The inci-
part explained by the lower incidence of m acrosomia. These dence of neonatal hypoglycemia was comparable in both
results are reason for further studies in GDM and, if any- groups (rtCGM 37% on rtCGM and 46% in the control
thing, underscore once the great importance of ongoing group); the incidence of severe hypoglycemia (plasma glu-
dietary counseling. At present, oGDM has no place in the cose <2.5 mmol/L [45 mg/dL]) and intravenous glucose was
peripartum management in GDM but may be a promising similar as well (11% vs. 17%). Analysis within the group in
tool in the correct setting to reduce neonatal hypoglycemia rtCGM showed that neonatal hypoglycemia was associated
by reducing macrosomia. with a higher mean maternal plasma glucose during the
Turning to rtCGM, this technique with direct, instanta- 8 hours up to delivery (either self-monitored or analyzed
neous display of interstitial glucose levels would theoretically from rtCGM) and that the infant birth weight z-scores were
offer a method to optimize glucose control in a specified inversely related to the 2-hour neonatal plasma glucose level.
period of time. No studies have been performed with isolated There was no evidence of a threshold in both associations.
use of rtCGM in the peripartum period. A major trial with This retrospective analysis of some of the data of a pro-
rtCGM during pregnancy was performed by Secher et al. In spective trial highlights some important issues on research
that trial, pregnant patients with type 1 diabetes were ran- and practical management of rtCGM in the peripartum
domized between intermittent use of rtCGM and no rtCGM arena: the background against which rtCGM is tested (fre-
(conventional care).6 No difference in neonatal hypoglyce- quency of self-measurement), self-selection for rtCGM (even
mia (defined as a plasma glucose <2.5 mmol/L [45 mg/dL]) in randomized trials, compliance is not a random phe-
was observed. This audacious trial has highlighted some of nomenon), management of rtCGM, insulin administration
the issues that explain why randomized as well as observa- technique (CSII vs. MDI), practical nutritional approach to
tional trials are so difficult to do in this patient group. The the neonate during the first hours postpartum, definition
trial showed that there was no difference in the primary of neonatal hypoglycemia, group size, and expected differ-
outcome macrosomia. However, this lack of effect may well ences between groups. This illustrates that only a carefully
be explained the intermittent use (use of at least 75% of the designed large trial may give more definite answers and that,
time is mandatory to see the effects of rtCGM) as shown in for now, common sense should prevail. Hopefully, the con-
the meta-analysis by Pickup et al. using the individual datas- tinuous glucose monitoring in women with type 1 diabetes
ets of clinical trials.7 In addition, late start during pregnancy in pregnancy trial (CONCEPTT) trial and the neonatal spin-
(patients require up to 6 weeks to get used to the device off trial may shed further light on the best peripartum and
and practical implementation) as well as the glucose targets neonatal policy.
directing insulin therapy (prevention of hypoglycemia vs. RtCGM could theoretically be used during delivery as an
optimal prevention of hyperglycemia) may have contributed add-on, managed by the medical team. This would obviate a
to the negative result and indicate avenues to improve the large role for the delivering woman and potentially help to
effects of rtCGM. Trials regarding glucose monitoring are achieve good maternal glucose levels. On the other hand, the
necessarily not blind (assessors may be blinded), and being risk of maternal hypoglycemia may increase. No data exist
in a trial usually has a beneficial effect in itself in the control on such a procedure (Figure 36.1).
group. This latest issue means that this placebo effect, which Logically, preventive efforts can best be directed at those
may be substantial, should be adequately accommodated in individuals at highest risk. The value of this approach has
group size calculations to prevent disappointment. An ear- been elegantly demonstrated by Pickup who demonstrated
lier small study from Macedonia randomized 25 patients that summarizing meta-analyses may be negative (in this
with type 1 diabetes on insulin pump to intermittent rtCGM case insulin pump therapy and incidence of severe hypo-
or usual care.8 No neonatal hypoglycemia (defined as a glycemia), whereas analyzing only those patients with prior
plasma neonatal glucose of 1.66 mmol/L [30 mg/dL] or less) severe hypoglycemia shows a dramatic positive effect.10
was observed. Neonatal hypoglycemia was not the primary The difficulty in neonatal hypoglycemia is to define
end point and group size was too small to detect relevant dif- those at highest risk. Preterm delivery is a well-known
ferences when rtCGM would have had a great impact. one. At the end of pregnancy, we could add a large fetus
The Secher trial may give us some insight in the effect (for example, with an abdominal circumference exceeding
on neonatal hypoglycemia. Cordua et al. have looked ret- the 90th percentile) and suboptimal glycemic control with
rospectively at the effect on neonatal hypoglycemia in this periods of definite hyperglycemia. A recent study by the
prospective trial in more detail.9 Neonatal hypoglycemia was Danish group may serve as a nice example, dealing with
8 intake, the actual plasma glucose level, CIR, insulin action

time, insulin sensitivity factor, and the target plasma glucose
Neonatal blood glucose (mM)
6
level. Use of this bolus calculator has been shown to be able
to achieve an improvement in glycemic control in patients
with type 1 diabetes.12,13 Logically, the bolus calculator could
4 be useful in pregnancy, but data on usefulness do not exist
yet. Theoretically, the ever-changing circumstances during
2 pregnancy, in contrast to the nonpregnant state, will mean
that the calculator has to be adapted very frequently. This
was shown in the aforementioned study by Mathiesen et al.
0
2 4 6 8 10 12 14
where prandial CIRs during pregnancy in pump-using type
Mean maternal blood glucose (mM)
1 diabetes patients fell considerably, most dramatically at
breakfast. These data are in line with another study and with
Figure 36.1 The association between maternal glucose levels common experiences in the change in insulin requirements
and neonatal plasma glucose levels: maternal and fetal plasma during pregnancy.14 They underscore once more the impor-
glucose levels measured in capillary blood. (Reproduced tance of ongoing dietary counseling throughout pregnancy
from Taylor, R. et al., Obstet. Gynecol., 99, 537, 2002. With
permission.)
as well as that of the care for women with diabetes during
pregnancy by a multidisciplinary team. No information on
the bolus calculator exists in the later stage of pregnancy
severe hypoglycemia during pregnancy.11 This nonran- and childbirth, but isolated use in that period does not seem
domized trial included women during the first trimester of to be a very good and practical idea. If the bolus calcula-
pregnancy who had had at least one episode of severe hypo- tor is used, this should begin much earlier, preferably before
glycemia the previous year. Of the 28 patients, 12 (43%) conception.
chose rtCGM, and 16 chose not to (mainly for logistical rea-
sons). The study period was the first half of pregnancy, the
period of highest risk. RtCGM was used 92% of the time (in Glycemic target levels at the end of
sharp contrast to the compliance in the earlier trial in preg-
nancy) and was associated with a dramatic and statistically pregnancy
significant decrease in severe hypoglycemia. This study
Part of the delivery plan is the definition of an acceptable
illustrates three important issues. First, it illustrates the
range of plasma glucose levels. An early study by Taylor
high compliance of patients when they have a tangible goal.
showed an inverse association between maternal glucose
Second, it illustrates the importance of selecting patients
levels in the time leading up to the actual delivery and the
at highest risk to demonstrate a positive effect on that spe-
neonatal plasma glucose level.15 Neonatal hypoglycemia was
cific end point. Otherwise, patients at risk are drowned in
relatively high in case maternal glucose values were above
a population at much lower risk, potentially diluting the
7–8 mmol/L (126–144 mg/dL). Using oCGM, Stenninger
positive effect into nonsignificance. Third, it illustrates the
et al. have reported that higher maternal glucose levels are
power of observational studies in glucose monitoring. This
associated with a higher risk for the need of intravenous
line of thinking opens new avenues of carefully controlled
glucose in the neonate.16 The protocol in that study was
prospective observational real-life studies as an added
already rather strict: target maternal glucose levels maxi-
research method in the scientific toolbox. Nationwide reg-
mally 7 mmol/L (126 mg/dL) with subcutaneous insulin for
istries with good study design and adequate governance of
correction of higher levels. The incidence of neonatal hypo-
the dataset and of the analysis may become powerful tools.
glycemia (plasma glucose <2.2 mmol/L [40 mg/dL]) was 10
The recently founded Dutch rtCGM registry (TRACING),
of the 15 analyzable subjects (60%) with an incidence of 5
including pregnant and preconceptional women with type
subjects (33%) of severe hypoglycemia (neonatal glucose
1 or type 2 diabetes on rtCGM, may provide real-life data
<2.2 mmol/L [40 mg/dL] and IV glucose). This led to the
on rtCGM use, results, and adverse events. Returning to
suggestion that perhaps we should aim at even lower mater-
the issue of neonatal hypoglycemia, one might consider
nal glucose levels. The retrospective analysis of peripartum
intervention in those women and neonates at highest risk,
glycemic control in a prospective rtCGM study also showed
i.e., preterm birth, large fetus, and/or suboptimal glycemic
such an association.9 There are some data that the incidence
control.
of neonatal hypoglycemia may be less when targeting mater-
nal glucose to <5.6 mmol/L (100 mg/dL).17 The consensus
however is generally that a range in maternal glucose levels
Bolus calculator between 4 and 7 mmol/L (72 and 126 mg/dL) is acceptable.
The absence of a threshold value leaves room for maneuver-
In the nonpregnant state, studies have indicated that the ing. Common sense would dictate that one can consider a
amount of prandial insulin can be calculated by an algorithm slightly higher lower level (4.5 or 5 mmol/L, 81 or 90 mg/dL)
taking into account the envisaged amount of carbohydrate in women prone to hypoglycemia.
Summary 303
Peripartum insulin management with breastfeeding should be discussed as well as the pos-
sibility of comorbidity or change in comorbidities (diabetic
There is no definite answer to the question of what the best complications like retinopathy but also thyroid disease or for
insulin treatment strategy is in preconceptional diabetes. example, preexistent rheumatoid arthritis).
Employed methods include repetitive doses of subcutane- Monitoring of maternal glucose levels should also be dis-
ous rapid-acting insulin above the maternal plasma glu- cussed including target levels and monitoring frequency.
cose, continuous intravenous insulin started above a preset Frequent measurement of maternal glucose levels should be
maternal plasma glucose level, and standard infusion of discussed. The frequency depends on the stage of delivery
glucose combined with varying intravenous insulin infu- and previous glucose levels. Some women (and partners)
sion speed to achieve and maintain target maternal plasma want to manage the glucose levels, but once the actual deliv-
glucose levels.16,18,19 The study by Lepercq et al. had a low ery has set in, experience has taught us that all of them gladly
frequency of neonatal hypoglycemia (13%). This low fre- leave it to the nursing staff. Different schemes can be used,
quency may have been due to the strict maternal glucose but once every 1–2 hours is advisable (except at the beginning
control achieved with a combination of 10% dextrose and of an induction of labor when a more liberal time scheme can
standard intravenous insulin to all women (although pre- be used). Such timing issues do not exist with rtCGM, but it
set target levels were in the usual range, 3.4–7.8 mmol/L is advisable to register glucose levels at specified regular time
[61–140 mg/dL]), to a more strict definition of neonatal points.
hypoglycemia but also to a rather aggressive approach to Neonatal management during the first days after birth is
early neonatal management. Neonates received intramus- described in another chapter. Regarding the mother, there is
cular glucagon at birth as well as oral suspension enriched no real need for strict glycemic control. On the contrary, as
with glucose and triglycerides during the first 3 hours after breastfeeding is associated with an increased risk of mater-
birth. Although apparently effective, it is questionable nal hypoglycemia, a more liberal approach to target glucose
whether this scheme will find a great welcome in pediatric levels is advisable. And with recurrent maternal hypogly-
circles. This study demonstrates that various avenues can be cemia, hypoglycemia unawareness may ensue and compli-
taken to manage peripartum glucose and demonstrates that cate the situation. A practical approach would be to advise
studies cannot easily be compared because of differences in plasma glucose levels between 6 and 12 mmol/L (108 and 216
maternal glucose target levels, definition of hypoglycemia, mg/dL) for a period of about 3 months.
and postnatal management.
Summary
Organizational issues
In this chapter, we have described some issues regarding
A well-organized structure of care and counseling is of physiology and pathophysiology in the diabetic pregnancy
prime importance when caring for these women (and their relevant to the period of childbirth. Maternal plasma glu-
partners and families). Good and well-based information cose levels in the period leading up to delivery and the levels
is a first requisite, and generally speaking, information on during delivery are inversely related to the degree of neo-
the latter stages of pregnancy, childbirth, and thereafter is natal hypoglycemia. Notwithstanding all potential benefits
best already discussed to some extent in the earlier stages of of advanced technology, adequate and ongoing nutritional
pregnancy. counseling of the pregnant woman is of paramount impor-
Many women find it difficult to combine a professional tance. Repetitive use of rtCGM is associated with a decreased
career with having type 1 diabetes and being pregnant. incidence of macrosomia but has no effect on the incidence of
Many of them have decreased working time or stop work- neonatal hypoglycemia in women with type 1 or type 2 dia-
ing before formal pregnancy leave starts. In the Netherlands, betes. There is evidence emerging that a positive effect may
for example, pregnancy leaves formally starts at 34 weeks of be seen in diet-treated gestational diabetes. The value and
gestation, but many women already start descending into possibilities offline CGM are still under debate. Audacious as
leave from 26 weeks onward. It is therefore wise to discuss these trials may be, there are many methodological issues to
this with the employer to prevent unpleasant surprises for all surmount. We hope that the current large-scale CONCEPTT
concerned. Likewise, maternity leave for women with diabe- trial (randomized trial with rtCGM started preconception-
tes may be too short for some women with type 1 diabetes, ally or early during pregnancy) will provide us with some
and this should be anticipated as well. answers. There is no single best insulin regimen during
On a more technical level, a multidisciplinary deliv- delivery, the choice having to be made locally. Safe target lev-
ery plan should be made at the most opportune moment els for maternal glucose levels are between 4 and 7 mmol/L
sometime after 30 weeks of gestation. A general plan can be (72–126 mg/dL) with maternal glucose levels being inversely
made for the different modes of delivery that can be indi- related to neonatal glucose levels without a threshold level.
vidualized per patient. It is useful and advisable to provide The incidence of neonatal hypoglycemia depends not only
the pregnant woman and the partner with a copy of the on the definition of hypoglycemia but also on the specific
plan. The reevaluation of comedication after delivery and protocol for neonatal care directly after birth.
REFERENCES
1. Mathiesen JM, Secher A, Rinholm L et al. Changes in basal rates 10. Pickup JC. The evidence base for diabetes technology:
and bolus calculator settings in insulin pumps during pregnancy Appropriate and inappropriate meta-analysis. J Diabetes Sci
in women with type 1 diabetes. J Matern Fetal Neonatal Med Technol 2013; 7: 1567–1574.
2014; 27: 724–728. 11. Secher A, Stage E, Ringholm L, Barferd C, Damm P, Mathiesen
2. De Valk HW, Visser GHA. Insulin therapy during pregnancy, ER. Real-time continuous glucose monitoring as a tool to prevent
labour and delivery. Best Pract Res Clin Obstet Gynaecol 2011; severe hypoglycaemia in selected pregnant women with type 1
25: 65–76. diabetes—An observational study. Diabet Med 2014; 31: 352–356.
3. Kerssen A, de Valk HW, Visser GHA. Do HbA1c levels and 12. Cukierman-Yaffe T, Konvalina N, Cohen O. Key elements for suc-
the self monitoring of glucose levels adequately reflect glycae- cessful intensive insulin pump therapy in individuals with type 1
mic control during pregnancy in women with type 1 diabetes? diabetes. Diabetes Res Clin Pract 2011; 92: 69–73.
Diabetologia 2006; 49: 25–28. 13. Schmidt S, Meldgaard M, Serifovski N, Storm C, Christensen TM,
4. Murphy HR, Rayman G, Lewis K et al. Effectiveness of continu- Gade-Rasmussen B, Nørgaard K. Use of an automated bolus
ous glucose monitoring in pregnant women with diabetes: A ran- calculator in MDI-treated type 1 diabetes: The BolusCal Study,
domised clinical trial. BMJ 2008; 337: a1680. a randomized controlled pilot study. Diabetes Care 2012; 35:
5. Yu F, Lv L, Liang Z et al. Continuous glucose monitoring effects on 984–990.
maternal; glycemic control and pregnancy outcomes in patients 14. Roeder HA, Moore TR, Ramos GA. Insulin pump dosing across
with gestational diabetes mellitus: A prospective cohort study. gestation in women with well-controlled type 1 diabetes mel-
J Clin Endocrinol Metab 2014; 99: 4674–4682. litus. Am J Obstet Gynecol 2012; 207: 324.e1–324.e5.
6. Secher A, Ringholm L, Andersen HU, Damm P, Mathiesen ER. 15. Taylor R, Lee C, Kyne-Grzebalski D, Marshall SM, Davidson JM.
The effect of real-time glucose monitoring in pregnant women Clinical outcomes of pregnancy in women with type 1 diabetes.
with diabetes: A randomized controlled-trial. Diabetes Care Obstet Gynecol 2002; 99: 537–541.
2013; 36: 1877–1883. 16. Stenninger E, Lindqvist A, Åman J, Östlund I, Schvrcz E.
7. Pickup JC, Freeman SC, Sutton AJ. Glycaemic control in type 1 Continuous subcutaneous glucose monitoring system in diabetic
diabetes during real time continuous glucose monitoring com- mothers during labour and postnatal glucose adaptation of their
pared with self monitoring of blood glucose: Meta-analysis of infants. Diabet Med 2008; 25: 450–454.
randomised controlled trials using individual patient data. BMJ 17. Curet LB, Izqueierdo LA, Gibson GJ, Schneider JM, Perelman R,
2011; 343: d3850. Converse J. Relative effects of antepartum and intrapartum mater-
8. Petrovski G, Dinitrovski C, Bogoev M, Milenkovic T, Ahmeti I, nal blood glucose levels on the incidence of neonatal hypoglyce-
Bitovska I. Is there a difference in pregnancy and glycemic out- mia. J Perinatol 1997; 17: 113–115.
come in patients with type 1 diabetes on insulin pump with con- 18. Kline GA, Edwards A. Antepartum and intra-partum insulin man-
stant or intermittent glucose monitoring? A pilot study. Diabetes agement of type 1 and type 2 diabetic women: Impact on clini-
Technol Ther 2011; 13: 1109–1113. cally significant neonatal hypoglycemia. Diabetes Res Clin Pract
9. Cordua S, Secher A, Ringholm L, Damm P, Mathiesen ER. Real- 2007; 77: 223–230.
time continuous glucose monitoring during labour and delivery 19. Lepercq J, Abbou H, Agostini C et al. A standardized protocol to
in women with type 1 diabetes—Observations form a random- achieve normoglycaemia during labour and delivery in women
ized trial. Diabet Med 2013; 30: 1374–1381. with type 1 diabetes. Diabetes Metab 2008; 34: 33–37.
37 Timing and mode of delivery
Salvatore Alberico and Gianpaolo Maso
Gestational diabetes is defined by the World Health Organi motivated by the evidence that diabetic mothers with infants
zation (WHO) as carbohydrate intolerance resulting in hyper- having a weight exceeding this limit exhibit a significant
glycemia of variable severity with onset or first recognition increase in dystocia upon birth.
during pregnancy.1 The implementation of systematic screen- GDM and even pregestational diabetes result in higher
ing as suggested by the evidence that emerged from the HAPO frequency of macrosomic newborns.
study2 and the conclusion by the International Association of Four studies also allowed assessment of the association
Diabetes and Pregnancy Study Groups (IADPSG), who spon- between untreated GDM according to the WHO criteria and
sored the International Workshop Conference on Gestational macrosomia.11–14
Diabetes in 2008, indicate that this is a frequent disease.3 The pooled relative risk (RR) was 1.81 (95% CI 1.47–2.22;
In effect, the simulation assumed a gestational diabetes p < 0.001), with very homogenous results across studies
mellitus (GDM) prevalence of 10% according to the WHO (I2 = 0%). No study was available to examine this association
criteria and a 50% higher prevalence (i.e., a 15% prevalence of using the IADPSG diagnostic criteria, but analysis using the
GDM) for the IADPSG criteria.4 evidence-based data group database showed an RR of 1.38
Before we go into the description of the obstetric o utcome (95% CI 1.14–1.68; p = 0.001). When using large for gesta-
relating to pregnancies complicated by GDM, we must tional age as the outcome, the magnitude of the association
remember that the data available from the literature quite for the WHO criteria was slightly lower (RR = 1.53, 95%
often do not distinguish outcomes resulting from pregnan- CI 1.39–1.69; p < 0.001; I2 = 0%). Applying the criteria of
cies complicated by GDM from those with pre-gestational IADPSG, derived from the three studies,15–17 we obtained a
diabetes type 1 or 2. higher RR, but with heterogeneous results (RR = 1.73, 95%
According to the Pedersen hypothesis,5 maternal hyper- 1.28–2.35; p < 0.001, I2 = 93%).
glycemia results in excess transfer of glucose to the fetus
resulting in fetal hyperinsulinemia. The effects of fetal
hyperinsulinemia include an overgrowth of insulin-sensitive Anthropometric characteristics of
tissues such as adipose tissues, especially around the chest,
shoulders, and abdomen, which increases the risk of mac- diabetic mothers with macrosomic
rosomic newborns and of shoulder dystocia, brachial plexus infants
injury, perinatal death, and birth trauma, as well as the need
for cesarean section.6–8 Studies of macrosomic infants of diabetic mothers revealed
Therefore, the prevention of excessive fetal growth during a greater amount of total body fat, thicker upper-extrem-
gestation, together with the proper management of maternal ity skin fold measurements, and smaller ratios of head to
glucose metabolism in the anticipation of the time of child- abdominal circumference than macrosomic infants of non-
birth, if possible, to lower the risk of foetal macrosomia, can diabetic mothers. This disproportion between abdominal
be considered as targets to be pursued by obstetricians.10 and cephalic diameter favors complications of childbirth,
which are obviously accentuated by the greater weight of
these newborns.18
Rate of macrosomia in diabetic The reasons are related to the fact that a normal or a pro-
pregnancy portionally lower cephalic circumference than the fetal body
allows a smooth progression of the fetus in the birth canal,
Fetal macrosomia is varyingly defined as either an absolute with normal disengagement of the head. The largest biacro-
birth weight greater than 4000 g, 4500 g, or 5000 g or a cus- mial diameter involves difficulty in making an internal
tomized birth weight centile greater than the 90th, 95th, or rotation, with the arrest of the anterior shoulder against the
97th percentile for the infant’s gestational age. inner face of the pubic bone, not allowing its fixation under
In our assessment, macrosomia will be considered in an the margin of the same bone, and then difficulty in making a
infant with a birth weight greater than 4000 g. This choice is normal disengagement.
305
The maneuvers required to reach a fetal extraction can The occurrence of shoulder dystocia involves risks not
then lead to an excessive traction on the nerves of the bra- only for the newborn, which can have a permanent damage
chial plexus leading to the paralysis of the same plexus. of the brachial plexus 25 or a greater incidence of neurologi-
cal disorders, resulting in permanent neurological dysfunc-
Macrosomia: Risk factors tion,26 but also for the mother. Maternal morbidity as a
result of infection is increased, particularly the incidence of
Macrosomia, even in the absence of a disease in a diabetic postpartum hemorrhage (11%) as well as third- and fourth-
mother, involves in each case a generic risk in the obstetrical degree perineal tears (3.8%).27
outcome. We can therefore say that shoulder dystocia is the most
A study investigating the effects of birth weight on fetal dangerous complication that the obstetrician is called to
mortality shows that higher fetal mortality rates are associated prevent, and the question that we must ask ourselves now is
with a birth weight greater than 4250 g in nondiabetic mothers what the risk factors of shoulder dystocia are.
and a birth weight greater than 4000 g in diabetic mothers.19 As we have said previously, there is a relationship
An interesting piece of information on the outcome of between fetal size and shoulder dystocia, 28 but it is not a
macrosomic newborns is derived from the Medical Birth good predictor, because the large majority of infants with a
Registry of Norway on 372,128 births, registered for the period birth weight of ≥4500 g do not develop shoulder dystocia 29
1999–2005. Especially in newborns weighing more than and equally important because 48% of births complicated
4000 g, the frequency of deliveries carried out by cesarean by shoulder dystocia occur in infants who weigh less than
section, particularly in emergency cesarean section, increases. 4000 g.7
According to the same registry, fetal macrosomia leads In a population of 2,014,956 deliveries, Overland et al.
to an increase in operative deliveries with a vacuum extrac- showed the correlation between neonatal weights and inci-
tor (OR = 1.7) or low forceps (OR = 1.4) and a risk of shoul- dence of shoulder dystocia: The adjusted OR (AdOR) for
der dystocia (OR = 6.8), clavicular fracture (OR = 3.4), and a weight of less than 3500 g is 1, AdOR = 3.92 for weight
plexus injury (OR = 5.6).20 between 3501 and 3999, AdOR = 17.10 for weight between
4499 and 4000 g, AdOR = 60.33 for weight between 4500
Fetal growth per week of pregnancy and 4999, and AdOR = 176.52 for weight greater than
5000 g.29 In the same study, the condition of maternal dia-
One of the conditions that complicates child birth is cer- betes resulted in a crude OR = 5.69 for a shoulder dystocia at
tainly macrosomia, which progresses over the course of ges- delivery. Numerous other studies available in the literature
tation. Anticipating the birth in a few weeks, after reaching have shown that the risk for shoulder dystocia has increased
the age defined as the end of gestation (≥37 weeks), one can still to 3.6- to 4.5-fold in infants of diabetic women when the
therefore avoid the weekly weight gain, which in fetuses of actual birth weight is ≥4000 g.28–33
diabetic mothers is obviously superior to a fetus with a nor- The debate on this issue is so open, and there are many
mal progression of growth. And in fact, it is estimated that tips that relate to the accomplishment of the delivery in
from the 37th to the 42nd week of pregnancy, weight gain pregnant women with GDM and suspected fetal macroso-
may be around 625 g ± 201 (37.73 ± 0.443 lb). mia. Some authors recommend elective cesarean at an esti-
In the case of acceleration of fetal growth, comparable to a mated fetal weight of 4000 g in diabetic patients;34,35 others
line coinciding with the 90th percentile, fetal weight exceeds suggest that a weight threshold of 4500 g should be used
4000 g at the end of the 39th week of gestation. to reduce unnecessary intervention because of ultrasono-
It is so easy to imagine that in anticipation of the birth at graphic error.36
38 weeks, with a safe fetal maturity, it can be easier to give The use of the general population estimation of fetal weight
birth to a fetus weighing less than 500 g than compared to as an indicator to perform an elective cesarean section is not
those who weigh more. cost-effective. The limits of predictivity of the weight using
the ultrasound method are still significant for macrosomic
Shoulder dystocia: Correlation with fetuses.
It has been calculated that by placing a threshold for fetal
fetal growth, GDM, and limit for weight >4500 g, obstetricians would have to perform elective
vaginal delivery 3700 cesarean sections to prevent one permanent brachial
plexus injury.
Shoulder dystocia is one of the major complications that can Thus, elective cesarean section for suspected macrosomia
occur during childbirth assistance of a diabetic mother’s alone is difficult to support.37
newborn. Shoulder dystocia is defined as a vaginal cephalic Different discourse appears to diabetic mothers, where
delivery that requires additional obstetric maneuvers to if the recommendation was an elective cesarean section
deliver the fetus after the head has been delivered and gentle for estimated fetal weight greater than 4500 g, it would
traction has failed.21,22 take 443 cesarean sections to prevent 1 brachial plexus
There is a wide variation in the reported incidence, and injury. 38
studies involving the largest number of vaginal deliveries Langer et al. instead suggested in women with GDM for
indicate a prevalence between 0.19% and 3.3%.23,24 elective cesarean section an estimated weight limit greater
Labor and delivery in GDM 307
than ≥4250 g during delivery. In this way, a 76% reduction Multivariate analysis found that being treated for gesta-
of shoulder dystocia was verified in their population, with tional diabetes was the most significant factor in determin-
an increase in the rate of cesarean sections by only 0.26%. ing a cesarean section (OR 2.1, 95% CI 1.3–3.6); untreated
A weight threshold of 4500 g yielded a smaller increase in gestational diabetes was not a significant risk factor (OR 1.6,
the cesarean section rate but failed to prevent approximately 95% CI 0.9–2.7) (EL = 2+).
40% of shoulder dystocia cases in diabetic patients.31 It makes you wonder why the cesarean section rate is
In any case, we can conclude that the class of birth weight so high in these pregnant women. Part of the justification
greater than 4000 g, gestational age greater than 38 weeks, is related to psychological conditioning that the obstetri-
and maternal diabetes correlate directly with an increased cian undergoes in the management of these pregnancies,
risk of shoulder dystocia at birth. when the fear of having to compare with the assistance of a
This is the reason why many groups of experts in the macrosomic child at birth associates with the forced imple-
management of childbirth in diabetic pregnancies establish mentation of difficult fetal extraction maneuvers in case of
a weight limit for vaginal birth in these women when the shoulder dystocia. We cannot ignore also the fear of having
estimated fetal weight at 38 weeks indicates a risk of exceed- to then undergo medicolegal disputes in the event of neona-
ing 4000 g. This recommendation takes into account the tal or maternal adverse outcomes. Another hypothesis is that
limitations of ultrasound evaluation methods to identify the many of these cesarean sections are due to failed induction
correct fetal weight at term gestation. Therefore, the same or labor arrest following prolonged attempts at induction.45
recommendation is less than the limit set in the ACOG bul- A systematic review on this topic was published by
letin in 2005 (which established a limit of no more than Witkop in 2009.46
4500 g birth weight) and NICE, 2008, who indicated that This study has selected 11,400 published papers on the
limit to 4250 g. aforementioned subject. After a review of their structure,
With regard, however, to the two societies mentioned ear- only five articles met the inclusion criteria for the key
lier, refer to the birth weight and not the fetal weight estima- questions related to delivery management, and of these,
tion assessed at 38 weeks.39,40 only one was a randomized controlled trial (RCT) that
compared the effect of two labor-induction protocols on
maternal and perinatal outcomes on a population of 200
Dystocia: Correlation with operative pregnancies.
Witkop says the following in his work:
vaginal delivery and GDM
The five studies were heterogeneous in their meth-
The same study by Øverland et al., cited earlier, clearly
odology, comparison groups, time period in which
indicates that an additional risk factor for shoulder dysto-
the study was conducted, length of the study period,
cia is represented by an operative vaginal delivery (forceps,
populations included, and outcome measures. Because
AdOR = 1.14, and vacuum extractor, AdOR = 2.86). This risk
of the heterogeneity, we were unable to conduct a
is even more pronounced in the case of macrosomic fetuses
meta-analysis
from diabetic mothers.32–41
In conclusion, we therefore could say that in pregnant and concludes
women with GDM, it is prudent to avoid a prolongation of
gestation, an excessive fetal weight at birth, and an operative it is difficult to fully assess the outcomes observed in
vaginal birth in order to reduce the risk of shoulder dystocia. this study on the basis of only one clinical trial of 200
participants.
Labor and delivery in GDM The study in question is that of Kjos, an RCT (n = 200) from
the United States involving women with insulin-requiring
The published studies on the mode of delivery in GDM indi- diabetes comparing the outcomes of active induced labor
cate a common denominator: an increase in the frequency of (accurate measurement of gestational development and
cesarean sections. In the same studies, this rate is very high, induction of labor with intravenous oxytocin, n = 100)
ranging in some series from 45% to 80%.42,43 with expectant management (close monitoring and insulin
A prospective case–control study (n = 3778) from Canada treatment, n = 100). Those enrolled had gestational diabe-
examined the relationship between cesarean section rates tes (n = 187) or preexisting diabetes (n = 13). There were
and gestational glucose intolerance (3-hour 100 g oral glu- no differences between the groups at baseline. In the active
cose challenge test).44 induction group, 70 women had induction of labor, 8 had
The study identified four groups: negative gestational dia- cesarean section, and 22 had spontaneous labor. In the
betes (n = 2940), false-positive gestational diabetes (n = 580), expectant management group, 49 women had induction, 7
untreated borderline gestational diabetes (n = 115), and had cesarean section, and 44 had spontaneous labor.47 This
known treated gestational diabetes (n = 143). Women with study has been shown that even after adjustment for gesta-
gestational diabetes had higher rates of macrosomia (28.7% tional age at delivery, maternal weight, and maternal age, the
vs. 13.7%, p < 0.001) and cesarean section (29.6% vs. 20.2%, average neonatal birth weight in the expectant management
p = 0.02). group was significantly greater than the mean neonatal birth
weight in the active induction group (p < 0.01). The findings section in diabetic pregnancies and the opportunity to lead or
of this RCT suggest that neonates born to women undergo- not an induction of labor.
ing induction at 38 weeks have a significantly lower risk of
macrosomia than those born to women treated with expect-
American Society of Obstetricians and Gynecologists
ant management. The absence of any difference in cesar-
ean delivery rates suggests that maternal morbidity among 1. A cesarean delivery may be considered if the estimated
women undergoing induction is similar to that of women fetal weight is greater than 4500 g in women with diabetes.
undergoing expectant management. 2. Expectant management beyond the estimated date gen-
In a review of randomized trials of elective delivery in erally is not recommended. Induction of labor in preg-
women with diabetes from the Cochrane database, includ- nancies with a fetus with suspected macrosomia has not
ing those with mild gestational diabetes, there was little evi- been found to reduce birth trauma and may increase the
dence to support elective labor induction.48,49 cesarean delivery rate.39
This last author forms part of the problems of management
(induction vs. expectant), an important clinical factor that Colorado clinical guidelines
might suggest whether or not the opportunity to induce labor ●● Timing of delivery remains relatively open. There are no
in women with GDM: the degree of ripening of the uterine data to support delivery prior to term or cesarean delivery
cervix. In fact, he says: “Objectively evaluating the risks and purely on the basis of GDM.
benefits of labor induction is potentially confounded by the ●● Well-controlled GDM pregnancies on medical nutrition
status of the cervix at the time of initiation of induction, early therapy have little indication for delivery prior to 38–39
determination of an arrest disorder and physician bias toward weeks gestation.
cesarean delivery for women who have diabetes.” ●● Delivery at ~39 weeks gestation has been shown to
The analysis of the studies mentioned seems to agree on decrease macrosomia in women with good dating criteria
the benefit that induction of labor at 38 weeks in case of GDM and a favorable cervix.
involves a lower average neonatal weight in the induced com- ●● To determine the mode of delivery when the estimated
pared to those with expectant management and a lower fre- fetal weight is 4000–4500 g, consider past delivery history
quency of macrosomic newborns. In this debate, it is a rather and clinical pelvimetry.51
recurrent fear that induction of labor, on the other hand, may
lead to an increase in the frequency of cesarean sections.35
In reference to this awe, Blackwell’s study seems to con- Collège National des Gynècologues et Obstètriciens
tradict that induction of labor is to determine this effect. In Françai
fact, in their experience after controlling confounding fac- 1. Elective cesarean section in the case in which the esti-
tors, labor induction was not associated with the delivery mated ultrasound fetal weight indicates a value greater
mode. In fact, 55.6% of gestational and 72.9% of pregesta- than >4250 to 4500 g.
tional diabetics who achieved vaginal delivery underwent 2. It is possible to propose an induction of labor at term
induction of labor. We agree with the conclusions of these (>39 weeks) taking into account a balance of benefit to
authors, which argue that nonmedical factors may play an the mother and the fetus.52
important role in the decision for cesarean delivery in dia-
betic pregnancies because they observed in their popula-
Austrian guidelines
tion the lack of effect of factors traditionally associated with
cesarean delivery: labor induction, bishop score, epidural 1. Timing of delivery was not determined in the Austrian
anesthesia, birth weight, and macrosomia.42 guidelines for gestational diabetes management.
Another support for this opinion came from the study of 2. Continue pregnancy until term when uncomplicated;
Levy, who, in a retrospective cohort study using routinely aim for vaginal delivery if possible.53
collected data (n = 108,487) undertaken in the United States,
examined whether induction of labor increased cesarean Australian guidelines
section rates in women with diabetes. The cesarean section 1. Induction of labor or operative delivery based on obstet-
rate was lower in women who had induction of labor than ric and/or fetal indications; otherwise deliver at term.54
those who did not have induction of labor (OR 0.77, 95% CI
0.50–0.89) (EL = 2+).50
Royal College of Obstetricians and Gynecologists
1. Induction of labor at term can reduce the incidence of
International societies: What shoulder dystocia in women with gestational diabetes
recommendations do they suggest? (Grade B).
2. A systematic review and meta-analysis of RCTs of the
Very briefly, we indicate the recommendations that the most effect of treatment in women with gestational diabe-
important societies of obstetrics and gynecology suggest in tes38 concluded that the incidence of shoulder dystocia is
reference to the weight limit for carrying an elective cesarean reduced with early induction of labor.
Special clinical conditions 309
3. The NICE diabetes guideline recommends that pregnant

Table 37.1 Patients in study
women with diabetes who have a normally grown fetus
should be offered elective birth through induction of
Induction Expectant p
labor, or by elective cesarean section if indicated, after 38
completed weeks. Number of 186 185
a. The NICE induction of labor guideline recommended recruited
that women with pregnancies complicated by diabetes pregnancies:
371
should be offered induction of labor before their esti-
Neonatal average 3200 g 3420 p < 0.001
mated date of delivery.40 weights
4 . Elective cesarean section should be considered to
Cesarean section 12.9% 11.4% n.s., p = 0.6
reduce the potential morbidity for pregnancies compli- (11.8%)
cated by preexisting or gestational diabetes, regardless CS + operative 19.4% 21.6% n.s., p = 0.6
of treatment, with an estimated fetal weight greater vaginal delivery
than 4500 g. 55
incidence of C-section in comparison to expectant manage-

Italian guidelines ment, particularly in case of maternal obesity.58
1. Pregestational and gestational diabetes are not in itself an The GINEXMAL RCT is expected to enroll women with
indication for cesarean section. GDM at 38 weeks, with singleton pregnancies in cephalic
2. In the case of estimated fetal weight ≥4500 g, cesarean presentation with no previous cesarean section, excluding
section from 38 + 0 weeks gestation is recommended. pregnant women with an estimated weight of more than
3. Pregestational and gestational diabetes are not contrain- 4000 g at 38 weeks. The assignment is in two arms: induction
dications to vaginal birth after previous cesarean section at 38 weeks vs. expectant management until 41 weeks, where
(VBAC).56 induction of labor is carried out.59
It was not easy to complete this trial, because as stated,
verbatim, by Witkop in his work: “We acknowledge the
Timing potential barriers to the conduction of clinical trials in
pregnant women. Clinical trials of labor management may
The timing of induction in the case of GDM must meet two be particularly difficult in the current highly litigious envi-
requirements: (1) must be performed at a time that the fetus ronment. The feasibility of trials might also be influenced
reaches lung maturity (delayed in diabetic pregnancy), and by patient preferences for care and provider perception of
(2) before fetal growth reaches a point that increases the risk patient risks.”46
of shoulder dystocia. The recruitment of patients was completed in April 2014.
Most of the studies have identified issues related to GDM The interim data measured at November 2013 are presented
in the 38th week of gestation. Moreover, this limit is the in Table 37.1.
one proposed by NICE, no. 63/2008, which reads as follows: Therefore, the conclusions that can be drawn on these
“Pregnant women with diabetes who have a normally gown preliminary assessments are as follows:
fetus should be offered elective birth through induction of
labor, or by elective cesarean section if indicated, after 38 ●● The averages of the weights of newborns in the induction
completed weeks.”40,47–57 group are significantly lower than those of the expectant
group.
●● Induction does not increase cesarean section.
Reduction of cesarean section + operative vaginal deliv-
Induction or expectant management ●●
ery in induction (n.s.).

Evidence in the literature about the best management ●● Low rate of cesarean section in all the recruited popula-
possible (induction vs. expectant) in women with GDM tion (11.8%).
are therefore neither strong nor unique. For this reason,
we decided to design a multicenter RCT on this topic,
GINEXMAL (Gestational Diabetes Mellitus: Induction Special clinical conditions
vs. Expectant Management Labor), in order to understand
whether the induction of labor at 38 weeks actually pro- Obesity
duces an increase in the rate of cesarean sections in front An obese patient (body mass index > 30) with GDM has spe-
of the undisputed benefit of reducing the frequency of mac- cific risk factors in labor that most often lead to an emer-
rosomic births. gency cesarean section.60
In a previous study, it had in fact occurred retrospectively It is also known that obesity is associated with an
that labor induction at 38 weeks in GDM patients with fetal increased risk of failed labor induction that appears to be
growth acceleration does not seem to determine an increased related directly to an increasing class of obesity.61
This is confirmed by several observational studies of fetal distress and neonatal hypoglycemia (NICE Clinical
available in the literature but has also been shown by some Guideline no. 63).
studies carried out by S. Quenby on the characteristics of
the muscle cells of the uterus. She noted that in obese sub-
jects, a higher concentration of cholesterol is present at Conclusion
the level of these fibrocells, which induces a dysfunction
in labor, of uterine contractions, which then lose their The recommendations concerning guidelines must be based
effectiveness.62 on randomized, controlled clinical trials. The two RCTs
available in this area (GL Kjos and GINEXMAL) indicate
with certainty that in women with GDM, the induction of
Vaginal birth after previous cesarean section labor produces an average of neonatal weights lower than
Vaginal delivery and the induction of labor are not contrain the group with expectant management with a lower fre-
indicated even in cases of GDM; a retrospective case series quency of macrosomic births.
(n = 10,110) from the United States examined the success Lowering the frequency of macrosomic births in women
of VBAC in women with previous obstetric complications, with GDM without increasing the rate of cesarean sections
including gestational diabetes.63 Sixty-two percent of women can already be considered an important positive result.
who attempted VBAC were successful. The factors associ- The selection of patients eligible for an active manage-
ated with unsuccessful VBAC were birth weight exceed- ment, which takes into account the degree of cervical ripe-
ing 4000 g, cephalopelvic disproportion, prolonged labor, ness, can help lower the frequency of cesarean sections in the
dysfunctional labor, preexisting diabetes and gestational induction group.
diabetes, hypertension, induced labor, sexually transmitted The recommendations that we can try to define with
disease, fetal distress, and breech birth. regard to the opportunity to induce labor in childbirth are
Another retrospective cohort study (n = 25,079) from therefore as follows:
the United States showed instead that gestational diabe-
tes is not an independent risk factor for VBAC.64 The study ●● The induction of labor at 38 weeks is justified in order to
involved 13,396 women who attempted VBAC. Data on 423 reduce the frequency of births with higher weight. The
women with diet-controlled gestational diabetes and 9437 effect leads to known benefits associated with the birth of
women without diabetes were analyzed. The success rate for a reduced number of macrosomia.
attempted VBAC among women with gestational diabetes ●● It is good to act with common sense according to the
was 70% compared to 74% for women without diabetes, and clinical characteristics of the patient, her Bishop’s score,
the proportion of babies weighing more than 4000 g was 18% and her motivations.
compared with 13% (p < 0.05). ●● It is also important to have one part of a comprehensive
informed consent from the patient and the other of a
full capacity to manage the complications of delivery by
Glycemic profile during labor consultants.
and risk of fetal death in GDM
Finally, it should be remembered that a clinical recommen-
We do not feel superfluous to recall once again the impor- dation is, however, a suggestion aimed at intelligent profes-
tance of maternal glycemic control during labor and birth sionals where they can customize the interests of patient
and early feeding of the baby in order to reduce the risk care, and not a military order to be executed at all costs.
REFERENCES
1. World Health Organization and Department of Non 6. Walsh JM, Kandamany N, Nihuibhne N et al. Neonatal bra-
communicable Disease Surveillance. Definition, diagnosis chial plexus injury: Comparison of incidence and antecedents
and classification of diabetes mellitus and its complications. between 2 decades. Am J Obstet Gynecol April 2011; 204(4):
Report of a WHO consultation. Part 1: Diagnosis and classifica- 324.e1–324.e6.
tion of diabetes mellitus. World Health Organization: Geneva, 7. Baskett F, Allen AC. Perinatal implications of shoulder dystocia.
Switzerland, 1999. Obstet Gynecol July 1995; 86(1): 14–17.
2. HAPO Study Cooperative Research Group. Hyperglycemia 8. Walsh JM, Mahony R, Byrne J et al. The association of mater-
and adverse pregnancy outcomes. N Engl J Med 2008; 358(19): nal and fetal glucose homeostasis with fetal adiposity and
1991–2002. birthweight. Eur J Obstet Gynecol Reprod Biol December 2011;
3. International Association of Diabetes and Pregnancy Study 159(2): 338–341.
Groups Consensus Panel. International association of diabetes 9. Hartling L, Dryden DM, Guthrie A et al. Benefits and harms of
and pregnancy study groups recommendations on the diagnosis treating gestational diabetes mellitus: A systematic review and
and classification of hyperglycemia in pregnancy. Diabetes Care meta-analysis for the U.S. Preventive Services Task Force and the
2010; 33(3): 676–682. National Institutes of Health Office of Medical Applications of
4. Diagnostic Criteria and Classification of Hyperglycaemia First Research. Ann Intern Med July 16, 2013; 159(2): 123–129.
Detected in Pregnancy, WHO: Geneva, Switzerland, September 10. Schmidt MI, Duncan BB, Reichelt AJ et al. Gestational diabetes
2013. mellitus diagnosed with a 2-h 75-g oral glucose tolerance test
5. Pedersen J. Weight and length at birth of infants of diabetic moth- and adverse pregnancy outcomes. Diabetes Care 2001; 24(7):
ers. Acta Endocrinol 1954; 16(4): 330–342. 1151–1155.
References 311
11. Berg A, Rydhstroem H, Frid A. Impaired glucose tolerance associ- 38. Lipscomb KR, Gregoy K, Shaw K. The outcome of macrosomic
ated with adverse pregnancy outcome: A population-based study infants weighing at least 4500 grams: Los Angeles Country, University
in southern Sweden. Am J Obstet Gynecol 2001; 184(2): 77–83. of Southern California experience. Obstet Gynecol 1995; 85: 558.
12. Khan KS, Syed AH, Hashmi FA, Rizvi JH. Relationship of fetal 39. ACOG. Pregestational diabetes mellitus, Practice Bulletin No. 60,
macrosomia to a 75g glucose challenge test in nondiabetic preg- March 2005. Obstet Gynecol March 2005; 105: 675–685.
nant women. Aust N Z J Obstet Gynaecol 1994; 34(1): 24–27. 40. NICE. Diabetes in pregnancy, Clinical Guideline No. 63, July
13. Shirazian N, Mahboubi M, Emdadi R et al. Comparison of differ- 2008; 4.10: 43.
ent diagnostic criteria for gestational diabetes mellitus based on 41. Gilbert WM, Nesbitt TS, Danielsen B. Associated factors in 1611
the 75-g oral glucose tolerance test: A cohort study. Endocr Pract cases of brachial plexus injury. Obstet Gynecol April 1999; 93(4):
2008; 14(3): 312–317. 536–540.
14. Forsbach G, Cantu-Diaz C, Vazquez-Lara J et al. Gestational dia- 42. Blackwell SC, Hassan SS, Wolfe HW et al. Why are cesarean
betes mellitus and glucose intolerance in a Mexican population. delivery rates so high in diabetic pregnancies? J Perinat Med
Int J Gynaecol Obstet 1997; 59(3): 229–232. 2000; 28: 316.
15. Metzger E, Lowe LP, Dyer AR et al. Hyperglycemia and adverse 43. Lapolla A, Dalfrà MG, Di Cianni G et al. A multicenter Italian
pregnancy outcomes. N Engl J Med 2008; 358(19): 1991–2002. Study on pregnancy outcome in women with diabetes. Nutr
16. Moses RG, Calvert D. Pregnancy outcomes in women without ges- Metab Cardiovasc Dis 2008; 18: 291.
tational diabetes mellitus related to the maternal glucose level. Is 44. Naylor CD, Sermer M, Chen E et al. Cesarean delivery in relation to
there a continuum of risk? Diabetes Care 1995; 18(12): 1527–1533. birth weight and gestational glucose tolerance. Pathophysiology
17. Khan KS, Syed AH, Hashmi FA et al. Relationship of fetal mac- or practice style? JAMA 1996; 275(15): 1165–1170.
rosomia to a 75g glucose challenge test in nondiabetic pregnant 45. Cousins L. Pregnancy complications among diabetic women:
women. Aust N Z J Obstet Gynaecol 1994; 34(1): 24–27. Review 1965–1985. Obstet Gynecol Surv 1987; 42: 140.
18. Spellacy WN, Miller S, Winegar A et al. Macrosomia—Maternal 46. Witkop CT, Neale D, Wilson LM et al. Active compared with
characteristics and infant complications. Obstet Gynecol, 1985; expectant delivery management in women with gestational dia-
66(2): 158–161. betes a systematic review. Obstet Gynecol 2009; 113: 206–217.
19. Mondestin MAJ, Ananth CV, Smulian JC, Vintzileos AM. Birth 47. Kjos SL, Henry OA, Montoro M et al. Insulin-requiring diabetes
weight and fetal death in the United States: The effect of maternal in pregnancy: A randomized trial of active induction of labor and
diabetes during pregnancy. Am J Obstet Gynecol 2002; 187(4): expectant management. Am J Obstet Gynecol 1993; 169: 611–615.
922–926. 48. Boulvain M, Stan C, Irion O. Elective delivery in diabetic preg-
20. Bjorstad AR, Irgens-Hansen K, Daltveit AK, Irgens LM. nant women. Cochrane Database Syst Rev 2001; 2: CD001997.
Macrosomia: Mode of delivery and pregnancy outcome. Acta 49. Sacks DA, Sacks A. Induction of labor versus conservative man-
Obstet Gynecol Scand 2010; 89: 661. agement of pregnant diabetic women. J Matern Fetal Neonatal
21. Royal College of Obstetricians and Gynaecologists. Green-Top Med 2002; 12: 438–441.
Guideline N. 42, 2nd ed., Royal College of Obstetricians and 50. Levy AL, Gonzalez JL, Rappaport VJ et al. Effect of labor induc-
Gynaecologists: London, U.K., March 2012. tion on cesarean section rates in diabetic pregnancies. J Reprod
22. Resnick R. Management of shoulder dystocia girdle. Clin Obstet Med 2002; 47(11): 931–932.
Gynecol 1980; 23: 559–564. 51. Colorado Clinical Guidelines. Gestational Diabetes Guidelines
23. Lurie S, Ben-Arie A, Hagay Z. Shoulder dystocia: Could it be 2006; 6.
deduced from the labor partogram? Am J Perinatol 1995; 12(1): 61. 52. Collège National Des Gynécologues Et Obstétriciens Françai.
24. Bofill JA, Rust OA, Devidas M et al. Shoulder dystocia and opera- Recommandations pour la pratique clinique Le diabète gestation-
tive vaginal delivery. J Matern Fetal Med July–August 1997; 6(4): nel. December 10, 2010; 680.
220–224. 53. Kautzky-Willer A, Bancher-Todesca D, Birnbacher R. Gestational
25. Gherman RB, Ouzounian JG, Goodwin TM. Obstetric maneu- diabetes mellitus. Acta Med Austriaca 2004; 31(5): 182–184.
vers for shoulder dystocia and associated fetal morbidity. Am J 54. Hoffman L, Nolan C, Wilson JD et al. Gestational diabetes
Obstet Gynecol 1998; 178: 1126–1130. mellitus—Management guidelines. The Australasian Diabetes in
26. Gherman RB, Ouzounian JG, Miller DA et al. Spontaneous vagi- Pregnancy Society. Med J Aust July 20, 1998; 169(2): 93–97.
nal delivery: A risk factor for Erb’s palsy? Am J Obstet Gynecol 55. Royal College of Obstetricians and Gynaecologists. Green-Top
1998; 178: 423–427. Guideline No. 42, 2nd ed., Shoulder Dystocia, Royal College of
27. Mazouni C, Menard JP, Porcu G et al. Maternal morbidity asso- Obstetricians and Gynaecologists: London, U.K., March 2012.
ciated with obstetrical maneuvers in shoulder dystocia. Eur J 56. Sistema Nazionale per le linee guida linee, ISS Taglio cesareo:
Obstet Gynecol Reprod Biol 2006; 129: 15–18. una scelta appropriata e consapevole Seconda parte linea guida
28. Acker DB, Sachs BP, Friedman EA. Risk factors for shoulder dys- 21, Gennaio 2011; 65.
tocia. Obstet Gynecol 1985; 66: 762–768. 57. Hod M, Bar J, Peled Y et al. Antepartum maagement protocol.
29. Gross TL, Sokol RJ, Williams T et al. Shoulder dystocia: A fetal- Timing and mode of delivery n gestational diabetes. Diabetes
physician risk. Am J Obstet Gynecol 1987; 156: 1408–1418. Care August 1998; 21 (Suppl. 2): B113–B117.
30. Øverland EA, Vatten LJ, Eskild A. Pregnancy week at delivery and 58. Alberico S, Businelli C, Maso G et al. Gestational diabetes and
the risk of shoulder dystocia: A population study of 2 014 956 fetal growth acceleration: Induction of labour versus expectant
deliveries. Br J Obstet Gynecol 2013; 121(1): 34–41. management. Minerva Ginecol December 2010; 62(6): 533–539.
31. Langer O, Berkus MD, Huff RW. Shoulder dystocia: Should the 59. Maso GM, Alberico S, Hod M et al. GINEXMAL RCT: Induction
fetus weighing >4000g delivered by cesarean section? Am J of labor versus expectant management in gestational diabetes
Obstet Gynecol 1991; 165: 831–837. pregnancies. BMC Pregnancy Childbirth 2011; 11: 31.
32. Nesbitt TS, Gilbert WM, Herrchen B. Shoulder dystocia and asso- 60. Dennedy MC, Avalos G, O’Reilly MW. ATLANTIC-DIP: Raised
ciated risk factors with macrosomic infants born in California. Am maternal body mass index (BMI) adversely affects maternal and fetal
J Obstet Gynecol 1998; 179: 476–480. outcomes in glucose-tolerant women according to international
33. Confidential into Maternal and Child Health. Pregnancy in association of diabetes and pregnancy study groups (IADPSG) cri-
Women with Type 1 and Types 2 Diabetes, CEMACH: London, teria. J Clin Endocrinol Metab April 2012, 97(4): E608–E612.
U.K., 2008. 61. Wolfe KB, Rossi RA, Warshak CR. The effect of maternal obesity
34. Acker DB, Gregory KD, Sachs BP et al. Risk factors for Erb- on the rate of failed induction of labor. Am J Obstet Gynecol
Duchenne palsy. Obstet Gynecol 1988; 71: 389–392. 2011; 205: xx.
35. Conway DL, Langer O. Elective delivery of infants with macrosomia 62. Quenby S, Zhang J, Bricker L et al. Poor contractility in obese
in diabetic women: Reduced shoulder dystocia versus increased women. BJOG March 2007; 114(3): 343.
cesarean deliveries. Am J Obstet Gynecol 1998; 178: 922–925. 63. Holt VL, Mueller BA. Attempt and success rates for vaginal birth
36. Elliott JP, Garite TJ, Freeman RK et al. Ultrasonic prediction of after caesarean section in relation to complications of the previous
fetal macrosomia in diabetic patients. Obstet Gynecol August pregnancy. Paediatr Perinat Epidemiol 1997; 11(Suppl. 1): 63–72.
1982; 60(2): 159–162. 64. Marchiano D, Elkousy M, Stevens E et al. Diet-controlled gesta-
37. Rouse DJ, Owen J, Goldenberg RL et al. The effectiveness of tional diabetes mellitus does not influence the success rates for
elective cesarean section delivery for fetal macrosomia diag- vaginal birth after cesarean delivery. Am J Obstet Gynecol 2004;
nosed by ultrasound. JAMA 1996; 276: 1480. 190(3): 790–796.
38 Management of the
macrosomic fetus
Federico Mecacci, Marianna Pina Rambaldi, and Giorgio Mello
than 4500 g.3,7–10 A review of 14 studies on the detection of

Introduction macrosomia (≥4000 g) using different methods for sono-
Despite major progress made in obstetrics over recent graphic EFW reported widely varying results: sensitivity of
decades, there is still a lack of knowledge on how to man- 12%–75%, specificity of 68%–99%, and posttest probability
age the macrosomic fetus. Attention has always been mostly of 17%–79% after a positive test.3 The diagnosis of macro-
focused on the opposite extreme of the scale of growth, thus somia defined as ≥4500 was even less accurate. The error
resulting in a gap in the literature on the management of is partially explained by the fact that ultrasound does not
large fetuses. Published studies do not even find a consensus take into account body composition; because of this factor,
on the definition and are designed primarily to detect com- it can be a substantial variation in birth weight even among
plications during labor, those that question the outcome of fetuses with the same biometric parameters. An additional
the pregnancy are dated, and there is no meta-analysis. error factor is the increased incidence of diabetes in moth-
ers of macrosomic fetuses; in fact, Hadlock’s formula is very
sensitive to abdominal circumference measurement that is
Definition and diagnosis increased in these fetuses.
Macrosomia is diagnosed in the presence of excessive intra-

uterine growth resulting in a birth weight that exceeds an Complications
established limit of either 4000 or 4500 g.1 Approximately 10%
of all newborns have a birth weight of >4000 g, 1.5% weigh Macrosomic fetuses are at increased risk of complications
>4500 g, and 0.1% weigh >5000 g.2,3 The American College of during pregnancy, childbirth, and postnatal life in the short
Obstetricians and Gynecologists recommends the use of the and long term (Table 38.2). In this chapter, we will only
4500 g threshold because morbidity increases sharply beyond review the risks related to pregnancy (for other complica-
this weight, but acknowledges that there is some increased risk tions, see Chapters 40 through 48).
of complications at weights >4000 g.1 This definition does not The birth weight–specific infant mortality curve shows
take into account the population distribution, in which large a decline in mortality with increasing birth weight until
fetuses are defined as those who weigh more than the 90th per- a point at which the slope reverses and the mortality rises
centile for gestational age. Weight percentile is the best means again for fetuses situated at the upper values of the curve.
for identifying the preterm macrosomic fetus (Table 38.1). High birth weight has been associated with a two to three
Regardless of the definition used, diagnosis of macro- times risk of intrauterine death.11–14 The most recent study
somia is difficult and often inaccurate. Symphysis–fundal published on this topic found that fetuses weighing 4500–
measurements combined with Leopold maneuvers showed 5000 g had an odds ratio of 2.7 (95% CI, 2.2–3.4) to die in utero
sensitivities of only 10%–43% and positive predictive values compared to fetuses of appropriate weight (3500–4500 g),
of 28%–53% for detecting macrosomia.4 Although these while fetuses who weigh >5000 g had a 13.2 times greater
results appear suboptimal, the use of ultrasonography exam risk (95% CI, 9.8–17.7). Furthermore, both categories were
ination does not improve the detection rate.1 When 2D at higher risk of morbidity, especially birth injury and low
ultrasound scan is used for the diagnosis of macrosomia, the 5-minute Apgar score.11
estimated fetal weight (EFW) obtained with the Hadlock’s Prevention of stillbirth in large fetuses is challenging and
formula was found to have a higher predictive value com- the most effective measure still remains the prevention of
pared to other methods.5,6 Nevertheless, EFW formulas macrosomia by a more careful management of maternal risk
perform better for appropriate-for-gestational-age fetuses; factors during the preconceptional period or as early as pos-
i.e., the Hadlock’s formula increases the absolute mean per- sible in pregnancy. It is debated whether it would be possible
centage of error from 8% to 13% when used for infants larger to prevent stillbirth, anticipating the deliveries of all at-risk
312
Conclusions 313
The macrosomic fetus is in a fragile balance; its large con-

Table 38.1 Risk factors for having a macrosomic
stitution and enhanced metabolism result in an increased
infant
oxygen need with respect to an average-weight fetus, while its
placenta is not equally developed to accommodate this request
Previous macrosomic infant
and may induce a chronic status of fetal hypoxemia.22–24
Maternal prepregnancy weight
The two main risk factors for macrosomia are obesity and
Excessive weight gain in pregnancy gestational diabetes and, when present, may complicate the
Maternal diabetes clinical management at term. On one hand, obese women are
Multiparity at increased risk of complication because the adipose layer
Male infant limits the maternal ability to perceive fetal movements and
Postterm pregnancy at the same time limits the diagnostic accuracy of ultraso-
Hispanic or African-American ethnicity nography investigation. In addition, if diabetes is the cause
Maternal birth weight over 4000 g of macrosomia, reactive fetal hyperinsulinism acts on the
Maternal height placenta. Insulin has a direct effect on the fetal placenta,
Maternal age younger than 17 inducing alterations in endothelial gene expression and stim-
Advanced maternal age ulating the synthesis of glycogen at this level, determining a
thickening of the membrane that worsens fetal hypoxia.
There is no data from large or randomized trials to for-
Table 38.2 Short-term complications mulate evidenced-based recommendations on which preg-
nancies complicated by macrosomia should undergo fetal
Shoulder dystocia surveillance, when to start it, what test to order, or how often
Birth injuries to perform it. The authors suggest antepartum monitoring
Mechanical ventilation for more than 30 min duration
using fetal movement counting, biophysical p rofile, and non-
stress test (NST) starting from 37 weeks of gestation and per-
A 5-minute Apgar score <3
formed at least weekly until 40 weeks.
Respiratory distress syndrome
Meconium aspiration
Perinatal mortality Induction of lung maturity
When considering an elective early delivery because of
pregnancies at 38 weeks: however, an analysis of the litera- fetal macrosomia, it must be taken into account whether
ture suggests that the majority of fetal deaths occur before there is sufficient fetal maturation. As the morbidity asso-
this term and therefore would not be prevented with early ciated with early term and late preterm deliveries in gen-
induction.15–17 eral obstetric population has become evident, elective
Short-term maternal complications reported in associa- delivery before 39 weeks is increasingly discouraged.25,26
tion with fetal macrosomia are increased rate of cesarean Furthermore, respiratory distress syndrome is more likely
section, perineal lacerations, uterine atonia, and postpar- to develop in infants of women with diabetes delivered
tum hemorrhage.18 The risk of emergency cesarean section early than in infants of women without diabetes; this risk
is reported to be nearly double and the risk of grades III does not become equivalent in the two groups until after
and IV lacerations can be three to six times higher when the 38.5 weeks of gestation.27 It is debated whether it is appro-
birth weight is above 4500 g.19,20 priate to test for lung maturity if delivery is anticipated
before 39 weeks, also because it has been shown that a posi-
Intervention tive result does not obviate all the complications associated
with late preterm birth.28
Currently, there is a lack of evidence in the literature on the
possibility of and on which could be the best way to treat
fetal macrosomia. For mothers without diabetes, no clini-
cal interventions have been reported to be useful. In this Conclusions
population, the best recommendation could be to prevent
the development of macrosomia with dietary intervention in
●● Diagnosis of macrosomia is often inaccurate and ultra-
women at risk for obesity or excessive weight gain, but no sonography does not perform better than clinical
randomized controlled trials have been conducted yet. For examination with symphysis–fundal measurements
pregnancies complicated by diabetes, it seems possible to combined with Leopold’s maneuver. Pregnancy affected
improve fetal outcome, adding insulin to the diet when mac- by m
acrosomia poses an increased risk for the fetus and
rosomia is diagnosed between 29 and 33 weeks of gestation.21 the mother.
●● No strategy has proved to be effective in reducing com-
plications once the macrosomia is present in nondiabetic
Antenatal monitoring pregnancies.
There is no standardized algorithm for performing antenatal ●● Further studies are needed to define the best way to moni-
monitoring of fetal macrosomia. tor macrosomic fetuses at term.
314 Management of the macrosomic fetus
REFERENCES
1. Chatfield J. ACOG issues guidelines on fetal macrosomia. 15. Macintosh MC et al. Perinatal mortality and congenital anomalies
American College of Obstetricians and Gynecologists. Am Fam in babies of women with type 1 or type 2 diabetes in England,
Physician July 1, 2001; 64(1): 169–170. Wales, and Northern Ireland: Population based study. BMJ 2006;
2. Ventura SJ, Martin JA, Curtin SC et al. Births: Final data for 1998. 333(7560): 177.
Natl Vital Stat Rep 2000; 48(3): 1–100. 16. Lowy C, Beard RW, Goldschmidt J. Congenital malformations in
3. Chauhan SP et al. Suspicion and treatment of the macrosomic babies of diabetic mothers. Diabet Med 1986; 3(5): 458–462.
fetus: A review. Am J Obstet Gynecol 2005; 193(2): 332–346. 17. Lowy C. Type I diabetes and pregnancy. Lancet 1995; 346(8980):
4. O’Reilly-Green C, Divon M. Sonographic and clinical methods 966–967.
in the diagnosis of macrosomia. Clin Obstet Gynecol 2000; 43(2): 18. Henriksen T. The macrosomic fetus: A challenge in current
309–320. obstetrics. Acta Obstet Gynecol Scand 2008; 87(2): 134–145.
5. Pinette MG et al. Estimation of fetal weight: Mean value from 19. Ju H et al. Fetal macrosomia and pregnancy outcomes. Aust N Z
multiple formulas. J Ultrasound Med 1999; 18(12): 813–817. J Obstet Gynaecol 2009; 49(5): 504–509.
6. Faschingbauer F et al. Fetal weight estimation in extreme mac- 20. Lipscomb KR, Gregory K, Shaw K. The outcome of macroso-
rosomia (≥4,500 g): Comparison of 10 formulas. Ultraschall Med mic infants weighing at least 4500 grams: Los Angeles County +
2012; 33(7): E62–E67. University of Southern California experience. Obstet Gynecol
7. Mongelli M, Benzie R. Ultrasound diagnosis of fetal macro- 1995; 85(4): 558–564.
somia: A comparison of weight prediction models using com- 21. Buchanan TA et al. Use of fetal ultrasound to select metabolic
puter simulation. Ultrasound Obstet Gynecol 2005; 26(5): therapy for pregnancies complicated by mild gestational diabe-
500–503. tes. Diabetes Care 1994; 17(4): 275–283.
8. Weiner Z et al. Clinical and ultrasonographic weight estimation 22. Philipps AF et al. Effects of chronic fetal hyperglycemia upon
in large for gestational age fetus. Eur J Obstet Gynecol Reprod oxygen consumption in the ovine uterus and conceptus. J Clin
Biol 2002; 105(1): 20–24. Invest 1984; 74(1): 279–286.
9. Sacks DA, Chen W. Estimating fetal weight in the management of 23. Cohn HE et al. The effect of hyperglycemia on acid-base and
macrosomia. Obstet Gynecol Surv 2000; 55(4): 229–239. sympathoadrenal responses in the hypoxemic fetal monkey.
10. Alsulyman OM, Ouzounian JG, Kjos SL. The accuracy of intra- J Dev Physiol 1992; 17(6): 299–304.
partum ultrasonographic fetal weight estimation in diabetic preg- 24. Hay WW Jr. et al. Effects of glucose and insulin on fetal glucose
nancies. Am J Obstet Gynecol 1997; 177(3): 503–506. oxidation and oxygen consumption. Am J Physiol 1989; 256(6
11. Huang DY et al. Determinants of unexplained antepartum fetal Pt. 1): E704–E713.
deaths. Obstet Gynecol 2000; 95(2): 215–221. 25. Consortium on Safe Labor et al. Respiratory morbidity in late pre-
12. Mondestin MA et al. Birth weight and fetal death in the United term births. JAMA 2010; 304(4): 419–425.
States: The effect of maternal diabetes during pregnancy. Am J 26. Tita AT et al. Timing of elective repeat cesarean delivery at term
Obstet Gynecol 2002; 187(4): 922–926. and neonatal outcomes. N Engl J Med 2009; 360(2): 111–120.
13. Rasmussen S et al. Risk factors for unexplained antepartum 27. Robert MF, Neff RK, Hubbell JP et al. Association between
fetal death in Norway 1967–1998. Early Hum Dev 2003; 71(1): maternal diabetes and the respiratory-distress syndrome in the
39–52. newborn. N Engl J Med 1976; 294: 357.
14. Zhang X et al. How big is too big? The perinatal consequences 28. Bates E et al. Neonatal outcomes after demonstrated fetal lung
of fetal macrosomia. Am J Obstet Gynecol 2008; 198(5): maturity before 39 weeks of gestation. Obstet Gynecol 2010;
517.e1–517.e6. 116(6): 1288–1295.
39 Congenital malformations in
diabetic pregnancy: Prevalence
and types
Paul Merlob
the control group (5.3%), with a relative risk of any malfor-

Introduction mation of 2.47. The frequency of malformations increased
The association between maternal diabetes mellitus and steadily with the severity of the maternal diabetes scaled by
congenital malformations in newborns may well be causal; White’s classification.
however, the teratogenic mechanism remains unclear.1 The The largest review of malformations associated with dia-
prevalence and the described type of malformations vary betic pregnancy, performed by Kucera3 in 1971, included the
among different studies, and a predictable malformation data available from the world literature between 1945 and
syndrome has not been identified.1 This chapter discusses the 1965, covering 7101 infants born to diabetic mothers and a
prevalence and types of structural congenital malformations control group of more than 400,000 healthy infants derived
in infants of mothers with diabetes mellitus (IDM) or gesta- from the registries of the World Health Organization. The
tional diabetes mellitus (IGDM). Their etiology and patho- prevalence of congenital malformations was 4.79% in the
genesis are presented in two other chapters of this book. study population compared with 0.65% in the controls.3
However, information regarding the criteria used for the
diagnosis and management of diabetes, and for the diagnosis
Historical data and classification of the malformations, was not consistently
available.
Many studies of small groups of infants (from 1930 to The results of the Collaborative Perinatal Project, using
1950), including malformed infants, were not conclusive. data from hospitals throughout the United States, were
It was only after 1960 that large, controlled, population- or reported in 1975.4 A total of 567 overt diabetic pregnancies,
hospital-based studies of offspring of mothers with diabetes 372 gestational diabetic pregnancies, and 47,408 nondiabetic
began to appear in the literature. pregnancies were analyzed. Congenital malformations were
In 1964, Molsted-Pedersen et al.2 studied 853 IDM born identified in a significantly higher proportion of infants
between 1926 and 1963 and weighing more than 1000 g. of mothers with overt diabetes (17.1%) than in infants of
They noted a 6.4% rate of congenital malformations, com- nondiabetic mothers (8.4%). The IGDM had an 8.9% rate of
pared to only 2.1% in a control group of 1212 infants born to malformations, almost equal to that of the control group.
nondiabetic mothers in the same hospital, but during 1958 In 1976, Soler et al.5 studied 701 IDM born in Bir
and 1960. mingham, United Kingdom, between 1950 and 1974, of
Further analysis revealed that the presence of maternal whom 8.1% were found to have a congenital malformation
vascular complications was associated with a significantly compared to 1.7% in the control group. In the same year,
greater risk of fatal, major (but not mild), and multiple Day and Insley,6 in a study of 205 IDM born in the same
malformations in the infants. area between 1969 and 1974, reported a malformation rate
Two years later, Naeve1 analyzed the frequency of con- of 12%, compared to 6% in the control nondiabetic group.
genital malformations in infants of 2592 diabetic mothers, After 1980, investigations in the field expanded w
orldwide.
892 women known to have become diabetic 5 or more years Although their specific figures differed considerably, in
after delivery, and 1262 infants of women married to dia- almost all of them, the prevalence rates were significantly
betic men and an equal number of infants of nondiabetic higher in IDM than in the general newborn population.
pregnancies. A significantly higher rate of congenital mal- However, even the largest studies7–11 comprised a relatively
formations was noted in the study group (13.1%) than in small absolute number of congenital malformations in
315
316 Congenital malformations in diabetic pregnancy
offspring of women with pregestational insulin-dependent malformation rates no different than the general n ondiabetic
diabetes.11 Recently (2012), the largest case series of mal- population. Nevertheless, it is important to remember that
formed babies (669 newborns) was published by Garne et al.12 gestational diabetes is a heterogeneous disorder that includes
confirming that the risk of major nonchromosomal congeni- previously unrecognized or newly diagnosed nongesta-
tal anomalies is at least twice as high in the pregestational tional diabetes mellitus.8,17 These patients (with preexist-
diabetic pregnancies compared to nondiabetic pregnancies. ing but undetected type 2 diabetes) represent a subgroup of
IGDM with an increased risk for occurrence of congenital
malformations.
Prevalence Type 2 diabetes (associated with both increasing insulin
resistance and abnormality of insulin secretion) is increas-
The prevalence of congenital malformations has been evalu- ing explosively.13,20 As in all diabetic pregnancies, infants
ated predominantly in women with type 1 diabetes (pre- born to mothers with type 2 diabetes are also at increased
gestational insulin-dependent diabetes), with rates ranging risk of having congenital malformations.16,20 The most
mainly from 4.1% to 17.1%. Despite improvements in obstet- common malformations20 were cardiac (53%) followed by
ric care, with strict control of diabetes and good pregnancy musculoskeletal (27%). These results are almost similar
surveillance, the rate of major malformations in infants of to rates previously identified in women with type 1 dia-
mothers with type 1 diabetes remains two to three times that betes. 20 The majority of malformations occurred in those
of the general population. Comparisons among the differ- with poor glycemic control who did not received prepreg-
ent studies are very difficult, if not impossible, for several nancy care. 20,21
reasons: The prevalence of congenital malformations in all types
of maternal diabetes is in good relationship with
●● The diagnostic classification of diabetes varies among and
even within countries. ●● The severity of the maternal diabetes (White’s class)
●● Screening for the detection of diabetes has not been ●● The control of maternal diabetes (expressed by HbA1c)
undertaken in the same manner and scale in all geo-
graphic regions. An increased prevalence of malformations occurring in
●● The timing and degree of preconception and postconcep- parallel with the severity and duration of the diabetes has
tion care differs. been described. 22 For example, the frequency of major
●● The diagnosis, definition, and classification of congenital malformations among offspring has been reported to be
malformations vary largely among studies. 4.4% in White’s classes B and C, 9.7% in class D, and 16.7%
●● The methods used to detect fetal and neonatal malforma- in class F, with an overall prevalence of 6.4% in all infants
tions are different. of diabetic mothers as compared with 2% in the general
●● Other confounding factors, such as maternal age, obesity, population. 22
ethnicity, and uptake of folic acid, are rarely taken into The same increase of prevalence of malformations was
account. observed in relation to the HbA1c levels. For example, a
frequency of 5.1% of malformations has been reported for
These confounding factors may explain the great variability a HbA1c of 7.0%–8.5%, 22.9% for a level of 8.6%–9.9% and
in the reported prevalence rates of congenital malformations 21.7% for a level of HbA1c above 10%.22 Even a slightly raised
in type 1 diabetic pregnancies, and they should be taken into of HbA1c during early pregnancy in women with type 1 dia-
account when prevalence of malformations is studied. betes carries an increased risk for fetal malformations.23,24
It is important to emphasize that the demographic pat-
tern of diabetes in pregnancy is changing: increasing num-
bers of young women are being diagnosed as having type 1 Types of malformations
diabetes, and the number of people diagnosed as having
type 2 diabetes is also increasing.13,14 The congenital malformations of IDM and IGDM consti-
With regard to gestational diabetes, several authors have tute a spectrum known as diabetic embryopathy (DE).12,22,25
reported an association with major malformations, and the This spectrum implies errors of morphogenesis that appear
same types of anomalies were found as in the offspring of between the third and the seventh week of embryonic devel-
women with pregestational type 1 diabetes.8,15–18 Although opment (end of blastogenesis and organogenesis).26 Within
the risk is apparently lower, being that gestational diabetes this spectrum of DE, cardiac, skeletal, central nervous system
accounts for approximately 90% of all cases of pregnancies (CNS), urogenital, gastrointestinal, facial, and multiple mal-
complicated with diabetes,19 the overall effect of this risk has formations were repeatedly described (Table 39.1). Congenital
important clinical and public health connotations. malformations in IGDM and offspring of women with type 2
Sheffield et al.10 found that women with pregestational diabetes affect the same organ systems that have been previ-
diabetes or gestational diabetes plus fasting hyperglycemia ously described in pregnancies with type 1 diabetes.16
had a threefold to fourfold greater risk of congenital mal- The most commonly affected organ systems were cardiac
formations in their offspring than the general population, (37.6%), musculoskeletal (14.7%), CNS (9.8%), and multiple
whereas infants of women with mild gestational diabetes had malformations (16%).16
Types of malformations 317
Table 39.1 Congenital malformations in infants of diabetic mothers
Malformations
Organ system Common Rare, occasional
Cardiac Corrected transposition Tetralogy of Fallot
Ventricular septal defect Hypoplastic left heart
Coarctation Single ventricle
Atrial septal defect Double-outlet right ventricle
Cardiomyopathy Pulmonic stenosis
Anomalous venous return
Skeletal Sacral agenesis Polydactyly
Vertebral and rib anomalies Syndactyly
Limb reduction defects Clinodactyly
Clubfoot
CNS Anencephaly Occipital encephalocele
Neural tube defects Holoprosencephaly
Microcephaly Septo-optic dysplasia
Hydrocephalus
Urogenital Hydronephrosis Hypoplastic genitalia
Renal agenesis Micropenis
Ureteral duplication Ambiguous genitalia
Multicystic dysplasia Megalo-urethra
Hypospadias
Gastrointestinal Duodenal atresia Malrotation
Anorectal atresia Volvulus
Esophageal atresia Omphalocele
Gastroschisis
Diaphragmatic hernia
Facial Cleft lip Choanal atresia
Cleft palate Absent depressor anguli oris muscle
Ears Fused orbits
Microtia
Anotia
Atresia of canal ear
Hairy ears
Hearing loss
Eyes
Cataract
Coloboma
Optic nerve hypoplasia
Others Single umbilical artery Laterality defects
Tracheal stenosis
Branchial arch anomalies
It has been debated whether maternal diabetes exerts a cardiac malformations, the strongest association with
nonspecific teratogenic effect expressed in a universally maternal diabetes was found in infants with defects of pri-
increased risk of all congenital malformations or whether mary congenital cardiogenesis, whereas most abnormalities
the disease should be regarded as a specific teratogen asso- arising later in cardiac development were not associated with
ciated with a distinct pattern of congenital abnormalities.11 diabetes.9 Others found a strong teratogenic effect on four
The spectrum in DE is large and highly variable; however, specific types of malformations: renal agenesis, obstructive
most studies have reported an increase of specific malforma- urinary tract, and cardiac and multiple abnormalities, as
tions especially involving the heart, the skeleton (particu- opposed to an unspecific increased general risk of congenital
larly sacral agenesis), the kidneys, and the CNS.7 Regarding malformations.11
Cardiac malformations malformations in IDM are renal agenesis, ureteral duplica-

Cardiac malformations are the most common congenital mal- tion, and hydronephrosis.1,11,25 Hypospadias is the most fre-
formations of IDM, and they occur significantly more often quent genital malformation in IDM and IGDM.11,25
in IDM than in infants of nondiabetic mothers.7,12,22 Rowland
et al.27 reported a 4% prevalence of cardiac malformations
in a series of 470 IDM, a fivefold higher rate than in the gen- Gastrointestinal malformations
eral population (0.8%). Becerra et al.7 found that IGDM who The abdominal malformations shown to occur with a higher
required insulin during the third trimester of pregnancy were prevalence in IDM include anorectal, duodenal, and lower
20.6 times more likely to have major cardiovascular malfor- intestine atresia.22,25 Malrotation, volvulus, and abdominal
mations than infants of nondiabetic mothers. No such differ- wall defects have also been described.
ence was noted in IGDM who did not require insulin.7
Loffredo et al.,9 in a population-based case–control study of
4390 IDM and 3572 healthy infants, observed that preconcep- Facial malformations
tional maternal diabetes was strongly associated with cardio- Facial anomalies in IDM and IGDM have been described
vascular malformations of early embryonic origin (OR = 4.7) in only a small number of reports.7,17,30–33 The most frequent
and cardiomyopathy (OR = 15.1), but not with obstructive were orofacial clefts7,17 and ear and eye abnormalities.1,7,31
and shunting defects (OR = 1.4). There was a strong associa- Interestingly, some studies reported an association of mater-
tion of cardiovisceral and cardiac chamber discordance, i.e., nal diabetes with certain facioskeletal syndromes, such as
“corrected” (levo-) transposition of the great arteries, but not femoral–facial syndrome and oculoauriculovertebral poly-
with “pure” transposition, i.e., intact ventricular septum or topic field defect.31–34 Therefore, IDM and IGDM should be
ventricular septal defect.9 Among outflow tract anomalies, carefully evaluated for facial malformations.31–34
the risk was strongly associated with normally related great
arteries (OR = 6.6) but not with simple transposition. These
findings imply a specific effect of maternal diabetes on certain Other anomalies
subtypes of cardiac malformations and may have important A single umbilical artery occurs in about 6.4% of IDM, a
clinical and preventive implications.9 fivefold higher rate than in the general population.22 This
mild malformation might be associated with other, major,
structural anomalies.
Skeletal malformations
Maternal diabetes has been associated with sacral agenesis,
also termed sacral dysgenesis or caudal regression.3 This is Multiple malformations
a complex malformation characterized by the absence or Many studies found a strong association between preges-
maldevelopment of the sacrum and coccyx, with or w ithout tational maternal diabetes and multiple-system malforma-
hypoplastic femurs, dislocated hips, defects in tibias or fib- tions (not defined as a syndrome) in the offspring.4,11,12 For
ulas, or other lower-limb malformations. Affected babies example, 27.5% of all malformed infants in the Collaborative
often have anomalies of other organ systems as well. Sacral Perinatal Project 4 had multiple anomalies. Aberg et al.17
agenesis occurs in about 0.2%–0.5% of IDM, representing observed that 6% of their malformed IDM had more than
a 200- to 400-fold higher rate than in the general popula- one anomaly compared to 0.57% of the control group. They
tion.5,12,22 Another skeletal malformation, proximally placed concluded that there is a clear-cut increase in the risk of
preaxial hallucal polydactyly, particularly when coupled multiple malformations in infants of mothers with preexist-
with segmentation anomalies of the spinal and tibial hemi- ing diabetes, but not of mothers with GDM.17 In the infants
melia, seems to be highly suggestive of DE.28 with multiple malformations, the same organ systems were
affected as in the whole group of IDM, with highest rates for
cardiac malformations, atresias, clefts, limb reduction, and
CNS malformations hypospadias.
Anencephaly is the most common CNS malformation asso-
ciated with diabetic pregnancy, affecting 0.57% of IDM, 22
which is threefold higher than the rate in the general Mild malformations
population (0.19%).22 IDM also have a high prevalence of Studies on mild malformations in IDM and IGDM are rela-
neural tube defects (1.95% vs. 0.2% in the general popula- tively scarce,35–37 and no randomized double-blind inves-
tion).29 One study of experimental diabetes induced after tigations have been performed to date. The results of the
the period of organogenesis noted no effect on the CNS of case–control studies are contradictory, with some showing
the offspring.22 significant differences but others not. No association was
observed between the severity of the metabolic derangement
(HbA1c) in the mother and the appearance of mild malfor-
Urogenital malformations mations in the offspring.36,37 This was also true for White’s
Kucera3 was the first to report an increased rate of uro- class, duration of diabetes, maternal age, and cigarette
logical malformations in IDM. The most frequent renal smoking.35
References 319
Preconception care and reduced risk of congenital the medical community are needed to promote education
malformations of the public, improve maternal access to PCC programs,
Many prospective controlled studies worldwide have dem- and maximize the interventions associated with improved
onstrated that strict metabolic control of maternal diabetes pregnancy outcome.21
before conception and during pregnancy can prevent most
neonatal complications, including congenital malforma-
tions.38 Ray et al.21 conducted a systematic review of all Summary
published studies on preconception care (PCC) (eight pro-
spective and eight retrospective studies were included in Maternal diabetes mellitus (types 1 and 2) and gestational
the final analysis) involving a total of 1192 infants whose diabetes with fasting hyperglycemia are associated with a
mothers had received PCC and 1459 infants of mothers who twofold to threefold increase in the risk of congenital mal-
had not. The pooled rate of major malformations was lower formations in offspring compared with the general popula-
in the infants of the PCC recipients (2.1%) than in nonre- tion. Although no specific malformation syndrome has been
cipients (6.5%), with a relative risk of 0.36. Interestingly, the identified, cardiac, skeletal, CNS, and multiple malforma-
risk of major malformations was lowest in the one study tions are the most frequently described. Up to 50% of all
in which folic acid was administered periconceptionally.39 perinatal deaths of offspring of diabetic mothers are due to
Recently, Correa et al. in a multicenter, population-based congenital malformations.42 Congenital malformations also
case–control study of birth defects, found that lack of peri- pose a serious social and financial burden to both the indi-
conceptional vitamins or supplements that contain folic acid vidual family and society at large.22 Therefore, the medical
may be associated with an excess risk for birth defects due to community must ensure that programs for PCC are made
maternal diabetes mellitus.40 The essential problem is that a available to diabetic mothers for strict control of their dis-
substantial percentage of diabetic women do not attend PCC ease before conception and during pregnancy. This could
programs21 and do not take prenatal folic acid and multivi- prevent the occurrence of most neonatal complications,
tamins, just as in the general population.41 Greater efforts by including congenital malformations.
REFERENCES
1. Naeve C. Congenital malformations in offspring of diabetics. 14. Hotu S, Carter B, Watson PD, Cutfield WS, Cundy T. Increasing
Perspect Pediatr Pathol 1984; 8: 213–222. prevalence of type 2 diabetes in adolescents. J Paediatr Child
2. Molsted-Pedersen L, Tygstrup I, Pedersen J. Congenital mal- Health 2004; 40: 201–204.
formations in newborn infants of diabetic women: Correlation 15. Kousseff BG. Gestational diabetes mellitus (class A): A human
with maternal diabetic vascular complications. Lancet 1964; I: teratogen? Am J Med Genet 1999; 83: 402–408.
1124–1126. 16. Shaefer-Graf UM, Buchanan TA, Xiang A et al. Patterns of con-
3. Kucera J. Rate and type of congenital anomalies among offspring genital anomalies and relationship to initial maternal fasting glu-
of diabetic mothers. J Reprod Med 1971; 7: 61–70. cose levels in pregnancies complicated by type 2 and gestational
4. Chung CS, Myrianthopoulos NC. Factors affecting risk of congen- diabetes. Am J Obstet Gynecol 2000; 182: 313–320.
ital malformations, II: Effect of maternal diabetes. Birth Defects 17. Aberg A, Westbom L, Kallen B. Congenital malformations among
Orig Artic Ser 1975; 11: 23–38. infants whose mothers had gestational diabetes or preexisting
5. Soler NG, Walsh CH, Malins JM. Congenital malformations in diabetes. Early Hum Dev 2001; 61: 85–95.
infants of diabetic mothers. QJM 1976; 45: 303–313. 18. Schaefer UM, Songster G, Xiang A et al. Congenital malforma-
6. Day RE, Insley J. Maternal diabetes mellitus and congenital tions in offspring of women with hyperglycemia first detected
malformations: Survey of 205 cases. Arch Dis Child 1976; 51: during pregnancy. Am J Obstet Gynecol 1997; 177: 1165–1171.
935–938. 19. Xiong X, Saunders LD, Wang FL, Demianczuk NN. Gestational
7. Becerra JE, Khoury MJ, Cordero JF, Erickson JD. Diabetes mel- diabetes mellitus: Prevalence, risk factors, maternal and infant
litus during pregnancy and the risks for specific birth defects: A outcomes. Int J Gynaecol Obstet 2001; 75: 221–228.
population-based case-control study. Pediatrics 1990; 85: 1–9. 20. Dunne F, Brydon P, Smith K, Gee H. Pregnancy in women with
8. Martinez-Frias ML, Bermejo E, Rodriguez-Pinilla E et al. type 2 diabetes: 12 years outcome data 1990–2002. Diabet Med
Epidemiological analysis of outcomes of pregnancy in gestational 2003; 30: 734–738.
diabetic mothers. Am J Med Genet 1998; 78: 140–145. 21. Ray JG, O’Brien TE, Chan WS. Preconception care and the risk
9. Loffredo CA, Wilson PD, Ferencz C. Maternal diabetes: An of congenital anomalies in the offspring of women with diabetes
independent risk factor for major cardiovascular malformations mellitus: A meta-analysis. QJM 2001; 94: 435–444.
with increased mortality of affected infants. Teratology 2001; 64: 22. Albert-Reece A, Hobbins JC. Diabetic embryopathy:
98–106. Pathogenesis, prenatal diagnosis and prevention. Obstet Gynecol
10. Sheffield JS, Butler-Koster EG, Casey BM et al. Maternal diabetes Surv 1986; 41: 325–335.
mellitus and infant malformations. Obstet Gynecol 2002; 100: 23. Miller E, Hare JW, Cloherty JP et al. Elevated maternal haemo-
925–930. globin A1c in early pregnancy and major congenital anomalies in
11. Nielsen GL, Norgard B, Puho E et al. Risk of specific congenital infants of diabetic mothers. N Engl J Med 1981; 304: 1331–1334.
abnormalities in offspring of women with diabetes. Diabet Med 24. Suhonen L, Hiilesmaa V, Teramo K. Glycaemic control during
2005; 22: 693–696. early pregnancy and fetal malformations in women with type 1
12. Garne E, Loane M, Dolk H et al. Spectrum of congenital anoma- diabetes mellitus. Diabetologia 2000; 43: 79–82.
lies in pregnancies with pregestational diabetes. Birth Defects 25. Merlob P, Reisner SH. Fetal effects from maternal diabetes. In:
Res A Clin Mol Teratol 2012; 94: 134–140. Buyse L, ed., Birth Defects Encyclopedia, 1st ed., Blackwell
13. MacIntosh MCM, Fleming KM, Bailey JA et al. Perinatal mortality Scientific: Cambridge, U.K., 1990, pp. 700–702.
and congenital anomalies in babies of women with type 1 or type 26. Mills JL, Baker L, Goldman AS. Malformations in infants of
2 diabetes in England, Wales and Northern Ireland: Population- diabetic mothers occur before the seventh gestational week:
based study. BMJ 2006; 333: 177–182. Implications for treatment. Diabetes 1978; 28: 292–293.
27. Rowland TW, Hubbell JP, Nadas AS. Congenital heart disease in 35. Rosenn B, Miodovnik M, St. John Dignan P et al. Minor con-
infants of diabetic mothers. J Pediatr 1973; 83: 815–820. genital malformations in infants of insulin-dependent diabetic
28. Adam MP, Hudgins L, Carey JC et al. Preaxial hallucal polydac- women: Association with poor glycemic control. Obstet Gynecol
tyly as a marker for diabetic embryopathy. Birth Defects Res A 1990; 76: 745–749.
Clin Mol Teratol 2009; 85: 13–19. 36. Hod M, Merlob P, Friedman S et al. Prevalence of minor con-
29. Milunsky A, Alpert E, Kitzmiller JL, Younger MD, Neff RK. genital anomalies in newborns of diabetic mothers. Eur J Obstet
Prenatal diagnosis of neural tube defects. The importance of Gynecol Reprod Biol 1992; 44: 111–116.
serum alpha-feto protein screening in diabetic pregnant women. 37. Simpson JL, Elias S, Martin AO et al. Diabetes in pregnancy,
Am J Obstet Gynecol 1982; 142: 1030–1032. Northwestern University Series (1977–1981), I. Prospective study
30. Ramos-Arroyo MA, Rodriguez-Pinnila E, Cordero JP. Maternal of anomalies in offspring of mothers with diabetes mellitus. Am J
diabetes: The risk of specific birth defects. Eur J Epidemiol 1992; Obstet Gynecol 1983; 146: 263–270.
8: 503–508. 38. Hod M, Merlob P. A meta-analysis of perinatal complications of
31. Lin HJ, Owens TR, Sinow RM et al. Anomalous inferior and supe- maternal diabetes. Can they be prevented? Early Pregnancy Biol
rior venae cavae with oculoauriculovertebral defect: Review of Med 1996; 2: 15–17.
Goldenhar complex and malformations of left-right asymmetry. 39. Kitzmiller JL, Gavin LA, Gin GD et al. Pre-conception care of
Am J Med Genet 1998; 75: 88–94. diabetes. Glycemic control prevents congenital anomalies. JAMA
32. Ewart-Toland A, Yankowitz J, Winder A et al. Oculoauriculo- 1991; 265: 731–736.
vertebral abnormalities in children of diabetic mothers. Am J 40. Correa A, Gilboa SM, Botto L et al. Lack of periconceptional
Med Genet 2000; 90: 303–309. vitamins or supplements that contain folic acid and diabetes
33. Spilson SV, Kim HJE, Chung KC. Association between maternal mellitus-associated birth defects. Am J Obstet Gynecol 2012;
diabetes mellitus and newborn oral clefts. Ann Plast Surg 2001; 206: 218.e1–218.e13.
47: 477–481. 41. Itikala PR, Ruuska SE, Oakley Jr. GP, Kloeblen-Tarver AS, Klein L.
34. Wang R, Martinez-Frias ML, Graham JM. Infants of diabetic Periconceptional intake of folic acid among low-income women.
mothers are at increased risk for the oculo-auriculo-vertebral JAMA 2000; 283: 3074.
sequence: A case-based and case–control approach. J Pediatr 42. Meyer BA, Palmer SM. Pregestational diabetes. Semin Perinatol
2002; 141: 611–617. 1990; 14: 12–23.
40 Diabetic embryopathy in the
preimplantation embryo
Asher Ornoy and Noa Bischitz
In the developing oocyte, the surrounding granulosa

Introduction cells support its growth and provide hormonal supplemen-
Diabetes mellitus can also cause, in addition to birth tation. There are several interactions between the oocyte
defects and/or intrauterine death, decreased fertility in and the granulosa cells such as paracrine signaling and gap
males and females. As maternal diabetes may affect embry- junctional communication. These interactions are critical
onic development at different developmental stages, an for the oocyte differentiation and growth. 3
important question relates to the stage at which the devel- Chang et al.4 suggested that maternal diabetes adversely
oping embryo may be affected. Does the first fault occur at affects preovulatory oocyte maturation, oocyte develop-
the zygote, morula, blastula, or perhaps even earlier, at the ment, and granulosa cell apoptosis in mice. Their main
oocyte phase? hypothesis was that reduction in paracrine communication
This chapter will discuss the influence of maternal diabe- between the granulosa cells and the oocyte in diabetic mice
tes on the embryo at the preimplantation and preorganoge- causes a maturational delay. Indeed, in mice with chronic
netic phase, i.e., the first week postfertilization in rodents and hypoinsulinemia and hyperglycemia (Akita diabetic mice)
the first 2 weeks in man. or in mice with acute hyperglycemia and hypoinsulinemia
The preimplantation period is a very initial process of (induced by streptozotocin [STZ] injection), the preovula-
pregnancy when most women do not know that they are tory follicles and oocytes were small with more apoptotic
pregnant. In the last two decades, due to the increased use follicular cells in comparison with controls. In addition,
of in vitro fertilization, our knowledge on the early develop- there was a delay in meiotic maturation, upregulation in the
mental events in the preimplantation and preorganogenetic expression of cell death signaling proteins, and a decrease
human embryos expanded. In spite of that, most of our in the expression level of key gap junction proteins such as
knowledge still comes from animal studies. Similarly in order connexin-43 that are necessary for communication.
to understand these early events in diabetic pregnancies, Recently, Aktug et al.5 studied the morphology of 279
most studies are still carried out on diabetic animal models. oocytes obtained from STZ-induced diabetic rats compared
It is important to note that insulin and insulin-like growth to 336 oocytes obtained from nondiabetic animals. They
factors (IGFs) seem to be essential for normal growth of the found a lower percentage of mature oocytes and a higher
early embryo in the blastocyst stage as well as for the process percentage of atretic and postmature oocytes among the dia-
of gastrulation.1 Hence, it is not surprising that diabetes may betic rats compared to controls. In addition, a lower in vitro
interfere with early embryonic development. fertilization rate was observed among oocytes from diabetic
Maternal diabetes in rodents influences embryonic animals and a decrease in the expression of beta-catenin and
and fetal development in a very similar manner to that of connexin gene family.
humans.2 Hence, most studies discussed in this chapter were
carried out on diabetic rodents or in in vitro models. We will
describe the effects of maternal diabetes starting from the In conclusion
ovum, gradually continuing to later preimplantation stages, Maternal diabetes, as evidenced from the studies in mice and
and then discuss the proposed mechanisms of action. rats, may interrupt the normal communication of the oocyte
and its surrounding cells increasing apoptosis, thus lead-
ing to abnormal development of the oocyte. This may lead
Changes in oocytes to increased atretic and postmature oocytes and therefore
to reduced fertilization ability. This might, if extrapolated
Most studies on the effects of maternal diabetes on early preg- to man, cause menstrual disturbances, reduced fertilization
nancy focused on the zygote, only few recent studies started rate, different embryonic insults, or very early miscarriage,
to examine the oocytes. as indeed found in diabetic women.5
321
322 Diabetic embryopathy in the preimplantation embryo
Zygote and morula damage is due to the diabetes and not to the possible action
of STZ. In the same line, other studies have shown that STZ-
The mechanism of the developmental delay of the zygote is induced diabetic rats had a 15%–20% decrease in the total cell
not clear. Several studies investigated the development of the number of the ICM, a 20%–25% decrease in the implantation
zygote in diabetic rats. Diabetes was induced by ablation of rate, and a similar reduction in the number of blastocysts.5
the insulin-secreting islet cells by STZ. These studies showed However, the implanted blastocysts seem to be normal, as
delay in early developmental progression of the zygotes and their protein synthesis was similar to that of implanted blas-
reduced implantation rates.6 Other studies, carried out on tocysts from nondiabetic animals.
embryos recovered from non obese diabetic (NOD) diabetic In contrast to the reduction in the number of ICM cells,
or nondiabetic mice 72 hours after fertilization, showed an high glucose concentrations did not change the number of
increased percentage of embryos at the one-cell stage, as these trophoblastic cells.5 Hinck et al.13 studied the differentiation
embryos could not continue their normal dividing process.7 of Rcho-1 rat trophoblastic cells into giant cells in a culture
When these embryos were further cultured in the presence medium supplemented with different concentrations of glu-
of high d-glucose concentrations, severe growth retardation cose. High glucose concentrations increased the number of
was demonstrated. No impairment of growth was observed trophoblastic cells; their nuclei were smaller and contained
following culture in high l-glucose concentrations. Similar more DNA compared to control cells. While the cells cultured
results were obtained when two-cell stage mouse embryos in control culture media increased their progesterone secre-
were obtained from nondiabetic mice but cultured on cul- tion, the addition of high glucose concentrations inhibited
ture media containing high concentrations of d-glucose.8 that increase. Apoptosis was not increased with the addition
Other studies have examined substrate utilization in the of glucose. These morphological and physiological changes
developing mouse embryos. Many zygotes that were incu- (reduction in progesterone secretion) could interfere with
bated in high concentrations of glucose or lactate failed to implantation and may possibly explain the increased rate of
progress to the two-cell stage.8,9 Moreover, it is known that early abortions in man.
mouse embryos begin to utilize glucose for several synthetic Thieme et al.1 recovered rabbit blastocysts 6 days postcopu-
processes at the eight-cell stage,9 and therefore the stages of lation. The rabbits were made diabetic by alloxan with postfeed-
morula and early blastocysts are highly susceptible to high ing blood glucose level ranging between 14 and 25 mmol/L.
glucose concentrations. In the diabetic animals, compared to nondiabetic ones, blas-
It can be concluded that high glucose levels at the earliest tocyst development was delayed and no embryos reached the
developmental stages, starting from 1 to 2 cells, can influ- pregastrulation stage. The authors explain the growth and
ence the future embryonic development and may even stop developmental delay of the diabetic blastocysts by the lack of
cell division and growth. insulin and IGF1 that are needed for normal embryonic devel-
opment. They indeed found in vitro that insulin is required
for Wnt3a, Wnt 4, and Brachyury gene expression and that, in
blastocysts removed from diabetic rabbits at the pregastrula-
Blastocyst tion stage, when only Wnt3a is normally expressed, the expres-
Leunda-Casi et al.10 examined how high levels of glucose sion of this gene was significantly downregulated.
influence the development of mouse blastocysts. Embryonic Cagnone et al.14 cultured in vitro fertilized bovine morulae
exposure to high levels of d-glucose for a short time impaired in control and hyperglycemic culture medium and studied the
trophectoderm differentiation, and the outgrowth of the gene expression profiles in the blastocyst stage. They found by
trophectoderm was increased. This effect was secondary to microarray analysis differences in gene expression between the
a deficiency in fibroblast growth factor-4 (FGF-4) protein embryos cultured on hyperglycemic culture medium and con-
in the inner cell mass (ICM). The addition of FGF-4 to the trols. In the “diabetic” blastocysts, they observed overexpres-
blastocysts pretreated with high glucose normalized tropho- sion of genes involved in metabolic control such as extracellular
blastic growth. The authors conclude that FGF-4 is impor- matrix signaling, calcium signaling, and energy metabolism.
tant in the normal differentiation of the trophoblast and that They also observed upregulation of HIF1 alpha signaling, a find-
these observations could explain some of the morphological ing described by us also in early somite rat embryos cultured in
changes detected in the placentae of diabetic pregnancies.10 diabetic culture medium.15 Hence, they have shown that hyper-
A comparison between blastocysts from STZ-induced glycemia induces at the blastocyst stage significant epigenetic
diabetic to normal rats indicated that the former contained a changes, similar to those described by several investigators dur-
lower number of cell nuclei in the ICM on day 5 postfertiliza- ing later developmental stages in “diabetic” embryos.
tion. The proportion of morulae versus blastocysts was also
different from controls, as the diabetic animals had more In conclusion
embryos persisting in the morula stage in comparison with High glucose increases apoptosis of the ICM and may cause
controls or with rats injected with subdiabetogenic doses embryonic death. Reduction of insulin, such as in type 1 diabetes,
of STZ, and hence not diabetic.11 Treatment of the diabetic may cause embryonic growth delay. Hyperglycemia-induced
rats with insulin during the preimplantation period (day changes in gene expression can be observed already at that
1–6 of pregnancy) normalized embryonic growth and devel- developmental stage. Trophoblastic cells are not diminished in
opment.12 These studies demonstrate that the embryonic number; on the contrary, their number may even increase.
Blastocyst 323
In vitro studies on preimplantation embryos At first, we cultured 2–4-cell mouse embryos for up to
72 hours to the early blastocyst stage in 30% serum from
1. Culture in “diabetic” culture medium: We cultured preim- pregnant women with PGD or with GD and found that a
plantation mouse blastocysts for 72 hours in Roswell Park high proportion of these embryos stopped from further
Memorial Institute (RPMI) medium with the addition of developing within 24–72 hours of culture (Table 40.1).
10% fetal bovine serum and found 20%–24% of embryonic The developmental stage of the living embryos was also
developmental arrest.16 Addition of high concentrations of reduced, but this reduction was significant only after
glucose, acetoacetate, β-hydroxybutyrate, glucagon, and 72 hours of culture (Table 40.1). When we cultured early
insulin were all embryotoxic, inducing a high percentage blastocysts in 50% diabetic serum, about half of the blas-
of embryonic death. However, while the concentrations tocysts died within the first 48 hours of culture, but from
of most substances were much higher than possibly found that stage, all living embryos had spread on the Petri dish
in diabetes, the embryotoxic glucose concentrations were and continued to develop normally. Hence, the develop-
only 300 mg%, concentrations often observed in diabetes. mental stages of the living embryos were not different
Moreover, a combination of β-hydroxybutyrate, acetoac- from that of controls (Table 40.2). There was a negative
etate, and glucose, in relatively low concentrations, was correlation between the degree of diabetic control and the
more embryotoxic than each of the substances individu- extent of damage to the embryos. The sera from women
ally. In addition, 50% serum from STZ-induced diabetic with good control had less deleterious effects on the
rats added to the culture medium caused 53% of embry- zygotes, morulae, or early blastocysts. Sera from women
onic developmental arrest, while 20% and 50% of rat and with GD were also less embryotoxic than sera from
human control serum did not reduce the number of nor- women with PGD.
mally developing embryos. These results imply that human diabetic serum may
2 . Culture in sera from diabetic pregnant women: We cul- stop mouse morulae and/or early blastocysts from fur-
tured, in two sets of experiments,17 mouse embryos on ther development and retard the development of living
RPMI culture medium with 30% or 50% serum obtained morulae. However, these sera do not retard the develop-
from women with pregestational diabetes (PGD) (type 1 ment of living late blastocysts and do not interfere with
or type 2) or gestational diabetes (GD), in comparison their hatching. The developmental stage specificity of the
with similar concentrations of serum from nondiabetic hyperglycemia-induced injuries is apparently related to
women. The results are shown in Tables 40.1 and 40.2. age-specific metabolic changes.
Table 40.1 Effects of 30% serum from diabetic pregnant women on the
developmental stages of two- to four-cell stage mouse embryos cultured for
72 hours
Number of living embryos and their developmental stage

Serum obtained from At beginning At 48 hours At 72 hours
Control 2.0 ± 0.05 (195) 8.1 ± 1.0 (177) 10.5 ± 0.7 (173)
Type 1 diabetes 2.2 ± 0.4 (170) 7.0 ± 0.7 (159) 8.9 ± 0.7a (128)a
Type 2 diabetes 2.3 ± 0.4 (128) 7.6 ± 0.8 (101) 9.2 ± 0.8a (84)a
Gestational diabetes 2.1 ± 0.1 (215) 7.8 ± 1.2 (188) 9.3 ± 0.5a (160)
Note: Values are mean ± SD (number of embryos).
a Significantly lower than control, p < 0.05.
Table 40.2 Effects of 50% serum from diabetic pregnant women on the
developmental stages of mouse blastocysts cultured for 72 hours
Number of living embryos and their developmental stage

Serum obtained from At beginning At 48 hours At 72 hours
Control 7.8 ± 0.6 (105) 10.5 ± 0.7 (98) 11.7 ± 0.7 (97)
Type 1 diabetes 7.6 ± 0.2 (144) 10.6 ± 0.7 (77)a 11.8 ± 0.1 (77)a
Type 2 diabetes 7.4 ± 0.5 (132) 10.6 ± 0.7 (63)a 11.8 ± 0.2 (62)a
Gestational diabetes 7.9 ± 0.5 (183) 10.6 ± 0.4 (125)a 11.9 ± 0.0 (123)a
Note: Values are mean ± SD (number of embryos).
a Significantly lower than control, p < 0.05.
In summary fragmentation. It was shown that these cells contain a large

The observations in diabetic mice and rats show that mater- amount of the clusterin transcripts, a gene associated with
nal diabetes may damage early embryonic development. apoptosis. The overexpression of clusterin in blastocysts of
This might lead to early embryonic death and, if pertinent to STZ-induced diabetic rats indicates that these embryos may
the human situation, to very early miscarriage, even before be affected by subtle disruptions in the expression pattern of
pregnancy is clinically recognized. Whether in such preim- critical developmental genes.20 Similar to the previous study,
plantations very early spontaneous abortions really occur in blastocysts from diabetic rats and mice showed increased
diabetic women, is currently unknown. nuclear chromatin degradation in the ICM cells.21,22 These
and other studies demonstrate that in preimplantation
embryogenesis, hyperglycemia triggers increased apoptosis,
Spontaneous abortions in diabetic especially at the blastocyst stage. Maternal hyperglycemia
women caused a decrease in the expression of facilitative glucose
transporter genes such as GLUT1 (glucose transporter),
If the data in animal models are relevant to humans, it is GLUT2, and GLUT3, which was associated with the reduc-
expected that diabetic women will have an increased rate of tion in glucose transport to the embryo and a decrease in
early miscarriages. Indeed, most human studies observed intraembryonic free glucose.22 This mechanism acts as a sig-
increased rate of spontaneous abortions in women with PGD. nal for cell death that triggers p53 activity. The hyperglyce-
For example, a retrospective sample of 164 pregnancies dur- mia-induced cell death signal increases the expression of the
ing 1956–1975 of 78 insulin-dependent diabetic women was proapoptotic protein BAX, which is a member of the Bcl-2
examined in order to evaluate the risk of clinically recogniz- family.22 In addition, mRNA levels and protein expression
able spontaneous abortions in diabetic women in compari- of BAX were higher in blastocysts cultured in hyperglyce-
son with the normal population.18 This study was done before mic culture medium or in blastocysts obtained from STZ-
the policy of meticulous control of diabetes in pregnancy. The induced diabetic mice.22 In rat blastocysts in culture, the
risk of spontaneous abortions among the diabetic females was transcription of the Bcl-2 gene and the activity of the protein
almost double compared to the n ondiabetic women. Later were also markedly increased in the presence of high glu-
retrospective studies from 1980 to 2000 showed lower differ- cose concentrations in the culture medium.23 These events
ences in the rate of spontaneous abortions between the dia- induced by hyperglycemia lead to the activation of caspases,
betic and the nondiabetic women. Jovanovic et al.19 followed to DNA fragmentation, and morphological changes consis-
389 diabetic and 429 nondiabetic pregnant women. They tent with apoptosis. This cascade (Figure 40.1), connecting
found that the mean pregnancy loss rates were 12% in dia- the expression of Bcl-2 and/or BAX with cell death signals
betic and 13% in normal pregnancies. When they examined involving p53, suggests that the hyperglycemia induces
the glucose blood levels, they found that both nondiabetic embryonic hypoglycemia due to the decrease in the glucose
and diabetic women had high glucose levels. However, the
level of hyperglycemia tolerated in the diabetic pregnancies
before reaching the threshold of fetal sensitivity was higher
than that in nondiabetic pregnancies, implying maternal, Glucose transporter expression
placental, or fetal defenses to the stress of hyperglycemia.
It can be assumed that these significant differences are
Embryonic free glucose
caused by the differences in the control of diabetes, and with
improved glycemic control, the rate of spontaneous abor-
tions is reduced. This is in accordance to the fact that serum Cell death signal
from women with well-treated diabetes was less deleterious
to mouse embryos in culture than serum from women with
poorly controlled diabetes.17 BAX expression
Possible mechanisms of action of diabetes on the

Clusterin expression
early embryo
Several mechanisms were shown to play a role in diabetes-
induced early embryopathy in animals. Elucidation of these Bcl-2 activity
mechanisms of action may help to understand the human
situation. We will, therefore, discuss these mechanisms
mainly based on experimental animal models. Caspase activity
Apoptosis Apoptosis
Hyperglycemia induces excessive cell death in the ICM of Figure 40.1 Hyperglycemia-induced apoptosis. (Modified
rat blastocysts, which was characterized mainly by nuclear from Keim, A.L. et al., Mol. Reprod. Dev., 60, 214, 2001.)
Growth factors and cytokines 325
transport expression and induces the p53 apoptotic cas- question is whether deranged oxidant antioxidant status
cade.22,23 It is important to note that P53 expression is neces- can occur at this early stage of preimplantation embryonic
sary for the induction of apoptosis by hyperglycemia.24 development. We found that serum from diabetic women
Of the different glucose transporters existing in the early can induce oxidative stress in the mouse blastocysts,17 appar-
embryo, GLUT8 was found to be one of the most impor- ently in a way similar to that induced in postimplantation
tant.25 This transporter is regulated by insulin. During early embryos.26 This was evidenced by reduced concentrations of
differentiation of the mouse blastocyst, there is a significant low-molecular-weight antioxidants (LMWA) such as gluta-
increase in glucose demand, and insulin causes GLUT8 to thione and vitamins C and E. The preimplantation mouse
move to the plasma membrane, thus increasing the uptake of embryos cultured in serum from diabetic pregnant women
glucose, which is then converted to lactic acid. It is presumed had a lower concentration of LMWA compared to embryos
that, similar to other glucose transporters, hyperglycemia cultured in serum from nondiabetic women. It seems, there-
decreases GLUT8, hence reducing the uptake of glucose by fore, that diabetic metabolic factors may induce embryo-
the ICM, inducing cell death. toxicity in preimplantation embryos through derangement
Mouse blastocysts, genetically BAX deficient (BAX −/−) of the antioxidant defense mechanism. Leunda-Casi et al.28
obtained from diabetic dams, showed lower chromatin found that hyperglycemia may increase ROS generation, and
degradation and apoptosis than BAX-positive (BAX +/+) this might be one of the reasons for increased cell apoptosis
embryos. Furthermore, the embryos from the BAX-deficient as evidenced in this study by the TUNEL (terminal deoxy-
diabetic mice had lower rates of malformations and resorp- nucleotidyl transferase dUTP Nick End Labeling) tech-
tion on day 14 of pregnancy.22 These data propose that nique. Similar findings were reported by others in mouse
increased apoptosis in the blastocysts might indicate future zygote and blastocysts.29 One of the key mediators that was
increased risk of early embryonic death or malformations, as suggested as the essence for apoptosis is hydrogen peroxide
observed in diabetic pregnancies. (H2O2). Violation of this balance by high glucose concentra-
The second apoptotic compound, Bcl-2, belongs to a fam- tions can cause massive cell damage, increase in apoptotic
ily of proteins that operate in the effector phase of apoptosis events, and defective embryonic development.30 When there
and may either promote or inhibit apoptosis. An increase are high levels of glucose, they need to be degraded, and the
in the expression of the antiapoptotic Bcl-2 mRNA was result is a high production of ketone bodies and an increase
observed in rat blastocysts that were cultured in 28 mM in the production of ROS.
glucose for 24 hours, compared to blastocysts incubated in Fertilization and embryonic development take place in
6 mM glucose.23 When the Bcl-2 expression was inhibited, an environment of low oxygen tension. Oxygen tension is
using antisense oligodeoxynucleotide, there was an increase gradually increasing with advanced gestation, once placen-
of chromatin degradation in blastocysts incubated in high tation is well established and maternal uterine arterioles are
glucose concentrations. The addition of specific inhibitors not obliterated by trophoblastic cells.27,28 Oxidative stress
to caspase-3 and caspase-activated deoxyribonuclease pre- also seems to play an important role in the early phases of
vented the degradation of rat blastocysts.23 embryonic development, and hence antioxidants may play a
A third apoptotic compound, clusterin, a disulfide-linked significant role in preventing damage to the embryos.17,26,30–32
heterodimeric protein associated with the clearance of cel- ROS mediate their action through many of the proinflam-
lular debris and apoptosis, was twice higher in embryos of matory cytokines. They can influence the oocyte, sperm, and
diabetic rats than in control embryos.21 When rat and mouse embryos. During pregnancy, there are increased numbers
blastocysts were incubated with high glucose concentra- of polymorphonuclear leukocytes that increase superoxide
tions, there was an increase in BAX, in clusterin expression, ions. This oxidative stress may regulate the expression of
and in nuclear chromatin degradation.22,23 cytokine receptors in the placenta, cytotrophoblastic cells,
It can be concluded that the significant loss of progenitor vascular endothelial cells, and smooth muscle cells. In addi-
cells from the ICM makes the embryos more sensitive to later tion, several studies have also demonstrated the significant
developmental deficiencies. Furthermore, it is known that role of free radicals in placental function. Oxidative damage
normal embryogenesis can occur only if a sufficient number to the trophoblastic cells early in pregnancy or to the pla-
of functional ICM cells are available.24 Increased apoptosis at centa during the establishment of its maternal circulation
this early stage of development may lead to spontaneous mis- may also cause early pregnancy loss.30–32
carriage or congenital malformations. Figure 40.1 summa-
rizes the steps leading to diabetes induced increased apoptosis.
Growth factors and cytokines
Oxidative stress and antioxidants Cytokines and growth factors play an active role in the nor-
mal implantation process. They also have important roles
Many studies implied that the causes of diabetic embryopa- in the pathogenesis of diabetes-induced organ damage, and
thy may be secondary cellular damage from overproduction those that interrupt the reproductive tract are able to cause
of reactive oxygen species (ROS) or/and decreased anti- preimplantation embryopathy.33–35 There are only few reports
oxidant defense mechanism in the embryonic cells.17,26–28 that refer to cytokine expression in the diabetic uterus.
The source of ROS is complex and nonspecific. The main IGF-2 synthesis is downregulated, and to the contrary,
tumor necrosis factor α (TNF-α) synthesis is upregulated fetuses were found to exhibit the same damage and anoma-
around the implantation sites in diabetic female rats.36,37 lies as observed in the offspring of diabetic animals,39 fur-
Wuu et al.38,39 showed a correlation between reduction in the ther demonstrating the importance of glucose transport to
concentrations of mRNA encoding IGF-2 and embryonic normal early embryonic development.
growth retardation 2 days after initiation of implantation in Downregulation of GLUT1 mRNA and protein expres-
C57B1/6J pregnant mice. However, later observations dem- sion in mouse trophoblasts and in human term placental
onstrated that maternal diabetes did not affect the uterine trophoblasts40–42 were observed after exposure to high
IGF-1 expression. Detection of TNF-α revealed overexpres- glucose concentrations. It has therefore been suggested
sion in the mRNA as well as in the protein concentrations in that the glucose transporter is regulated in the placenta
the preimplantation uterus of STZ-induced diabetic Wistar by glucose and that, as a protective mechanism for the
rats.32 The majority of TNF-α protein synthesis was located embryo, the transporter is being moved from the cell sur-
in the epithelium lining the uterine lumen. Despite normal- face into an intracellular position, thus decreasing glucose
ization of glycemia by addition of insulin to the diabetic ani- uptake in hyperglycemia.43–46 This mechanism seems to
mals, it did not prevent the overproduction of TNF-α in the protect the embryo from the toxic effects of high glucose
uterus.37 Incubation of rat and mouse uterine cells in different concentrations.
glucose concentrations induced the stimulation of TNF-α
secretion in uterine epithelial cells, apparently mediating the Nitric oxide and prostaglandins
release of other cytokines, i.e., interleukin 1β—from the sub- In the preimplantation period, when the trophoblastic
epithelial population of macrophages after their direct acti- cells penetrate the decidua, there is an increase in vascu-
vation by hyperglycemia. There is evidence that high levels lar permeability and other inflammation-like changes. This
of TNF-α in utero can be harmful to the embryonic develop- increase in vascular permeability is related to vasoactive
ment at the implantation phase. Embryos exposed to high agents like nitric oxide (NO) and prostaglandins (PGs),
levels of TNF-α and surviving to term were significantly as both are associated with increased vascular permeabil-
smaller than control embryos.34–40 Mouse embryonic stem ity, vasodilation, and increased blood flow in the uterus.
cells indicate that TNF-α inhibits cell proliferation in the At the beginning of implantation in the rat uterus, there is
ICM lineage and decreased their differentiation potential.35 an increase in the NO synthase activity, synthesizing more
Further evidence for the hypothesis that TNF-α contributes NO, and in the production of PGE, PGF2α, and PGI2.47,48
to the harmful influence of maternal diabetes on preimplan- Moreover, when the synthesis of NO or PGs in the uterus
tation development was found in blastocysts from Wistar is blocked, there is a decrease in the number of implanted
rats. Culture medium was produced from normal and dia- rat embryos. During preimplantation, there is an oversen-
betic uterine cells; the blastocysts of Wistar rats incubated sitivity to metabolic disturbances, and the external envi-
in these media showed diminished growth in the diabetic ronment can easily affect embryonic development. Novaro
medium, but improved significantly (not completely) by pre- et al.48 studied the uterine synthesis and temporal pattern
treating the blastocysts with antisense oligonucleotide that of two vasoactive agents, NO and PGE, both modulating
blocked the embryonic TNF-α receptors.40 the implantation process in rats with non-insulin-depen-
To conclude, there is an increased secretion of TNF-α in dent diabetes mellitus, and found that the activity of NO
diabetic rats and mice at the preimplantation period that synthase in the rat uterus and the production of PG E were
may harm the embryos, interfering with normal embryonic increased in diabetes. In addition, the temporal profile of
growth. A similar mechanism may be responsible for the their activity was different from the control animals, and
increased risk of early embryonic loss in diabetic women. the number of implanted embryos was reduced. However,
the implantation rate was not different between the diabetic
and control animals. While the NO increase was observed
Metabolic factors in different tissues, PGE production was located at the
Glucose metabolism implantation sites and remained high after the implanta-
Glucose supplement in mouse blastocysts can be provided tion. The authors suggest that the increase of PGE synthesis
by the glucose transporters (i.e., GLUT1, GLUT8), by passive in diabetic uterine tissue is a secondary effect. First, there
diffusion, and by an active transport system.24,41,42 In human is an increase in NO production; NO regulates the synthe-
and mouse blastocysts, the uptake of deoxyglucose is very sis of PGs; together they have a positive influence on glu-
active, and its concentrations in the mouse embryos were cose metabolism and on the process of vasodilatation in the
found to be 30 times higher than in the culture medium.42 uterus. The rate and timing of the beginning of implanta-
This transport was blocked by cytochalasin B but not by tion are not altered by diabetes, but diabetes may damage
substances that usually block the active transport of glucose the early blastocysts, reducing the number of embryos that
in the gut or kidney, emphasizing that this mechanism is are able to implant normally.48
unique to the early developing embryo.
Suppression of GLUT1 in preimplantation embryos by Epigenetic changes
antisense GLUT1 is known to produce ICM apoptosis, in a Many studies on embryos, fetuses, and placentae from dia-
way similar to high glucose concentrations. In addition, in betic animals have demonstrated a variety of epigenetic
the GLUT1 transporter–deficient mouse, the embryos and changes, especially changes in gene expression and in DNA
References 327
methylation induced by diabetic maternal environment.15,49 fertility of men with long-standing diabetes, by affecting
Several recent studies have shown that the changes in gene spermatogenesis, seems to be similar to the effects of dia-
expression can already be observed at the earliest stages of betes on the ova.
embryonic development.1,5,14,50 In spite of the fact that these
epigenetic changes (i.e., modifications of histones and DNA
methylation) occur very early in development, many of them
Conclusions
persist and can have transgenerational effects.50 Hence, the As there are practically no data in man except for reduced
different epigenetic changes observed in the early embryos fertility and increased spontaneous abortions in diabetics,
might be responsible for many of the long-term and trans- most of the research was done on animal models. A direct
generational effects commonly found in the offspring of dia- extrapolation of these studies to man must be with great
betic mothers. caution, as it may lead to wrong conclusions. In addition, in
most in vivo animal studies, diabetes was induced by abla-
Effects on male fertility tion of insulin production, and it is possible that some of the
Studies suggest that sexual dysfunction is frequently asso- results stem from how diabetes develops. However, many
ciated with diabetes in men and in experimental animals.51 of the results were obtained also in vitro, including studies
Many of these problems result from diabetes-induced on human diabetic serum that served as culture medium,
changes in the vascular system as well as in the peripheral strengthening the simulation to man. There is a need for
and central nervous system that are related to changes in human additional studies to verify the different proposed
endocrine function.52 Studies on different diabetic male mechanisms of diabetic-induced early embryopathy.
animals have documented abnormalities in testicular func-
tion and spermatogenesis.53 Recently, Aktug et al.5 ana-
lyzed sperm from STZ-induced diabetic rats compared to To summarize
nondiabetic animals. While the sperm number in the dia- Glucose blood levels can influence most processes related
betic rats was markedly reduced, there was no difference in to oocyte maturation, fertilization, and early embryonic
sperm abnormalities. In addition, studies demonstrated a development. Good glycemic control in diabetic women,
reduction in the activity of ATPases and phosphatases in even before pregnancy is diagnosed, can decrease or possi-
the epidermis and in spermatozoa.54 Impotence, infertility, bly prevent most of these effects that, if not prevented, may
and retrograde ejaculation have been described in diabetic have long-term consequences on early and late embryonic
men, but the etiology remains unclear. The reduction in the development.
REFERENCES
1. Thieme R, Ramin N, Fischer S, Puschel B, Fischer B, Santos AN. 11. De Hertogh R, Vanderheyden I, Pampfer S et al. Stimulatory and
Gastrulation in rabbit blastocysts depends on insulin and insulin- inhibitory effects of glucose and insulin on rat blastocysts devel-
like-growth-factor 1. Mol Cell Endocrinol 2012; 348: 112–119. opment in vitro. Diabetes 1991; 40: 641–647.
2. Polanco Ponce AC, Revilla Monsalve MC, Palomino Garibay 12. De Hertogh R, Vanderheyden I, Pampfer S, Robin D,
MA, Islas Andrade S. Effect of maternal diabetes on human and Delcourt J. Maternal insulin treatment improves pre-implanta-
rat fetal development. Ginecol Obstet Mex 2005; 73: 544–552. tion embryo development in diabetic rats. Diabetologia 1992;
3. Ackert CL, Gittens JE, O’Brien MJ, Eppig JJ, Kidder GM. 35: 406–408.
Intercellular communication via connexin 43 gap junctions is 13. Hinck L, Thissen JP, Pampfer S, De Hertogh R. Effect of high con-
required for ovarian folliculogenesis in the mouse. Dev Biol 2001; centrations of glucose on differentiation of rat trophoblast cells
233: 258–270. in vitro. Diabetologia 2003; 46: 276–283.
4. Chang AS, Dale AN, Moley KH. Maternal diabetes adversely 14. Cagnone GL, Dufort I, Vigneault C, Sirard MA. Differential gene
affects preovulatory oocyte maturation, development, and gran- expression profile in bovine blastocysts resulting from hypergly-
ulosa cell apoptosis. Endocrinology 2005; 146: 2445–2453. cemia exposure during early cleavage stages. Biol Reprod 2012;
5. Aktug H, Cetintas VB, Uysal A, Oltulu F, Yavasoglu A, Akarca 86: 1–12.
SO, Kosova B. Evaluation of the effects of STZ—Induced diabe- 15. Ornoy A, Livshitz A, Ergaz Z, Stodgell CJ, Miller RK.
tes on in vitro fertilization and early embryogenesis processes. Hyperglycemia, hypoxia and their combination exert oxidative
J Diabetes Res 2013; 603813. stress and changes in antioxidant gene expression: Studies on cul-
6. Pampfer S. Peri-implantation embryopathy induced by maternal tured rat embryos. Birth Defects Res B Dev Reprod Toxicol 2011;
diabetes. J Reprod Fertil 2000; 55(Suppl.): 129–139. 92: 231–219.
7. Diamond MP, Moley KH, Pellicer A, Vaughn WK, DeCherney 16. Ornoy A, Zusman I. Embryotoxic effects of diabetes on pre-
AH. Effects of streptozotocin- and alloxan-induced diabetes implantation embryos. Isr J Med Sci 1991; 27: 487–492.
mellitus on mouse follicular and early embryo development. 17. Ornoy A, Kimyagarov D, Yeffee P, Abir R, Raz I, Kohen R. Role of
J Reprod Fertil 1989; 86: 1–10. reactive species in diabetes-induced embryotoxicity: Studies on
8. Moley KH, Vaughn WK, DeCherney AH, Diamond MP. Effect of preimplantation mouse embryos cultured in serum from diabetic
diabetes mellitus on mouse pre-implantation embryo develop- pregnant women. Isr J Med Sci 1996; 32: 1066–1073.
ment. J Reprod Fertil 1991; 93: 325–332. 18. Greene MF. Spontaneous abortions and major malformations in
9. Biggers JD, Whittingham DG, Donahue RP. The pattern of energy women with diabetes mellitus. Semin Reprod Endocrinol 1999;
metabolism in the mouse oocyte and zygote. Proc Natl Acad Sci 17: 127–136.
USA 1967; 58: 560–567. 19. Jovanovic L, Knopp RH, Kim H, Cefalu WT, Zhu XD, Lee YJ,
10. Leunda-Casi A, De Hertogh R, Pampfer S. Decreased expres- Simpson JL, Mills JL. Elevated pregnancy losses at high and low
sion of fibroblast growth factor-4 and associated dysregulation of extremes of maternal glucose in early normal and diabetic preg-
trophoblast differentiation in mouse blastocysts exposed to high nancy: Evidence for a protective adaptation in diabetes. Diabetes
D-glucose in vitro. Diabetologia 2001; 44: 1318–1325. Care 2005; 28: 1113–1117.
20. Pampfer S, Vanderheyden I, McCracken JE, Vesela J, De Hertogh 38. Wuu YD, Pampfer S, Becquet P, Vanderheyden I, Lee KH, De
R. Increased cell death in rat blastocysts exposed to maternal Hertogh R. Tumor necrosis factor alpha decreases the viability
diabetes in utero and to high glucose or tumor necrosis factor-α of mouse blastocysts in vitro and in vivo. Biol Reprod 1999; 60:
in vitro. Development 1997; 124: 4827–4836. 479–483.
21. Moley KH, Chi MM, Knudson CM et al. Hyperglycemia induces 39. Wuu YD, Pampfer S, Vanderheyden I, Lee KH, De Hertogh R.
apoptosis in pre-implantation embryos through cell death effec- Impact of tumor necrosis factor alpha on mouse embryonic stem
tor pathways. Nat Med 1998; 4: 1421–1424. cells. Biol Reprod 1998; 58: 1416–1424.
22. Chi MM, Pingsterhouse J, Carayannopoulos M et al. Decrease 40. Pampfer S, Vanderheyden I, Vesela J, De Hertogh R. Neutralization
glucose transporter expression triggers Bax-dependent apoptosis of tumor necrosis factor alpha (TNF alpha) action on cell prolif-
in the murine blastocysts. J Biol Chem 2000; 275: 40252–40257. eration in rat blastocysts by antisense oligodeoxyribonucleotides
23. Pampfer S, Cordi S, Vanderheyden I et al. Expression and role directed against TNF alpha p60 receptor. Biol Reprod 1995; 52:
of Bcl-2 in rat blastocysts exposed to high D-glucose. Diabetes 1316–1326.
2001; 50: 143–149. 41. Gardner DK, Leese HJ. The role of glucose and pyruvate trans-
24. Keim AL, Chi MY, Moley KH. Hyperglycemia-induced apoptotic port in regulating nutrient utilization by preimplantation mouse
cell death in the mouse blastocyst is dependent on expression of embryos. Development 1988; 104: 423–429.
p53. Mol Reprod Dev 2001; 60: 214–224. 42. Chi MM, Manchester JK, Basuray R, Mahendra S, Strickler RC,
25. Carayannopoulus MO, Chi MM, Cui Y et al. Glut 8 is a glucose McDougal DB Jr., Lowry OH. An unusual active hexose transport
transporter responsible for insulin-stimulated glucose uptake in system in human and mouse preimplantation embryos. Proc Natl
the blastocyst. Proc Natl Acad Sci USA 2000; 97: 7313–7318. Acad Sci USA 1993; 90: 10023–10025.
26. Zaken V, Kohen R, Ornoy A. Vitamins C and E improve rat 43. Heilig CW, Saunders T, Brosius FC, Moley K, Heilig K, Baggs R.
embryonic antioxidant defense mechanism in diabetic culture Glucose transporter-1-deficient mice exhibit impaired develop-
medium. Teratology 2001; 64: 33–44. ment and deformities that are similar to diabetic embryopathy.
27. Parchment RE. The implications of a unified theory of pro- Proc Nat Acad Sci USA 2003; 100: 15613–15618.
grammed cell death, polyamines, oxyradicals and histogenesis in 44. Ogura K, Sakata M, Yamaguchi M, Kurachi H, Murata Y. High
the embryo. Int J Dev Biol 1993; 37: 75–83. concentration of glucose decreases glucose transporter-1 expres-
28. Leunda-Casi A, Genicot G, Donnay I, Pampfer S, De Hertogh sion in mouse placenta in vitro and in vivo. J Endocrinol 1999;
R. Increased cell death in mouse blastocysts exposed to high 160: 443–452.
D-glucose in vitro: Implications of an oxidative stress and altera- 45. Hahn T, Barth S, Weiss U, Mosgoeller W, Desoye G. Sustained
tions in glucose metabolism. Diabetologia 2002; 45: 571–579. hyperglycemia in vitro down-regulates the GLUT1 glucose trans-
29. Burton GJ, Hempstock J, Jauniaux E. Oxygen, early embryonic port system of cultured human term placental trophoblast: A mech-
metabolism and free radical-mediated embryopathies. Reprod anism to protect fetal development? FASEB J 1998; 12: 1221–1231.
Biomed Online 2003; 6: 84–96. 46. Hahn T, Hahn D, Blaschitz A, Korgun ET, Desoye G, Dohr G.
30. Agarwal A, Gupta S, Sikka S. The role of free radicals and anti- Hyperglycemia-induced sub-cellular redistribution of GLUT1
oxidants in reproduction. Curr Opin Obstet Gynecol 2006; 18: glucose transporters in cultured human term placental tropho-
325–332. blast cells. Diabetologia 2000; 43: 173–180.
31. Jauniaux E, Watson AL, Hempstock J, Bao YP, Skepper JN, Burton 47. Kennedy TG. Evidence for a role for prostaglandins in the initia-
GJ. Onset of maternal arterial blood flow and placental oxidative tion of blastocyst implantation in the rat. Biol Reprod 1977; 16:
stress. A possible factor in human early pregnancy failure. Am J 286–291.
Pathol 2000; 157: 2111–2122. 48. Novaro V, Jawerbaum A, Faletti A, Gimeno MA, Gonzalez
32. Jauniaux E, Hempstock J, Greenwold N, Burton GJ. Trophoblastic ET. Uterine nitric oxide and prostaglandin E during embryonic
oxidative stress in relation to temporal and regional differences implantation in non-insulin-dependent diabetic rats. Reprod
in maternal placental blood flow in normal and abnormal early Fertil Dev 1998; 10: 217–223.
pregnancies. Am J Pathol 2003; 162: 115–125. 49. Ergaz Z, Guillemin C, Neeman-Azulay M, Fudim L, Stodgell
33. Pankewycz OG, Guan JX, Benedict JF. Cytokines as mediators JC, Miller KR, Szyf M, Ornoy A. Placental oxidative stress and
of autoimmune diabetes and diabetic complications. Endocr Rev decreased global DNA methylation are corrected by copper in
1995; 16: 164–176. the Cohen diabetic rat. Toxicol Appl Pharmacol 2014; 276(3):
34. Chernicky CL, Redline RW, Tan HQ, Gwatkin RB, Johnson TR, 220–230.
Ilan J, Ilan J. Expression of insulin-like growth factors I and II in 50. Chason RJ, Csokmay J, Segars JH, DeCherney AH, Armant DR.
conceptuses from normal and diabetic mice. Mol Reprod Dev Environmental and epigenetic effects upon preimplantation
1994; 37: 382–390. embryo metabolism and development. Trends Endocrinol Metab
35. Banerjee S, Smallwood A, Moorhead J, Chambers AE, 2011; 22: 412–421.
Papageorghiou A, Campbell S, Nicolaides K. Placental expression 51. Amador A, Steger RW, Bartke A, Johns A, Hayashi RH, Stallings
of interferon-gamma (IFN-gamma) and its receptor IFN-gamma R2 MH. Pituitary and testicular function in spontaneously hyperten-
fail to switch from early hypoxic to late normotensive development sive rats. J Androl 1983; 4: 67–70.
in preeclampsia. J Clin Endocrinol Metab 2005; 90: 944–952. 52. Calvo JC, Baranao JL, Tesone M, Charreau EH. Hypothalamic-
36. Pampfer S, Vanderheyden I, Wuu YD, Baufays L, Maillet O, De hypophyseal-gonadal axis in the streptozotocin-induced diabetic
Hertogh R. Possible role for TNF-alpha in early embryopathy male rat. J Steroid Biochem 1984; 20: 769–772.
associated with maternal diabetes in the rat. Diabetes 1995; 44: 53. McVary KT, Rathnau CH, McKenna KE. Sexual dysfunction in the
531–536. diabetic BB/WOR rat: A role of central neuropathy. Am J Physiol
37. Pampfer S, Vanderheyden I, De Hertogh R. Increased synthesis 1997; 272(1 Pt. 2): R259–R267.
of tumor necrosis factor-alpha in uterine explants from pregnant 54. Kuhn-Velten N, Schermer R, Staib W. Effect of streptozotocin-
diabetic rats and in primary cultures of uterine cells in high glu- induced hyperglycaemia on androgen-binding protein in rat
cose. Diabetes 1997; 46: 1214–1224. testis and epididymis. Diabetologia 1984; 26: 300–303.
41 Postimplantation diabetic
embryopathy
Ulf J. Eriksson and Parri Wentzel
to predict the possible future means of (prophylactic) antit-

Introduction eratological treatment of diabetic pregnancy.
Despite increased clinical efforts to improve glycemic con- Several reviews have recently characterized varied
trol during diabetic pregnancy, the rate of congenital mal- aspects of the complex teratological etiology of diabetic
formations remains increased in studies of type 11–12 and pregnancy41–46 and also compared diabetic teratogenesis
type 2 gestational diabetes,7–16 whereas the situation in GDM with other teratogenic conditions.47
pregnancies remains controversial, the consensus leaning
toward no, or marginally increased, occurrence of congenital
malformations.13,17–24 This marginal incidence of congenital First line of teratogenesis:
malformations may be related to the presence of undiag- Diabetes-induced alterations of
nosed type 2 diabetic women among women with diagnosed
gestational diabetes (Table 41.1). maternal metabolites/nutrients
In recent studies, the occurrence of congenital malfor-
Several teratological factors in maternal serum have been
mations was estimated to 4%–6% in type 1 diabetic preg-
suggested, often from clinical experience, and have been
nancy8–10,12 and to 3%–5% in type 2 diabetic pregnancy.8–10,12
subsequently characterized in various experimental sys-
In two recent large meta-studies, it was found that the mal-
tems. The maternal teratogenic factors most often indicated
formation rates in type 1 diabetic pregnancy did not differ
are hyperglycemia and hyperketonemia. Also, alterations
from that of type 2 diabetic pregnancy25,26; both rates were
in maternal triglycerides/fatty acid and (branched-chain)
estimated to be around 5%–6%. The similar rates of malfor-
amino acid levels have been suggested to have teratogenic
mation may relate to the increased age and concomitant adi-
roles, based on data from experimental studies (Table 41.2).
posity in type 2 diabetic women, both of which may increase
the incidence in this group.27–29
The cell biological reason for the teratogenic effect of the Glucose
diabetic state is not known. However, both environmental An increased level of glucose (Figure 41.1) is the hallmark of
factors (maternal diabetic state and intrauterine conditions) the diabetic state, and there is ample clinical evidence that
and genetic predisposition seem to be of importance in dia- increased glucose/HbA1c levels correlate with increased risk
betic embryopathy, i.e., this is an environment–gene interac- of congenital malformation in the offspring.31–34,36,48–50 In
tion situation. The congenital malformations are likely to be experimental studies, the same correlation between increased
induced in early gestation,30–32 and the risk of giving birth serum glucose levels and increased risk for fetal malfor-
to a child with a malformation is increased by an increased mations has been demonstrated in diabetic rodents.51–87 In
degree of maternal dysregulation.33–36 addition, injections of glucose to pregnant, nondiabetic, ani-
It has been suggested that the absence of a specific malfor- mals have also yielded teratological effects.88,89 Furthermore,
mation pattern for diabetic embryopathy signals the presence in vitro culture of rodent embryos in increased glucose con-
of several teratological factors and mechanisms in diabetic centrations,59,68,75,90–106 as well as in diabetic serum,39,40,94,107–114
pregnancy.37 Likewise, the number of different teratogenic yielded disturbed development.
agents identified would indicate that diabetic embryopathy The alleged teratological effect of hyperglycemia is
is of complex etiology.38–40 In the following, we will review likely to be correlated with the metabolism of glucose,
the various etiological suggestions for diabetic embryopathy, since exposure to l-glucose in vitro has no negative devel-
in order to be able to present a reasonable overview of the opmental effect.90 Increased embryonic uptake of glucose
major teratological pathway(s) and, consequently, to be able by existing,115 often upregulated,116,117 glucose transporters
329
330 Postimplantation diabetic embryopathy
CH3
Table 41.1 Malformation rates O
Type 1 DM—5%–6% congenital malformation H OH

OH
Type 2 DM—5%–6% congenital malformation
GDM—Background rate of congenital malformation Figure 41.2 β-Hydroxybutyrate.
major compounds, beta-hydroxybutyrate (Figure 41.2) and

Table 41.2 First line teratogens acetoacetate, traverse the placenta124 and are utilized as a
substrate (acetoacetyl-CoA) in the embryonic mitochondria.
Glucose—Major teratogen; hyperglycemia correlates with There is experimental evidence of the teratogenic role of
congenital malformation risk other diabetes-associated metabolites than increased glucose
Ketone bodies—Teratogenic in experimental studies concentration, thus serum from diabetic rats with normal-
Triglycerides/fatty acids—Associated with increased ized glucose levels is still teratogenic in in vitro cultures,40,113
malformation rate in experimental studies
and the direct exposure to increased ketone body levels
Branched-chain amino acids—Associated with increased in vitro yields disturbed development in cultured rodent
malformation rate in experimental studies
embryos.125–132 The mechanism for ketone body teratogenesis
should also include metabolism of the compounds.133
H OH
HO H
Triglycerides/fatty acids
H The maternal non esterified fatty acids (NEFA) and fatty acids
O
HO from (hydrolyzed) triglycerides (Figure 41.3) cross the pla-
centa and reach the embryo and fetus.124,134 Preimplantation
H OH rabbit embryos exposed to a diabetic environment in vivo or
OH
high glucose and lipid concentration in vitro have increased
Figure 41.1 Glucose. intracellular lipid accumulation and enhanced expression
of key genes for lipid storage, fatty acid transport, metabo-
lism, and lipogenesis.135 Increased levels of triglycerides/
also seems to be necessary for the teratogenic effect. The
fatty acids72,113,136 have been associated with congenital mal-
increased influx of glucose would yield increased glyco-
formations in experimental diabetic pregnancy. In fact, cul-
lytic flux, as well as increased mitochondrial activity of the
ture of rodent embryos in serum of rats given a high-fat diet
citric acid cycle and enhanced oxidative phosphorylation.
yields developmental defects (Wentzel et al., manuscript).
Several consequences of increased glucose metabolism have
The likely mechanism would be associated with the beta-
been suggested,118 where increased mitochondrial produc-
oxidation of the fatty acids, leading to increased flux in the
tion of superoxide may possess the most pronounced tera-
citric acid cycle and enhanced oxidative phosphorylation.
togenic potential. However, increased sorbitol production,
increased hexose monophosphate shunt activity, reactive
oxygen species (ROS)-mediated inhibition of glyceraldehyde- Branched-chain amino acids
3-phosphate dehydrogenase (GAPDH), and increased for- Branched-chain amino acids are transported to the embryo/
mation of reactive alpha-oxoaldehydes may also play roles fetus137 and used as fuel and for protein synthesis. This class
in diabetes-induced embryonic dysmorphogenesis. In addi- of amino acids is increased in the serum of diabetic individ-
tion, increased formation of advanced glycosylation end uals, as well as in the serum of diabetic pregnant rats, where
products (AGEs) and decreased intracellular availability of this increase correlates with embryo dysmorphogenesis.72,138
inositol and arachidonic acid (and its products, the prosta- In vitro culture with the leucine analogue (Figure 41.4), alpha-
glandins) are suspected players in the teratogenic drama of ketoisocaproic acid (KIC), yields disturbed development in
diabetic pregnancy. rat embryos.101 The branched-chain amino acids are metabo-
Of note, the preimplantation embryo follows another lized to citric acid metabolites (acetyl-CoA or succinyl-CoA)
route toward maldevelopment, since a diabetes-induced and utilized for oxidative phosphorylation.
decrease of the inner cell mass 66,119 is likely to be induced
by hypoglycemia, due to the downregulation of glu- O
cose transporters in the blastocyst, leading to decreased
intraembryonic glucose levels120–122 and enhanced H2C O
O
apoptosis.119,123
HC O
O
9 12 15
Ketone bodies H2C O
ω
α
The ketone bodies are synthesized in increased amounts in
the (maternal) liver in response to the diabetic state. The two Figure 41.3 Triglyceride.
Second line of teratogenesis: Diabetes-induced alterations of fetal metabolites/nutrients 331
NH2 CH3
H
Table 41.3 Second line teratogens
CH3 Inositol—Deficiency is teratogenic in experimental studies

O
Sorbitol—Accumulation parallel to hyperglycemia-induced
OH malformations
Arachidonic acid/prostaglandins—Alterations are
Figure 41.4 Leucine. teratogenic in experimental studies
Folate—Deficiency is teratogenic in experimental studies
Combination effect
Culture of rat embryos in a combination of subteratogenic
levels of glucose, beta-hydroxybutyrate, and KIC yields dis- Inositol
turbed in vitro embryonic development, thereby supporting Inositol (Figure 41.6) is taken up by cells via a saturable
the notion of a synergistic relationship between the diabetes- mechanism, which is inhibitable by glucose, in preim-
associated metabolites.101 plantation embryos,151 as well as in adult individuals. This
All of the proximal teratogens identified can serve as has led to the speculation that glucose-induced inositol
metabolic substrates in the embryo (and mother), and the deficiency may be an integral part of the pathogenesis of
most prominent teratogen, glucose, has the most complex diabetic complications,152 as well as of diabetic embryopa-
metabolism involving glycolysis. Furthermore, the metabolic thy.92 The immediate effect of lowered inositol concentra-
pathways of the primary (maternal) teratogens converge at tion would be decreased levels of the phosphoinositides
the citric acid cycle and oxidative phosphorylation in the (PI, PIP, and PIP2) and their products in the embryonic
mitochondrion, which has led to the speculation that an tissue.100 A lack of PIP2 would subsequently yield less IP3
overactivity of the oxidative phosphorylation pathway may and diacylglycerol (DAG), both of which are stimulators
be at the core of diabetic teratogenicity (Figure 41.5). of protein kinase C (PKC) activity. A lowered PKC activity
would exert a number of effects, including lowered activity
of phospholipase A2,153 which would subsequently dimin-
Second line of teratogenesis: ish the availability of free arachidonic acid and thereby
hamper the production and metabolism of prostaglandins,
Diabetes-induced alterations of as discussed later.
fetal metabolites/nutrients Rat embryos cultured in inositol-depleted serum develop
neural tube defects (NTDs),154 which demonstrates the ter-
Major teratogenic processes in embryonic tissues so far identi- atogenic capacity of inositol deficiency. In addition, when
fied include alterations of signaling systems such as the metab- curly tail mouse embryos, who have a predisposition for
olism of inositol,92,97,98,139–141 sorbitol,92,98,140,142,143 arachidonic developing folate-resistant NTDs, are cultured in inositol-
acid/prostaglandins,84,97,105,144–148 and folate (Table 41.3).149,150 depleted serum, their NTD rate increases further.155 If
the inositol-depleted culture medium is successively sup-
plemented with inositol, the NTD rate in the curly tail
embryos decreases proportionally155 and results in a lower
NTD rate than in nontreated curly tail embryos.156 Also,
administering myo-inositol or d-chiro-inositol to curly tail
mice, both in vivo and in vitro, diminishes the NTD/spina
bifida rate.157
High glucose concentration in vitro causes decreased
levels of inositol in the embryo due to impaired uptake,141
yielding an embryonic deficiency of inositol92,100,140,158
concomitant with an increased rate of embryonic dys-
morphogenesis. Supplementation of myo-inositol to high
glucose–cultured embryos,97,98,139 or dietary addition to
diabetic pregnant rodents,67,158–160 diminishes both the
HO H
H OH
HO H
H OH
H
H OH
Figure 41.5 Day 11 embryos from normal (left) and diabetic HO
(right) rats; the latter is malformed (growth retarded, failed rota-
tion, and NTD). Figure 41.6 Inositol.
inositol deficiency and the rate of embryonic maldevelop- Arachidonic acid/prostaglandins

ment. Furthermore, adding the inositol uptake inhibitor Arachidonic acid (Figure 41.8) is a polyunsaturated fatty
scyllo-inositol to the culture medium of rodent embryos causes acid present in the phospholipids of cell membranes. The
similar changes as excess glucose to the embryos, i.e., both metabolism of arachidonic acid and its products, the pros-
inositol deficiency and embryonic maldevelopment.100,161,162 taglandins, is crucial for cellular life. In particular, arachi-
In addition, similarly to the glucose-induced damage, both donic acid is involved in cellular signaling as a lipid second
inositol deficiency and embryo maldevelopment caused by messenger.
scyllo-inositol can be diminished by the addition of inositol Disturbed metabolism of arachidonic acid and prostaglan-
to the culture medium.100,161,162 These findings identify inositol dins has been found in previous studies of experimental dia-
deficiency as a likely component of diabetic teratogenesis.163 betic pregnancy. Intraperitoneal injections of arachidonic acid
However, the addition of antioxidants diminishes both to pregnant diabetic rats diminished the rate of neural tube
glucose-induced and scyllo-inositol-induced embryonic dys- damage,59 as did enriching the diet of the pregnant diabetic
morphogenesis in vitro114,162; therefore inositol deficiency rats with arachidonic acid.160,169,170 The addition of arachidonic
appears to induce embryonic oxidative stress, which, in turn, acid to the culture medium was shown to block the embry-
affects embryonic development; see the following. onic dysmorphogenesis elicited by high glucose concentra-
tion.59,93,105 Addition of PGE2 to the culture medium also
blocked glucose-induced teratogenicity in vitro,97,105 as well
Sorbitol as maldevelopment of embryos cultured in diabetic serum.111
Sorbitol (Figure 41.7) is formed from glucose with the addi- Measurements of PGE2 have indicated that this prostaglandin
tion of aldose reductase (AR) and further metabolized is decreased in embryos of diabetic rodents during neural tube
to fructose by sorbitol dehydrogenase. In the presence of closure,147,171 in high glucose–cultured embryos,147 as well as in
increased intracellular levels of free, nonphosphorylated the yolk sac of embryos of diabetic women.172
glucose, an increased amount of sorbitol is formed, and Previous studies have shown that the uptake of arachi-
accumulated, due to the low capacity of sorbitol dehydroge- donic acid by embryonic yolk sacs is increased in a hyper-
nase. This accumulation of sorbitol is likely to be part of the glycemic environment.145 This finding would preclude an
pathogenesis of diabetic complications,164,165 e.g., neuropa- uptake deficiency of arachidonic acid in the conceptus of
thy, nephropathy,166 and retinopathy, but it has also been diabetic pregnancy, a result supported by the demonstra-
suggested to have a role in diabetic embryopathy. tion of unchanged concentration of arachidonic acid in the
Exposure to a diabetic environment yields enhanced membranes of high glucose–cultured embryos in vitro.173
embryonic formation of sorbitol.92,98,140,142,143 Several studies Measurements in day 12 embryos, however, indicated a
of AR inhibitors in pregnant diabetic animals have man- decreased arachidonic acid concentration in offspring from
aged to lower the increased sorbitol levels, however, with- diabetic rats.158 Disturbances in the availability or metabo-
out diminishing the increased malformation rates.92,98,167 lism of arachidonic acid affect the synthesis of prostaglan-
However, in a recent in vitro study of cultured neural stem dins. Alteration in the activity of the rate-limiting enzyme
cells (NSCs), it was found that high glucose induced ROS cyclooxygenase (COX), which converts arachidonic acid
production, decreased NSC viability and proliferation, and to prostaglandin H2, may be of major importance. There
enhanced mRNA expression of AR.168 Administration of are two isoforms of COX, COX-1 (constitutive) and COX-2
fidarestat, an AR inhibitor, decreased the oxidative stress, (inducible). A glucose-induced downregulation of the gene
restored cell viability and proliferation, and reduced apop- expression of COX-2 and a GSH-dependent decrease of the
totic cell death in NSCs exposed to high glucose levels, sug- conversion of the precursor PGH2 to PGE2 (PGE synthase)
gesting that the activation of the polyol pathway may play have been demonstrated.147 Thus, the PGE2 concentration of
a role in the induction of oxidative stress resulting in dis- day 10 embryos and membranes was decreased after expo-
turbed development of the neural tube in embryos of preg- sure to high glucose in vitro or diabetes in vivo. In vitro
nant animals with diabetes.168 addition of N-acetylcysteine (NAC) to high-glucose cultures
Taken together, the experimental data provide only a
limited support for the role of sorbitol accumulation in dia-
betic embryopathy, and it appears, at the present state of the
research, to be a side phenomenon to the teratogenic pathway.
HO H
H OH OH
HO H
O
HO OH
OH CH3
Figure 41.7 Sorbitol. Figure 41.8 Arachidonic acid.

Second line of teratogenesis: Diabetes-induced alterations of fetal metabolites/nutrients 333
– synthetic form of folate, which is stable and often used in

Arachidonic acid
food fortification. The 5-methyltetrahydrofolate is the active
Glucose form of folate and plays a substantial role in metabolism,
IND – – in vitro particular in rapidly proliferating cells, in both pyrimi-
ASA
COX-2 Diabetes dine and purine synthesis and the formation of methionine
in vivo from homocysteine and the conversion of phenylalanine to
tyrosine.
PGG2
– Oxidative Functional deficiency of folate during early pregnancy
GSH
+
stress may result in NTDs, such as spina bifida, and conse-
GSSG quently, administration of folic acid in the periconceptional
–
period protects the offspring from developing NTDs.174–176
PGH2
NAC Fortification of the U.S. diet with folic acid in 1996 coincided
GSH
PGE synthase + SOD with a reduced incidence of NTDs from 0.38 to 0.31 per 1000
GSSG live births.177 In Canada, folic acid fortification of all cereal
PGE2 grain products started in January 1998, and the NTD inci-
dence has been compared before and after the fortification.
Figure 41.9 Schematic outline of arachidonic acid
In the province of Ontario (336,963 births), the NTD preva-
teratogenesis.
lence decreased from 1.13 to 0.58 per 1000 live births,178 and
in whole Canada (1.9 million births), the decrease was from
restored the PGE2 concentration.147 Hyperglycemia-/dia- 1.58 to 0.86 per 1000 live births.179
betes-induced downregulation of embryonic COX-2 gene It has been suggested that disturbed folate metabolism
expression may be an early event in diabetic embryopathy, is involved in diabetic teratogenesis. This is based on the
leading to lowered PGE2 levels and dysmorphogenesis, pre- increased incidence of NTD in diabetic pregnancy180,181 com-
sumably because this pathway plays an important role in pared with nondiabetic pregnancy. A recent report from the
neural tube development. Antioxidant treatment does not U.S. National Birth Defect Prevention Study (1997–2004)
prevent the decrease in COX-2 mRNA levels but restores suggested that a periconceptional intake of vitamins or
PGE2 concentrations, suggesting that diabetes-induced supplements that contain folic acid may diminish the risk
oxidative stress aggravates the loss of COX-2 activity. From for NTDs in diabetic pregnancy,182 which is in line with a
these data, it may be concluded that decreased availability of previous report from the Atlanta Birth Defects Case-Control
arachidonic acid and the resulting decrease in several pros- Study where diabetic women who took multivitamins dur-
taglandins, in particular PGE2, are likely to be involved in ing the periconceptional period had diminished the risk of
the teratogenicity of diabetic pregnancy.170 having a baby with a birth defect.183 Similarly, in a study
Other studies have shown that a diabetes-like environ- from northern England, diabetic women with no folate
ment decreases embryonic PGE2 concentration146,147,171 in supplement before pregnancy had a doubled risk for a baby
embryonic tissues (Figure 41.9). Affected arachidonic acid with a birth defect.11 However, after adjustment for HbA1c,
metabolism disturbs prostaglandins and embryogenesis.41 the statistical significance of this finding disappeared. Also, a
recent Dutch study reported that periconceptional folic acid
supplement decreases the risk for congenital cardiac mal-
Folate
formation,184 indicating that folate deficiency may hamper
Vitamin B9 (folate) has to be supplied via the diet and is embryonic development in other organs than the neural tube.
essential for numerous cellular functions, such as to synthe- In experimental studies of folic acid administration to
size, repair, and methylate DNA, as well as to act as a cofac- rat embryos exposed to diabetes in vivo and hyperglycemia
tor in several biological reactions. It is therefore needed for in vitro, it was shown that folic acid treatment increased folic
rapid cell division and growth and has an important role in acid concentration in the embryos and almost completely
the closure of the neural tube. Folic acid (Figure 41.10) is a abolished diabetes-/glucose-induced dysmorphogenesis, i.e.,
both the growth retardation and somatic maldevelopment
H2N N N in the embryos.149,150 This beneficial effect of folic acid treat-
ment was thus present both in vivo and in vitro. The possi-
H
N N HO bility that the cellular uptake of folate may be hampered by a
N
H
H
diabetic/hyperglycemic environment was inspired by earlier
O
O N studies where Folbp1–/– embryos display severe morphoge-
netic abnormalities and die in utero by embryonic day 10.185
H
O The cellular uptake of 5-methyltetrahydrofolate is mediated
by clusters of membrane-bound transport proteins (folbps).
In both in vivo and in vitro studies of embryos, the folbp
HO
gene expression was decreased on day 10 and supplementa-
O
tion of folic acid normalized the embryonic mRNA expres-
Figure 41.10 Folic acid. sion of folbp.149
In subsequent experiment, it was reported that folic rat embryos does not block dysmorphogenesis.162 In sub-
acid supplementation to pregnant diabetic mice reduced sequent work, it was shown that a diabetic environment
the malformation rate (NTD, cardiovascular and skeletal in vivo caused a general increase in PKC activity, as well as
anomalies) in fetuses.186 Also, folic acid supplementation, increased DAG levels, concomitant with disturbed embryo-
in conjunction with safflower oil, decreased resorptions and genesis.190 A detailed study of the activity of individual PKC
malformations in diabetic rats.187 It was also reported that isoenzymes in embryos of diabetic rats showed that PKC-
folic acid administration diminishes apoptosis and differ- α, PKC-βI, PKC-γ, PKC-δ, and PKC-ζ were increased.191
entiation of embryonic stem cells (ESCs) exposed to diabe- In addition, the PKC-γ and PKC-δ activities were further
tes188; however, in further experiments it was found that folic increased in malformed embryos.191
acid administration may also change the differentiation of It has been reported that PKC signaling is associated with
neural progenitor cells in embryos of diabetic mice,189 an apoptosis, especially the isoforms PKC-δ202 and PKC-ζ.203
observation that will need further investigation. It has further been suggested that PKC-δ is involved in sta-
The experimental reports suggest that folic acid has a bilizing p53 proteins204 and that it is related to ROS produc-
marked antiteratogenic effect on embryonic development in tion,205 both of which would ultimately lead to apoptotic
a diabetic environment. cell death. Since enhanced apoptosis is a feature of diabetic
embryopathy (see later), one study used addition of specific
inhibitors of PKC-δ and PKC-ζ to high-glucose culture
Diabetes-induced fetal processes with medium and found diminished dysmorphogenesis.192 In a
subsequent study with specific inhibitors of PKC-α, PKC-
suspected teratological impact βII, and PKC-δ added to the high-glucose culture medium,
Several enzyme systems have been reported as disturbed in it was shown that all of these inhibitors diminish embry-
the offspring of diabetic animals, and, consequently, have onic dysmorphogenesis and decrease the glucose-enhanced
been assumed to be part of the teratogenic process. The two activity of their respective PKC isoform.193 In a subsequent
most often reported are varied activation/inhibition of PKC study, inhibition of PKC-βII in high glucose–cultured mouse
isoforms162,190–195 and activation of c-Jun N-terminal kinases embryos diminished the rate of NTD and also decreased
(JNKs) (Table 41.4).196–200 several glucose-induced embryonic effects, such as activa-
tion of Bid, caspase 8 and 3, and cytochrome C release from
mitochondria.194 Furthermore, in a recent study, female mice
Protein kinase C activity heterozygous for PKC-δ (pkcδ+/−) were made diabetic with
PKC is a family of protein kinase enzymes that phosphory- streptozotocin (SZ) and mated with heterozygous males. The
lates hydroxyl groups of serine and threonine amino acids NTD rate in the diabetes-exposed homozygous (pkcδ−/−)
in target proteins. PKC enzymes are activated by increased embryos was markedly diminished compared with the rate
concentration of DAG, or calcium ions (Ca2+). The PKC fam- of diabetes-exposed wild-type (pkcδ+/+) embryos.195 Also,
ily consists of 15 isoenzymes in humans. They are divided the diabetes-exposed pkcδ−/− embryos had lower markers of
into three subfamilies, based on their second messenger oxidative stress, endoplasmic reticulum (ER) stress, as well
requirements: conventional (or classical), novel, and atypi- as normalized expression of phosphorylated JNK (p-JNK).195
cal. Conventional (c)PKCs contain the isoforms α, βI, βII, These findings implicate PKC activity changes in dia-
and γ. These require Ca2+, DAG, and a phospholipid such as betic embryopathy; however, the exact relationships between
phosphatidylserine for activation. Novel (n)PKCs include the maternal diabetes, embryonic PKC isoforms activity, and
δ, ε, η, and θ isoforms and require DAG, but do not require embryonic developmental disturbances will be further
Ca2+ for activation. Thus, conventional and novel PKCs are investigated.
activated through the same signal transduction pathway
as phospholipase C. On the other hand, atypical (a)PKCs
(including protein kinase ζ and ι/λ isoforms) require neither JNK activation
Ca2+ nor DAG for activation. The family of JNKs has three isoforms, JNK1, JNK2, and
Investigations of a possible teratological role for PKC in JNK3, where JNK1 and JNK2 are found in all cells and tis-
diabetic pregnancy initially showed that nonspecific inhi- sues and JNK3 is found in the brain, heart, and testes. JNK
bition of PKC activity in cultured mouse embryos causes is activated by several stress stimuli, e.g., changes in levels of
malformations and growth retardation,201 and that addi- ROS, ultraviolet radiation, inflammatory signals, and pro-
tion of a nonspecific PKC inhibitor to high glucose–cultured tein synthesis inhibitors. JNK phosphorylates and thereby
modifies the activity of several mitochondrial and nuclear
proteins. Downstream molecules that are activated by JNK
Table 41.4 Suspected processes include c-Jun and p53. By activating and inhibiting other
cellular proteins, JNK thus regulates cell growth, differentia-
PKC activity—Alterations in activities correlate with tion, survival, and apoptosis.
malformations in experimental studies
Support for the notion that embryonic activation of
JNK activity—Increased activity correlates with mal JNK enzymes may be a constituent of diabetic embryopa-
formation in experimental studies
thy was initially found in the studies of embryos of diabetic
Teratological processes provoked by diabetes 335
rats where the malformed embryos had increased yolk sac O H

levels of p-JNK1/2 and decreased levels of phosphorylated
ERK1/2 (p-ERK1/2) compared with yolk sacs of nonmal- Figure 41.11 Hydroxyl radical.
formed embryos from diabetic rats and from normal control
embryos.196,197 In later work, this finding has been confirmed normal cellular signaling. ROS are produced through mul-
and expanded.206 Thus, inhibition of JNK with an inhibi- tiple mechanisms, e.g., by NADPH oxidase (NOX) enzymes,
tor reduces embryopathy and yolk sac vasculopathy in high and in mitochondria, where about 1%–2% of electrons pass-
glucose–cultured mouse embryos,198,199 as does targeted ing through the electron transport chain are incompletely
disruption of the JNK2 enzyme where JNK2-KO diabetic reduced and give rise to the superoxide radical (·O2−). The
mice have less malformed offspring than diabetic mice with ROS with the highest capacity to cause cellular damage is
intact JNK2.198 Enhanced JNK activity has subsequently the hydroxyl radical (Figure 41.11), which, once formed, will
been demonstrated in rodent embryos exposed to diabetes alter DNA, RNA, proteins, lipids, or carbohydrates without
in vitro and hyperglycemia in vivo and has been found to being removed by any scavenging enzyme system.
be associated with oxidative stress,206 apoptosis,200,206,207 and The notion that diabetes is associated with oxidative
nitrosative stress.208 stress has been suggested by several authors.217–221 For
JNK activation thus appears to be one component in the instance, increased lipid peroxidation and ROS generation
teratogenic cascade of diabetic embryopathy. were found in diabetic rats, measured as increased serum
8-epi-PGF2a levels222 and increased electron spin clearance
rate.223 Cyclic voltammetric studies have also indicated
Teratological processes provoked by increased levels of lipid peroxidation in diabetic rats, 224 and
the isoprostane 8-epi-PGF2a is increased in embryos exposed
diabetes to high glucose levels in vitro147 and diabetes in vivo.150
Several studies have suggested that diabetic embryopa- Examination of litters of diabetic rats demonstrated lowered
thy is associated with alterations of various signaling a-tocopherol (vitamin E) concentration in day 11 embryos
systems that result in disturbed intracellular conditions, and in the liver of day 20 fetuses.80
such as oxidative stress,73,99,101,209 nitrosative stress, 210,211 The first evidence of an involvement of oxidative stress
ER stress,195,211,212 and hexosamine stress.46,213 In addition, in the pathogenesis of diabetic embryopathy was the
enhanced embryonic apoptosis88,89,104,123,200,207,214–216 has demonstration that treatment of rodent embryos with
been regarded as a component of diabetic embryopathy antioxidative agents largely normalizes increased glucose-
(Table 41.5). induced malformation rates in vitro,99,101 observations that
were repeated102,132,225,226 and extended to in vivo stud-
ies.73,76,77,80,81,105,114,160,227–229 Furthermore, antioxidative treat-
Oxidative stress ment was found to normalize several markers of oxidative
Oxidative stress reflects an imbalance between the produc- stress, such as serum 8-epi-PGF2a levels,222 electron spin
tion of ROS and an ability to detoxify the reactive interme- clearance rate,223 and the concentration of embryonic iso-
diates or to repair the resulting damage. ROS can damage prostanes in vitro147 and in vivo.150 Evidence for diabetes-
all components of the cell, including proteins, lipids, and induced oxidative stress has subsequently been found in
DNA. Some ROS act as cellular messengers in redox sig- several rat models of diabetic pregnancy.230
naling and a state of oxidative stress can therefore disturb Several different compounds with antioxidative proper-
ties have been shown to diminish embryonic maldevelop-
ment resulting from exposure to high glucose levels in vitro
Table 41.5 Teratogenic processes or to a diabetic intrauterine environment in vivo. Thus, add-
ing scavenging enzymes, e.g., superoxide dismutase (SOD),99
Oxidative stress—Confirmed teratogen in experimental catalase,99 or glutathione peroxidase,99 to the culture medium
studies
protects rat embryos from dysmorphogenesis induced by
Hypoxic stress—Condition suggested to precede high glucose concentration in vitro. The antioxidant NAC
oxidative stress
(Figure 41.12) blocks dysmorphogenesis in high glucose–
Isoprostane formation—Confirmed teratogen in
experimental studies cultured rodent embryos105,114,147,162,192,231 and neural crest
cells (NCCs),232,233 and addition of glutathione ester to high-
AGE/RAGE activity—Suspected teratogen in experimental
studies glucose medium diminishes embryonic maldevelopment.102
Nitrosamine stress—Suspected teratogenic process in
experimental studies
SH
ER stress—Suspected teratogenic process in experimental O
studies H
Hexosamine stress—Suspected teratogenic process in O
H3C N
experimental studies H
OH
Apoptosis—Enhanced embryonic process both before and
after implantation
Figure 41.12 N-Acetylcysteine.
Actually, teratogenic concentrations of beta-hydroxybutyr- the apoptotic machinery, and previous studies have sug-
ate or the branched-chain amino acid analog KIC can also gested that an altered apoptotic rate may affect the maldevel-
be blocked by the addition of SOD to the culture medium.101 opment of embryos subjected to a diabetic milieu.104,214
Analogously, dietary supplementations with antioxi- Furthermore, diabetic transgenic mice overexpressing
dative compounds have been shown to diminish diabetic thioredoxin-1 have a lower incidence of malformations and
embryopathy in vivo. Thus, administration of BHT,73 decreased oxidative stress markers than diabetic wild-type
v itamin E,77,80,229 vitamin C,81,229 and folic acid,149 dimin- mice,251 demonstrating a parallel relationship between dys-
ishes the malformation rate in offspring of diabetic rats and morphogenesis and degree of oxidative stress. Also, it has
largely improves embryonic and fetal growth in vivo. Alpha- been suggested that diabetes-induced oxidative stress in
lipoic acid supplementation has been found to diminish the embryos may cause maldevelopment via altered TNF-alpha
high diabetes-induced incidence of resorptions and mal- levels.252
formations in offspring of diabetic rodents.234–236 Embryos In a recent study of the activity of AMP-activated kinase
exposed to a diabetic intrauterine milieu have demonstrated (AMPK) in embryos of hyperglycemic mice, it was dem-
diminished malformation rates after maternal supplemen- onstrated that maternal hyperglycemia stimulated AMPK
tation of NAC.237 Combined supplementation of antioxida- activity and that stimulation of AMPK with 5-aminoimid-
tive compounds, e.g., vitamin E and C228 or folic acid and azole-4-carboxamide-1-beta-4-ribofuranoside (AICAR)
vitamin E, 216 to pregnant diabetic rats also diminished increased the rate of NTD in the embryos.253 In addition,
diabetes-induced dysmorphogenesis. In a study of glucose- stimulation of AMPK by hyperglycemia, hypoxia, or antimy-
induced cardiac malformations in a mouse model, the cin A could be inhibited by antioxidants. The AMPK inhibi-
administered antioxidants diminished all negative effects tor compound C blocked the effects of hyperglycemia or
of hyperglycemia/oxidative stress, such as hampered migra- antimycin A on NTD occurrence, suggesting that diabetes-/
tion and increased apoptosis of NCCs, and prevented out- glucose-induced stimulation of embryonic AMPK activity
flow tract defects.238 is a teratogenic consequence of oxidative stress in diabetic
In addition, pregnant diabetic mice, transgenic for the embryopathy.253 In support of that notion, it was reported in
CuZnSOD gene (SOD1), have fewer malformed offspring a subsequent study that sole stimulation of AMPK disrupts
than their diabetic wild-type mice,209,210,212,239 illustrating the embryonic gene expression and causes NTDs.254
antiteratogenic capacity of increased ROS scavenging activity. The bulk of data implicates oxidative stress and ROS
Embryonic neural tissue subjected to high glucose con- excess as an important component in the etiology of diabetic
centration shows increased superoxide production, as mea- embryopathy. The data also suggest that long-term expo-
sured in a Cartesian diver system.240 One effect of increased sure to high glucose creates embryonic ROS excess either
intracellular ROS production would be inhibition of the from increased ROS production240 or from diminished anti-
rate-limiting enzyme of glycolysis, GAPDH, since this oxidant defense capacity.102,225,255 ROS excess may be small,
enzyme has displayed sensitivity to ROS in several differ- restricted to particular cell populations,256,257 and likely vary
ent conditions of oxidative stress.241 This sensitivity resides with gestational time and nutritional status, making direct
in the thiol group of cysteine residue 149 in the active site ROS determinations difficult.
of the enzyme.242,243 Oxidation of the thiol group by NO Increasing ROS levels in embryos leads to malforma-
or ROS leads to decreased enzyme activity,244 and block- tions,258,259 suggesting that ROS excess may also have a role
ing of this process by antioxidants protects the activity of in the teratogenic process(es) of phenytoin medication,260,261
the enzyme.245 Another mechanism for GAPDH inhibition ethanol abuse,256,257,262 and, possibly, thalidomide adminis-
also results from mitochondrial production of ROS, acti- tration.263 Therefore, ROS excess may constitute a common
vating poly {ADP-ribose} polymerase 1 (PARP 1) by dam- element in a number of teratogenic situations, including dia-
aging DNA. PARP 1, in turn, induces ADP ribosylation of betic pregnancy.264
GAPDH, leading to its inactivation and an accumulation of
metabolites earlier in the metabolism pathway. In line with Hypoxic stress
these considerations, decreased GAPDH activity was found In early organogenesis, oxygen levels are likely to be very low
in rat embryos subjected to a diabetic environment both in the embryonic environment, and excess glucose metabo-
in vivo and in vitro,246 and furthermore, addition of the anti- lism could accelerate the rate of O2 consumption, thereby
oxidant NAC prevented the decrease in activity.246 exacerbating the hypoxic state. Since hypoxia can increase
In addition, fetuses and embryos of diabetic rodents dis- mitochondrial superoxide production, excessive hypoxia
play increased rates of DNA damage,71,86,247 another indica- may contribute to oxidative stress. In a study of O2 availabil-
tion of enhanced ROS activity and damage in the embryonic ity in embryos of glucose-injected hyperglycemic mice, it was
tissues. High-amplitude mitochondrial swelling was demon- found that O2 availability was reduced by 30% in embryos of
strated in the embryonic neuroectoderm of embryos exposed hyperglycemic mice.265 When pregnant hyperglycemic mice
to a diabetic environment,248,249 a swelling diminished by were housed in 12% O2, the NTD rate increased eightfold in
antioxidative treatment of the mother,85 implicating an the offspring. Conversely, housing pregnant hyperglycemic
embryonic ROS imbalance, with conceivable consequences mice in 30% O2 significantly suppressed the effect of mater-
for the rate of apoptosis in susceptible cell lineages in the nal diabetes to increase NTD.265 These observations sug-
embryo.231,250 The mitochondrion has an important role in gest that maternal hyperglycemia depletes O2 in the embryo
and that this contributes to oxidative stress and the adverse the detrimental effects of hyperglycemia in neuropathy and
effects of maternal hyperglycemia on embryo development. nephropathy.271
It has also been suggested that AGE–RAGE activation
Isoprostane formation is involved in diabetic embryopathy. Thus, the embryonic
Isoprostanes, e.g., 8-epi-PGF2a (Figure 41.13), are prostaglandin- formation of glycated proteins96,272,273 has been suggested to
like compounds formed in situ from peroxidation of ara- influence the teratological events in diabetic pregnancy. It has
chidonic acid by nonenzymatic, free radical–catalyzed been shown in rodent embryos cultured in high glucose that
reactions and therefore serve as indicators of lipid peroxi- the levels of the AGE precursor 3-Deoxyglucosone (3-DG)
dation.266–268 These nonclassical eicosanoids possess potent increase in embryonic tissue, and the addition of 3-DG to
biological activity as inflammatory mediators. Also, the for- the culture medium with physiologic concentrations of glu-
mation of isoprostanes may, in itself, consume arachidonic cose induces malformations, an effect that is reversible with
acid and therefore diminish the available pool of arachidonic the addition of SOD.96
acid in the embryo (see earlier). It has been shown that a In a recent study of diabetic embryopathy with RAGE
diabetes-like environment increases isoprostane levels147 in knockout mice, it was found that maternal diabetes induced
embryonic tissues. In addition, supplementation of 8-epi- more fetal resorptions, malformations (facial skeleton, neu-
PGF2a to the culture medium caused malformations in the ral tube), and weight retardation in the wild-type fetuses than
whole embryo culture, thereby illustrating the independent in the RAGE−/− fetuses, despite similar maternal hypergly-
teratogenic activity of isoprostanes.269 Furthermore, add- cemia. In wild-type offspring, maternal diabetes increased
ing SOD or NAC to the culture medium with isoprostane fetal hepatic levels of 8-iso-PGF2α and activated NFκB in
excess normalized almost all morphological and biochemi- the embryos, in contrast to unchanged 8-iso-PGF2α levels
cal parameters, including the elevated tissue concentration and NFκB activity in diabetes-exposed RAGE−/− offspring.
of 8-epi-PGF2a, thereby illustrating the teratogenic potential These findings suggest that RAGE activation and oxidative
of diabetes-/glucose-induced oxidative stress.269 stress are associated phenomenon in diabetic embryopathy
(Ejdesjö et al., submitted).
AGE formation and RAGE activation
The biochemical process of AGE formation, which is accel-
erated in diabetes as a result of chronic hyperglycemia and Nitrosative stress
increased oxidative stress, has been postulated to play a cen- Nitrosamine overproduction, or stress, has been implicated
tral role in the development of diabetic complications.270 in diabetic embryopathy, as a downstream event to oxidative
AGEs are known to accelerate oxidative damage to cells in stress. Thus, in a study of yolk sacs of CuZnSOD-(SOD1)-
a diabetic environment. Examples of AGE-modified sites overexpressing embryos from normal and diabetic mice, it
are carboxymethyllysine (CML), carboxyethyllysine (CEL), was found that the SOD1-transgenic embryos were largely
and argpyrimidine. Under oxidative stress due to hypergly- protected from several of the negative effects of diabetes. 210
cemia in patients with diabetes, AGE formation is increased Thus, diabetes-induced elevated markers of oxidative stress
beyond normal levels. (4-hydroxynonenal and malondialdehyde reductions) were
RAGE, the receptor for AGE, is a transmembrane pat- diminished in the SOD1-transgenic embryos compared
tern recognition receptor. Except for AGEs, RAGE also has to the wild-type embryos. Furthermore, hyperglycemia-
other agonistic ligands: the high-mobility group protein B1 increased iNOS expression and nitrosylated protein were
(HMGB1), S100/calgranulins, amyloid-beta, and MAC-1. also diminished, and caspase-3 and caspase-8 cleavages
The interaction between RAGE and its ligands is thought to were blocked, in the SOD1-transgenic embryos. This find-
result in pro-inflammatory gene activation. Ligand stimula- ing suggests that oxidative stress induces iNOS expression,
tion of RAGE initiates a signaling cascade resulting in acti- nitrosative stress, and apoptosis in diabetic embryopa-
vation of NFκB and activation of NADPH oxidase, thereby thy.210 In another study, diabetic pregnant mice were fed
yielding increased intracellular oxidative stress. via gavage, an inhibitor of nitric oxide (Figure 41.14) (NO)
An increased accumulation of AGE has been found in synthase (NOS) 2, L-N6-(1-iminoethyl)-lysine (L-NIL;
the pathogenesis of several diabetic complications, such 80 mg/kg), once a day from embryonic (E) day 7.5–9.5 dur-
as cataract, retinopathy, atherosclerosis, neuropathy, and ing early stages of neurulation. The treatment significantly
nephropathy. Inhibition or knockout of RAGE attenuates reduced the NTD rate in the embryos, compared with that
in the vehicle (normal saline)-treated diabetic group.211
In addition to alleviation of nitrosative stress, ER stress
OH was also ameliorated, assessed by quantification of associ-
ated factors. Apoptosis was reduced, indicated by caspase
COOH 8 activation. These results show that nitrosative stress is
HO O
OH N
Figure 41.13 8-epi-PGF2a. Figure 41.14 Nitric oxide.

important in diabetes-induced NTDs via exacerbating ER addition, regulate embryonic cell proliferation, differentia-
stress, leading to increased apoptosis.211 tion, and apoptosis. Furthermore, blockage of this process
The combined observations support a role for nitrosative in embryos causes NTDs reminiscent of those observed in
stress in diabetic embryopathy, apparently as a consequence diabetic pregnancies. In a study of NTD induction in dia-
of oxidative stress. betic mice with or without 2%–5% trehalose water, the role
of autophagy was investigated. Maternal diabetes suppressed
autophagy in neuroepithelial cells and altered autophagy-
ER stress related gene expression. Trehalose treatment reversed
ER stress, also named unfolded protein response (UPR), autophagy impairment and prevented NTDs in the embryos
is a cellular stress response related to the ER, which is of diabetic pregnancies. The study demonstrates that mater-
induced by an accumulation of misfolded proteins in the nal diabetes suppresses autophagy in neuroepithelial cells of
ER lumen. The ER stress/UPR diminishes protein transla- the developing neural tube, leading to NTD formation.276
tion, degrades misfolded proteins, and produces chaper- In a recent study of diabetes-induced cardiac malforma-
ones involved in protein folding in order to restore normal tions, it was found that the rate of atrioventricular septal
ER function. If this is not achieved, or the disruption is defects (AVSDs) was increased concomitant with enhanced
prolonged, the ER stress/UPR shifts toward promoting ER stress in embryonic hearts. Blocking of glucose-induced
apoptosis. ER stress with 4-PBA in an endocardial cushion explant
In a study of the effects of maternal diabetes on the devel- culture restored endocardial cell migration. The findings
opment of oocytes and early embryos by using time-lapse suggest that the development of endocardial cushions is sus-
live cell imaging confocal microscopy, the ER displayed an ceptible to the insult of maternal hyperglycemia, and that
increased percentage of homogeneous distribution patterns diabetes-induced ER stress in the developing heart mediates
throughout the entire ooplasm during oocyte maturation the negative effect on endocardial cell migration.277
and early embryo development. In addition, a higher fre- The studies suggest that ER stress, downstream of oxi-
quency of large ER aggregations was detected in oocytes and dative stress, is involved in the teratogenesis of neural and
two cell embryos from diabetic mice. These results suggest cardiac malformations in embryos exposed to diabetes or
that the diabetic condition adversely affects the ER distri- hyperglycemia.
bution pattern during mouse oocyte maturation and early
embryo development.274
In a study of oxidative and ER stress in SOD1- Hexosamine stress
overexpressing mice, it was found that maternal diabetes Increased ambient glucose concentration yields increased
causes increased levels of ER stress markers, e.g., C/EBP uptake, phosphorylation, and metabolism of glucose, pri-
homologous protein (CHOP), calnexin, phosphorylated marily by enhanced flux in the glycolytic pathway and,
(p)-eIF2alpha, p-PERK, and p-IRE1alpha; triggered XBP1 also, in the hexosamine biosynthetic pathway.118 This is
mRNA splicing; and enhanced ER chaperone gene expres- a pathway that converts glucose to uridine diphosphate
sion in wild-type embryos, whereas all these changes were N-acetylglucosamine (UDP-GlcNAc) (Figure 41.15).
blocked in the embryos of diabetic transgenic mice. This Under normoglycemic conditions, approximately
supports the notion that diminishing diabetes-induced oxi- 1%–3% of total glucose consumed by somatic cells is
dative stress, e.g., by SOD1 overexpression, blocks ER stress directed down the hexosamine pathway.278 UDP-GlcNAc
in embryos.212 is the substrate for the majority of glycosylation in the
A downstream effect of the ER stress would be an acti- cell, producing mucopolysaccharides. 279 In this process,
vation of JNK. The possible relationship between JNK1/2 UDP-GlcNAc is attached to serine or threonine residues of
activation and ER stress in diabetic embryopathy was inves- proteins, thus becoming a posttranslational modification
tigated in mice. Maternal diabetes increased ER stress mark- (beta-O-linked glycosylation), which regulates protein
ers and induced swollen/enlarged ER lumens in embryonic function in an analogous manner to phosphorylation. 280
neuroepithelial cells during neurulation. Deletion of both Altered beta-O-linked glycosylation has been associated
JNK1 or JNK2 genes diminished hyperglycemia-increased with a number of disease states, including cancer, inflam-
ER stress markers and ER chaperone gene expression. In high matory conditions, and neurodegenerative diseases. 281
glucose–cultured embryos, the addition of the ER chaperone Notably, it is also implicated as a primary mechanism
4-phenylbutyric acid (4-PBA) diminished ER stress markers
and abolished the activation of JNK1/2 and its downstream
transcription factors, caspase 3 and caspase 8, as well as Sox1 OH
neural progenitor apoptosis. Consequently, 4-PBA blocked HO O
O O O
high glucose–induced NTD in vivo. It was concluded, there-
O NH
fore, that hyperglycemia induces ER stress, which yields the O P O P O
N
HO
activation of proapoptotic JNK1/2 pathway, which yields O HO HO
NH O
induced neural tube apoptosis and thereby NTD.275 HO OH
Autophagy is an intracellular process to degrade dysfunc-
tional proteins and damaged cellular organelles, which, in Figure 41.15 UDP-GlcNAc.
behind the development of insulin resistance and pancre- of diabetes-induced malformation is markedly lower in
atic beta-cell destruction in type 2 diabetes.278,281 diabetic transgenic mice compared with diabetic wild-type
It has been suggested that hexosamine stress may have a mice. Furthermore, the diabetes-induced increased mark-
role in diabetic teratogenesis.46,213,282 Indeed, defect devel- ers for oxidative stress, apoptosis-promoting proteins, and
opment has been demonstrated in preimplantation mouse cleaved caspase-3 production are all diminished in the off-
embryos, treated with glucose (27 mM) or glucosamine spring of diabetic transgenic mice compared with the off-
(0.2 mM) that was added to embryo culture media. Both spring of diabetic wild-type mice.251
treatments disturbed embryo development, increased A new apoptotic pathway was recently suggested to be
apoptosis, and decreased cell number in the resulting blas- activated by a diabetic/hyperglycemic environment, involv-
tocysts.282 Addition of benzyl-2-acetamido-2-deoxy-a- ing the ASK1–FoxO3a–TRADD–caspase 8 gene products.207
d-galactopyranoside (BADGP), an inhibitor of O-linked Blocking components of this pathway diminished the NTD
beta-N-acetylglucosaminyltransferase (OGT), the enzyme rate in diabetic mice. Thus, hyperglycemia-induced apopto-
that adds O-GlcNAc to proteins, rescued all these pheno- sis and the development of NTDs were reduced with genetic
types in the hyperglycemia treatment group, although only blockage of either FoxO3a or Casp8, or by inhibition of ASK1
mild improvement was seen in the glucosamine group.282 by thioredoxin. In addition, examination of human neu-
This may reflect the relative potencies of each hexose in their ral tissues affected by NTDs revealed increased activation
capacity to stimulate UDP-GlcNAc production.278 In another or abundance of the genes in the cascade. The conclusion
study, pregnant mice were injected with glucose to induce was that activation of the ASK1–FoxO3a–TRADD–caspase
hyperglycemia, or glucosamine, to directly activate the 8 pathway participates in the development of NTDs, which
hexosamine pathway. Both treatments increased the NTD could be prevented by inhibiting intermediates in this
rate in the embryos, decreased GSH levels, and increased cascade.207
oxidative stress, as indicated by increased 2,7-dichloro- There is, evidently, strong experimental evidence for the
dihydrofluorescein fluorescence. Glucose and glucosamine role of apoptosis in diabetic embryopathy (Figure 41.16).
also inhibited expression of Pax-3; however, all these effects
were prevented by GSH ethyl ester administration.213
These findings suggest a role for hexosamine stress in dia-
betic embryopathy. Genetics and epigenetics of diabetic dysmorphogenesis
(Table 41.6)
Genetic predisposition
Apoptosis Despite similar teratological exposure, the effect of any
The notion that apoptosis may be a component of the tera- teratogen, including maternal diabetes/hyperglycemia, var-
togenic process of diabetic pregnancy has been studied ies between individuals. In addition to stochastic conditions,
thoroughly. Thus, there are several reports of an increased genetic predisposition determines the effect of each terato-
rate of apoptosis in embryos exposed to a diabetic environ- gen on a particular individual.285,286 Although predisposing
ment.88,89,200,206,207,215,216,231,250,283,284 In particular, there are genetic conditions for diabetes are clearly present in offspring
findings indicating increased apoptotic rate already in pre- of diabetic parents,287,288 as the offspring of a diabetic father
implantation embryos.104,123,214 has a higher risk of developing the disease than the off-
In a study of early postimplantation embryos, the expres- spring of a diabetic mother,289–293 it has been established
sion of Bcl-2 mRNA was decreased, and the number of that diabetic men do not have an increased risk of father-
deoxynucleotidyl transferase–mediated nick end labeling ing malformed offspring.294,295 This indicates that the genes
(TUNEL)-positive cells increased in embryos of diabetic predisposing to diabetes do not induce congenital malforma-
rats compared to control. These results suggest that a Bax- tions. In contrast, maternal diabetes has been suggested to
regulated mitochondrial cytochrome c–mediated caspase-3 be associated with Down’s syndrome296–298 and has also been
activation pathway might be involved in diabetic embryopa- suggested to predispose for optic nerve hypoplasia in female
thy.215 In another early study, it was reported that combined offspring.299 A genetic element may be present in the etiology
supplementation of folic acid and vitamin E to pregnant of diabetic embryopathy,300 a notion supported by experi-
diabetic rats diminished diabetes-induced dysmorphogen- mental data.69,95,209,301–304
esis and normalized apoptotic-associated protein levels.216 The contribution of the fetal genome and maternal
In another study of rodent embryos subjected to high glu- (diabetic) environment was evaluated in a rat model were
cose in vitro or diabetes in vivo, disturbed development was the outcome of diabetic pregnancy in two outbred sub-
found, concomitant with increased activation of caspase 3 strains of Sprague Dawley rats (with low incidence {H} and
and other markers of apoptosis. Supplementation of NAC high incidence {U} of skeletal malformations in the off-
or an apoptosis inhibitor diminished both dysmorphogen- spring) and in F1 hybrids between them.95 The fetuses of
esis and apoptosis.231 Exposure to a diabetic milieu during diabetic H mothers had no skeletal malformations, regard-
organogenesis thus increases dysmorphogenesis and apop- less of embryo type (H/H or H/U). When the diabetic
tosis in embryos. mother was U or from the hybrid strain (H/U) and the off-
In mice, transgenic for the thioredoxin-1 gene with over- spring of the mixed H/U type, increased skeletal malforma-
production of the antioxidant thioredoxin-1, the incidence tion (3%–5%) rates resulted. When the embryos contained a
Diabetes
Glucose
Ketone bodies Glycolytic flux ER stress PKC activity NT stress

Fatty acids/TG
BC amino acids
JNK activity
Arachidonic acid/PG
Glycosylation
Inositol Hexosamine
Sorbitol stress
Mitochondrial activity Hypoxic stress AGE
Oxidative stress
ROS GSH
Anti ox enzymes
Apoptosis
Malformation
Figure 41.16 Schematic outline of the development of diabetic embryopathy. Blue color marks increased activity/amount and
red color decreased activity/amount of compounds or processes (mixed blue–red color marks mixed effect—both increased and
decreased). Note that more interactions between the items are likely to be present than those denoted here and that the epigen-
etic/genetic alterations are not included.
pseudopregnant female recipients and were evaluated at

Table 41.6 Conditions of teratogenesis
embryonic day 14.305 The offspring from diabetic rats had
higher rates of malformations than the controls, and the
Genetic predisposition—Individual teratological susceptibility
conclusion was that exposure to maternal diabetes during
Gene expression—Part of the teratogenic process
oogenesis, fertilization, and the first 24 hours is enough to
Epigenetics—Part of the teratogenic process
program permanently the fetus to develop significant mor-
phological changes.305 In order to characterize the relative
importance of the maternal and paternal genome in relation
major U genome (either U/U or U/(H/U)), further increased to the teratogenicity of the maternal (intrauterine) milieu,
skeletal malformations (17%–19%) were found. These find- rats from two different strains were cross mated. Thus, male
ings indicate that both the maternal and fetal genome are rats from a malformation-prone (L) and a malformation-
involved in the etiology of diabetes-induced (skeletal) mal- resistant (W) strain were mated with diabetic females from
formations in rodent diabetic pregnancy.95 Also, when pre- the other strain to produce F1 offspring—LW (L male × W
implantation embryos were transferred from diabetic NOD female) and WL (W male × L female)—which would be
mice to nondiabetic Institute for Cancer Research (ICR) genetically identical with exception for imprinted genes and
mice and allowed to develop until gestation day 13, as was mitochondrial types. However, the malformation rate was
embryos from the reverse transfer, as well as from an ICR- 0% in the LW and 9% in the WL offspring,138 demonstrating
to-ICR transfer, there were 8/58 (14%), 18/45 (40%), and 0/73 both a teratologic dilution effect and the importance of the
(0%) malformed embryos in the NOD–ICR, ICR–NOD, and maternal environment. Based on metabolic data from the
ICR–ICR transfers, respectively. The result thus suggests two types of pregnancy (indicating a more disturbed meta-
that both the embryo genotype and the maternal environ- bolic state in the diabetic L rats), the study suggested that
ment are of importance for diabetic embryopathy.69 Also, in the fetal genome controls the embryonic dysmorphogenesis
a recent study, one-cell mouse zygotes and blastocysts were in diabetic pregnancy by instigating a threshold level for the
transferred from diabetic or control mice to nondiabetic teratological insult and that the maternal genome controls
the teratogenic insult by (dys)regulating maternal metabo- Pax-3 expression in embryos of diabetic mice could be nor-
lism.138 Furthermore, when the F1 crosses were mated in a malized by treatment of the mother with antioxidants,265
subsequent study, the malformation rates of the F2 combi- thereby demonstrating a coupling between ROS excess and
nations, WL × WLdiabetic and LW × LWdiabetic, were around a teratologically important change in gene expression. In
5%,306 a further dilution of the teratogenic induction; how- a later study, NCCs of Pax3-deficient embryos displayed
ever, the malformed WL × WL offspring only had agnathia/ impaired migration and increased apoptosis. Suppression
micrognathia, whereas the malformed LW × LW offspring of p53, by either null mutation of the p53 gene or admin-
had 60% agnathia/micrognathia and 40% cleft lip and pal- istration of a p53 inhibitor, pifithrin-alpha, prevented the
ate. Thus, despite identical autosomal genotypes, the diabetic defective NCC migration and apoptosis in Pax3-deficient
WL and LW female rats gave birth to offspring with mark- embryos and also restored proper development of cardiac
edly different malformation patterns. This study suggested outflow tracts. Pax3 thus appears to be required for cardiac
a teratological mechanism in diabetic pregnancy influenced outflow tract septation because it blocks p53 expression dur-
by maternal metabolism and parental strain epigenetics.306 ing NCC migration.317
In a previous study, it could be demonstrated that a spe- The embryonic expression of genes controlling the
cific variant of the catalase enzyme is present in rats that defense against oxidative stress is sensitive to maternal
are malformation-prone (Cs-1a), whereas another variant of diabetes. Thus, in a study of pregnant diabetic rats, the
the catalase protein was present in rats that do not develop embryos demonstrated decreased expression of CuZnSOD
malformations in response to maternal diabetes (Cs-1b).303 and MnSOD.327 In addition, the expression of Gpx-1 and
Thus, embryonic catalase activity was lower in embryos Gpx-2 was decreased compared with embryos from normal
from normal U rats than in embryos from normal H rats, rats, enzymes that function in the detoxification of hydrogen
and maternal diabetes augments this difference.304 The cata- peroxide, an important antioxidant system. The decrease in
lase cDNA and the promotor region of the catalase gene in gene expression of Gpx-1 was further enhanced in the mal-
the U and H rat were sequenced307 and yielded one nucleo- formed embryos. The immunostaining of Gpx-1 displayed a
tide mutation in the 5′-UTR-region of the U rat cDNA and general accumulation of positive cells in the cardiac tissue
a heterozygocity in the U rat gene promoter. Therefore, the of all embryos. However, the nonmalformed embryos had
decreased catalase mRNA levels may result from a different less staining than embryos from normal rats and malformed
regulation of transcription (promotor), and the difference in embryos from diabetic rats had almost no staining at all.327
the electrophoretic mobility in zymograms303 may be a result In a study of cardiac malformations in diabetic mouse
of posttranslational modifications of the catalase protein. pregnancy demonstrating dilated heart tube, smaller ven-
Using an inbred Sprague Dawley strain (L) with about tricles, conotruncal stenosis, and abnormal heart loop-
20% skeletal malformations when the mother is diabetic ing, ventricular septal defects were observed and actors in
and inbred Wistar Furth rats (no diabetes-inducible skeletal the TGF-β signaling that regulate heart development were
malformations), a global gene linkage analysis of the skel- downregulated by maternal diabetes. It was concluded that
etal malformations was performed with microsatellites, a TGF-β signaling is involved in cardiac malformations in dia-
study that yielded strong coupling of the malformations to betic embryopathy.311 Also, in a study of global gene expres-
7 regions on chromosomes 4, 10, 14, 18, and 19 and a weaker sion in a transgenic mouse model of caudal dysgenesis and
coupling to 14 other loci in the genome; altogether we found in a pharmacological model using in situ hybridization and
loci on 16 chromosomes. Searching for candidate genes quantitative real-time PCR, altered expression of several
within a distance of 10 cM from each microsatellite yielded molecules that control developmental processes and embry-
18 genes that had been implicated in previous studies of dia- onic growth was noted. The most pronounced finding was
betic embryopathy. These genes were involved in embryonic that of altered Wnt signaling, which suggests that impaired
development/morphogenesis (Map1b, Shh, Tgfb3, Vegfa, signaling in this pathway may be involved in diabetic embry-
Dvl2, Nf1, Gsk3b, Gap43, Tgfbr3, Gdf1, Csf1r),308–314 regula- opathy.313 A genome-wide investigation of gene expression in
tion of DNA/RNA metabolism (En2, Brcc3, Tp53),192,308,315–317 embryos from normal and diabetic mice followed by quanti-
regulation of apoptosis (Nol3, Bak1),308 and cellular metabo- tative RT-PCR yielded several genes with altered expression.
lism (Folr1, Akr1b1).149,168,315 Sequence motifs in the promoters of diabetes-affected genes
suggest potential binding of transcription factors that are
Gene expression involved in responses to oxidative stress and/or to hypoxia,
Altered gene expression in the offspring has been found in and furthermore, around 30% of the diabetes-affected genes
several studies308,311,313–315,318–322 and appears to be an integral encoded transcription factors and chromatin-modifying
component of diabetic embryopathy.323 proteins or components of signaling pathways that affect
The Pax-3 gene expression has been found to be reduced transcription.318 In a genome-wide expression profiling in
in embryos of diabetic mice,88,89 and this transcription fac- the developing heart of embryos from diabetic and control
tor may regulate the gene expression of the licensing factor mice, it was found that a total of 878 genes exhibited more
cdc-46324 and a gene, Dep-1,325 as well as p53,316 all of which than 1.5-fold changes in the expression level in hearts of
may be of importance for a correct neural tube closure. Null experimental embryos compared with their respective con-
mutation of the Pax-3 gene yields the Splotch mouse display- trols. Several genes involved in a number of molecular sig-
ing NTDs.88,326 It has also been shown that the decreased naling pathways such as apoptosis, proliferation, migration,
and differentiation in the developing heart were differen- as well as increased expression of Dcx and Pafah1b1 and
tially expressed in embryos of diabetic pregnancy.322 decreased expression of four microRNAs targeting these
Several different genes and pathways have been demon- genes. This study suggested that hyperglycemia alters the
strated to be affected in diabetic pregnancy; however, there epigenetic mechanisms in NSCs, resulting in the altered
is not, as yet, any universal gene identified to be respon- expression of developmental genes.333 In a genome-wide sur-
sible for enhanced (or decreased) susceptibility to diabetic vey of histone acetylation in neurulation stage embryos from
embryopathy. diabetic mouse pregnancies, it was found that exposure to
maternal diabetes and, independently, exposure to a high-fat
Epigenetics diet were associated with increases and decreases of H3 and
There are considerable indications that epigenetic processes H4 histone acetylation, respectively, in the embryo. These
have a role in diabetic embryopathy.319,328–334 data suggest that epigenetic changes in response to the diet
In a seminal study, it was found that transient hypergly- and metabolic condition may contribute to the increased
cemia induced long-lasting activating epigenetic changes in risk for NTD in diabetic and obese pregnancies.330
the promoter of the NFκB subunit p65 in aortic endothe- In a study of embryos of pregnant hyperglycemic mice
lial cells both in vitro and in nondiabetic mice, resulting in and mouse ESCs, the methylation of a Pax3 CpG island was
increased p65 gene expression. Both the epigenetic changes decreased in embryos and ESCs. Use of shRNA in ESCs
and the gene expression changes persisted for at least 6 days demonstrated that DNA methyltransferase 3b (Dnmt3b)
of subsequent normal glycemia. Furthermore, the hypergly- was responsible for the methylation and silencing of Pax3
cemia-induced epigenetic changes and increased p65 expres- prior to differentiation and by oxidative stress. These results
sion were prevented by reducing mitochondrial superoxide indicate that hyperglycemia-induced oxidative stress stimu-
production or superoxide-induced alpha-oxaldehydes.328 lates Dnmt3b activity, thereby inhibiting chromatin modi-
Analysis of gene expression data from two sets of embryos fications necessary for induction of Pax3 expression and,
of diabetic mice suggested that maternal diabetes may increase thus, providing a molecular mechanism for defects caused
the overall variability of gene expression levels in embryos. by Pax3 insufficiency in diabetic pregnancy.334
The suggestion was that altered gene expression and increased Epigenetic changes in the embryo caused by maternal
variability of gene expression together constitute the molecular diabetes are likely to be transferring the teratogenic input
correlates for the incomplete phenotype penetrance in diabetic of the diabetic environment. Identifying these changes is
pregnancy. Based on this model, it was suggested that maternal important both for the increased knowledge generated and
diabetes reduces the precision of gene regulation in exposed also for the possible antiteratological treatment that may
individuals. Loss of precision in embryonic gene regulation emerge from the identified mechanisms.
may include changes to the epigenome via deregulated expres-
sion of chromatin-modifying factors.319,329
In a study of maternal diabetes effect on DNA methylation Conclusions and future directions
of imprinted genes in oocytes, it was found in SZ-induced dia-
betic and NOD mice that the methylation pattern of Peg3 dif- From the previous discussion, it is evident that diabetic
ferential methylation regions (DMR) was altered, and in the embryopathy has a complex etiology and pathogenesis. The
SZ-induced mice, demethylation was observed on day 35 after studies of etiologic factors in the pathogenesis of congenital
SZ injection. The expression level of DNA methyltransferases malformations have revealed a scenario in which the dia-
(DNMTs) was also decreased in diabetic oocytes. These results betic state simultaneously induces alterations in a series of
indicate that maternal diabetes has adverse effects on the DNA teratogenically capable pathways. These pathways are inter-
methylation of maternally imprinted gene Peg3 in oocytes, but twined, and several of them result in an imbalance of the
also that methylation in the oocytes of the offspring is nor- ROS metabolism, yielding ROS excess in teratogenically
mal.331 Also, the expression was increased and the methylation sensitive cell populations, an imbalance ultimately causing
level of H19 was decreased, whereas the expression and meth- congenital malformations. Blocking the ROS excess may be
ylation levels of Peg3 was completely opposite in placentas of one valid way to diminish the disturbed development caused
diabetic mice. When embryos of normal females were trans- by the diabetic environment (Table 41.7).
ferred to normal/diabetic pseudopregnant females, the meth- Downstream of the oxidative stress, however, several con-
ylation and expression of Peg3 in placentas were also clearly ditions of cellular stress are present, such as nitrosative, ER,
altered in the normal-to-diabetic group compared to the nor- hypoxic, and hexosamine stress, as well as a possible terato-
mal-to-normal group. However, when the embryos of diabetic genic involvement of AGEs. In this area of metabolites and
female were transferred to normal pesudopregnant female
mice, the methylation and expression of Peg3 and H19 in pla-
centas were similar between the two groups. The data suggest Table 41.7 Conclusions and future directions
that the effects of maternal diabetes on imprinted genes may
primarily be caused by the adverse uterus environment.332 What is the current level of understanding of the
teratogenic process of diabetic embryopathy?
In a study, NSCs were exposed to high glucose/hypergly-
cemia and alterations were found in chromatin reorganiza- Are there any putative target molecules/pathways that we
could possibly utilize to prevent malformations?
tion, global histone H3 status, and global DNA methylation,
References 343
pathways, there may be possibilities of finding a molecule to Acknowledgments

use for intervention therapy.
There is also a growing understanding of the diabetes- The authors wish to gratefully acknowledge the support from
induced alterations in genetic and epigenetic systems, which the Swedish Research Council, the Family Ernfors Fund,
will, again, increase our knowledge and, hopefully, inspire to the Novo Nordisk Foundation, and the Swedish Diabetes
develop new ways to block diabetic embryopathy in the future. Association.
REFERENCES
1. Mølsted-Pedersen L, Tygstrups I, Pedersen J. Congenital mal- 21. Aberg A, Westbom L, Kallen B. Congenital malformations among
formations in newborn infants of diabetic women: Correlation infants whose mothers had gestational diabetes or preexisting
with maternal diabetic vascular complications. Lancet 1964; i: diabetes. Early Hum Dev 2001; 61: 85–95.
1124–1126. 22. Correa A et al. Diabetes mellitus and birth defects. Am J Obstet
2. Kucera J. Rate and type of congenital anomalies among offspring Gynecol 2008; 199: 237 e231–e239.
of diabetic women. J Reprod Med 1971; 7: 61–70. 23. Mitanchez D. Foetal and neonatal complications in gestational
3. Mills JL. Malformations in infants of diabetic mothers. Teratology diabetes: Perinatal mortality, congenital malformations, macro-
1982; 25: 385–394. somia, shoulder dystocia, birth injuries, neonatal complications.
4. Ray JG, O’Brien TE, Chan WS. Preconception care and the risk Diabetes Metab 2010; 36: 617–627.
of congenital anomalies in the offspring of women with diabetes 24. Balsells M, Garcia-Patterson A, Gich I, Corcoy R. Major congeni-
mellitus: A meta-analysis. QJM 2001; 94: 435–444. tal malformations in women with gestational diabetes mellitus:
5. Platt MJ, Stanisstreet M, Casson IF et al. St Vincent’s Declaration A systematic review and meta-analysis. Diabetes Metab Res Rev
10 years on: Outcomes of diabetic pregnancies. Diabet Med 2012; 28: 252–257.
2002; 19: 216–220. 25. Balsells M, Garcia-Patterson A, Gich I, Corcoy R. Maternal and
6. Evers IM, de Valk HW, Visser GH. Risk of complications of preg- fetal outcome in women with type 2 versus type 1 diabetes mel-
nancy in women with type 1 diabetes: Nationwide prospective litus: A systematic review and metaanalysis. J Clin Endocrinol
study in the Netherlands. BMJ 2004; 328: 915–924. Metab 2009; 94: 4284–4291.
7. Verheijen EC, Critchley JA, Whitelaw DC, Tuffnell DJ. Outcomes 26. Gizzo S et al. An update on diabetic women obstetrical out-
of pregnancies in women with pre-existing type 1 or type 2 dia- comes linked to preconception and pregnancy glycemic profile:
betes, in an ethnically mixed population. Br J Obstet Gynaecol A systematic literature review. ScientificWorldJournal 2013; 2013:
2005; 112: 1500–1503. 254901.
8. Macintosh MC et al. Perinatal mortality and congenital anomalies 27. Blomberg MI, Kallen B. Maternal obesity and morbid obesity:
in babies of women with type 1 or type 2 diabetes in England, The risk for birth defects in the offspring. Birth Defects Res A Clin
Wales, and Northern Ireland: Population based study. BMJ 2006; Mol Teratol 2010; 88: 35–40.
333: 177. 28. Rankin J et al. Maternal body mass index and congenital anomaly
9. Murphy HR et al. Obstetric and perinatal outcomes in pregnan- risk: A cohort study. Int J Obes (Lond) 2010; 34: 1371–1380.
cies complicated by Type 1 and Type 2 diabetes: Influences of 29. Brite J, Laughon SK, Troendle J, Mills J. Maternal overweight and
glycaemic control, obesity and social disadvantage. Diabet Med obesity and risk of congenital heart defects in offspring. Int J
2011; 28: 1060–1067. Obes (Lond) 2014; 38: 878–882.
10. Knight KM, Thornburg LL, Pressman EK. Pregnancy outcomes in 30. Mills JL, Baker L, Goldman AS. Malformations in infants of
type 2 diabetic patients as compared with type 1 diabetic patients diabetic mothers occur before the seventh gestational week.
and nondiabetic controls. J Reprod Med 2012; 57: 397–404. Implications for treatment. Diabetes 1979; 28: 292–293.
11. Bell R, Glinianaia SV, Tennant PW, Bilous RW, Rankin J. Peri- 31. Ylinen K, Aula P, Stenman UH, Kesaniemi-Kuokkanen T, Teramo
conception hyperglycaemia and nephropathy are associated with K. Risk of minor and major fetal malformations in diabetics with
risk of congenital anomaly in women with pre-existing diabetes: A high haemoglobin A1c values in early pregnancy. Br Med J 1984;
population-based cohort study. Diabetologia 2012; 55: 936–947. 289: 345–346.
12. Vinceti M et al. Risk of birth defects associated with maternal 32. Fuhrmann K, Reiher H, Semmler K, Glöckner E. The effect of
pregestational diabetes. Eur J Epidemiol 2014; 29: 411–418. intensified conventional insulin therapy before and during preg-
13. Schaefer-Graf UM et al. Patterns of congenital anomalies and nancy on the malformation rate in offspring of diabetic mothers.
relationship to initial maternal fasting glucose levels in pregnan- Exp Clin Endocrinol 1984; 83: 173–177.
cies complicated by type 2 and gestational diabetes. Am J Obstet 33. Leslie RDG, Pyke DA, John PN, White JM. Hemoglobin A1 in
Gynecol 2000; 182: 313–320. diabetic pregnancy. Lancet 1978; ii: 958–959.
14. Brydon P et al. Pregnancy outcome in women with type 2 dia- 34. Miller E et al. Elevated maternal hemoglobin A1c in early preg-
betes mellitus needs to be addressed. Int J Clin Pract 2000; 54: nancy and major congenital anomalies in infants of diabetic
418–419. mothers. N Engl J Med 1981; 304: 1331–1334.
15. Dunne F, Brydon P, Smith K, Gee H. Pregnancy in women with 35. Ylinen K, Raivio K, Teramo K. Haemoglobin AIc predicts the
Type 2 diabetes: 12 years outcome data 1990–2002. Diabet Med perinatal outcome in insulin-dependent diabetic pregnancies. Br
2003; 20: 734–738. J Obstet Gynaecol 1981; 88: 961–967.
16. Knight KM, Pressman EK, Hackney DN, Thornburg LL. Perinatal 36. Suhonen L, Hiilesmaa V, Teramo K. Glycaemic control during
outcomes in type 2 diabetic patients compared with non-diabetic early pregnancy and fetal malformations in women with type I
patients matched by body mass index. J Matern Fetal Neonatal diabetes mellitus. Diabetologia 2000; 43: 79–82.
Med 2012; 25: 611–615. 37. Khoury MJ, Becerra JE, Cordero JF, Erickson JD. Clinical-
17. Becerra JE, Khoury MJ, Cordero JF, Erickson JD. Diabetes mel- epidemiologic assessment of pattern of birth defects associated
litus during pregnancy and the risks for specific birth defects: A with human teratogens: Application to diabetic embryopathy.
population-based case-control study. Pediatrics 1990; 85: 1–9. Pediatrics 1989; 84: 658–665.
18. Janssen PA, Rothman I, Schwartz SM. Congenital malformations 38. Sadler TW, Hunter ES, Wynn RE, Phillips LS. Evidence for mul-
in newborns of woman with established and gestational diabetes tifactorial origin of diabetes-induced embryopathies. Diabetes
in Washington State, 1984–91. Paediatr Perinat Epidemiol 1996; 1989; 38: 70–74.
10: 52–63. 39. Zusman I, Yaffe P, Raz I, Bar-On H, Ornoy A. Effects of human
19. Martinez-Frias M, Bermejo E, Rodriguez-Pinilla E, Prieto L, Frias J. diabetic serum on the in vitro development of early somite rat
Epidemiological analysis of outcomes of pregnancy in gestational embryos. Teratology 1989; 39: 85–92.
diabetic mothers. Am J Med Genet 1998; 78: 140–145. 40. Buchanan TA, Denno KM, Sipos GF, Sadler TW. Diabetic terato-
20. Persson B, Hanson U. Neonatal morbidities in gestational diabe- genesis. In vitro evidence for a multifactorial etiology with little
tes mellitus. Diabetes Care 1998; 21(Suppl. 2): B79–B84. contribution from glucose per se. Diabetes 1994; 43: 656–660.
41. Jawerbaum A, Gonzalez E. The role of alterations in arachidonic 65. Chartrel NC, Clabaut MT, Boismare FA, Schrub JC. Uteroplacental
acid metabolism and nitric oxide homeostasis in rat models hemodynamic disturbances in establishment of fetal growth
of diabetes during early pregnancy. Curr Pharm Des 2005; 11: retardation in streptozotocin-induced diabetic rats. Diabetes
1327–1342. 1990; 39: 743–746.
42. Miller RR. Hyperglycemia-induced oxidative-stress, apoptosis, 66. Pampfer S, de Hertogh R, Vanderheyden I, Michiels B, Vercheval
and embryopathy. J Pediatr Biochem 2011; 1: 309–324. M. Decreased inner cell mass proportion in blastocysts from dia-
43. Jawerbaum A, White V. Animal models in diabetes and preg- betic rats. Diabetes 1990; 39: 471–476.
nancy. Endocr Rev 2010; 31: 680–701. 67. Akashi M et al. Effects of insulin and myo-inositol on embryo
44. Castori M. Diabetic embryopathy: A developmental perspective growth and development during early organogenesis in strepto-
from fertilization to adulthood. Mol Syndromol 2013; 4: 74–86. zotocin-induced diabetic rats. Diabetes 1991; 40: 1574–1579.
45. Zhao Z, Reece EA. New concepts in diabetic embryopathy. Clin 68. Moley KH, Vaughn WK, DeCherney AH, Diamond MP. Effect of
Lab Med 2013; 33: 207–233. diabetes mellitus on mouse pre-implantation embryo develop-
46. Frank LA, Sutton-McDowall ML, Gilchrist RB, Thompson JG. The ment. J Reprod Fertil 1991; 93: 325–332.
effect of peri-conception hyperglycaemia and the involvement of 69. Otani H, Tanaka O, Tatewaki R, Naora H, Yoneyama T. Diabetic
the hexosamine biosynthesis pathway in mediating oocyte and environment and genetic predisposition as causes of congeni-
embryo developmental competence. Mol Reprod Dev 2014; 81: tal malformations in NOD mouse embryos. Diabetes 1991; 40:
391–408. 1245–1250.
47. Zhu H, Kartiko S, Finnell RH. Importance of gene-environment 70. De Hertogh R, Vanderheyden I, Pampfer S, Robin D, Delcourt
interactions in the etiology of selected birth defects. Clin Genet J. Maternal insulin treatment improves pre-implantation embryo
2009; 75: 409–423. development in diabetic rats. Diabetologia 1992; 35: 406–408.
48. Rose BI, Graff S, Spencer R, Hensleigh P, Fainstat T. Major con- 71. Lee AT, Plump A, DeSimone C, Cerami A, Bucala R. A role for
genital anomalies in infants and glycosylated hemoglobin lev- DNA mutations in diabetes-associated teratogenesis in transgenic
els in insulin-requiring diabetic mothers. J Perinatol 1988; 8: embryos. Diabetes 1995; 44: 20–24.
309–311. 72. Styrud J, Thunberg L, Nybacka O, Eriksson UJ. Correlations
49. Mills JL et al. Lack of relation of increased malformation rates in between maternal metabolism and deranged development in
infants of diabetic mothers to glycemic control during organo- the offspring of normal and diabetic rats. Pediatr Res 1995; 37:
genesis. N Engl J Med 1988; 318: 671–676. 343–353.
50. Hanson U, Persson B, Thunell S. Relationship between haemo- 73. Eriksson UJ, Simán CM. Pregnant diabetic rats fed the antioxidant
globin A1c in early type 1 (insulin-dependent) diabetic pregnancy butylated hydroxytoluene show decreased occurrence of malfor-
and the occurrence of spontaneous abortion and fetal malforma- mations in the offspring. Diabetes 1996; 45: 1497–1502.
tion in Sweden. Diabetologia 1990; 33: 100–104. 74. Moley K, Chi M, Manchester J, McDougal D, Lowry O.
51. Kim JN, Runge W, Wells LJ, Lazarow A. Effects of experimental Alterations of intraembryonic metabolites in preimplantation
diabetes on the offspring of the rat. Fetal growth, birth, weight, mouse embryos exposed to elevated concentrations of glucose:
gestation period and fetal mortality. Diabetes 1960; 9: 396–404. A metabolic explanation for the developmental retardation seen
52. Lazarow A, Kim JN, Wells LJ. Birth weight and fetal mortality in in preimplantation embryos from diabetic animals. Biol Reprod
pregnant subdiabetic rats. Diabetes 1960; 9: 114–117. 1996; 54: 1209–1216.
53. Endo A. Teratogenesis in diabetic mice treated with alloxan prior 75. Ornoy A et al. Role of reactive oxygen species in diabetes-
to conception. Arch Environ Health 1966; 12: 492–500. induced embryotoxicity: Studies on pre-implantation mouse
54. Deuchar EM. Embryonic malformation in rats, resulting from embryos culture in serum from diabetic pregnant women. Isr J
maternal diabetes: Preliminary observations. J Embryol Exp Med Sci 1996; 32: 1066–1073.
Morphol 1977; 41: 93–99. 76. Sivan E et al. Dietary vitamin E prophylaxis and diabetic embry-
55. Ellington SKL. In vivo and in vitro studies on the effects of mater- opathy: Morphologic and biochemical analysis. Am J Obstet
nal fasting during embryonic organogenesis in the rat. J Reprod Gynecol 1996; 175: 793–799.
Fertil 1980; 60: 383–388. 77. Viana M, Herrera E, Bonet B. Teratogenic effects of diabetes mel-
56. Eriksson UJ, Dahlstrom E, Larsson KS, Hellerstrom C. Increased litus in the rat. Prevention with vitamin E. Diabetologia 1996;
incidence of congenital malformations in the offspring of diabetic 39: 1041–1046.
rats and their prevention by maternal insulin therapy. Diabetes 78. Hunter SK, Wang Y, Weiner CP, Niebyl J. Encapsulated beta-islet
1982; 31: 1–6. cells as a bioartificial pancreas to treat insulin-dependent diabe-
57. Brownscheidle M, Wootten V, Mathieu MH, Davis DL, Hofman tes during pregnancy. Am J Obstet Gynecol 1997; 177: 746–752.
IA. The effects of maternal diabetes on fetal maturation and neo- 79. Malaisse-Lagae F, Vanhoutte C, Rypens F, Louryan S, Malaisse
natal health. Metabolism 1983; 32(Suppl. 1): 148–155. WJ. Anomalies of fetal development in GK rats. Acta Diabetol
58. Funaki K, Mikamo K. Developmental-stage-dependent terato- 1997; 34: 55–60.
genic effects of maternal spontaneous diabetes in the Chinese 80. Simán CM, Eriksson UJ. Vitamin E decreases the occurrence of
hamster. Diabetes 1983; 32: 637–643. malformations in the offspring of diabetic rats. Diabetes 1997;
59. Goldman AS et al. Hyperglycemia-induced teratogenesis is medi- 46: 1054–1061.
ated by a functional deficiency of arachidonic acid. Proc Natl 81. Simán CM, Eriksson UJ. Vitamin C supplementation of the mater-
Acad Sci USA 1985; 82: 8227–8231. nal diet reduces the rate of malformation in the offspring of dia-
60. Eriksson UJ, Dahlstrom VE, Lithell HO. Diabetes in pregnancy: betic rats. Diabetologia 1997; 40: 1416–1424.
Influence of genetic background and maternal diabetic state 82. Sivan E, Lee Y, Wu Y, Reece E. Free radical scavenging enzymes
on the incidence of skeletal malformations in the fetal rat. Acta in fetal dysmorphogenesis among offspring of diabetic rats.
Endocrinol Suppls (Copenh) 1986; 277: 66–73. Teratology 1997; 56: 343–349.
61. Giavini E, Broccia ML, Prati M, Roversi GD, Vismara C. Effects of 83. Torchinsky A, Toder V, Carp H, Orenstein H, Fein A. In vivo evi-
streptozotocin-induced diabetes on fetal development of the rat. dence for the existence of a threshold for hyperglycemia-induced
Teratology 1986; 34: 81–88. major fetal malformations: Relevance to the etiology of diabetic
62. Zusman I, Yaffe P, Ornoy A. The effects of high-sucrose diets and teratogenesis. Early Pregnancy 1997; 3: 27–33.
of maternal diabetes on the ultrastructure of the visceral yolk sac 84. Jawerbaum A et al. Increased prostaglandin E generation and
endoderm in rat embryos developing in vivo and in vitro. Acta enhanced nitric oxide synthase activity in the non-insulin-
Anat 1987; 128: 11–18. dependent diabetic embryo during organogenesis. Reprod Fertil
63. Diamond MP, Moley KH, Pellicer A, Vaughn WK, DeCherney Dev 1998; 10: 191–196.
AH. Effects of streptozotocin- and alloxan-induced diabetes 85. Yang X, Borg LAH, Simán CM, Eriksson UJ. Maternal antioxi-
mellitus on mouse follicular and early embryo development. dant treatments prevent diabetes-induced alterations of mito-
J Reprod Fertil 1989; 86: 1–10. chondrial morphology in rat embryos. Anat Rec 1998; 251:
64. Eriksson UJ, Bone AJ, Turnbull DM, Baird JD. Timed interruption 303–315.
of insulin therapy in diabetic BB/E rat pregnancy: Effect on mater- 86. Lee AT, Reis D, Eriksson UJ. Hyperglycemia induced embryonic
nal metabolism and fetal outcome. Acta Endocrinol (Copenh) dysmorphogenesis correlates with genomic DNA mutation fre-
1989; 120: 800–810. quency in vitro and in vivo. Diabetes 1999; 48: 371–376.
References 345
87. Jawerbaum A et al. Diminished PGE2 content, enhanced PGE2 110. Mulder EJ et al. Whole rat embryo culture in serum from insu-
release and defects in 3H-PGE2 transport in embryos from lin-dependent (Type 1) diabetic women. Toxicol Vitro 1989; 3:
overtly diabetic rats. Reprod Fertil Dev 2000; 12: 141–147. 221–226.
88. Phelan SA, Ito M, Loeken MR. Neural tube defects in embryos 111. Goto MP, Goldman AS, Uhing MR. PGE2 prevents anomalies
of diabetic mice: Role of the Pax-3 gene and apoptosis. Diabetes induced by hyperglycemia or diabetic serum in mouse embryos.
1997; 46: 1189–1197. Diabetes 1992; 41: 1644–1650.
89. Fine EL, Horal M, Chang TI, Fortin G, Loeken MR. Evidence that 112. Styrud J, Eriksson UJ. Development of rat embryos in culture
elevated glucose causes altered gene expression, apoptosis, and media containing different concentrations of normal and diabetic
neural tube defects in a mouse model of diabetic pregnancy. rat serum. Teratology 1992; 46: 473–483.
Diabetes 1999; 48: 2454–2462. 113. Wentzel P, Eriksson UJ. Insulin treatment fails to abolish the tera-
90. Cockroft DL, Coppola PT. Teratogenic effect of excess glucose on togenic potential of serum from diabetic rats. Eur J Endocrinol
head-fold rat embryos in culture. Teratology 1977; 16: 141–146. 1996; 134: 459–466.
91. Sadler TW. Effects of maternal diabetes on early embryogenesis. 114. Wentzel P, Thunberg L, Eriksson UJ. Teratogenic effect of diabetic
II. Hyperglycemia-induced exencephaly. Teratology 1980; 21: serum is prevented by supplementation of superoxide dismutase
349–356. and N-acetylcysteine in rat embryo culture. Diabetologia 1997;
92. Hod M, Star S, Passonneau JV, Unterman TG, Freinkel N. Effect of 40: 7–14.
hyperglycemia on sorbitol and myo-inositol content of cultured 115. Takao Y et al. Glucose transporter gene expression in rat con-
rat conceptus: Failure of aldose reductase inhibitors to modify ceptus during high glucose culture. Diabetologia 1993; 36:
myo-inositol depletion and dysmorphogenesis. Biochem Biophys 696–706.
Res Commun 1986; 140: 974–980. 116. Trocino RA et al. Cellular-tissue localization and regulation of the
93. Pinter E et al. Arachidonic acid prevents hyperglycemia- Glut-1 protein in both the embryo and the visceral yolk sac from
associated yolk sac damage and embryopathy. Am J Obstet normal and experimental diabetic rats during the early postim-
Gynecol 1986; 155: 691–702. plantation period. Endocrinology 1994; 134: 869–879.
94. Ornoy A, Zusman I, Cohen AM, Shafrir E. Effects of sera from 117. Li R, Thorens B, Loeken MR. Expression of the gene encoding
Cohen, genetically determined diabetic rats, streptozotocin dia- the high-Km glucose transporter 2 by the early postimplantation
betic rats and sucrose fed rats on in vitro development of early mouse embryo is essential for neural tube defects associated with
somite rat embryos. Diabetes Res 1986; 3: 43–51. diabetic embryopathy. Diabetologia 2007; 50: 682–689.
95. Eriksson UJ. Importance of genetic predisposition and maternal 118. Brownlee M. Biochemistry and molecular cell biology of diabetic
environment for the occurrence of congenital malformations in complications. Nature 2001; 414: 813–820.
offspring of diabetic rats. Teratology 1988; 37: 365–374. 119. Pampfer S, Vanderheyden I, McCracken JE, Vesela J, De Hertogh
96. Eriksson UJ, Wentzel P, Minhas HS, Thornalley PJ. Teratogenicity R. Increased cell death in rat blastocysts exposed to maternal dia-
of 3-deoxyglucosone and diabetic embryopathy. Diabetes 1998; betes in utero and to high glucose or tumor necrosis factor-alpha
47: 1960–1966. in vitro. Development 1997; 124: 4827–4836.
97. Baker L, Piddington R, Goldman A, Egler J, Moehring J. Myo- 120. Moley KH. Diabetes and preimplantation events of embryogen-
inositol and prostaglandins reverse the glucose inhibition of neu- esis. Semin Reprod Endocrinol 1999; 17: 137–151.
ral tube fusion in cultured mouse embryos. Diabetologia 1990; 121. Chi MM, Pingsterhaus J, Carayannopoulos M, Moley KH.
33: 593–596. Decreased glucose transporter expression triggers BAX-
98. Hashimoto M et al. Effects of hyperglycaemia on sorbitol and dependent apoptosis in the murine blastocyst. J Biol Chem 2000;
myo-inositol contents of cultured embryos: Treatment with 275: 40252–40257.
aldose reductase inhibitor and myo-inositol supplementation. 122. Wang Q, Chi MM, Schedl T, Moley KH. An intercellular pathway
Diabetologia 1990; 33: 597–602. for glucose transport into mouse oocytes. Am J Physiol Endocrinol
99. Eriksson UJ, Borg LAH. Protection by free oxygen radical scav- Metab 2012; 302: E1511–E1518.
enging enzymes against glucose-induced embryonic malforma- 123. Pampfer S. Apoptosis in rodent peri-implantation embryos:
tions in vitro. Diabetologia 1991; 34: 325–331. Differential susceptibility of inner cell mass and trophectoderm
100. Strieleman PJ, Connors MA, Metzger BE. Phosphoinositide cell lineages—A review. Placenta 2000; 21: S3–S10.
metabolism in the developing conceptus. Effects of hyperglyce- 124. Herrera E. Lipid metabolism in pregnancy and its consequences
mia and scyllo-inositol in rat embryo culture. Diabetes 1992; 41: in the fetus and newborn. Endocrine 2002; 19: 43–55.
989–997. 125. Sadler TW, Horton WEJ. Effects of maternal diabetes on early
101. Eriksson UJ, Borg LA. Diabetes and embryonic malformations. embryogenesis: The role of insulin and insulin therapy. Diabetes
Role of substrate-induced free-oxygen radical production for 1983; 32: 1070–1074.
dysmorphogenesis in cultured rat embryos. Diabetes 1993; 42: 126. Horton WEJ, Sadler TW. Effects of maternal diabetes on early
411–419. embryogenesis. Alterations in morphogenesis produced by the
102. Trocino RA et al. Significance of glutathione depletion and oxida- ketone body B-hydroxybutyrate. Diabetes 1983; 32: 610–616.
tive stress in early embryogenesis in glucose-induced rat embryo 127. Lewis NJ, Akazawa S, Freinkel N. Teratogenesis from beta-
culture. Diabetes 1995; 44: 992–998. hydroxybutyrate during organogenesis in rat embryo organ cul-
103. Ellington S. Effects of excess glucose on mammalian post-implan- ture and enhancement by subteratogenic glucose. Diabetes 1983;
tation embryos. Int J Dev Biol 1997; 41: 299–306. 32(Suppl. 1): 11A.
104. Moley K, Chi M, Knudson C, Korsmeyer S, Mueckler M. 128. Hunter ES, Sadler TW, Wynn RE. A potential mechanism of
Hyperglycemia induces apoptosis in pre-implantation dl-beta-hydroxybutyrate-induced malformations in mouse
embryos through cell death effector pathways. Nat Med 1998; embryos. Am J Physiol 1987; 253: E72–E80.
4: 1421–1424. 129. Shum L, Sadler TW. Embryonic catch-up growth after exposure
105. Wentzel P, Eriksson UJ. Antioxidants diminish developmental to the ketone body d,l,-beta-hydroxybutyrate in vitro. Teratology
damage induced by high glucose and cyclooxygenase inhibitors 1988; 38: 369–379.
in rat embryos in vitro. Diabetes 1998; 47: 677–684. 130. Moore DCP, Stanisstreet M, Clarke CA. Morphological and phys-
106. Moley KH, Chi MM, Mueckler MM. Maternal hyperglycemia iological effects of beta-hydroxybutyrate on rat embryos grown
alters glucose transport and utilization in mouse preimplantation in vitro at different stages. Teratology 1989; 40: 237–251.
embryos. Am J Physiol 1998; 275: E38–E47. 131. Shum L, Sadler TW. Biochemical basis for d,l,beta-hydroxybutyrate-
107. Sadler TW. Effects of maternal diabetes on early embryogenesis. induced teratogenesis. Teratology 1990; 42: 553–556.
I. The teratogenic potential of diabetic serum. Teratology 1980; 132. Ornoy A, Zaken V, Kohen R. Role of reactive oxygen species
21: 339–347. (ROS) in the diabetes-induced anomalies in rat embryos in vitro:
108. Sadler TW, Phillips LS, Balkan W, Goldstein S. Somatomedin Reduction in antioxidant enzymes and low-molecular-weight
inhibitors from diabetic rat serum alter growth and development antioxidants (LMWA) may be the causative factor for increased
of mouse embryos in culture. Diabetes 1986; 35: 861–865. anomalies. Teratology 1999; 60: 376–386.
109. Rashbass P, Ellington SKL. Development of rat embryos cultured 133. Yang X, Borg LAH, Eriksson UJ. Metabolic alterations in neural
in serum prepared from rats with streptozotocin-induced diabe- tissue of rat embryos exposed to beta-hydroxybutyrate during
tes. Teratology 1988; 37: 51–61. organogenesis. Life Sci 1998; 62: 293–300.
134. Nandakumaran M et al. Kinetics of palmitic acid transport in 158. Khandelwal M, Reece EA, Wu YK, Borenstein M. Dietary
insulin-dependent diabetic pregnancies: In vitro study. Pediatr myo-inositol therapy in hyperglycemia-induced embryopathy.
Int 2000; 42: 296–301. Teratology 1998; 57: 79–84.
135. Schindler M et al. Maternal diabetes leads to unphysiologi- 159. Reece EA, Khandelwal M, Wu YK, Borenstein M. Dietary intake
cal high lipid accumulation in rabbit preimplantation embryos. of myo-inositol and neural tube defects in offspring of diabetic
Endocrinology 2014; 155: 1498–1509. rats. Am J Obstet Gynecol 1997; 176: 536–539.
136. Eriksson RS, Thunberg L, Eriksson UJ. Effects of interrupted insu- 160. Reece AE, Wu YK. Prevention of diabetic embryopathy in off-
lin treatment on fetal outcome of pregnant diabetic rats. Diabetes spring of diabetic rats with use of a cocktail of deficient substrates
1989; 38: 764–772. and an antioxidant. Am J Obstet Gynecol 1997; 176: 790–798.
137. Battaglia FC. Placental transport: A function of permeability and 161. Strieleman PJ, Metzger BE. Glucose and scyllo-inositol impair
perfusion. Am J Clin Nutr 2007; 85: 591S–597S. phosphoinositide hydrolysis in the 10.5-day cultured rat concep-
138. Ejdesjo A, Wentzel P, Eriksson UJ. Genetic and environmental tus: A role in dysmorphogenesis? Teratology 1993; 48: 267–278.
influence on diabetic rat embryopathy. Am J Physiol Endocrinol 162. Wentzel P, Wentzel CR, Gareskog MB, Eriksson UJ. Induction
Metab 2011; 300: E454–E467. of embryonic dysmorphogenesis by high glucose concentration,
139. Hod M, Star S, Passonneau J, Unterman TG, Freinkel N. Glucose- disturbed inositol metabolism, and inhibited protein kinase C
induced dysmorphogenesis in the cultured rat conceptus: activity. Teratology 2001; 63: 193–201.
Prevention by supplementation with myo-inositol. Isr J Med Sci 163. Baker L, Piddington R. Diabetic embryopathy: A selective review
1990; 26: 541–544. of recent trends. J Diabetes Complications 1993; 7: 204–212.
140. Sussman I, Matschinsky FM. Diabetes affects sorbitol and myo- 164. Gabbay KH. The sorbitol pathway and the complications of dia-
inositol levels of neuroectodermal tissue during embryogenesis in betes. N Engl J Med 1973; 288: 831–836.
rat. Diabetes 1988; 37: 974–981. 165. Greene DA, Lattimer SA, Sima AAF. Sorbitol, phosphoinositides,
141. Weigensberg MJ, Garcia-Palmer F-J, Freinkel N. Uptake of myo- and sodium-potassium-ATPase in the pathogenesis of diabetic
inositol by early-somite rat conceptus. Transport kinetics and complications. N Engl J Med 1987; 316: 599–606.
effects of hyperglycemia. Diabetes 1990; 39: 575–582. 166. Hodgkinson AD et al. Aldose reductase expression is induced
142. Eriksson UJ, Naeser P, Brolin SE. Increased accumulation of sor- by hyperglycemia in diabetic nephropathy. Kidney Int 2001; 60:
bitol in offspring of manifest diabetic rats. Diabetes 1986; 35: 211–218.
1356–1363. 167. Eriksson UJ, Andersson A, Efendic S, Elde R, Hellerström C.
143. Eriksson UJ, Brolin SE, Naeser P. Influence of sorbitol accumula- Diabetes in pregnancy: Effects on the fetal and newborn rat with
tion on growth and development of embryos cultured in elevated particular regard to body weight, serum insulin concentration
levels of glucose and fructose. Diabetes Res 1989; 11: 27–32. and pancreatic contents of insulin, glucagon and somatostatin.
144. Pinter E et al. Yolk sac failure in embryopathy due to hyperglyce- Acta Endocrinol (Copenh) 1980; 94: 354–364.
mia: Ultrastructural analysis of yolk sac differentiation associated 168. Fu J, Tay SS, Ling EA, Dheen ST. Aldose reductase is implicated in
with embryopathy in rat conceptuses under hyperglycemic con- high glucose-induced oxidative stress in mouse embryonic neu-
ditions. Teratology 1986; 33: 73–84. ral stem cells. J Neurochem 2007; 103: 1654–1665.
145. Engström E, Haglund A, Eriksson UJ. Effects of maternal diabetes 169. Reece EA et al. Dietary polyunsaturated fatty acid prevents mal-
or in vitro hyperglycemia on uptake of palmitic and arachidonic formations in offspring of diabetic rats. Am J Obstet Gynecol
acid by rat embryos. Pediatr Res 1991; 30: 150–153. 1996; 175: 818–823.
146. Jawerbaum A et al. Diminished levels of prostaglandin E in type 170. Wiznitzer A, Furman B, Mazor M, Reece EA. The role of pros-
I diabetic oocyte-cumulus complexes. Influence of nitric oxide tanoids in the development of diabetic embryopathy. Semin
and superoxide dismutase. Reprod Fertil Dev 1999; 11: 105–110. Reprod Endocrinol 1999; 17: 175–181.
147. Wentzel P, Welsh N, Eriksson UJ. Developmental damage, 171. Piddington R, Joyce J, Dhanasekaran P, Baker L. Diabetes mellitus
increased lipid peroxidation, diminished cyclooxygenase-2 gene affects prostaglandin E2 levels in mouse embryos during neurula-
expression, and lowered PGE2 levels in rat embryos exposed to a tion. Diabetologia 1996; 39: 915–920.
diabetic environment. Diabetes 1999; 48: 813–820. 172. Schoenfeld A et al. Yolk sac concentration of prostaglandin E2
148. White V et al. Oxidative stress and altered prostanoid production in diabetic pregnancy: Further clues to the etiology of diabetic
in the placenta of streptozotocin-induced diabetic rats. Reprod embryopathy. Prostaglandins 1995; 50: 121–126.
Fertil Dev 2002; 14: 117–123. 173. Pinter E et al. Fatty acid content of yolk sac and embryo in
149. Wentzel P, Gareskog M, Eriksson UJ. Folic acid supplementation hyperglycemia-induced embryopathy and effect of arachidonic

diminishes diabetes- and glucose-induced dysmorphogenesis in acid supplementation. Am J Obstet Gynecol 1988; 159: 1484–1490.
rat embryos in vivo and in vitro. Diabetes 2005; 54: 546–553. 174. MRC. Prevention of neural tube defects: Results of the Medical
150. Wentzel P, Eriksson UJ. A diabetes-like environment increases Research Council Vitamin Study. MRC Vitamin Study Research
malformation rate and diminishes prostaglandin E(2) in rat Group. Lancet 1991; 338: 131–137.
embryos: Reversal by administration of vitamin E and folic acid. 175. Czeizel AE, Dudas I. Prevention of the first occurrence of neural-
Birth Defects Res A Clin Mol Teratol 2005; 73: 506–511. tube defects by periconceptional vitamin supplementation. N
151. Kane MT, Norris M, Harrison RA. Uptake and incorporation of Engl J Med 1992; 327: 1832–1835.
inositol by preimplantation mouse embryos. J Reprod Fertil 1992; 176. Lumley J, Watson L, Watson M, Bower C. Periconceptional
96: 617–625. supplementation with folate and/or multivitamins for prevent-
152. Greene DA, MacKway AM. Decreased myo-inositol content ing neural tube defects. Cochrane Database Syst Rev 2001; (3):
and Na+-K+-ATPase activity in superior cervical ganglion of CD001056.
STZ-diabetic rat and prevention by aldose reductase inhibition. 177. Honein MA, Paulozzi LJ, Mathews TJ, Erickson JD, Wong
Diabetes 1986; 35: 1106–1108. LY. Impact of folic acid fortification of the US food supply
153. Lapetina EG. Regulation of arachidonic acid production: Role on the occurrence of neural tube defects. JAMA 2001; 285:
of phospholipases C and A2. Trends Pharmacol Sci 1982; 3: 2981–2986.
115–118. 178. Ray JG et al. Association of neural tube defects and folic acid
154. Cockroft DL. Vitamin deficiencies and neural-tube defects: food fortification in Canada. Lancet 2002; 360: 2047–2048.
Human and animal studies. Hum Reprod 1991; 6: 148–157. 179. De Wals P et al. Reduction in neural-tube defects after folic acid
155. Cockroft DL, Brook FA, Copp AJ. Inositol deficiency increases fortification in Canada. N Engl J Med 2007; 357: 135–142.
the susceptibility to neural tube defects of genetically predis- 180. Martínez-Frias ML. Epidemiological analysis of outcomes of preg-
posed (curly tail) mouse embryos in vitro. Teratology 1992; 45: nancy in diabetic mothers: Identification of the most character-
223–232. istic and most frequent congenital anomalies. Am J Med Genet
156. Greene MF, Hare JW, Cloherty JP, Benacerraf BR, Soeldner JS. 1994; 51: 108–113.
First-trimester hemoglobin A1 and risk for major malformation 181. Ray JG, Vermeulen MJ, Meier C, Wyatt PR. Risk of congenital
and spontaneous abortion in diabetic pregnancy. Teratology anomalies detected during antenatal serum screening in women
1989; 39: 225–231. with pregestational diabetes. QJM 2004; 97: 651–653.
157. Cogram P et al. d-Chiro-inositol is more effective than myo- 182. Correa A et al. Lack of periconceptional vitamins or supplements
inositol in preventing folate-resistant mouse neural tube defects. that contain folic acid and diabetes mellitus-associated birth
Hum Reprod 2002; 17: 2451–2458. defects. Am J Obstet Gynecol 2012; 206: 218.e211–218.e213.
References 347
183. Correa A, Botto L, Liu Y, Mulinare J, Erickson JD. Do multivita- 206. Yang P, Zhao Z, Reece EA. Activation of oxidative stress signaling
min supplements attenuate the risk for diabetes-associated birth that is implicated in apoptosis with a mouse model of diabetic
defects? Pediatrics 2003; 111: 1146–1151. embryopathy. Am J Obstet Gynecol 2008; 198: 130.e131–130.
184. van Beynum IM et al. Protective effect of periconceptional folic e137.
acid supplements on the risk of congenital heart defects: A reg- 207. Yang P et al. Maternal hyperglycemia activates an ASK1-FoxO3a-
istry-based case-control study in the northern Netherlands. Eur caspase 8 pathway that leads to embryonic neural tube defects.
Heart J 2010; 31: 464–471. Sci Signal 2013; 6: ra74.
185. Piedrahita JA et al. Mice lacking the folic acid-binding protein 208. Yang P, Cao Y, Li H. Hyperglycemia induces inducible nitric
Folbp1 are defective in early embryonic development. Nat Genet oxide synthase gene expression and consequent nitrosative stress
1999; 23: 228–232. via c-Jun N-terminal kinase activation. Am J Obstet Gynecol
186. Oyama K et al. Folic acid prevents congenital malformations in 2010; 203: 185.e5–185.e11.
the offspring of diabetic mice. Endocr J 2009; 56: 29–37. 209. Hagay ZJ et al. Prevention of diabetes-associated embryopathy
187. Higa R et al. Folic acid and safflower oil supplementation inter- by overexpression of the free radical scavenger copper zinc
acts and protects embryos from maternal diabetes-induced dam- superoxide dismutase in transgenic mouse embryos. Am J Obstet
age. Mol Hum Reprod 2012; 18: 253–264. Gynecol 1995; 173: 1036–1041.
188. Jia DY et al. Folic acid supplementation affects apoptosis and 210. Weng H, Li X, Reece EA, Yang P. SOD1 suppresses maternal
differentiation of embryonic neural stem cells exposed to high hyperglycemia-increased iNOS expression and consequent
glucose. Neurosci Lett 2008; 440: 27–31. nitrosative stress in diabetic embryopathy. Am J Obstet Gynecol
189. Yuan Q et al. Folic acid supplementation changes the fate of neu- 2012; 206: 448.e441–448.e447.
ral progenitors in mouse embryos of hyperglycemic and diabetic 211. Zhao Z, Eckert RL, Reece EA. Reduction in embryonic malfor-
pregnancy. J Nutr Biochem 2012; 24: 1202–1212. mations and alleviation of endoplasmic reticulum stress by nitric
190. Hiramatsu Y et al. Diacylglycerol production and protein kinase oxide synthase inhibition in diabetic embryopathy. Reprod Sci
C activity are increased in a mouse model of diabetic embryopa- 2012; 19: 823–831.
thy. Diabetes 2002; 51: 2804–2810. 212. Wang F, Reece EA, Yang P. Superoxide dismutase 1 overexpres-
191. Gareskog M, Wentzel P. Altered protein kinase C activation asso- sion in mice abolishes maternal diabetes-induced endoplasmic
ciated with rat embryonic dysmorphogenesis. Pediatr Res 2004; reticulum stress in diabetic embryopathy. Am J Obstet Gynecol
56: 849–857. 2013; 209: 345.e341–345.e347.
192. Gareskog M, Wentzel P. N-Acetylcysteine and {alpha}-cyano- 213. Horal M, Zhang Z, Stanton R, Virkamaki A, Loeken MR.
4-hydroxycinnamic acid alter protein kinase C (PKC)-{delta} Activation of the hexosamine pathway causes oxidative stress
and PKC-{zeta} and diminish dysmorphogenesis in rat embryos and abnormal embryo gene expression: Involvement in dia-
cultured with high glucose in vitro. J Endocrinol 2007; 192: betic teratogenesis. Birth Defects Res A Clin Mol Teratol 2004;
207–214. 70: 519–527.
193. Zhiyong Z, Wu YK, Reece EA. Demonstration of the essential 214. Moley KH. Hyperglycemia and apoptosis: Mechanisms for con-
role of protein kinase C isoforms in hyperglycemia-induced genital malformations and pregnancy loss in diabetic women.
embryonic malformations. Reprod Sci 2008; 15: 349–356. Trends Endocrinol Metab 2001; 12: 78–82.
194. Cao Y, Zhao Z, Eckert RL, Reece EA. Protein kinase Cbeta2 215. Sun F et al. Apoptosis and its pathway in early post-implantation
inhibition reduces hyperglycemia-induced neural tube defects embryos of diabetic rats. Diabetes Res Clin Pract 2005; 67:
through suppression of a caspase 8-triggered apoptotic pathway. 110–118.
Am J Obstet Gynecol 2011; 204: 226.e221–226.e225. 216. Gareskog M, Eriksson UJ, Wentzel P. Combined supplementation
195. Cao Y, Zhao Z, Eckert RL, Reece EA. The essential role of protein of folic acid and vitamin E diminishes diabetes-induced embryo-
kinase C delta in diabetes-induced neural tube defects. J Matern toxicity in rats. Birth Defects Res A Clin Mol Teratol 2006; 76:
Fetal Neonatal Med 2012; 25: 2020–2024. 483–490.
196. Reece EA, Ma XD, Wu YK, Dhanasekaran D. Aberrant patterns 217. Oberley LW. Free radicals and diabetes. Free Radic Biol Med
of cellular communication in diabetes-induced embryopathy. I. 1988; 5: 113–124.
Membrane signalling. J Matern Fetal Neonatal Med 2002; 11: 218. Gillery P, Monboisse JC, Maquart FX, Borel JP. Does oxygen free
249–253. radical increased formation explain long term complications of
197. Reece EA, Wu YK, Zhao Z, Dhanasekaran D. Dietary vitamin diabetes mellitus? Med Hypotheses 1989; 29: 47–50.
and lipid therapy rescues aberrant signaling and apoptosis and 219. Baynes JW. Role of oxidative stress in development of complica-
prevents hyperglycemia-induced diabetic embryopathy in rats. tions in diabetes. Diabetes 1991; 40: 405–412.
Am J Obstet Gynecol 2006; 194: 580–585. 220. Schmidt AM et al. Advanced glycation endproducts (AGEs)
198. Yang P, Zhao Z, Reece EA. Involvement of c-Jun N-terminal induce oxidant stress in the gingiva: A potential mechanism
kinases activation in diabetic embryopathy. Biochem Biophys underlying accelerated periodontal disease associated with dia-
Res Commun 2007; 357: 749–754. betes. J Periodontal Res 1996; 31: 508–515.
199. Yang P, Zhao Z, Reece EA. Blockade of c-Jun N-terminal kinase 221. West IC. Radicals and oxidative stress in diabetes. Diabet Med
activation abrogates hyperglycemia-induced yolk sac vascu- 2000; 17: 171–180.
lopathy in vitro. Am J Obstet Gynecol 2008; 198: 321.e321– 222. Palmer AM et al. Dietary antioxidant supplementation reduces
321.e327. lipid peroxidation but impairs vascular function in small mes-
200. Li X, Weng H, Xu C, Reece EA, Yang P. Oxidative stress-induced enteric arteries of the streptozotocin-diabetic rat. Diabetologia
JNK1/2 activation triggers proapoptotic signaling and apop- 1998; 41: 148–156.
tosis that leads to diabetic embryopathy. Diabetes 2012; 61: 223. Sano T, Umeda F, Hashimoto T, Nawata H, Utsumi H. Oxidative
2084–2092. stress measurements by in vivo electron spin resonance spectros-
201. Ward KW, Rogers EH, Hunter ES. Dysmorphogenic effects of a copy in rats with streptozotocin-induced diabetes. Diabetologia
specific protein kinase C inhibitor during neurulation. Reprod 1998; 41: 1355–1360.
Toxicol 1998; 12: 525–534. 224. Elangovan V, Shohami E, Gati I, Kohen R. Increased hepatic lipid
202. Santiago-Walker AE et al. Protein kinase C delta stimulates apop- soluble antioxidant capacity as compared to other organs of
tosis by initiating G1 phase cell cycle progression and S phase streptozotocin-induced diabetic rats: A cyclic voltammetry study.
arrest. J Biol Chem 2005; 280: 32107–32114. Free Radic Res 2000; 32: 125–134.
203. Leroy I, de Thonel A, Laurent G, Quillet-Mary A. Protein kinase 225. Ishibashi M et al. Oxygen-induced embryopathy and the sig-
C zeta associates with death inducing signaling complex and nificance of glutathione-dependent antioxidant system in the rat
regulates Fas ligand-induced apoptosis. Cell Signal 2005; 17: embryo during early organogenesis. Free Radic Biol Med 1997;
1149–1157. 22: 447–454.
204. Cao Y et al. XBP1 forms a regulatory loop with BMP-4 and 226. Zaken V, Kohen R, Ornoy A. Vitamins C and E improve rat
suppresses mesodermal and neural differentiation in Xenopus embryonic antioxidant defense mechanism in diabetic culture
embryos. Mech Dev 2006; 123: 84–96. medium. Teratology 2001; 64: 33–44.
205. Domenicotti C et al. Role of PKC-delta activity in glutathione- 227. Sakamaki H et al. Significance of glutathione-dependent anti-
depleted neuroblastoma cells. Free Radic Biol Med 2003; 35: oxidant system in diabetes-induced embryonic malformations.
504–516. Diabetes 1999; 48: 1138–1144.
228. Cederberg J, Siman CM, Eriksson UJ. Combined treatment with 251. Kamimoto Y et al. Transgenic mice overproducing human thio-
vitamin E and vitamin C decreases oxidative stress and improves redoxin-1, an antioxidative and anti-apoptotic protein, prevents
fetal outcome in experimental diabetic pregnancy. Pediatr Res diabetic embryopathy. Diabetologia 2010; 53: 2046–2055.
2001; 49: 755–762. 252. Torchinsky A, Toder V. TNFalpha in the pathogenesis of diabetes-
229. Cederberg J, Eriksson UJ. Antioxidative treatment of pregnant induced embryopathies: Functions and targets. Rev Diabet Stud
diabetic rats diminishes embryonic dysmorphogenesis. Birth 2007; 4: 200–209.
Defects Res A Clin Mol Teratol 2005; 73: 498–505. 253. Wu Y, Viana M, Thirumangalathu S, Loeken MR. AMP-activated
230. Ornoy A. Embryonic oxidative stress as a mechanism of terato- protein kinase mediates effects of oxidative stress on embryo
genesis with special emphasis on diabetic embryopathy. Reprod gene expression in a mouse model of diabetic embryopathy.
Toxicol 2007; 24: 31–41. Diabetologia 2012; 55: 245–254.
231. Gareskog M, Cederberg J, Eriksson UJ, Wentzel P. Maternal dia- 254. Lee HY, Wei D, Loeken, MR. Lack of metformin effect on mouse
betes in vivo and high glucose concentration in vitro increases embryo AMPK activity: Implications for metformin treatment
apoptosis in rat embryos. Reprod Toxicol 2007; 23: 63–74. during pregnancy. Diabetes Metab Res Rev 2014; 30: 23–30.
232. Suzuki N, Svensson K, Eriksson UJ. High glucose concentra- 255. Menegola E, Broccia ML, Prati M, Ricolfi R, Giavini E. Glutathione
tion inhibits migration of rat cranial neural crest cells in vitro. status in diabetes-induced embryopathies. Biol Neonate 1996;
Diabetologia 1996; 39: 401–411. 69: 293–297.
233. Roest PA et al. Exposure of neural crest cells to elevated glucose 256. Davis WL et al. Ethanol induces the generation of reactive free
leads to congenital heart defects, an effect that can be prevented radicals by neural crest cells in vitro. J Craniofac Genet Dev Biol
by N-acetylcysteine. Birth Defects Res A Clin Mol Teratol 2007; 1990; 10: 277–293.
79: 231–235. 257. Chen SY, Sulik KK. Free radicals and ethanol-induced cytotoxicity
234. Wiznitzer A et al. Lipoic acid prevention of neural tube defects in neural crest cells. Alcohol Clin Exp Res 1996; 20: 1071–1076.
in offspring of rats with streptozotocin-induced diabetes. Am J 258. Jenkinson PC, Anderson D, Gangolli SD. Malformations induced
Obstet Gynecol 1999; 180: 188–193. in cultured rat embryos by enzymically generated active oxygen
235. Al Ghafli MH, Padmanabhan R, Kataya HH, Berg B. Effects of species. Teratog Carcinog Mutagen 1986; 6: 547–554.
alpha-lipoic acid supplementation on maternal diabetes-induced 259. Anderson D, Francis AJ. The modulating effects of antioxidants
growth retardation and congenital anomalies in rat fetuses. Mol in rat embryos and Sertoli cells in culture. Basic Life Sci 1993; 61:
Cell Biochem 2004; 261: 123–135. 189–200.
236. Sugimura Y et al. Alpha-lipoic acid reduces congenital malforma- 260. Winn LM, Wells PG. Phenytoin-initiated DNA oxidation in
tions in the offspring of diabetic mice. Diabetes Metab Res Rev murine embryo culture, and embryoprotection by the antioxida-
2009; 25: 287–294. tive enzymes superoxide dismutase and catalase: Evidence for
237. Moazzen H et al. N-Acetylcysteine prevents congenital heart reactive oxygen species mediated DNA oxidation in the molecu-
defects induced by pregestational diabetes. Cardiovasc Diabetol lar mechanism of phenytoin teratogenecity. Mol Pharmacol 1995;
2014; 13: 46. 48: 112–120.
238. Morgan SC, Relaix F, Sandell LL, Loeken MR. Oxidative stress 261. Winn LM, Wells PG. Maternal administration of superoxide dis-
during diabetic pregnancy disrupts cardiac neural crest migration mutase and catalase in phenytoin teratogenicity. Free Radic Biol
and causes outflow tract defects. Birth Defects Res A Clin Mol Med 1999; 26: 266–274.
Teratol 2008; 82: 453–463. 262. Kotch LE, Chen S-E, Sulik KK. Ethanol-induced teratogenesis:
239. Li X, Weng H, Reece EA, Yang P. SOD1 overexpression in vivo Free radical damage as a possible mechanism. Teratology 1995;
blocks hyperglycemia-induced specific PKC isoforms: Substrate 52: 128–136.
activation and consequent lipid peroxidation in diabetic embry- 263. Parman T, Wiley MJ, Wells PG. Free radical-mediated oxidative
opathy. Am J Obstet Gynecol 2011; 205: 84.e81–84.e86. DNA damage in the mechanism of thalidomide teratogenicity.
240. Yang X, Borg LAH, Eriksson UJ. Altered metabolism and super- Nat Med 1999; 5: 582–585.
oxide generation in neural tissue of rat embryos exposed to high 264. Eriksson UJ. Oxidative DNA damage and embryo development
glucose. Am J Physiol 1997; 272: E173–E180. (Letter). Nat Med 1999; 5: 715.
241. Janero DR, Hreniuk D, Sharif HM. Hydroperoxide-induced oxi- 265. Li R, Chase M, Jung SK, Smith PJ, Loeken MR. Hypoxic stress in
dative stress impairs heart muscle cell carbohydrate metabolism. diabetic pregnancy contributes to impaired embryo gene expres-
Am J Physiol 1994; 266: C179–C188. sion and defective development by inducing oxidative stress. Am
242. Rivera-Nieves J, Thompson WC, Levine RL, Moss J. Thiols J Physiol Endocrinol Metab 2005; 289: E591–E599.
mediate superoxide-dependent NADH modification of glycer- 266. Morrow JD, Harris TM, Roberts 2nd LJ. Noncyclooxygenase oxi-
aldehyde-3-phosphate dehydrogenase. J Biol Chem 1999; 274: dative formation of a series of novel prostaglandins: Analytical
19525–19531. ramifications for measurement of eicosanoids. Anal Biochem
243. Ishii T et al. Critical role of sulfenic acid formation of thiols in the 1990; 184: 1–10.
inactivation of glyceraldehyde-3-phosphate dehydrogenase by 267. Morrow JD, Awad JA, Boss HJ, Blair IA, Roberts II LJ. Non-
nitric oxide. Biochem Pharmacol 1999; 58: 133–143. cyclooxygenase-derived prostanoids (F2-isoprostanes) are
244. Morgan PE, Dean RT, Davies MJ. Inhibition of glyceraldehyde- formed in situ on phospholipids. Proc Natl Acad Sci USA 1992;
3-phosphate dehydrogenase by peptide and protein peroxides 89: 10721–10725.
generated by singlet oxygen attack. Eur J Biochem 2002; 269: 268. Davi G et al. In vivo formation of 8-iso-prostaglandin f2alpha and
1916–1925. platelet activation in diabetes mellitus: Effects of improved meta-
245. McKenzie SM, Doe WF, Buffinton GD. 5-Aminosalicylic acid bolic control and vitamin E supplementation. Circulation 1999;
prevents oxidant mediated damage of glyceraldehyde-3-phos- 99: 224–229.
phate dehydrogenase in colon epithelial cells. 1999; Gut 44: 269. Wentzel P, Eriksson UJ. 8-Iso-PGF(2alpha) administration gener-
180–185. ates dysmorphogenesis and increased lipid peroxidation in rat
246. Wentzel P, Ejdesjo A, Eriksson UJ. Maternal diabetes in vivo and embryos in vitro. Teratology 2002; 66: 164–168.
high glucose in vitro diminish GAPDH activity in rat embryos. 270. Goh SY, Cooper ME. Clinical review: The role of advanced glyca-
Diabetes 2003; 52: 1222–1228. tion end products in progression and complications of diabetes. J
247. Viana M, Aruoma OI, Herrera E, Bonet B. Oxidative damage in Clin Endocrinol Metab 2008; 93: 1143–1152.
pregnant diabetic rats and their embryos. Free Radic Biol Med 271. Coughlan MT et al. RAGE-induced cytosolic ROS promote mito-
2000; 29: 1115–1121. chondrial superoxide generation in diabetes. J Am Soc Nephrol
248. Horton WE, Sadler TW. Mitochondrial alterations in embryos 2009; 20: 742–752.
exposed to beta-hydroxybutyrate in whole embryo culture. Anat 272. Wolff SP, Jiang ZY, Hunt JV. Protein glycation and oxidative stress
Rec 1985; 213: 94–101. in diabetes mellitus and ageing. Free Radic Biol Med 1991; 10:
249. Yang X, Borg LAH, Eriksson UJ. Altered mitochondrial morphol- 339–352.
ogy of rat embryos in diabetic pregnancy. Anat Rec 1995; 241: 273. Thornalley PJ. Advanced glycation and the development of
255–267. diabetic complications. Unifying the involvement of glucose,
250. Forsberg H, Eriksson UJ, Welsh N. Apoptosis in embryos of dia- methylglyoxal and oxidative stress. Endocrinol Metab 1996; 3:
betic rats. Pharmacol Toxicol 1998; 83: 104–111. 149–166.
References 349
274. Zhang CH et al. Maternal diabetes causes abnormal dynamic 297. Narchi H, Kulaylat N. High incidence of Down’s syndrome in
changes of endoplasmic reticulum during mouse oocyte matu- infants of diabetic mothers. Arch Dis Child 1997; 77: 242–244.
ration and early embryo development. Reprod Biol Endocrinol 298. Pelz J, Kunze J. Down’s syndrome in infants of diabetic mothers
2013; 11: 31. [letter; comment]. Arch Dis Child 1998; 79: 199–200.
275. Li X, Xu C, Yang P. c-Jun NH2-terminal kinase 1/2 and endoplas- 299. Nelson M, Lessell S, Sadun AA. Optic nerve hypoplasia and
mic reticulum stress as interdependent and reciprocal causation maternal diabetes mellitus. Arch Neurol 1986; 43: 20–25.
in diabetic embryopathy. Diabetes 2013; 62: 599–608. 300. Van Allen MI, Myhre S. New multiple congenital anomalies syn-
276. Xu C, Li X, Wang F, Weng H, Yang P. Trehalose prevents neural drome in a stillborn infant of consanguinous parents and a pre-
tube defects by correcting maternal diabetes-suppressed autoph- diabetic pregnancy. Am J Med Genet 1991; 38: 523–528.
agy and neurogenesis. Am J Physiol Endocrinol Metab 2013; 305: 301. Endo A, Ingalls TH. Chromosomal anomalies in embryos of dia-
E667–E678. betic mice. Arch Environ Health 1968; 16: 316–325.
277. Zhao Z. Endoplasmic reticulum stress in maternal diabetes- 302. Yamamoto M, Endo A, Watanabe G, Ingalls TH. Chromosomal
induced cardiac malformations during critical cardiogenesis aneuploidies and polyploidies in embryos of diabetic mice. Arch
period. Birth Defects Res B Dev Reprod Toxicol 2012; 95: 1–6. Environ Health 1971; 22: 468–475.
278. Marshall S, Bacote V, Traxinger RR. Discovery of a metabolic 303. Eriksson UJ, den Bieman M, Prins JB, van Zutphen LFM.
pathway mediating glucose-induced desensitization of the glu- Differences in susceptibility for diabetes-induced malformations
cose transport system. Role of hexosamine biosynthesis in the in separated rat colonies of common origin. In: Fourth FELASA
induction of insulin resistance. J Biol Chem 1991: 266: 4706–4712. Symposium, Fondation Marcel Mérieux: Lyon, France, 1990,
279. Anderson JW, Nicolosi RJ, Borzelleca JF. Glucosamine effects in pp. 53–57.
humans: A review of effects on glucose metabolism, side effects, 304. Cederberg J, Eriksson UJ. Decreased catalase activity in mal-
safety considerations and efficacy. Food Chem Toxicol 2005; 43: formation-prone embryos of diabetic rats. Teratology 1997; 56:
187–201. 350–357.
280. Whelan SA, Hart GW. Proteomic approaches to analyze the 305. Wyman A, Pinto AB, Sheridan R, Moley KH. One-cell zygote
dynamic relationships between nucleocytoplasmic protein gly- transfer from diabetic to nondiabetic mouse results in congenital
cosylation and phosphorylation. Circ Res 2003; 93: 1047–1058. malformations and growth retardation in offspring. Endocrinology
281. Yang YR, Suh PG. O-GlcNAcylation in cellular functions and 2008; 149: 466–469.
human diseases. Adv Biol Regul 2013; 54: 68–73. 306. Ejdesjo A, Wentzel P, Eriksson UJ. Influence of maternal metab-
282. Pantaleon M, Tan HY, Kafer GR, Kaye PL. Toxic effects of hyper- olism and parental genetics on fetal maldevelopment in dia-
glycemia are mediated by the hexosamine signaling pathway betic rat pregnancy. Am J Physiol Endocrinol Metab 2012; 302:
and o-linked glycosylation in early mouse embryos. Biol Reprod E1198–E1209.
2010; 82: 751–758. 307. Cederberg J, Galli J, Luthman H, Eriksson UJ. Increased mRNA
283. Reece EA, Ma XD, Zhao Z, Wu YK, Dhanasekaran D. Aberrant levels of Mn-SOD and catalase in embryos of diabetic rats from
patterns of cellular communication in diabetes-induced embry- a malformation-resistant strain. Diabetes 2000; 49: 101–107.
opathy in rats: II, apoptotic pathways. Am J Obstet Gynecol 2005; 308. Jiang B et al. Global gene expression analysis of cranial neural
192: 967–972. tubes in embryos of diabetic mice. J Neurosci Res 2008; 86:
284. Zhao Z, Yang P, Eckert RL, Reece EA. Caspase-8: A key role in the 3481–3493.
pathogenesis of diabetic embryopathy. Birth Defects Res B Dev 309. Fu J, Tay SS, Ling EA, Dheen ST. High glucose alters the expres-
Reprod Toxicol 2009; 86: 72–77. sion of genes involved in proliferation and cell-fate specifica-
285. Kapron C, Trasler D. Genetic determinants of teratogen-induced tion of embryonic neural stem cells. Diabetologia 2006; 49:
abnormal development in mouse and rat embryos in vitro. Int J 1027–1038.
Dev Biol 1997; 41: 337–344. 310. Liao DM, Ng YK, Tay SS, Ling EA, Dheen ST. Altered gene
286. Morrison K et al. Susceptibility to spina bifida; an associa- expression with abnormal patterning of the telencephalon in
tion study of five candidate genes. Ann Hum Genet 1998; 62: embryos of diabetic Albino Swiss mice. Diabetologia 2004; 47:
379–396. 523–531.
287. Degnbol B, Green A. Diabetes mellitus among first and second- 311. Zhao Z. Cardiac malformations and alteration of TGFbeta sig-
degree relatives of early onset diabetes. Ann Hum Genet 1978; naling system in diabetic embryopathy. Birth Defects Res B Dev
42: 25–47. Reprod Toxicol 2010; 89: 97–105.
288. Tillil H, Köbberling J. Age-corrected empirical genetic risk esti- 312. Madri JA, Enciso J, Pinter E. Maternal diabetes: Effects on embry-
mates for first degree relatives of IDDM patients. Diabetes 1982; onic vascular development—A vascular endothelial growth
36: 93–99. factor-A-mediated process. Pediatr Dev Pathol 2003; 6: 334–341.
289. Warram JH, Krolewski AS, Gottlieb MS, Kahn CR. Differences in 313. Pavlinkova G, Salbaum JM, Kappen C. Wnt signaling in caudal
risk of insulin-dependent diabetes in offspring of diabetic moth- dysgenesis and diabetic embryopathy. Birth Defects Res A Clin
ers and diabetic fathers. N Engl J Med 1984; 311: 149–152. Mol Teratol 2008; 82: 710–719.
290. Dahlquist G et al. The Swedish childhood diabetes study— 314. Reece EA, Ji I, Wu YK, Zhao Z. Characterization of differential
Results from nine year case register and a one year case-referent gene expression profiles in diabetic embryopathy using DNA
study indicating that type 1 (insulin-dependent) diabetes mellitus microarray analysis. Am J Obstet Gynecol 2006; 195: 1075–1080.
is associated with both type 2 (non-insulin-dependent) diabetes 315. Sato N et al. Identification of genes differentially expressed in
mellitus and autoimmune disorders. Diabetologia 1989; 32: 2–6. mouse fetuses from streptozotocin-induced diabetic pregnancy
291. McCarthy BJ, Dorman JS, Aston CE. Investigating genomic by cDNA subtraction. Endocr J 2008; 55: 317–323.
imprinting and susceptibility to insulin-dependent diabetes mel- 316. Pani L, Horal M, Loeken MR. Rescue of neural tube defects in
litus: An epidemiological approach. Genet Epidemiol 1991; 8: Pax-3-deficient embryos by p53 loss of function: Implications for
177–186. Pax-3-dependent development and tumorigenesis. Genes Dev
292. Pociot F et al. A nationwide population-based study of the 2002; 16: 676–680.
familial aggregation of insulin-dependent diabetes in Denmark. 317. Morgan SC et al. Cardiac outflow tract septation failure in Pax3-
Diabetologia 1993; 36: 870–875. deficient embryos is due to p53-dependent regulation of migrat-
293. Tuomilehto J, Podar T, Tuomilehto-Wolf E, Virtala E. Evidence ing cardiac neural crest. Mech Dev 2008; 125: 757–767.
for importance of gender and birth cohort for risk of IDDM in 318. Pavlinkova G, Salbaum JM, Kappen C. Maternal diabetes alters
offspring of IDDM parents. Diabetologia 1995; 38: 975–982. transcriptional programs in the developing embryo. BMC
294. Comess LJ, Bennett PH, Burch TA, Miller M. Congenital anoma- Genomics 2009; 10: 274.
lies and diabetes in the Pima Indians of Arizona. Diabetes 1969; 319. Salbaum JM, Kappen C. Neural tube defect genes and maternal
18: 471–477. diabetes during pregnancy. Birth Defects Res A Clin Mol Teratol
295. Chung CS, Myrianthopoulos NC. Factors affecting risks of con- 2010; 88: 601–611.
genital malformations. II. Effect of maternal diabetes on congeni- 320. Liang J et al. Elevated glucose induces congenital heart defects
tal malformations. Birth Defects 1975; 11: 23–38. by altering the expression of tb×5, tb×20, and has2 in developing
296. Stoll C, Alembik Y, Dott B, Roth MP. Study of Down syndrome in zebrafish embryos. Birth Defects Res A Clin Mol Teratol 2010; 88:
238,942 consecutive births. Ann Genet 1998; 41: 44–51. 480–486.
321. Nordquist N, Luthman H, Pettersson U, Eriksson UJ. Linkage 328. El-Osta A et al. Transient high glucose causes persistent epi-
study of embryopathy—Polygenic inheritance of diabetes- genetic changes and altered gene expression during subsequent
induced skeletal malformations in the rat. Reprod Toxicol 2012; normoglycemia. J Exp Med 2008; 205: 2409–2417.
33: 297–307. 329. Salbaum JM, Kappen C. Diabetic embryopathy: A role for the
322. Vijaya M et al. Differential gene expression profiles during epigenome? Birth Defects Res A Clin Mol Teratol 2011; 91:
embryonic heart development in diabetic mice pregnancy. Gene 770–780.
2013; 516: 218–227. 330. Salbaum JM, Kappen C. Responses of the embryonic epigenome
323. Kappen C, Salbaum JM. Gene expression in teratogenic expo- to maternal diabetes. Birth Defects Res A Clin Mol Teratol 2012;
sures: A new approach to understanding individual risk. Reprod 94: 770–781.
Toxicol 2014; 45: 94–104. 331. Ge ZJ et al. Maternal diabetes causes alterations of DNA meth-
324. Hill AL, Phelan SA, Loeken MR. Reduced expression of Pax-3 is ylation statuses of some imprinted genes in murine oocytes. Biol
associated with overexpression of cdc46 in the mouse embryo. Reprod 2013; 88: 117.
Dev Genes Evol 1998; 208: 128–134. 332. Ge ZJ et al. Diabetic uterus environment may play a key role
325. Cai H et al. Role of diacylglycerol-regulated protein kinase C iso- in alterations of DNA methylation of several imprinted genes at
types in growth factor activation of the Raf-1 protein kinase. Mol mid-gestation in mice. Reprod Biol Endocrinol 2013; 11: 119.
Cell Biol 1997; 17: 732–741. 333. Shyamasundar S et al. Analysis of epigenetic factors in mouse
326. Conway S, Henderson D, Copp A. Pax3 is required for cardiac embryonic neural stem cells exposed to hyperglycemia. PLOS
neural crest migration in the mouse: Evidence from the splotch ONE 2013; 8: e65945.
(Sp2H) mutant. Development 1997; 124: 505–514. 334. Wei D, Loeken MR. Increased DNA methyltransferase 3b
327. Wentzel P, Gareskog M, Eriksson UJ. Decreased cardiac glu- (Dnmt3b)-mediated CpG island methylation stimulated by oxida-
tathione peroxidase levels and enhanced mandibular apopto- tive stress inhibits expression of a gene required for neural tube
sis in malformed embryos of diabetic rats. Diabetes 2008; 57: and neural crest development in diabetic pregnancy. Diabetes
3344–3352. 2014; 63(10): 3512–3522.
42 Fetal malformations detected
with magnetic resonance
imaging in the diabetic mother
Tuangsit Wataganara
Prenatal detection of anomaly can lead to a more targeted

Introduction counseling, timely intervention, fetal therapy, and appro-
Diabetes mellitus is one of the most common medical dis- priate neonatal management. Most of the fetal anoma-
orders during pregnancy, affecting approximately 3%–10% lies are detected by ultrasound in the second trimester.
of women. Gestational diabetes mellitus accounts for Recently, magnetic resonance imaging (MRI) has been
80%–90% of cases of diabetes mellitus in pregnancy, while used as an adjunct when limitations of fetal ultrasound
preexisting type 2 diabetes accounts for approximately 10%– are encountered. This chapter is focused on the additional
20%.1 Abnormal glucose metabolism results in persistent benefit of MRI in the detection of fetal malformations
or postprandial elevation of blood sugar level. Elevation of related to diabetes mellitus.
blood sugar level during pregnancy is linked to suboptimal
perinatal o utcomes, including miscarriage, fetal malforma-
tions, stillbirth, and birth asphyxia. Diabetes mellitus is con- Comparison of fetal ultrasound and
sidered a major risk factor for abnormal fetal development. magnetic resonance imaging
More women are expected to have diabetes mellitus during
pregnancy due to older age at the time of conception, a grow- Ultrasound and MRI are considered as complementing
ing rate of obesity, and the change in diagnostic criteria for imaging technologies. Ultrasound has been the p rincipal
gestational diabetes. imaging modality in fetal medicine. In theory, ultrasound
Congenital anomalies are more commonly found in beam may cause thermal and nonthermal changes at the
babies born from diabetic mothers. This risk has been subcellular level. This phenomenon has never been proven
thought to be higher in pregnancy complicated with type to be of clinical significance.6 Ultrasound is safe, conve-
1 (insulin dependent) diabetes mellitus.2 Recently, type 2 nient, affordable, and widely available in most places. Real-
(non-insulin-dependent) diabetes mellitus was proven to be time interpretation of fetal anatomy can be accomplished
equally embryopathic.3 The chance of fetal malformation in by the perinatologist. Fetal well-being can also be readily
pregnant women with diabetes mellitus is as high as 13.3%, assessed by the observation of fetal movement pattern and
compared to 1%–2% in low-risk groups.4 The pathogenesis Doppler interrogation of fetal, placental, and uterine ves-
of fetal malformations in diabetic mothers is multifacto- sels. Off-line analysis is possible with a 3D computerized
rial. Poor glycemic control in peri- and early conceptional ultrasound system.7 Fetal ultrasound has some limitations.
period is a strong determination factor for fetal anomaly. Most fetal anomalies or damages will be seen on the ultra-
Optimal blood sugar control can be monitored by the level sound only when they are at their later stages. View during
of hemoglobin A1C (Hb A1C). Two-thirds of diabetes-related examination may be obstructed by fetal position, reduced
birth defects involve the cardiovascular system and central liquor volume, maternal adiposity, or bony structures.
nervous system (CNS).1 The risk of fetal malformation is Assessment of fetal brain may not be adequate if the head
not increased when HbA1C at 14 weeks is less than 7%. The is engaged deep down in the pelvic canal. Obesity and oli-
chance of fetal malformation can be as high as 22% if Hb A1C gohydramnios always compromise the accuracy of ultra-
level is higher than 8.5%.5 sound diagnosis.8 Ultrasound has a smaller field of view;
Due to the higher chance of fetal malformations, therefore, the interpretation heavily relies on the experi-
detailed routine fetal ultrasound examination is offered ence of the operator.
in pregnant women with diabetes mellitus, especially for With these limitations of the fetal ultrasound exami-
those with poor glycemic control in the first trimester. nation, MRI has been advocated in selected cases of fetal
351
352 Fetal malformations detected with magnetic resonance imaging in the diabetic mother
anomaly suspected from the ultrasound. MRI has an advan- The comparison between ultrasound and MRI imaging in
tage of excellent tissue contrast. This quality allows for the the fetus is summarized in Table 42.1.
detection of subtle fetal malformation and early changes
that may lead to fetal damage. The MRI field of view is rela-
tively large. Surrounding maternal structures can be simul- Principles and techniques of fetal
taneously evaluated to determine their relationship with the magnetic resonance imaging
uterus and placenta, especially when surgical intervention
is planned. Unlike fetal ultrasound, maternal adiposity and MRI is an alternative imaging method that can confirm or
bony structure do not affect the quality of MRI image. MRI complement abnormal ultrasound findings. Its use in preg-
data can be interpreted off-line by specialists. For instance, nant women was first described in 1983.9 The original obstet-
neurogeneticists can help determine the significance of the ric indication for MRI was to assess myometrial invasion of
subtle findings in the developing fetal brain at various ges- placenta accreta. Its application in fetal medicine was pre-
tational ages. MRI can also improve the understanding of viously limited because the original MRI sequence protocol
the whole fetal care team, including nurses, genetic coun- was subjected for a significant motion artifact. Maternal and
selors, and social workers. Limitations of fetal MRI exist. fetal sedation were commonly practiced to reduce move-
Its high cost naturally limits broader availability. Strict ment and to compensate for a long examination time. Fetal
indications for fetal MRI have been implemented in many MRI has regained broader application after the develop-
places where both fetal ultrasound and fetal MRI are offered. ment of ultrafast MRI sequences in the 1990s. The single-
MRI is very sensitive to fetal movement. Motion artifact, as shot rapid acquisition with focused echoes is a high-quality
shown in Figure 42.1, can occur, even with the ultrafast MRI T2-weighted sequence that acquires a slice in less than a sec-
protocol that takes only 20 seconds to complete the scan. ond.10 The preferred system for fetal MRI consists of 1.5 T
T1 T2
Figure 42.1 Magnetic resonance imaging (MRI) is extremely sensitive to fetal movement. The motion artifact can equally affect
the image quality of both T1- and T2-weighted gradient. Development of ultrafast fetal MRI protocol as well as optimal patient
preparation can reduce the artifact and improve the image quality.
Table 42.1 Comparison between ultrasound and magnetic resonance imaging for an assessment of fetal anatomy
Ultrasound MRI
Advantages Safe Better soft tissue contrast, especially intracranial and intrathoracic lesion
Convenient Large field of view
Inexpensive Off-line interpretation
More experienced
Disadvantages Maternal obesity Expensive
Oligohydramnios Not widely available
Operator dependent Subjected to motion artifact
Less experienced
Limitation in fetal cardiac assessment and intravenous contrast study
Additional benefits of magnetic resonance imaging in the detection of fetal anomalies related to gestational diabetes 353
resonators with multichannel coils. MRI can measure vol- Fetal anomalies related to maternal
ume of fetal organs in three dimensions. For instance, total
lung volume in case of congenital diaphragmatic hernia can diabetes
be used to predict neonatal survival.11 Assessment of fetal
Poor glycemic control in periconceptional period and early
intracranial and intrathoracic lesions has also been benefited
gestation has been linked to malformation in various fetal
from the superb soft tissue delineation of MRI.
organ systems, especially the cardiovascular system and
Most literatures suggest the safety of MRI used in preg-
CNS. Anomalies in fetal gastrointestinal, renal, and skel-
nancy.12,13 Many professional organizations discourage its use
etal systems are also linked to maternal diabetes mellitus.
during the first trimester.14 By nature, spontaneous miscar-
Fetal cardiac defects commonly associated with maternal
riage occurs more frequently in the first trimester. It could be
gestational diabetes include ventricular septal defect and
perceived as iatrogenic by the patients if the incidence occurs
conotruncal anomalies (such as transposition of the great
soon after the MRI session. The patient can be assured that
arteries and truncus arteriosus). Cardiomyopathy that
there has been no known risk to the fetus if MRI is offered in
leads to intrauterine congestive heart failure has also been
the late second or third trimester.12 Maternal claustrophobia,
observed in diabetic mothers with poor glycemic control.18
pacemaker, and ferromagnetic implants remain the contra-
The overall prevalence of neural tube defect (NTD) is 2–5
indications of MRI during pregnancy.15
per 1000 live births. Maternal diabetes mellitus is associated
Good preparation can help improve the image quality from
with a 16-fold increase incidence of fetal NTD. Anencephaly
MRI sequence. The mother is advised not to eat or drink at
and spina bifida are the two most common NTD related
least 4 hours prior to the examination. This is to reduce motion
to maternal diabetes mellitus.19 There is also a significant
artifacts from bowel movement and to prevent excessive
increase in the incidence of encephalocele in infants born
fetal activity after meal. Written informed consent should be
from diabetic mothers.20 The incidence of fetal spina bifida
obtained. Fetal sedation is not mandatory. Screening ultra-
is even higher when the pregnancy is complicated with obe-
sound examination should be performed immediately before
sity.21 Obesity and diabetes mellitus is strongly interrelated.
the MRI examination to confirm fetal viability. The mother is
Neural tube is a primitive bundle of nerves that folds
usually in supine position. First shots can be done to adjust for
to form brain at the front end and spinal cord at the rear
the signal density, as shown in Figure 42.2. Single-shot rapid
end. Failure of this structure to close on the 28th concep-
acquisition T2-weighted images with refocused echoes in the
tion day results in a broad spectrum of NTDs. Anencephaly
axial, coronal, and sagittal planes are obtained. T1-weighted
and spina bifida are the most common variants of NTD. The
images are also acquired, since it can better depict fat and
former begins with maldevelopment of the skull (acrania)
hemorrhage.16 Intravenous contrast study with gadolinium is
with protrusion of the brain (exencephaly). The brain is then
discouraged, because it rapidly crosses the placenta and its fetal
destroyed by its exposure to amniotic fluid.
safety profile has not been clearly shown.17
Fetal caudal regression, or sacral agenesis, syndrome is an
anomaly that strongly related to maternal diabetes mellitus.
The risk of caudal dysplasia increases remarkably, up to 600
times higher in infants born from diabetic mothers.19 This mal-
formation consists of agenesis of the lumbar vertebrae, sacrum,
and coccyx. Fetal lower extremities are poorly developed. Renal
malformations, such as hydronephrosis, ureteral duplication,
and renal agenesis, are also commonly associated with caudal
regression syndrome. It can lead to oligohydramnios that sig-
nificantly compromise ultrasound evaluation of the fetus.
Additional benefits of magnetic

resonance imaging in the detection of
fetal anomalies related to gestational
diabetes
Most fetal anomalies related to gestational diabetes can be
detected by ultrasound, especially in fetal cardiovascular
system. Fetal echocardiography provides both spatial and
temporal information, which is very important to make a
diagnosis of cardiac disease in utero. Ultrasound is also a
Figure 42.2 First magnetic resonance imaging shots can help
in the formation of an overall position of the fetus. Note the great imaging modality for following up the progression of
spina bifida with protruded meningeal sac at the sacral end of fetal cardiac disease. MRI may complement the ultrasound
this baby. when certain fetal conditions related to maternal diabetes
Figure 42.3 Magnetic resonance image of a second-trimester Figure 42.5 Magnetic resonance (MR) image of a second-
fetus with small spina bifida at the lumbosacral level. The fetus trimester fetus with dysplastic kidney displayed in sagittal
is displayed in coronal plane. Multiplanar image display can plane. Fetal kidney (arrow) is small and irregular in shape.
help associate the defect with its surrounding fetal anatomy. Multiple hyposignaling small cavities are found in the subcorti-
cal area of the kidney, which is highly suggestive for dysplastic
kidney. Note the small amount of amniotic fluid that can affect
the image quality of ultrasound. MRI has superb soft tissue
delineation even in the presence of low liquor level.
Figure 42.4 Sagittal magnetic resonance image of a fetus Figure 42.6 Ultrafast sagittal magnetic resonance (MR)
with open neural tube defect at the sacral level. Note the image shows a fetus with neural tube defect and c loacal
protrusion of meningeal sac from the spinal canal at fetal sacral malformation. The cloacal malformation resulted in severe
level. No visible spinal cord tissue herniated in this meningeal oligohydramnios that makes ultrasound examination
sac. A favorable prognosis is expected in this low-level lesion suboptimal. MRI is a second-tier imaging modality in this case,
without spinal tissue herniation. when limitation of ultrasound is encountered.
One of the most unique benefits of MRI study in the fetus

is the detailed evaluation of NTD. Fetal NTDs are detected
by ultrasound. Direct sonographic finding of NTD is verte-
bral defect. Indirect sonographic findings, such as ventricu-
lomegaly or Arnold–Chiari type II malformation, can raise
a suspicion that lead to a final diagnosis. MRI can confirm
the presence of NTD, especially in small lesions, as shown
in Figure 42.3. Defect level can be clearly visualized, as in
Figure 42.4. MRI can help detect associated anomalies that
can significantly impact the prognosis, especially when the
liquor volume is low or when maternal adiposity precludes
adequate ultrasound examination, as shown in Figures 42.5
and 42.6. It is not uncommon for a fetus with NTD to have
multiple malformations. MRI can help demonstrate these
associated malformations with a much better clarity, as
shown in Figure 42.7. This benefit of MRI is important for
prenatal counseling, especially now that in utero surgical
repair is possible. The invasiveness of open fetal closure of
the NTD that put a significant risk to the mother may not be
justified if associated serious malformations are found.
In order for fetal NTD to be adequately assessed with
Figure 42.7 Sagittal magnetic resonance image of a second- MRI, basic knowledge of fetal structure as it appears on the
trimester fetus with a large omphalocele. Note that this fetus MRI is crucial. T2-weighted MRI is preferred to visualize
also had small spina bifida at the sacral level with the skin
covering the defect (closed neural tube defect). No hindbrain
fetal brain, spinal cord, and their integuments. The compari-
herniation is noted. son between T1- and T2-weighted images for spina bifida is
shown in Figure 42.8. Fetal skull usually shows hypointen-
mellitus are encountered. Examples of these conditions are sity, or “dark” signaling, on T2-weighted images. The ver-
as follows: tebrae also show hypointensity. Intervertebral discs show
hyperintensity, or “bright” signaling. The spinal canal shows
1. Fetal spina bifida with suspicion of hindbrain herniation hyperintensity because it is filled with cerebrospinal fluid.
2. Maternal morbid obesity precluding adequate ultrasound The spinal cord can be seen as a structure with hypointen-
examination sity signaling. The relationship between fetal spinal cord and
3. Oligohydramnios precluding adequate ultrasound its canal, the defect of vertebral column, and the presence
examination or absence of skin covering can be depicted with MRI, as
T1 T2
Figure 42.8 Comparison between T1- and T2-weighted sagittal fetal magnetic resonance images. Note that T2-weighted image
provides a clearer tissue definition for the fetal central nervous system compared to the T1-weighted image.
Figure 42.9 Sagittal magnetic resonance (MR) image of a large spina bifida at the lumbosacral level. Note the continuity of
the spinal canal into the meningeal sac. The sac did not contain spinal cord tissue. The skin covering the defect is visible on the
MR image. Postnatal image of the same infant shows a large defect with full skin cover (closed neural tube defect).
shown in Figure 42.9. Axial cut at the defect level can dem- always found with Arnold–Chiari type II malformation. This
onstrate the presence or absence of neural tissue in the sac, is a herniation of the brain stem, the cerebellar vermis, and the
as shown in Figure 42.10. fourth ventricle through the foramen magnum due to loss of
NTD is actually a broad variety of malformations related to pressure from leakage of cerebrospinal fluid. The degree of cer-
failure of closure of neural tube. It can be categorized into open ebellar herniation is related to seizures, bladder dysfunction,
and closed defect. Common open NTD include rachischisis and ambulation restriction.22 Cerebellum and other intracra-
and myelomeningocele. Rachischisis is the most severe form nial structures in the posterior fossa have limited visibility
of NTD and is invariably lethal. The absence of an extensive from standard ultrasound examination. MRI can yield more
segment of vertebral column leads to a significant exposure of details in the posterior fossa due to its superb soft tissue con-
neural tissue. The more common form of open NTD is myelo- trast, as shown in Figure 42.12.23 Associated higher cerebral
meningocele. Open NTD, as shown in Figure 42.11, is almost abnormalities, as a consequence of NTD, can also be clearly
demonstrated with MRI, as shown in Figure 42.13.
Spina bifida (the so-called meningomyelocele) is a com-
mon fetal malformation of the CNS. It can cause a signifi-
cant neurological handicap. In utero surgical restoration
of meningomyelocele has been increasingly offered after
the result of a randomized trial of prenatal versus postna-
tal repair of myelomeningocele (MoMs trial) suggested its
benefit of decrease postnatal shunt placement and improved
motor outcomes.24 Fetal MRI has been advocated in this
trial to detect hindbrain herniation, as shown in Figure 42.14.
This is used not only for prognostic purposes but also to
monitor the possibility of reversal of hindbrain herniation in
cases that opted for in utero repair. Comparison of advantages
and disadvantages between ultrasound and MRI in imaging
of fetal NTD are summarized in Table 42.2.
If defect of the neural tube closure occurs on the cranial
end, it will result in meningocele or meningoencephalocele,
as shown in Figure 42.15. Most of the time, the defect is on
the occipital part of the skull. In meningoencephalocele or
encephalocele, there is some brain tissue herniated in the
protruded meningeal sac. Encephalocele is associated with
a much worse neurological prognosis. Encephalocele can
Figure 42.10 Axial magnetic resonance image of the fetus
with open neural tube defect at the level of vertebral defect.
readily be detected on the ultrasound. MRI complements the
Note a large protrusion of meningeal sac without spinal tissue ultrasound by its ability to detect the herniated brain with
herniation. higher accuracy. This is particularly important in smaller
Figure 42.11 Sagittal magnetic resonance image of a fetus with large spina bifida. Large vertebral defect was seen extending
from upper lumbar down to the sacral level. There was no skin coverage, and the spinal cord is fully exposed, as confirmed by the
postnatal picture. Also note an obliteration of fetal cisterna magna and dilated cerebral ventricle, which indicate fetal hindbrain
herniation.
Figure 42.13 Axial fetal magnetic resonance image shows

bilateral dilatation of lateral cerebral ventricles in case with
Figure 42.12 Axial fetal magnetic resonance image
large open neural tube defect.
shows intact cerebellar vermis with cisterna magna.
Structures in posterior fossa are harder to visualize with
ultrasound.
to assess the severity of neural tissue herniation that can lead
to a more precise counseling.25
or deeper lesions, such as lesion in the base of the skull. Fetal caudal regression syndrome is usually detected by
Associated anomalies of the posterior fossa can be evaluated screening ultrasound examination. It has a typical sono-
with higher clarity, as shown in Figure 42.16. graphic finding of a sudden interruption of the spine with
Since NTD is actually a broad spectrum of diseases, the absent or hypoplastic sacrum. Associated malformations of
prognosis in neonatal and childhood period can be var- the lower extremities, such as hypoplastic limbs and sireno-
ied significantly. Larger size of the defect, higher vertebral melia, are common. The legs of the fetus may be consistently
level, and associated Arnold–Chiari type II malformation crossed. MRI can confirm these ultrasound findings by its
are among the important ultrasound prognostic factors for clearer soft tissue definition. It can also be used to assess the
infants born with NTD. The size of meningeal sac itself is not stenosis of the spinal canal, which can predict the degree of
necessarily associated with bad outcomes. Detailed charac- neurological deprivation.26
terization of the fetal cystic lesion is not possible with opti- Cardiac and great vessel malformations are consistently
mal MRI sequence. Since the presence of neural tissue in the linked to poor periconceptional maternal glycemic control.
sac can result in worse prognosis, MRI has been advocated Most of these cardiac defects can be detected on screening
(a) (b)
Figure 42.14 (a) Sagittal magnetic resonance images of a fetus with normal hindbrain and (b) a fetus with hindbrain herniation.
Table 42.2 Advantages and disadvantages of ultrasound and magnetic resonance imaging in imaging of fetal
neural tube defect
Spina bifida Ultrasound MRI

Advantage First-tier diagnostic imaging for NTD Second-tier confirmatory imaging for NTD
Convenient to follow up on the size and degree of Adequate visualization of posterior fossa
ventriculomegaly Assessment of intracranial involvement
Inexpensive and decision for in utero repair
Multidisciplinary off-line interpretation
Widely available
for neurological outcomes
Disadvantage Cannot adequately visualize posterior fossa and hindbrain Expensive
herniation Not widely available
Cannot adequately visualize associated fetal intracranial Requires an optimized MRI sequences
pathology, especially when the fetus is not in vertex position and ultrafast protocol
Requires an experienced perinatologist to perform and interpret
fetal neurosonogram
Detect only most severe lesion
Abbreviation: NTD, neural tube defect.
fetal echocardiography. Aside from structural defects, fetal ultrasound in the detection of fetal malformations when the
echocardiography can also show fetal heart rate, rhythm, and amniotic fluid is low.29 Anomalies in fetal gastrointestinal
wall motion. MRI is not widely used for the study of fetal heart system may be better visualized in the ultrasound.
and great vessel. Because the fetal heart is always moving,
the demonstration of fetal cardiac structures using standard
fetal MRI sequence is always suboptimal, as shown in Figure Rationale and application of magnetic
42.17. Gadolinium, a commonly used intravenous contrast
for MRI angiography, is contraindicated during pregnancy. resonance imaging in pregnancy
Gadolinium can readily cross the placenta and its mutagenic complicated with diabetes
effects are not clearly elucidated.27 However, with the develop-
ment of MRI sequence optimization, the image quality of the Diabetes mellitus is a strong risk factor for congenital malfor-
fetal heart and great vessels is improving without the need for mations, especially those with poor glycemic control in peri-
intravenous contrast agent.28 conceptional period. These women should be offered dedicated
In addition to anomalies in the CNS, MRI may have ultrasound as the primary imaging modality to detect and to
a complementary role in some other fetal malformations monitor the progress of most fetal malformations related to
related to maternal diabetes mellitus. Fetal renal agenesis maternal diabetes mellitus. MRI can complement ultrasound
related to maternal diabetes mellitus can result in anhy- examination, especially in certain malformation, such as NTD.
dramniotic state. Decreased amount of amniotic fluid makes Using MRI as a first-tier screening tool in all pregnancy com-
fetal ultrasound examination suboptimal. MRI is superior to plicated with diabetes mellitus is currently not advisable.
Rationale and application of magnetic resonance imaging in pregnancy complicated with diabetes 359
Figure 42.17 Axial magnetic resonance image of a fetus at

four-chamber plane. Note that fetal cardiac chambers and
great vessels are poorly delineated. The heart is constantly
moving, and it requires special MR sequence to adequately
Figure 42.15 Axial magnetic resonance image of a fetus with demonstrate the structures.
occipital meningoencephalocele. Note the large defect size
with a minimal protrusion of occipital lobe of the brain into the
hernia sac. abdominal adiposity can limit the acoustic window for fetal
ultrasound examination; therefore, a significant number of
fetal anomalies will be missed from the standard ultrasound.
It has been estimated that about one-third of women in the
United States are obese.32 MRI has no limitation in obesity
and can complement the clinical care if fetal anomalies are
suspected, but cannot be confirmed, on the ultrasound.
In fact, for an obese diabetic mother, one may choose to
evaluate the fetus in the first trimester using high-resolution
transvaginal ultrasound. Using transvaginal approach,
maternal adiposity and/or bony structures are less likely
to compromise the image quality, even in mothers with
high body mass index. Figure 42.18 shows the ultrasound
Figure 42.16 Sagittal magnetic resonance image of a fetus

with large occipital meningoencephalocele. Note that the sac
is large and covered by the skin. Herniation of the brain is
noted, along with dilatation of cerebral ventricle suggestive for
obstructive hydrocephalus.
Maternal obesity, defined as maternal body mass index

over 30 kg/m2, is frequently associated with abnormal glu-
cose tolerance and diabetes.30 Obesity is an independent risk
factor for congenital malformations due to imbalances in a
number of metabolic pathways.8 The chance of fetal cardiac
Figure 42.18 Fetal brain can be sonographically examined
malformation in the group of mothers with gestational dia- using transvaginal approach through one of the fetal fonta-
betes mellitus and obesity (BMI ≥ 30 kg/m2) was higher than nels. This approach can obviate the sonographic hindrance by
those of nonobese women (odd ratio, OR 2.82).31 Excessive maternal adiposity and fetal skull plate.
Figure 42.19 An example of using “niche mode”, a novel function of three-dimensional sonography that may improve the q uality
of fetal brain imaging. This ultrasound picture shows an image of porencephaly (destructive brain lesion) in three-orthogonal
multiplanar view.
Figure 42.20 Fetus at 11 weeks with spina bifida at the thoracic level diagnosed by transvaginal ultrasound during routine first
trimester screening. At this early gestational age, ultrasound is superior to magnetic resonance imaging (MRI) in the diagnosis of
spina bifida because of an excessive fetal movement and unclear safety profile of MRI to the first-trimester fetus.
examination of fetal brain using transvaginal approach Summary and future directions
through fetal fontanel. Ultrasound examination of fetal
brain has been enhanced by some novel 3D sonographic Diabetes mellitus is a common medical condition found dur-
modalities, as shown in Figure 42.19. Fetal movement can ing pregnancy. Those with poor glycemic control during the
be readily observed on ultrasound, and it could lead to an periconceptional period should be offered dedicated screen-
early diagnosis of some conditions. Figure 42.20 shows fetal ing ultrasound examination. Certain malformations, such as
spina bifida at thoracic levels, where the fetal contour is mal- NTD and heart defects, are strongly associated with mater-
formed and the movement of lower limbs seen from trans- nal diabetes mellitus. Ultrasound is the first-tier screening
vaginal ultrasound examination was limited. On the other imaging tool of choice in most cases. MRI can confirm the
hand, fetal MRI is not advisable in the first trimester because diagnosis in difficult cases, such as in obese mothers or oli-
excessive movement of the fetus at this stage can cause sig- gohydramnios. MRI can provide more details regarding
nificant motion artifact on the MRI. Diabetes-related poly- hindbrain herniation and neural tissue prolapsed in NTD
hydramnios can increase motion artifact in MRI. It takes that can be used for a proper counseling for postnatal out-
several factors to be considered for the choice of optimal fetal comes. The quality of fetal cardiac MRI has been improv-
imaging modality in each setting. ing in recent years, and it may be clinically applicable soon
References 361
in fetal cardiac assessment.33 Fetal neurobehavior could be using dynamic MRI protocol.34 It can be expected that after
affected by secondary intracranial lesion of NTD that can be perinatologists and radiologists gain more experiences with
studied by ultrasound. Fetal movement can be demonstrated MRI, its fetal indication will be broaden.
REFERENCES
1. Nold JL, Georgieff MK. Infants of diabetic mothers. Pediatr Clin 19. DeBoer T, Wewerka S, Bauer PJ, Georgieff MK, Nelson CA.
North Am 2004; 51: 619–637, viii. Explicit memory performance in infants of diabetic mothers at
2. Evers IM, de Valk HW, Visser GH. Risk of complications of 1 year of age. Dev Med Child Neurol 2005; 47: 525–531.
pregnancy in women with type 1 diabetes: Nationwide pro- 20. Garne E, Loane M, Dolk H et al. Spectrum of congenital anom-
spective study in the Netherlands. BMJ 2004; 328: 915. alies in pregnancies with pregestational diabetes. Birth Defects
3. Macintosh MC, Fleming KM, Bailey JA et al. Perinatal mortal- Res A Clin Mol Teratol 2012; 94: 134–140.
ity and congenital anomalies in babies of women with type 1 21. Agopian AJ, Tinker SC, Lupo PJ, Canfield MA, Mitchell LE,
or type 2 diabetes in England, Wales, and Northern Ireland: National Birth Defects Prevention Study. Proportion of neural
Population based study. BMJ 2006; 333: 177. tube defects attributable to known risk factors. Birth Defects
4. Lucas MJ, Leveno KJ, Williams ML, Raskin P, Whalley PJ. Early Res A Clin Mol Teratol 2013; 97: 42–46.
pregnancy glycosylated hemoglobin, severity of diabetes, and 22. Chao TT, Dashe JS, Adams RC, Keefover-Hicks A, McIntire DD,
fetal malformations. Am J Obstet Gynecol 1989; 161: 426–431. Twickler DM. Central nervous system findings on fetal mag-
5. Guerin A, Nisenbaum R, Ray JG. Use of maternal GHb con- netic resonance imaging and outcomes in children with spina
centration to estimate the risk of congenital anomalies in the bifida. Obstet Gynecol 2010; 116: 323–329.
offspring of women with prepregnancy diabetes. Diabetes care 23. Blondiaux E, Garel C. Fetal cerebral imaging—Ultrasound vs.
2007; 30: 1920–1925. MRI: An update. Acta Radiol 2013; 54: 1046–1054.
6. Stratmeyer ME, Greenleaf JF, Dalecki D, Salvesen KA. Fetal 24. Adzick NS, Thom EA, Spong CY et al. A randomized trial of
ultrasound: Mechanical effects. J Ultrasound Med 2008; 27: prenatal versus postnatal repair of myelomeningocele. N Engl J
597–605: quiz 6–9. Med 2011; 364: 993–1004.
7. Yagel S, Cohen SM, Rosenak D et al. Added value of three-/ 25. Russell NE, Chalouhi GE, Dirocco F, Zerah M, Ville Y. Not
four-dimensional ultrasound in offline analysis and diagnosis of all large neural tube defects have a poor prognosis: A case of
congenital heart disease. Ultrasound Obstet Gynecol 2011; 37: prenatally diagnosed limited dorsal myeloschisis. Ultrasound
432–437. Obstet Gynecol 2013; 42: 238–239.
8. Paladini D. Sonography in obese and overweight pregnant 26. Duczkowska A, Bekiesinska-Figatowska M, Herman-Sucharska
women: Clinical, medicolegal and technical issues. Ultrasound I et al. Magnetic resonance imaging in the evaluation of the
obstet Gynecol 2009; 33: 720–729. fetal spinal canal contents. Brain Dev 2011; 33: 10–20.
9. Smith FW, Adam AH, Phillips WD. NMR imaging in pregnancy. 27. Webb JA, Thomsen HS, Morcos SK, Members of Contrast
Lancet 1983; 1: 61–62. Media Safety Committee of European Society of Urogenital
10. Levine D, Barnes PD, Sher S et al. Fetal fast MR imaging: Radiology. The use of iodinated and gadolinium contrast
Reproducibility, technical quality, and conspicuity of anatomy. media during pregnancy and lactation. Eur Radiol 2005; 15:
Radiology 1998; 206: 549–554. 1234–1240.
11. Alfaraj MA, Shah PS, Bohn D et al. Congenital diaphragmatic 28. Dong SZ, Zhu M, Li F. Preliminary experience with cardiovas-
hernia: Lung-to-head ratio and lung volume for prediction of cular magnetic resonance in evaluation of fetal cardiovascular
outcome. Am J Obstet Gynecol 2011; 205: 43.e1–43.e8. anomalies. J Cardiovasc Magn Reson 2013; 15: 40.
12. Mevissen M, Buntenkotter S, Loscher W. Effects of static and 29. Bulas D, Egloff A. Benefits and risks of MRI in pregnancy.
time-varying (50-Hz) magnetic fields on reproduction and fetal Semin Perinatol 2013; 37: 301–304.
development in rats. Teratology 1994; 50: 229–237. 30. Yu CK, Teoh TG, Robinson S. Obesity in pregnancy. BJOG
13. Beers GJ. Biological effects of weak electromagnetic fields 2006; 113: 1117–1125.
from 0 Hz to 200 MHz: A survey of the literature with special 31. Martinez-Frias ML, Frias JP, Bermejo E, Rodriguez-Pinilla E,
emphasis on possible magnetic resonance effects. Magn Reson Prieto L, Frias JL. Pre-gestational maternal body mass index
Imaging 1989; 7: 309–331. predicts an increased risk of congenital malformations in
14. Shellock FG, Kanal E. Policies, guidelines, and recommenda- infants of mothers with gestational diabetes. Diabet Med 2005;
tions for MR imaging safety and patient management. SMRI 22: 775–781.
Safety Committee. J Magn Reson Imaging 1991; 1: 97–101. 32. Ogden CL, Carroll MD, McDowell MA, Flegal KM. Obesity
15. Glenn OA. Fetal central nervous system MR imaging. among adults in the United States—No statistically significant
Neuroimaging Clin N Am 2006; 16: 1–17, vii. chance since 2003–2004. NCHS Data Brief 2007; (1): 1–8.
16. Barkovich AJ, Atlas SW. Magnetic resonance imaging of intra- 33. Manganaro L, Savelli S, Di Maurizio M et al. Potential role
cranial hemorrhage. Radiol Clin North Am 1988; 26: 801–820. of fetal cardiac evaluation with magnetic resonance imaging:
17. Garcia-Bournissen F, Shrim A, Koren G. Safety of gadolinium Preliminary experience. Prenat Diagn 2008; 28: 148–156.
during pregnancy. Can Family Physician 2006; 52: 309–310. 34. Prayer D. Fetal MR. Eur J Radiol 2006; 57(2): 171.
18. Reis ZS, Miranda AP, Lage EM, Braganca RD, Costa CR, Cabral
AC. Echocardiographic findings of congenital cardiopathies
among fetuses of diabetic pregnant women and their relation-
ship with plasma fructosamine levels. J Matern Fetal Neonatal
Med 2011; 24: 943–947.
43 Continuous glucose monitoring
in pregnancy
Marlon Pragnell and Aaron Kowalski
around mealtimes and before going to sleep at night. The

Introduction optimal frequency of finger-stick blood glucose monitoring
Despite improvements in care, women with diabetes who remains unclear with the Diabetes in Pregnancy Study Group
become pregnant remain at high risk of adverse maternal of North America reporting in 2002 that “more measurements
and fetal outcomes including congenital malformations, are better than fewer, both for patient compliance and in order
perinatal mortality, spontaneous preterm delivery, and mac- to assess outcomes.”12 Continuous glucose monitoring (CGM)
rosomia.1–3 It is now well accepted that exposure to maternal is a technology that allows frequent glucose measurements
hyperglycemia, in pregnancy complicated by type 1 (T1D), (typically every 5 minutes) and the ability to monitor glucose
type 2 (T2D), or gestational diabetes mellitus (GDM), is the trends in real time. Although CGM is currently more expen-
primary cause of these potentially devastating complica- sive and mildly invasive, the use of CGM has demonstrated
tions.4–6 Hyperglycemia is toxic from the earliest stages of improved glucose control as a stand-alone device, integrated
embryo development with the risk of congenital malforma- with an insulin pump or ultimately as a component of a closed-
tions and fetal death rising dramatically with increasing loop artificial pancreas system in pilot studies.13–16 Already
maternal HbA1c. A recent U.K. population–based cohort CGM technology has provided unprecedented insights into
study of over 400,000 singleton pregnancies indicated glucose control in pregnancy and its pathophysiology when
that every 1% increase in HbA1c was associated with a complicated by diabetes.17–20 For the pregnant woman with
30% increase in the risk of major congenital abnormality.7 T1D and her healthcare provider, CGM provides a more com-
Maternal glucose levels equilibrate with the fetal circulation, prehensive view of how glucose levels fluctuate in response to
and this elevated glucose drives insulin secretion leading to her meals and other activities and may enable more informed
abnormal growth acceleration and increased risk of shoul- decision making to improve glucose control.
der dystocia, labor induction, and cesarean section. The
consequences of these intrauterine metabolic abnormali-
ties may extend well beyond pregnancy as offspring are at CGM
increased risk for obesity, impaired glucose tolerance, and
the development of T2D later in life.4,5,8,9 The risks of mater- CGMs in clinical use include devices manufactured by
nal complications also increase and include worsening and/ Medtronic (Northridge, CA), DexCom (San Diego, CA),
or development of nephropathy and retinopathy.10,11 Efforts and Abbott (Alameda, CA—not marketed in the United
to avoid these devastating complications can be particularly States) (Figure 43.1). These can be worn for 5–7 days and
challenging for women early in pregnancy as increased insu- consist of a disposable glucose sensor that is inserted
lin sensitivity often results in a three- to fivefold increase in through the skin into the interstitial tissue, a small trans-
the rate of severe hypoglycemia, although the consequences mitter that is fixed onto the body surface, and a separate
on the developing fetus have not yet been determined. Severe display/storage device. The transmitter contains the power
hypoglycemia remains the leading cause of death in preg- source for the sensor, the amplifier of the sensor signal, and
nant women with T1D. the wireless transmitter. The data generated by the sensor
The risks of hyperglycemia and hypoglycemia to both can be stored on the unit for later downloading (termed
mother and fetus can be mitigated by maintaining tight gly- “retrospective CGM” or “professional GGM”) or transmit-
cemic control through the diet and insulin and attentive ted for display on a handheld monitor in real time (termed
monitoring of blood glucose levels. Fingerstick blood glucose “RT-CGM”). The disposable glucose sensor consists of a
monitoring (also known as self-monitoring of blood glucose or very thin metallic filament coated with glucose oxidase
SMBG) is the universally accepted method for monitoring that generates an electric current by catalytic oxidation
glucose control in diabetic pregnancy, and pregnant women of glucose in the interstitial fluid. The signal generated is
with diabetes are advised to test seven times per day including filtered to remove noise, and the corresponding glucose
362
Efficacy of CGM in type 1 diabetes 363
the market, and insulin delivery systems are anticipated to

be available within the next 2–5 years.
Efficacy of CGM in type 1 diabetes

A number of well-controlled clinical trials in the T1D
(a)
population have established the efficacy of RT-CGM for
improving metabolic control and decreasing time spent in
hypoglycemia.23–27 The landmark JDRF CGM trial, which
studied 322 patients with T1D with baseline HbA1c >7.0%
over 26 weeks, showed that adults aged over 25 years using
(b)
multiple daily injections (MDIs) and RT-CGM experienced
a 0.5% reduction in HbA1c (from 7.6% to 7.1%) compared to
Figure 43.1 Various types of real-time continuous g lucose
monitors. (a) Stand-alone devices. From left to right: Abbott patients using MDI and SMBG.28 No significant reduction
Freestyle Navigator®, Dexcom Gen4®, and Medtronic in HbA1c was observed in subjects younger than 25 years,
Guardian®. (b) Continuous glucose monitors integrated with and there was no significant difference in hypoglycemia in
insulin pumps. Left: Medtronic MiniMed Paradigm Revel and any group. However, the greatest predictor of HbA1c lower-
Right: Animas Vibe. ing for all age groups was frequency of sensor use, which
was lower in younger age groups. In a smaller trial of 129
concentrations are displayed every 5 minutes (typically adults and children who were already well controlled with
288 values/day). Calibration by way of finger-stick blood baseline HbA1c <7.0%, outcomes combining HbA1c and
glucose readings is required at least twice per day. There hypoglycemia favored the group using RT-CGM, suggest-
have been debates in the field regarding the lag between the ing it is also beneficial for individuals with T1D who have
interstitial fluid glucose from capillary blood. While the already achieved good control.24 Similarly, the STAR 3
physiologic lag is about 5 minutes, there may be additional study, which included 485 subjects who switched from MDI
lag due to diffusion of interstitial fluid glucose across the with SMBG to sensor-augmented pump therapy, showed
sensor membrane (1–2 minutes) and signal filtering by the improved HbA1c without increased rates of severe hypo-
CGM software (3–12 minutes).21 This limitation is most glycemia or diabetic ketoacidosis (DKA) in both adults and
evident when glucose concentrations are changing rapidly children.29
(>2 mg/dL/minutes) and may be different during a rising Clinical studies in the nonpregnant diabetic population
vs. a falling phase. show that adherence to RT-CGM, with a minimum usage
In addition to recent improvements in both accuracy and of 4–5 days per week, is a key determinant of improved
form factor,22 these devices may include the option to display glucose control. A meta-analysis of six randomized con-
the data on a smartphone or smartwatch or to share with trolled trials (RCTs) consisting of 449 patients randomized
significant others and caregivers via the cloud. to RT-CGM and 443 to SMBG showed that RT-CGM was
RT-CGM systems can be set to alert the wearer when associated with a significant reduction in HbA1c, which
glucose values are too high or low via audible and/or vibra- was greatest in those with the highest HbA1c at baseline
tion alarms. These are particularly useful when the user and who most frequently used the sensors. A best fit regres-
does not usually perform SMBG finger sticks, such as the sion model showed that for every 1-day increase of sensor
first few hours following a meal and overnight—a particu- usage per week, the effect of RT-CGM vs. SMBG increased
larly vulnerable time for hypoglycemic seizures. Alerts are by 0.150% (95% confidence interval −0.194% to −0.106%),
also optional for the rate of glucose rise and fall. To avoid and every 1% increase in baseline HbA1c increased the
alarm fatigue, these can be customized based on individual effect by 0.126% (−0.257% to 0.0007%). 30
safety, tolerance for alerts, and the time of day (using differ- A Cochrane review of CGM studies reported until
ent thresholds during the daytime and nighttime). June 2011 performed a meta-analysis of 22 RCTs. The
Continuous glucose monitors are also available linked results indicated the benefit of CGM for patients using the
with an insulin pump, known as “sensor-augmented pump device compared with those using MDI and SMBG. After
systems,” in which glucose concentrations are conveniently 6 months, there was a significantly larger decline in HbA1c
displayed on the pump’s screen. Significant progress has for RT-CGM users starting insulin pump therapy compared
also been made in the development of future closed-loop with patients using MDIs of insulin and SMBG (mean differ-
artificial pancreas systems where CGM data will be con- ence in change in HbA1c level, −0.7% [95% confidence inter-
tinually relayed to a software controller or algorithm on a val −0.8% to −0.5%]). For patients starting with RT-CGM
smartphone or in the pump that calculates and drives the only, the average decline in HbA1c level 6 months after base-
delivery of the appropriate amount of insulin to maintain line was also statistically significantly larger for RT-CGM
glucose control. The first generations, which automatically users compared to SMBG users, but much smaller than for
suspend insulin infusion when hypoglycemia is predicted patients starting using an insulin pump and RT-CGM at the
and resume after the event has been avoided, are already on same time (mean difference change in HbA1c level, −0.2%
364 Continuous glucose monitoring in pregnancy
[95% confidence interval −0.4% to −0.1%]).31 None of these excursions was found to better correlate with birth weight
studies, however, included pregnant women with T1D. percentile than from a single glucose value from a 1-hour
RT-CGM use is not without challenges, and the increased oral glucose challenge test.36
burden perceived by some wearers is an important factor CGM has also provided insights into maternal glycemia
in adherence. Clinical psychologists have identified three in obese pregnancies; however, these have been limited by
major psychosocial themes in responders who have worn the relatively small numbers and short duration of retro-
RT-CGMs and improved their HbA1c or reduced hypogly- spective CGM data collection. Harmon and colleagues
cemia. These include (1) coping with frustrations and using compared 16 obese and 22 normal-weight pregnancies dur-
self-
controlled rather than emotion-based coping when ing early (14–18) and late (26–30) gestation. Obese women
faced with frustrations while using the device; (2) using were found to have higher 24-hour and nocturnal glucose
information in retrospective pattern analysis, not just profiles compared with normal-weight subjects and that
minute-by-minute data analysis, in glycemic management; there was little difference to any of the measured glucose
and (3) “significant other”/spousal involvement with inter- parameters whether the women were on ad libitum or con-
est, encouragement, and participation by their loved ones. trolled diets.37
Indeed, it was noted that CGM use promoted collaborative
diabetes management and increased spousal understanding
of diabetes especially when planning and managing preg- Physiological insights into diabetic pregnancy
nancy. Both responders and nonresponders expressed body CGM technology has proved invaluable for understanding
image concerns when using RT-CGM.32,33 the pathophysiology of pregnancy complicated by diabetes.
Mazze and colleagues performed a retrospective CGM study
in 82 pregnant women of whom 51 were healthy, 6 had T1D,
and 25 were diagnosed with GDM. Their study, which also
CGM in pregnancy: Physiological included 21 nonpregnant women, showed healthy pregnan-
insights and clinical utility cies were characterized by tight glucose control, whereas
increasing glycemic variability was observed in the diabetic
While many small studies have utilized CGM to gain a women. The authors suggested that the use of CGM through-
deeper understanding of glucose control in pregnancy and out diabetic pregnancy is “critical if mimicking normal glu-
its pathophysiology with diabetes, the evidence base for cose patterns is to be achieved.”19
effectiveness in clinical practice is weak, and additional Dalfra and colleagues assessed glycemic variability in 50
large well-designed randomized CGM trials are required to pregnant women (20 T1D, 20 GDM, and 10 healthy controls)
inform choices of glucose monitoring techniques.34 who wore retrospective CGM for 2 days in each trimester of
pregnancy. Women with T1D showed higher glucose vari-
ability, with a twofold higher risk of hyperglycemic spikes
Physiological insights in normal and obese pregnancy during the day than healthy women or those with GDM. The
Prior to the advent of CGM, very little was known even about latter have only slightly higher CV parameters than healthy
the “normal” glucose profile in pregnancy. The first study to controls.20
address this, conducted in 57 nondiabetic pregnant women Most recently, Law and colleagues applied functional data
on 72 hours of retrospective CGM, reported mean and noc- analysis (FDA) to data from a previously published RCTs in
turnal glucose levels of 84 ± 18 mg/dL and 68 ± 10 mg/dL, a total of 117 pregnant women with T1D (n = 89) and T2D
respectively, with a peak postprandial glucose level of 110 ± (n = 28) who used repeated CGM during pregnancy. They
16 mg/dL that occurred at an average of 70.5 ± 13 minutes demonstrated that lower, less variable glucose levels in the
following the start of the meal.35 first trimester and higher, more variable glucose levels in
The first detailed longitudinal CGM profiles of normal the second and third trimesters were significantly associ-
pregnancy demonstrated a tendency of postprandial glucose ated with large-for-gestational-age (LGA) babies. Temporal
levels to rise even during the course of normal pregnancies profiles in the first trimester indicated that the lower-average
with similar caloric intake. These profiles were obtained in glucose levels associated with LGA were driven by distinct
a study of 32 women where 72 hours of retrospective CGM dips in glucose levels midmorning and midevening, whereas
data were recorded at 16, 22, 30, and 36 weeks gestation and higher-average glucose levels associated with LGA in the
then at 6 weeks postpartum. Detailed food diaries showed second and third trimesters were driven by significantly
no significant change in caloric intake, and fasting glucose higher glucose levels that occur during the early hours of the
levels did not change significantly although an upward trend morning and afternoon in the second trimester and during
was observed at 36 weeks and was highest 6 weeks after late evening in the third trimester. The authors also noted
delivery. The 2-hour postprandial glucose levels tended to that the magnitude of the transient excursions detected by
increase throughout pregnancy and then decrease signifi- FDA of CGM were substantively larger than the differences
cantly 6 weeks postpartum. in summary statistical indices of average glucose levels and
More recently, in a study in 55 normal pregnancies using glucose variability, suggesting that the FDA approach may
retrospective CGM over 7 days between 24 and 28 weeks provide greater sensitivity in the analysis of glycemic control
of gestation, the area under the curve of hyperglycemic in pregnancy.17
CGM in pregnancy: Physiological insights and clinical utility 365
Clinical utility of retrospective CGM in diabetic and 2011, failed to show improvement in glycemic control
pregnancy or pregnancy outcomes in women with pregestational dia-
Retrospective CGM has been shown to facilitate therapeutic betes.41 This Danish study randomized 154 unselected preg-
changes as a patient educational tool during T1D and T2D nant women with T1D (n = 123) and T2D (n = 31) to either
pregnancy. In a trial of 71 women (46 T1D, 25 T2D) where the intermittent RT-CGM or SMBG. Participants in this study
baseline HbA1c was relatively high at 7.3% (SD 1.2%), improve- already had good glycemic control with baseline median
ments were demonstrated in maternal glucose control, birth HbA1c of 6.7% (range 5.3%–10.7%), and so the primary ther-
weight, and the risk of macrosomia. Study participants were apeutic focus was on preventing hypoglycemia. Women in
allocated to standard antenatal care with or without retrospec- both groups were advised to perform at least seven finger-
tive CGM. Women in the retrospective CGM group wore the stick tests daily (before and after meals and before bedtime),
device over four 6-weekly intervals. Women returned to the and those randomized to RT-CGM were advised to wear
clinic each time for their data to be downloaded and inter- the device for at least 6 days at 8, 12, 21, 27, and 33 weeks
preted with a clinician to guide modifications to diet, exer- gestation. Women in the RT-CGM group were additionally
cise, and insulin dosing with a focus on reducing postprandial encouraged to use the device continuously. However, only 49
hyperglycemia. Women who wore the device showed reduction out of 76 women (64%) reported using CGM as per protocol,
in HbA1c that first appeared between 28 and 32 weeks g estation and only 5 women (7%) wore the device near continuously
leading to a sustained HbA1c reduction of 5.8% vs. 6.4% (at least 60% of the time). The baseline HbA1c of 6.7% at
(p = 0.007) in controls. The infants of mothers wearing retro- randomization improved to 6.0% and 6.1% in RT-CGM and
spective CGM were found to benefit from a lower birth weight control groups, respectively, at 33 weeks with no apparent
standard deviation score (0.9 ± 2.0 vs. 2.6 ± 1.4; p = 0.05), lower RT-CGM benefits on glucose control (hyper- or hypoglyce-
birth weight percentile (69 vs. 93; p = 0.02), and reduced risk mia) or neonatal outcomes, either in MDI or insulin pump
of macrosomia (35% vs. 60%; p = 0.05) compared to controls.38 users. Neonatal morbidity was high, with 40% of infants
A recent study with retrospective CGM in 340 pregnant being macrocosmic and 25% of infants being delivered pre-
Chinese women with GDM reported CGM use was associ- term or treated for hypoglycemia. The key recruitment bar-
ated with better glycemic control and improved pregnancy rier for women entering the trial was identified as allocation
outcomes with reduced risk of preeclampsia and cesarean to the RT-CGM arm, and adherence was poor. A patient sat-
delivery, decreased birth weight, and improved neonatal isfaction questionnaire showed that adherence was limited
complications. In these studies, the women were guided, by inaccuracy compared with SMBG, technical challenges,
through modified dietary, activity, or insulin regimens, to and skin irritation. Additionally, women experienced up to
limit their postprandial glucose excursions over 140 mg/dL to 12 alarms/day during their first period of RT-CGM use, and
no more than 10 minutes/day and duration of asymptomatic one-third of these interrupted sleep.42
hypoglycemia to no more than 30 minutes/day. As a result, The lack of benefit of RT-CGM in this trial may be due to
more women in the CGM group received insulin regimen a relatively low baseline HbA1c compared with the retrospec-
than in the routine care group (27.9% compared with 6.9%) tive CGM trial by Murphy et al.38 and the RT-CGM trials that
and experienced a shorter duration of both hyperglycemia demonstrated efficacy in the nonpregnant population.28,29
and hypoglycemia.39 Indeed, the authors noted the lack of HbA1c reduction
GlucoMoms, a large multicenter open-label randomized was more in line with a study on well-controlled nonpreg-
clinical trial of retrospective CGM study, including a decision nant patients who saw no additional HbA1c improvement
and cost-effectiveness assessment, is nearing completion in the with RT-CGM.43 Very poor CGM adherence may also have
Netherlands (TrialRegister.nl NTR2996). This trial is enroll- been a contributing factor as the meta-analysis suggests
ing 300 women with T1D or T2D before 16 weeks gestational near-continuous sensor use (6–7 days/week) in addition to
age or with GDM requiring insulin therapy before 30 weeks higher baseline HbA1c (≥7%) is required for CGM benefit in
gestational age. Subjects have been randomized to stan- nonpregnant diabetic population.30 It was also noted that the
dard care or retrospective CGM with the device being worn intervention focused primarily on minimizing hypoglycemia,
5–7 days every 6 weeks. Based on their CGM profiles, sub- whereas the primary outcome, macrosomia, is usually associ-
jects receive dietary advice and insulin therapy adjustments ated with maternal hyperglycemia. Women in this trial still
as needed. The primary outcome is macrosomia (birth weight had 25% of SMBG readings above 144 mg/dL, with the major-
above 90%) with birth weight, composite neonatal morbidity, ity of these presumably due to postprandial excursions. Fine-
maternal outcome, and costs as secondary outcomes.40 tuning of carbohydrate intake and prandial insulin has been
identified as key to achieving stringent postmeal targets, and
therefore more robust carbohydrate-based insulin algorithms
Clinical utility of real-time CGM in diabetic pregnancy and/or regular dietary guidance could have helped women
While multiple randomized controlled clinical trials have achieve the level of postprandial glucose control necessary to
shown that sustained use of RT-CGM improves glucose reduce macrosomia and improve HbA1c in the trial.44
control in nonpregnant T1D populations, there remains a CONCEPTT, a large international trial of RT-CGM in
dearth of evidence in T1D pregnancy. women with T1D funded by JDRF, is currently underway to
The only randomized trial of RT-CGM in diabetic preg- evaluate the role of continuous RT-CGM before and during
nancy reported to date, conducted in Denmark between 2009 pregnancy (NCT01788527). The trial is enrolling 324 women
366 Continuous glucose monitoring in pregnancy
with T1D in two parallel studies, one in women planning also provide the most comprehensive assessment of pump vs.
pregnancy (n = 110) and the other in pregnant women up MDI on glucose control before and during pregnancy.
to 13-week 6-day gestation (n = 214). The primary outcome
is the change in HbA1c from baseline with secondary out-
comes that include CGM time in target, episodes of severe Conclusions
hypoglycemia, measures of glucose variability, and quality of
life. The baseline HbA1c for eligibility is >6.5% or <10% in Tight glycemic control in diabetic pregnancy is essential for
the prepregnant arm and >7.0% or <10% in the pregnant arm. achieving better maternal and fetal outcomes. Clinical stud-
To ensure patients randomized into the study can under- ies have shown that retrospective CGM in pregnancy can
stand the device and adhere to its use, a key challenge in the improve glycemic control and thereby reduce macrosomia.
Danish study, the CONCEPTT protocol includes a 6-day run RT-CGM may further improve control by enabling insu-
in phase during which time women will wear a retrospective lin therapy adjustments in real time; however, more clini-
CGM. CONCEPTT is anticipated to provide unprecedented cal data are needed before its use in diabetic pregnancy can
details of maternal control from conception until delivery be fully endorsed. CONCEPTT, a large international trial
and will be the first trial to evaluate the role of RT-CGM of RT-CGM in pregnant women with T1D, may provide a
metrics in pregnancy (time in target, time spent above/ better understanding of the clinical utility of RT-CGM in
below target with maternal/fetal outcomes). The study will improving glycemic control.
REFERENCES
1. Lapolla A et al. A multicenter Italian study on pregnancy outcome 19. Mazze R, Yogev Y, Langer O. Measuring glucose exposure and vari-
in women with diabetes. Nutr Metab Cardiovasc Dis 2008; 18(4): ability using continuous glucose monitoring in normal and abnor-
291–297. mal glucose metabolism in pregnancy. J Matern Fetal Neonatal
2. Jensen DM et al. Outcomes in type 1 diabetic pregnancies: A Med 2012; 25(7): 1171–1175.
nationwide, population-based study. Diabetes Care 2004; 27(12): 20. Dalfra MG et al. Glucose fluctuations during gestation: An addi-
2819–2823. tional tool for monitoring pregnancy complicated by diabetes. Int J
3. Negrato CA, Mattar R, Gomes MB. Adverse pregnancy outcomes Endocrinol 2013; 279021(10): 11.
in women with diabetes. Diabetol Metab Syndr 2012; 4(1): 41. 21. Davey RJ et al. Contribution of an intrinsic lag of continuous glu-
4. Ali S, Dornhorst A. Diabetes in pregnancy: Health risks and man- cose monitoring systems to differences in measured and actual glu-
agement. Postgrad Med J 2011; 87(1028): 417–427. cose concentrations changing at variable rates in vitro. J Diabetes
5. Ballas J, Moore TR, Ramos GA. Management of diabetes in preg- Sci Technol 2010; 4(6): 1393–1399.
nancy. Curr Diab Rep 2012; 12(1): 33–42. 22. Kropff J et al. Accuracy of two continuous glucose monitoring
6. Metzger BE et al. Hyperglycemia and adverse pregnancy out- systems: A head-to-head comparison under clinical research cen-
comes. N Engl J Med 2008; 358(19): 1991–2002. tre and daily life conditions. Diabetes Obes Metab 2015; 17(4):
7. Bell R et al. Peri-conception hyperglycaemia and nephropa- 343–349.
thy are associated with risk of congenital anomaly in women 23. Bergenstal RM et al. Sensor-augmented pump therapy for A1C
with pre-existing diabetes: A population-based cohort study. reduction (STAR 3) study: Results from the 6-month continuation
Diabetologia 2012; 55(4): 936–947. phase. Diabetes Care 2011; 34(11): 2403–2405.
8. Yessoufou A, Moutairou K. Maternal diabetes in pregnancy: 24. Juvenile Diabetes Research Foundation Continuous Glucose
Early and long-term outcomes on the offspring and the concept Monitoring Study Group et al. Sustained benefit of continuous
of “metabolic memory.” Exp Diabetes Res 2011; 2011: 218598. glucose monitoring on A1C, glucose profiles, and hypoglycemia in
9. Vlachova Z et al. Increased metabolic risk in adolescent offspring adults with type 1 diabetes. Diabetes Care 2009; 32(11): 2047–2049.
of mothers with type 1 diabetes: The EPICOM study. Diabetologia 25. Ruedy KJ, Tamborlane WV, Juvenile Diabetes Research Foundation
2015; 58(7): 1454–1463. Continuous Glucose Monitoring Study Group. The landmark JDRF
10. Vestgaard M et al. Pregnancy-induced sight-threatening diabetic continuous glucose monitoring randomized trials: A look back
retinopathy in women with Type 1 diabetes. Diabet Med 2010; at the accumulated evidence. J Cardiovasc Transl Res 2012; 5(4):
27(4): 431–435. 380–387.
11. Bhatnagar A et al. Diabetic retinopathy in pregnancy. Curr 26. Juvenile Diabetes Research Foundation Continuous Glucose
Diabetes Rev 2009; 5(3): 151–156. Monitoring Study Group et al. Factors predictive of use and of
12. Reece EA et al. A consensus report of the Diabetes in Pregnancy benefit from continuous glucose monitoring in type 1 diabetes.
Study Group of North America Conference, Little Rock, Arkansas, Diabetes Care 2009; 32(11): 1947–1953.
May 2002. J Matern Fetal Neonatal Med 2002; 12(6): 362–364. 27. Slover RH et al. Effectiveness of sensor-augmented pump therapy in
13. Lodwig V et al. Current trends in continuous glucose monitoring. children and adolescents with type 1 diabetes in the STAR 3 study.
J Diabetes Sci Technol 2014; 8(2): 390–396. Pediatr Diabetes 2012; 13(1): 6–11.
14. Lodwig V et al. What are the next steps in continuous glucose 28. Tamborlane WV et al. Continuous glucose monitoring and inten-
monitoring? J Diabetes Sci Technol 2014; 8(2): 397–402. sive treatment of type 1 diabetes. N Engl J Med 2008; 359(14):
15. Thabit H, Hovorka R. Bringing closed-loop home: Recent 1464–1476.
advances in closed-loop insulin delivery. Curr Opin Endocrinol 29. Bergenstal RM et al. Effectiveness of sensor-augmented insulin-
Diabetes Obes 2014; 21(2): 95–101. pump therapy in type 1 diabetes. N Engl J Med 2010; 363(4):
16. Cengiz E et al. New-generation diabetes management: Glucose 311–320.
sensor-augmented insulin pump therapy. Expert Rev Med Devices 30. Pickup JC, Freeman SC, Sutton AJ. Glycaemic control in type 1
2011; 8(4): 449–458. diabetes during real time continuous glucose monitoring compared
17. Law GR et al. Analysis of continuous glucose monitoring in preg- with self monitoring of blood glucose: Meta-analysis of randomised
nant women with diabetes: Distinct temporal patterns of glucose controlled trials using individual patient data. BMJ 2011; 343:
associated with large-for-gestational-age infants. Diabetes Care d3805.
2015; 38(7): 1319–1325. 31. Langendam M et al. Continuous glucose monitoring systems
18. Siegmund T et al. Longitudinal changes in the continuous glucose for type 1 diabetes mellitus. Cochrane Database Syst Rev 2012;
profile measured by the CGMS in healthy pregnant women and 18(1): CD008101.
determination of cut-off values. Eur J Obstet Gynecol Reprod Biol 32. Ritholz MD et al. Psychosocial factors associated with use of con-
2008; 139(1): 46–52. tinuous glucose monitoring. Diabet Med 2010; 27(9): 1060–1065.
References 367
33. Ritholz MD et al. Impact of continuous glucose monitoring on 39. Yu F et al. Continuous glucose monitoring effects on maternal gly-
diabetes management and marital relationships of adults with cemic control and pregnancy outcomes in patients with gestational
Type 1 diabetes and their spouses: A qualitative study. Diabet diabetes mellitus: A prospective cohort study. J Clin Endocrinol
Med 2014; 31(1): 47–54. Metab 2014; 99(12): 4674–4682.
34. Moy FM, Ray A, Buckley BS. Techniques of monitoring blood 40. Voormolen DN et al. Effectiveness of continuous glucose moni-
glucose during pregnancy for women with pre-existing diabetes. toring during diabetic pregnancy (GlucoMOMS trial); a ran-
Cochrane Database Syst Rev 2014; 4: CD009613. domised controlled trial. BMC Pregnancy Childbirth 2012; 12:
35. Yogev Y et al. Diurnal glycemic profile in obese and normal 164.
weight nondiabetic pregnant women. Am J Obstet Gynecol 41. Secher AL et al. The effect of real-time continuous glucose moni-
2004; 191(3): 949–953. toring in pregnant women with diabetes: A randomized con-
36. Sung JF et al. Correlation of continuous glucose monitoring pro- trolled trial. Diabetes Care 2013; 36(7): 1877–1883.
files with pregnancy outcomes in nondiabetic women. Am J 42. Secher AL et al. Patient satisfaction and barriers to initiating
Perinatol 2014; 28: 28. real-time continuous glucose monitoring in early pregnancy in
37. Harmon KA et al. Continuous glucose profiles in obese and women with diabetes. Diabet Med 2012; 29(2): 272–277.
normal-weight pregnant women on a controlled diet: Metabolic 43. Beck RW et al. The effect of continuous glucose monitoring
determinants of fetal growth. Diabetes Care 2011; 34(10): in well-controlled type 1 diabetes. Diabetes Care 2009; 32(8):
2198–2204. 1378–1383.
38. Murphy HR et al. Effectiveness of continuous glucose monitoring 44. Murphy HR. Continuous glucose monitoring in pregnancy: We
in pregnant women with diabetes: Randomised clinical trial. BMJ have the technology but not all the answers. Diabetes Care 2013;
2008; 337: a1680. 36(7): 1818–1819.
44 Insulin infusion pumps
in pregnancy
Ilana Jaye Halperin and Denice S. Feig
CSII is the therapy of choice for a growing number of

Introduction to continuous patients with type 1 diabetes mellitus (T1DM) because of
subcutaneous insulin infusion therapy: the flexibility and precision that it offers over multiple daily
injections (MDIs). However, high costs and the intensive self-
Components, evolution, and use in management requirements mean that it is still only used in a
the nonpregnant population minority of patients. A 2012 report from the Australian Institute
of Health and Welfare showed that 20% of patients with
Continuous subcutaneous insulin infusions (CSIIs), more T1DM used CSII, with more than 50% of the patients being
commonly known as “insulin pumps,” were first developed less than 25 years old and 60% of the users being female.4 In
as a research tool in the late 1970s.1 Insulin pumps use a bat- Canada, a pump costs approximately 7000 Canadian dollars
tery driven motor to continuously deliver insulin through a and supplies about 6000 Canadian dollars/year. Increasingly,
subcutaneous port. In an effort to mimic endogenous insulin Western governments are paying for pumps and supplies for
production by a healthy pancreas, prespecified rates of insu- patients who are compliant and felt to benefit by their caregiv-
lin are infused continuously as the basal rate. The patient, via ers, and this may increase the use of pumps substantially.
the pump, delivers spurts of insulin, known as bolus doses,
when carbohydrates are eaten. Components of CSII therapy
include the pump, which has a battery, motor, plunger, and Comparison of CSII to multiple daily
insulin reservoir, tubing that delivers the insulin from the
reservoir to the insertion site, and the insertion port that injections for glycemic control in the
adheres to the skin and keeps the subcutaneous cannula nonpregnant population
in place.
Pump technology continues to evolve at a rapid pace, and A 2002 meta-analysis of 12 randomized controlled tri-
the options available in North America today include pumps als (RCTs) comparing 301 patients on CSII to 299 patients
without tubing, glucose meters that communicate directly on MDIs showed that CSII patients had an average blood
with pumps via wireless technology, and devices that link sugar level of 0.9 mmol/L less (95% confidence interval [CI]
with continuous glucose monitors (CGMs) enabling “low 0.5–1.2), a mean A1C of 0.5% less (95% CI 0.2–0.7), and a 14%
glucose suspend mode,” which temporarily suspends the lower total daily dose (TDD) of insulin when compared to
infusion if the patient’s CGM senses hypoglycemia.2 Most patients on MDI (7.3 unit/day range 4.1–10.6).5 It is important
newer-generation pumps automatically calculate bolus to note that these trials were done prior to the widespread
doses based on prespecified insulin-to-carbohydrate ratios use of insulin analogues such as glargine and detemir. The
(i.e., how much insulin the patient needs to cover a certain MDI regimens included neutral protamine hagedorn (NPH)
amount [grams] of carbohydrate) and insulin sensitivity insulin and ultralente insulin, both of which are more often
factors (how much one unit of insulin will lower the blood associated with nocturnal hypoglycemia and thus possibly
glucose level) that have been programmed into the pump limiting the glycemic control that could be achieved. A more
by the patient and/or caregiver. For the bolus doses to be recent study comparing CSII and a glargine-based MDI regi-
accurate and effective, however, patients are required to men showed no difference in A1C but a significant decrease
enter their current glucose measured by capillary blood and in glucose variability.6 Other commonly touted advantages
have a good grasp on carbohydrate counting. Despite these to insulin pump therapy is the flexibility of basal dosing to
advancements, the elusive “closed-loop” system in which the target the dawn phenomenon (high fasting blood sugar lev-
insulin pump functions as an extracorporeal pancreas and els that begin around 4 a.m.), without having to risk over-
the patient is not required to actively self-manage is not yet night hypoglycemia. A number of randomized trials have
available commercially.3 concluded that with comparable or even improved glycemic
368
Starting and adjusting CSII therapy during pregnancy 369
control, CSII is associated with up to 50% less hypoglycemia American Journal of Obstetrics and Gynecology meta-analy-
than MDI.1 However, the meta-analyses only looked at glu- sis of 6 studies with 107 women on CSII and 106 women on
cose variability and not hypoglycemia. The Ontario Health MDI showed no significant difference in pregnancy outcomes
Technology Assessment Series reviewed the literature in 2009 and glycemic control between users of MDI and CSII. Higher
and felt there was conflicting data on hypoglycemia in CSII number of ketoacidosis episodes and diabetic retinopathy was
compared with MDI.7 One concern for users of CSII is that found in the CSII group but did not reach statistical signifi-
without any long-acting insulin on board, pump malfunc- cance.11 In their discussion, both authors stressed the need
tion can quickly result in ketoacidosis. Studies have shown, for large, multicenter, RCTs that also assessed the quality of
however, that with proper education and resources, the rates life and cost-effectiveness.
of ketoacidosis are similar between the users of CSII or MDI.1 Unfortunately, since that time there have been no further
The key is that patients who use CSII have access to a needle RCTs. Since 2007, over six observational studies have been
and syringe so that they can give a subcutaneous injection published.12–18 None show a significant difference in patient
in the case of pump malfunction. Most pump suppliers will important outcomes such as cesarean delivery and perina-
provide a backup pump within 24 hours. tal morbidity or mortality. One study showed significantly
Insulin pumps are generally recommended to any patient higher Apgar scores in babies born to mothers who used
with T1DM who has one or more of the following: difficulty CSII during pregnancy.12 Two studies showed a lower HBA1c
with labile (high and low) blood sugars, frequent or severe (6.2% vs. 6.5% p = 0.0214 and 6.4% vs. 6.8% p = 0.00219) and
lows, hypoglycemia unawareness, variable work schedules, lower insulin doses at parturition. Chen et al. showed a
intensive exercise program, gastroparesis, or a desire for bet- higher rate of diabetic ketoacidosis in the CSII group (13%
ter control to prevent complications.8 vs. 2% p = 0.04) and neonatal hypoglycemia (35% vs. 13%
CSII therapy is not right for all patients with T1DM. In p = 0.01).17 However, Kallas-Koeman et al. showed no differ-
order to benefit from pump therapy, patients need to be ence in the rate of DKA or hypoglycemia between women on
engaged with a team of diabetes educators, be willing to do fre- MDI and on pump. They did show a higher rate of large-for-
quent self-monitoring of blood glucose, count carbohydrates gestational-age babies born to women on pump; however,
adequately, and be comfortable with the physical aspects the study was not powered to determine if this was a true
of insertion sites, tubing, and wearing a pump. In Ontario, association or due to some unknown confounder.19
Canada, the province provides funding only to patients with The current body of evidence suggests that pumps can be
T1DM who have demonstrated an ongoing commitment to safely used in pregnancy. Similar to the literature in the gen-
blood glucose monitoring, the safe and appropriate use of the eral population, it remains unclear if pump therapy offers
insulin pump, participation in an insulin pump education significant benefits over MDI in terms of hard outcomes
program, and regular diabetes clinic attendance. If patients such as diabetes complications, severe hypoglycemia, and
do not have hemoglobin A1C measurements done regularly pregnancy outcomes. However, there are certainly individ-
or do not attend visits with physicians and the diabetes edu- ual patients for whom CSII offers more flexibility, decreased
cation team, they lose their funding.9 glucose variability, and improved quality of life.
CSII compared to MDI in the Starting and adjusting CSII therapy

pregnant population during pregnancy
There have been two meta-analyses of RCTs comparing CSII In 2000, Gabbe et al. demonstrated that initiating CSII dur-
and MDI therapy in women with T1DM immediately prior to ing pregnancy was safe and effective. They published a retro-
and during pregnancy.10,11 Interestingly, despite having similar spective review of three groups: the first started CSII during
inclusion criteria and search strategies, the two meta-analysis pregnancy, the second used MDI during pregnancy, and the
pooled slightly different studies. Most of the studies were done third started CSII prior to pregnancy; there was no differ-
in the 1980s and 1990s, consisted of very small sample sizes ence in A1C across all three groups throughout pregnancy. Of
and were of poor methodological quality. The Cochrane meta- those that started CSII during pregnancy, 8 of 24 had severe
analysis10 of 5 trials and 154 pregnancies did include one study hypoglycemia prior to starting the pump, whereas only 1
done in 2005; however, this was only published as an abstract had a severe hypoglycemic episode afterward. Compared to
and results are no longer available online. The final results no ketoacidosis in those patients on MDI or who started the
for the Cochrane meta-analysis showed no significant differ- pump prior to pregnancy, 70% started as an outpatient and 8%
ence in maternal outcomes of cesarean birth (risk ratio [RR] (2/24) experienced ketoacidosis. At 12 months postpartum,
1.09, 95% CI 0.66–1.77), maternal hypoglycemia (RR 3.00, 95% of the patients continued with CSII and they had a signifi-
95% CI 0.35–25.87), or maternal hyperglycemia (RR 7.00, 95% cantly better A1C compared to those on MDI (7.2% vs. 9.1%).20
CI 0.39–125.44). Also no significant differences were found in When starting patients on an insulin pump prior to or
perinatal mortality (RR 2.33, 95% CI 0.38–14.32), macroso- during pregnancy, a multidisciplinary team of certified
mia (birth weight greater than 4000 g) (RR 3.20, 95% CI 0.14– pump trainers, dietitians, nurse educators, and physicians
72.62), fetal anomaly (RR 1.07, 95% CI 0.07–15.54), or small for is needed.21 A TDD of insulin is calculated from the cur-
gestational age (RR 1.40, 95% CI 0.10–18.71).10 Similarly, the rent MDI regimen and reduced by 10%–20% depending on
370 Insulin infusion pumps in pregnancy
the current glycemic control. At the start of pregnancy, basal the more traditional IV insulin and glucose infusion.
insulin makes up 50% of the TDD—to get an hourly basal One retrospective study from Italy of 65 women on CSII
rate, divide this by 24. With increasing pregnancy, bolus rates reported that starting at 28 weeks gestation, women and
surpass basal rates in making up the TDD.22 Most patients their caregivers were given weekly education on how to
on CSII have 2–4 basal rates per day to reflect differences in manage the pump during labor and delivery. 24 Women
insulin sensitivity throughout the day. These can be prepro- were instructed to enter a profile B into their pump that
grammed if there is a known dawn phenomenon or adjusted was 50% of their lowest basal rate at the end of pregnancy
after basal rate testing is completed. Basal rate testing involves and a profile C that was 0.1 unit/hour. If women were in
skipping a meal and refraining from vigorous activity dur- good metabolic control leading up to the early stages of
ing that time. Frequent capillary blood glucose checks (every labor, they were instructed to continue with their usual
2 hours or more) are required. If the basal rate is appropriate, basal rates (profile A) until the initiation of anesthesia
then the blood sugar level should remain constant throughout or the beginning of active labor at which point they were
the period of testing. If the capillary glucose readings drift to switch to profile B. If they had a low blood sugar level
up, the basal rate is considered to be too low. If the capillary (<3.8 mmol/L), they were to switch to profile C until
glucose readings drift down, then the basal rate is too high. the BG returned to normal. If the blood glucose level
The TDD is also used to determine the carbohydrate-to- was about 7.8 mmol/L, a correction bolus of 0.5–2 unit
insulin ratio that is preprogrammed into the pump. The tra- was administered. If the blood sugar level was above
ditional teaching is to divide 500 by the TDD to get a ratio of 10 mmol/L, CSII was stopped and IV insulin was initi-
X g for 1 unit of insulin. For example, a TDD of 50 units/day ated. The average blood glucose level across the 65 women
results in an insulin-to-carbohydrate ratio of 10, meaning that was within the target of 3.8–7.8 mmol/L, regardless of
for every 10 g of carbohydrate, 1 unit of insulin is required. mode of delivery; none of the women experienced a sig-
This will also be refined over time with postprandial blood nificant hypoglycemia (<2.8 mmol/L) or needed to switch
glucose monitoring. Finally, the sensitivity factor will be cal- to intravenous insulin. Neonatal hypoglycemia occurred
culated by dividing 100 by the TDD; this will give a correction in 17% of the deliveries; all were preterm. 24 The authors
factor in mmol/L, for example, a TDD of 50 units/day yields concluded from this study that CSII can be used safely
a sensitivity factor of 2, meaning that each 1 unit of insulin during labor and delivery in highly motivated patients.
lowers the blood glucose level by 2 mmol/L. It is worth noting that in this protocol, the upper target
It is important to provide women with T1DM anticipa- of 7.8 mmol/L is higher than the target recommended by
tory guidance at the start of pregnancy. Regardless of the some guidelines, for example, the 2013 Canadian Diabetes
mode of insulin delivery, they should be aware that their Association Guidelines, which recommends a target of
insulin requirements may decrease in the first trimester and 4.0–7.0 mmol/L to prevent neonatal hypoglycemia (grade
then slowly increase after 12–14 weeks, with a leveling off or D, based on observational data and no prospective, con-
even a decrease in requirements in the last 6–8 weeks prior trolled studies). 25 The American Diabetes Association
to delivery. Titration of insulin doses will be required every makes no recommendation on target blood sugar lev-
2–7 days to maintain blood sugar levels within the tight tar- els intrapartum. The National Institute for Health and
gets of pregnancy. Two studies have now shown that bolus Care Excellence in the United Kingdom recommends
insulin doses increase twice as much as basal doses do.22,23 maintaining the intrapartum blood sugar level between
One study of 9 women followed closely during pregnancy 4.0 and 7.0. 26 The patient and physician should be given
showed that the average TDD increased from 33 to 94 units. the option of switching to an intravenous insulin infu-
Interestingly, basal rates doubled but bolus doses quadru- sion if the patient become medically unstable or unable
pled, and the basal/bolus split went from 50/50 preconcep- or unwilling to continue with the pump during labor and
tion to 25/75 at delivery.22 The second study of 26 women delivery. Some pump companies advise disconnecting the
showed that the most marked increase in the bolus was seen pump when undergoing a surgical procedure where elec-
with breakfast, where the insulin-to-carbohydrate ratio went trocautery is performed. Cesarean sections often involve
from median 12 to 13 over the course of the pregnancy.23 some degree of cautery; however, our experience and that
As the consequences of diabetic ketoacidosis in preg- of the Italian center24 is that there have been no reported
nancy are dire to the fetus and insulin precision is para- cases of pump malfunction; this needs to be validated in
mount, trainers recommend that women change their a larger survey.
infusion sites every 48 hours (as opposed to 72 hours in the Postpartum insulin requirements drop considerably, and
nonpregnant population), although there is no evidence to women are usually instructed to take between 50% and 75%
back up this recommendation. of their prepregnancy basal rates and decrease to two-thirds
of their prepregnancy carbohydrate ratio and sensitivity fac-
tor. Breastfeeding should be encouraged, but women should
CSII in labor, delivery, postpartum, be informed that they may require even less insulin while
and breastfeeding breastfeeding because the glucose shunts to the breast milk.
In a small study of 18 women, those who breastfed used
While feasible, it is still not common practice to continue on average 0.2 unit/kg/day compared to 0.3 unit/kg/day in
CSII during labor and delivery; many centers still favor women who did not breastfeed.27
References 371
Table 44.1 Advantages to pump therapy over Table 44.2 Disadvantages of pump therapy
conventional multiple daily injections compared to multiple daily injections
• Provides more precision in calculating basal rates and • Patients always have a device attached to them.
bolus doses that should result in tighter glycemic • To achieve optimal control, patients need to have skills
control. in carbohydrate counting and must monitor blood
• Allows for changes in basal rates to account for sugar levels frequently (this is actually required for MDI
changes in activity. as well, but some patients on MDI with fixed
• Can adjust basal rate to account for the dawn carbohydrate content in their meals can achieve good
phenomenon. control).
• Easier to give boluses for meals and snacks. • Pump malfunction can lead to diabetic ketoacidosis
quickly.
• Bolus delivery can be rapid or extended depending on
fat and carbohydrate content of meals. • More expensive.
• Absorption of insulin is more reliable. • Need to have skills to troubleshoot device when
malfunction occurs.
• Only need to change site every 24–48 hours as
opposed to multiple injections daily. • Requires a multidisciplinary team who are skilled at
managing pump therapy.
• Allows for pairing with continuous glucose sensor with
potential for low glucose suspend feature.
Table 44.3 Common definitions in pump therapy

Conclusions
Carbohydrate ratio: Expressed as 1 unit of insulin per X g
There is ample evidence that CSII therapy can be used safely of carbohydrate. Indicates how many grams of
in pregnancy; whether it results in improved outcomes for carbohydrate are covered by one unit of insulin. Used
mother and baby over conventional MDI remains unclear. to calculate bolus dose based on patient’s carbohydrate
Women on CSII need to be supported by an experienced team counting. Initially calculated by dividing 500 over the
of diabetes nurses, dietitians, and physicians and should be total daily dose of insulin.
seen frequently during pregnancy as titration of basal rates, Sensitivity factor: Expressed as 1 unit for X mmol/L.
Indicates how many mmol/L drop in capillary blood
insulin-to-carbohydrate ratios, and sensitivity factors are
glucose can be expected with one unit of insulin given
required every week or two. Insulin pumps can probably as a correction bolus for a high sugar.
be used safely in labor and delivery with a caveat that labor Bolus insulin: Surge of insulin given for a high blood sugar
and delivery nursing staff, obstetricians, and partners should level or to cover the carbohydrate content of the
be aware that women can be switched to an intravenous meal—made up of both the carbohydrate ratio and
insulin infusion at any time if the situation necessitate this. sensitivity factor and based on the preset glycemic
Postpartum, it is recommended that doses of both basal and targets.
bolus insulin be reduced by one- to two-thirds of the pre- Basal rates: Slow infusion of insulin usually 0.5–2 units/
hour can be adjusted through the day to accommodate
pregnancy settings. Breastfeeding should be encouraged, but changes in insulin sensitivity such as the dawn
pump settings may need to be further reduced to prevent phenomenon.
associated hypoglycemia (see Tables 44.1 through 44.3).
REFERENCES
1. Pickup J, Keen H. Continuous subcutaneous insulin infusion at 6. Bruttomesso D, Crazzolara D, Maran A et al. In Type 1 diabetic
25 years: Evidence base for the expanding use of insulin pump patients with good glycaemic control, blood glucose variability is
therapy in type 1 diabetes. Diabetes Care March 2002; 25(3): lower during continuous subcutaneous insulin infusion than dur-
593–598. ing multiple daily injections with insulin glargine. Diabetic Med
2. Fry A. Insulin delivery device technology 2012: Where are March 2008; 25(3): 326–332.
we after 90 years? J Diabetes Sci Technol July 2012; 6(4): 7. Health Quality Ontario. Diabetes strategy evidence platform:
947–953. A summary of evidence-based analyses. Ont Health Technol
3. Pickup J. Insulin pumps. Diabetes Technol Ther February 2013; Assess Ser 2009; 9(19): 1–42.
15(Suppl. 1): S24–S28. 8. Walsh J, Roberts R. Pumping Insulin: Everything You Need to
4. Australian Institute of Health and Welfare. Insulin Pump Use in Succeed on an Insulin Pump. 5th ed., San Diego, CA: Torrey
Australia. Diabetes Series No. 18. Australian Institute of Health Pines Press, 2012.
and Welfare: Canberra, Australian Capital Territory, Australia, 9. Assistive Devices Program: Insulin Pump Therapy. 2012. Available
2012, p. 58. from: http://www.health.gov.on.ca/en/public/programs/adp/
5. Pickup J, Mattock M, Kerry S. Glycaemic control with contin- insulin_pamp.aspx.
uous subcutaneous insulin infusion compared with intensive 10. Farrar D, Tuffnell DJ, West J. Continuous subcutaneous insulin
insulin injections in patients with type 1 diabetes: Meta- infusion versus multiple daily injections of insulin for pregnant
analysis of randomised controlled trials. BMJ March 23, 2002; women with diabetes. Cochrane Database Syst Rev 2007; (3):
324(7339): 705. 005542.
372 Insulin infusion pumps in pregnancy
11. Mukhopadhyay A, Farrell T, Fraser RB, Ola B. Continuous subcuta- 19. Kallas-Koeman M, Kong J, Klinke J et al. Insulin pump use in
neous insulin infusion vs intensive conventional insulin therapy in pregnancy is associated with lower HbA1C without increas-
pregnant diabetic women: A systematic review and metaanalysis ing the rate of severe hypoglycemia or diabetic ketoacidosis in
of randomized, controlled trials. Am J Obstet Gynecol November women with type 1 diabetes. Diabetologia 2014; 57: 681–689.
2007; 197(5): 447–456. 20. Gabbe SG, Holing E, Temple P, Brown ZA. Benefits, risks,
12. Wender-Ozegowska E, Zawiejska A, Ozegowska K et al. Multiple costs, and patient satisfaction associated with insulin pump
daily injections of insulin versus continuous subcutaneous insu- therapy for the pregnancy complicated by type 1 diabetes mel-
lin infusion for pregnant women with type 1 diabetes. Aust N Z J litus. Am J Obstet Gynecol June 2000; 182(6): 1283–1291.
Obstet Gynaecol April 2013; 53(2): 130–135. 21. Bernasko J. Insulin pump therapy for pregnancy: A primer.
13. Talaviya PA, Saboo BD, Joshi SR et al. Pregnancy outcome and gly- J Matern Fetal Neonatal Med June 2012; 25(6): 552–557.
cemic control in women with type 1 diabetes: A retrospective com- 22. Roeder HA, Moore TR, Ramos GA. Insulin pump dosing across
parison between CSII and MDI treatment. Diabetes Metab Syndr gestation in women with well-controlled type 1 diabetes melli-
April–June 2013; 7(2): 68–71. tus. Am J Obstet Gynecol October 2012; 207(4): 324.e1–324.e5.
14. Bruttomesso D, Bonomo M, Costa S et al. Type 1 diabetes con- 23. Mathiesen J, Secher A, Ringholm L et al. Changes in basal rates
trol and pregnancy outcomes in women treated with continuous and bolus calculator settings in insulin pumps during preg-
subcutaneous insulin infusion (CSII) or with insulin glargine and nancy in women with type 1 diabetes. J Matern Fetal Neonatal
multiple daily injections of rapid-acting insulin analogues (glargine- Med 2013; early online: 1–5.
MDI). Diabetes Metab November 2011; 37(5): 426–431. 24. Fresa R, Visalli N, Di Blasi V et al. Experiences of continuous
15. Gonzalez-Romero S, Gonzalez-Molero I, Fernandez-Abellan M subcutaneous insulin infusion in pregnant women with type 1
et al. Continuous subcutaneous insulin infusion versus multiple diabetes during delivery from four Italian centers: A retrospec-
daily injections in pregnant women with type 1 diabetes. Diabetes tive observational study. Diabetes Technol Ther April 2013;
Technol Ther April 2010; 12(4): 263–269. 15(4): 328–334.
16. Cypryk K, Kosinski M, Kaminska P, Kozdraj T, Lewinski A. Diabetes 25. Thompson D, Berger H, Feig DS et al. Diabetes and pregnancy.
control and pregnancy outcomes in women with type 1 diabetes Can J Diabetes 2013; 37(Suppl. 1): S61–S68.
treated during pregnancy with continuous subcutaneous insulin 26. National Collaborating Center for Women and Children’s
infusion or multiple daily insulin injections. Pol Arch Med Wewn Health. Diabetes in pregnancy: Management of diabetes and
June 2008; 118(6): 339–344. its complications from pre-conception to the post-natal period.
17. Chen R, Ben-Haroush A, Weismann-Brenner A, Melamed N, Hod Clinic Guidelines, March 2008. Available from: http://www.
M, Yogev Y. Level of glycemic control and pregnancy outcome in nice.org.uk/nicemedia/live/11946/41320/41320.pdf, accessed
type 1 diabetes: A comparison between multiple daily insulin injec- on March 28, 2014.
tions and continuous subcutaneous insulin infusions. Am J Obstet 27. Riviello C, Mello G, Jovanovic LG. Breastfeeding and the basal
Gynecol October 2007; 197(4): 404.e1–404.e5. insulin requirement in type 1 diabetic women. Endocr Pract
18. Gimenez M, Conget I, Nicolau J, Pericot A, Levy I. Outcome of April 2009; 15(3): 187–193.
pregnancy in women with type 1 diabetes intensively treated with
continuous subcutaneous insulin infusion or conventional therapy.
A case-control study. Acta Diabetol March 2007; 44(1): 34–37.
45 Closed-loop insulin delivery in
type 1 diabetes pregnancy
Zoe A. Stewart and Helen R. Murphy
Introduction Insulin pumps in pregnancy

The complications of diabetes in pregnancy are widely rec- In nonpregnant cohorts, insulin pumps that deliver insulin
ognized for both mother and infant and include higher rates continuously via a subcutaneous catheter have been demon-
of congenital anomaly, stillbirth, neonatal death, and mac- strated to achieve better glucose control with less frequent
rosomia.1 These complications can be partially mitigated hypoglycemia than is achieved with conventional multiple
with good glycemic control and minimization of time spent daily injections (MDIs) of insulin.15,16 However, the benefits
hyperglycemic.2–5 of pump therapy are less clear in pregnancy. There have
However, women with type 1 diabetes face a number been relatively few studies comparing pump therapy with
of challenges in achieving and maintaining tight glycemic MDIs during pregnancy, and most have been retrospective
control during pregnancy. Hormonal and other factors observational studies examining small numbers of patients.
mean that insulin requirements change with advanc- Further, the majority of extant data derive from older studies
ing gestation and can be difficult to accurately predict.6,7 conducted when clinicians were less familiar with the use of
Despite regular glucose monitoring, intensive insulin insulin pumps and before the development of rapid-acting
therapy, and “safe” HbA1c levels, women with type 1 dia- insulin analogues such as aspart and lispro that are used
betes spent an average of 10 hours daily with glucose levels routinely in current clinical practice.
outside the recommended range.8 Furthermore, women Extant literature suggests that use of insulin pumps during
with type 1 diabetes are at particular risk of hypoglycemia pregnancy is safe and achieves similar metabolic and neonatal
during pregnancy,6,9–11 and so the benefits of tight glyce- outcomes as conventional treatment,17–25 perhaps with lower
mic control must be weighed against the increased risk of rates of small-for-gestational-age infants,26 fewer episodes
hypoglycemia. of hypoglycemia,27 and lower insulin requirements.24,27–29
Advances in technology for glucose monitoring and A 2011 Cochrane Collaboration review of prospective
insulin delivery such as continuous glucose monitoring trials of CSII and MDI in pregnancy found a paucity of
(CGM), insulin pumps, and sensor-augmented pump ther- randomized controlled trials. Meta-analysis of included
apy offer the potential of improved glycemic control; how- t rials demonstrated a clinically insignificant increase in

ever, their efficacy in pregnancy and impact on obstetric mean birth weight when CSII was used, but no difference in
and neonatal outcomes need further evaluation.12 macrosomia, cesarean section rates, perinatal mortality, fetal
Closed-loop systems use glucose measurements obtained anomaly, maternal hypoglycemia, maternal hyperglycemia,
via CGM and a control algorithm to adjust insulin pump or small-for-gestational-age infants.30 Another meta-analysis
delivery in real time. Initial data suggest that closed-loop of 6 studies (107 women on CSII vs. 106 on MDI) showed
systems may be able to maintain near-normal glucose lev- comparable glucose control and pregnancy outcomes.23
els and prevent nocturnal hypoglycemia in pregnant women These studies had very small sample sizes (mean 18 pregnan-
with diabetes.13,14 cies per treatment arm) and lacked power to detect differ-
ences in maternal/infant outcomes.
However, some observational studies have suggested
Components of closed-loop systems increased rates of severe maternal hypoglycemia,24 keto-
acidosis, neonatal hypoglycemia,31 large-for-gestational-age
Closed-loop systems comprise three components: a CGM, infants, and perinatal mortality26 when insulin pump ther-
an insulin pump that delivers a continuous subcutaneous apy is used in pregnancy. Given the highly selected patient
insulin infusion (CSII), and a control algorithm that is populations and small sample sizes, the validity and repro-
administered via a computer system (see Figure 45.1). ducibility of these results are unclear.
373
374 Closed-loop insulin delivery in type 1 diabetes pregnancy
CGM and sensor-augmented pump

therapy in pregnancy
(a) It is widely accepted that in order to reduce the fetal impacts
of maternal diabetes, periods of hyperglycemia should be
(a) minimized. However, in doing so, the risk of hypoglycemia
is increased.37 Given that pregnant women are especially
prone to hypoglycemia,9,11 fear of hypoglycemia can limit the
ability to achieve near normoglycemia.
CGM devices comprise a transcutaneous sensor that con-
tinuously measures interstitial glucose as a measure of blood
(b)
glucose and a transmitter that either stores the data obtained
(c) by the sensor for retrospective analysis or transmits the
glucose values to a receiver so that they can be seen by the
patient in real time. CGM, therefore, provides a more detailed
Figure 45.1 Closed-loop systems comprise (a) a continuous description of glucose levels than can be achieved with self-
glucose monitor (sensor, transmitter, and receiver), (b) an monitoring of blood glucose via finger-stick measurements.
insulin pump, and (c) a control algorithm device. Patients using CGM can see trends in glucose readings
and set alerts for hyperglycemia or hypoglycemia that may
The majority of insulin pump studies in pregnancy have be missed when relying on clinical symptoms and finger-
been retrospective and observational in nature. Women who stick measurements alone.7,38 This allows for earlier inter-
are commenced on insulin pump therapy often have a lon- vention and potential avoidance of more severe glucose
ger duration of diabetes and are more likely to already have disturbances. In nonpregnant populations, CGM has been
diabetic complications.17,22,28 Therefore, it may be that the demonstrated to reduce HbA1c, glucose excursions, and
increased rates of complications noted previously reflect a exposure to hyper- and hypoglycemia.39–42
more severe glycemic disturbance rather than the mode of Sensor-augmented pump therapy involves a patient wear-
insulin delivery. ing both a CGM and an insulin pump. The glucose measure-
A retrospective analysis of 387 consecutive pregnancies ments obtained via CGM are entered into the pump (either
of women with type 1 diabetes from three Canadian cen- automatically or manually), and using inbuilt calculators,
ters found improved glycemic control with pump therapy the pump can recommend appropriate insulin bolus and/or
compared with MDI and no difference in frequency of correction doses.
severe hypoglycemia, diabetic ketoacidosis, or cesarean Treatment of type 1 diabetes using sensor-augmented
sections. 32 In this study, women on pump therapy started pump therapy has been demonstrated to improve glycemic
with lower prepregnancy HbA1cs and were more likely to control and reduce hypoglycemia when compared with
have had preconception care. Women treated with pump treatment using MDIs of insulin or insulin pump therapy
therapy were more likely to have large-for-gestational-age alone.43–47 The benefits of sensor-augmented pump therapy
infants with a trend toward a higher frequency of neo- appear to be greatest with increased sensor use42,44,47,48 and in
natal hypoglycemia but no difference in other neonatal patients who begin with higher HbA1c values.39,45,48
outcomes. Very few studies have examined the use of CGM devices
Insulin pumps can offer easier titration of doses, which or sensor-augmented pump therapy in pregnancy. To date,
is particularly relevant in pregnancy when insulin require- studies have demonstrated that CGM is accurate in preg-
ments are highly variable6,33 and increased user satisfac- nancy13,49 and can assist in the identification of hypergly-
tion and quality of life when compared with MDIs.25,34,35 cemia and nocturnal hypoglycemia50,51 and allow more
Therefore, some benefit may be conferred by using insulin targeted treatment,51 which results in lower HbA1c, lower
pumps even if glycemic control and maternal–fetal outcomes birth weight percentiles, and lower risk of macrosomia.52 As
remain unchanged.36 in nonpregnant populations, the devices are less accurate in
Given advances in our understanding of physiology, the periods of hypoglycemia or when glucose levels are changing
development of new insulin analogues, and improved prac- rapidly (e.g., during exercise).13,49,53
titioner familiarity with the use of insulin pump therapy, As a result of ongoing limitations of CGM, current rec-
many argue that published literature regarding the use of ommendations advise that these systems should only be
insulin pumps in pregnancy is outdated and that, if used used as an adjunct tool and should not replace the use of
to best effect, CSII may confer benefits including improved finger-stick blood glucose measurement. An international
glycemic control, similar to what is seen in nonpregnant Continuous Glucose Monitoring in Women with Type 1
populations. Large randomized controlled trials using new- Diabetes in Pregnancy Trial (CONCEPTT, ClinicalTrials.
generation pumps and fast-acting insulin analogues are gov NCT01788527) is currently evaluating the role of con-
needed. tinuous real-time CGM before and during pregnancy.
Progress to date 375
Another randomized controlled trial (GLUCOMOMS, Portable insulin pumps that deliver continuous subcu-
TrialRegister.nl NTR2996) is evaluating the role of retro- taneous infusions of insulin have been used in clinical care
spective CGM. since the early 1980s.60 Since then, pump technology has
continued to advance and is becoming increasingly popu-
lar. Improvements in insulin pump devices, availability of
Control algorithms portable CGM systems (CGMS), development of new quick-
acting insulin analogues, and more sophisticated control
Proportional integral derivative algorithms have accelerated the development of closed-loop
Proportional integral derivative (PID) algorithms control systems.
insulin administration using a combination of the propor- One approach to closed-loop systems uses intraperitoneal
tional, integral, and derivative terms.54 These terms adjust insulin delivery with a glucose sensor either implanted in
insulin doses according to real-time glucose levels, the area the superior vena cava61 or inserted subcutaneously.62 This
under the curve between the actual and target glucose levels, approach has the potential to reduce inherent delays in sub-
and rate of change of glucose to adjust insulin, respectively. cutaneous insulin absorption and delivery insulin in a way
The three terms and three individualized constants deter- that more closely mimics normal physiology. However, intra-
mine the insulin dose administered. The PID algorithm peritoneal devices are expensive, require surgical implanta-
reacts solely to measured glucose levels and cannot predict tion, and carry a higher risk of infection and failure than
future changes in glucose. subcutaneous pumps. Intraperitoneal pumps would be inap-
propriate for routine use during pregnancy.
All other closed-loop prototypes currently being inves-
Model predictive control
tigated use a subcutaneous glucose measurement device
Model predictive control (MPC) algorithms predict future coupled with CSII via a portable pump. This subcutaneous–
glucose levels based on patients’ predicted glycemic responses subcutaneous approach is attractive and likely represents
to insulin and meals.55 The difference between the glucose the most feasible model because it is minimally invasive and
concentration predicted by the model and a preset insulin uses devices that are already commercially available and
target is used to determine the rate of insulin delivery. The acceptable by patients.
majority of closed-loop systems currently under investigation
use an MPC algorithm because these systems can account
for meals, patient-initiated insulin boluses, and individual Progress to date
differences in insulin absorption more easily than systems
using other algorithms. Low-glucose suspend
The simplest forms of automated insulin delivery are low-
Fuzzy logic glucose suspend systems. In these systems, a message is sent
to the insulin pump to temporarily stop basal insulin deliv-
Fuzzy logic models recognize that the same dose of insulin
ery when the CGM device detects or predicts hypoglycemia
may not always cause the same change in glucose level. These
and there is no response to warning alarms. This feature has
models aim to emulate the decision-making processes and
been demonstrated to decrease frequency, duration, and
expertise of clinicians and try to account for multiple pos-
severity of hypoglycemia without causing hyperglycemia or
sible responses to the same insulin dose.56
ketoacidosis.63–68
Sensor-augmented pump systems with a low-glucose sus-
pend feature have been commercially available since 2009
History of closed-loop glucose control (Paradigm Veo and MiniMed 530G with Enlite; Medtronic
Diabetes, Northridge, CA). They suspend insulin delivery
The concept of automated glucose control is not new. In the for up to 2 hours based on measured hypoglycemia but not
1960s, a relatively simple “on-off” system of closed-loop glu- predicted hypoglycemia. In selected cohorts of high-risk
cose management using intravenous glucose and insulin was patients, the Paradigm Veo can reduce exposure to hypogly-
developed for use in research settings.57 This concept was cemia without inducing hyperglycemia.68,69
developed further to incorporate a computational algorithm, The effects of low-glucose suspend systems in pregnancy
and the first commercial closed-loop system, the Biostator have not been evaluated.
(Miles Laboratories, Elkhart, IN), was released in 1977.58
This system obtains blood via an intravenous catheter and
measures glucose levels before discarding the sample. Blood Overnight closed-loop insulin delivery
glucose results are then fed to a computer algorithm that Approximately 50% of severe hypoglycemic episodes occur
controls administration of intravenous insulin and dextrose. overnight and can therefore go unnoticed.70 The proportion
This large device is only suitable for use in closely monitored of severe hypoglycemic episodes that occur overnight is sim-
inpatient settings, and despite multiple attempts, optimal glu- ilar in pregnant and nonpregnant populations.71 However, in
cose control could not be achieved with available algorithms.59 pregnancy, the risk of adverse outcome may be particularly
high because counterregulatory responses to hypoglycemia insulin dose was given than when a fully automated closed-
are impaired and severe hypoglycemia is more frequent.10 loop system was used. 82 Other studies have utilized
One potential early application of closed-loop technol- closed-loop technology for basal insulin delivery but have
ogy is overnight, when meals and exercise are less relevant, relied on standard calculation and manual administration
so models for glucose control are less complex. Overnight of premeal insulin boluses to control meal-related glucose
closed-loop control could also offer substantial clinical ben- excursions with good effect.62,73,83
efits including tighter glycemic control and prevention of A randomized crossover trial of 12 pregnant women
nocturnal hypoglycemia. examined the use of 24-hour closed-loop glucose control
The short-term safety of overnight closed-loop glucose with standardized meals and exercise and manual admin-
control using an MPC-based algorithm has been demon- istration of premeal boluses.14 In this study, closed-loop
strated under research conditions in multiple nonrandom- insulin delivery achieved near normoglycemia with less fre-
ized and proof-of-concept trials72,73 and in randomized quent severe hypoglycemia than conventional insulin pump
crossover studies examining a total of 17 children, adoles- therapy.
cents, and adults.74 In these studies, overnight closed-loop Although preliminary, data from these studies suggest that
control resulted in better glycemic control and less expo- stable glycemic control can be achieved with day and night
sure to hypoglycemia, including after a large carbohydrate closed-loop insulin delivery. Meal times and exercise con-
meal and alcohol.75 Two small pilot studies have demon- tinue to pose the greatest challenges for closed-loop systems.
strated that overnight closed-loop control using an MPC- Day and night use of closed-loop technology has been dem-
based algorithm is safe in early and late pregnancy and can onstrated to be feasible in the outpatient setting72 although
achieve near normoglycemia while reducing exposure to efficacy has not been established outside of research settings
hypoglycemia.13,14 and further studies are ongoing. We previously described
Closed-loop systems using PID and fuzzy logic algo- a significant gestational delay of approximately 30 minutes
rithms have also been shown to be safe and able to maintain in time-to-peak plasma aspart concentration from early to
good overnight glycemic control in small nonrandomized late pregnancy (taking 80 ± 30 minutes in late compared to
inpatient studies76–78 but have not been tested in pregnancy. 50 ± 10 minutes in early pregnancy).84 We recently high-
lighted the marked interoccasion variability of insulin phar-
macokinetics during type 1 diabetes pregnancy, showing
Day and night (24-hour) closed-loop insulin delivery that while basal insulin delivery was stable, the day-to-day
Ideally, closed-loop technology would be able to achieve variability of premeal boluses was particularly striking.85
and continuously maintain normoglycemia autonomously
including during exercise and meal times. However, this
remains an ambitious goal. In conventional treatment, Dual-hormone closed-loop systems
patients give a premeal insulin bolus to account for the food Under normal physiological conditions, glycemia is largely
they are about to eat, allowing time for insulin absorption regulated by feedback systems using insulin and gluca-
and action before glucose from the meal has reached the gon. In patients with type 1 diabetes, glucagon function is
bloodstream. However, closed-loop control around meal impaired,86 which increases the risk of hypoglycemia. Many
times would rely on changes in interstitial glucose to detect argue that an artificial pancreas using both insulin and glu-
meals, resulting in an inherent delay. These delays, together cagon to control glucose levels would more closely mimic
with the pharmacokinetics of currently available insulins, normal physiology and would therefore be better able to pre-
can result in postmeal hypoglycemia. vent hypoglycemia. Some investigators have demonstrated,
Four proof-of-concept trials of day and night closed- under research conditions, that dual-hormone systems can
loop control without meal announcement demonstrated provide good glycemic control and reduced exposure to
that closed loop could achieve acceptable glycemic control hypoglycemia.87–91 However, current formations of gluca-
with low rates of hypoglycemia.56,76,79,80 As expected, in fully gon are unstable and would not be viable for use in outpa-
automated closed-loop control, prolonged postprandial tient settings.92 There have been no studies of dual-hormone
hyperglycemia persisted and postprandial hypoglycemia closed-loop systems in pregnant women.
occurred frequently. Hypoglycemia could be avoided in the
first 2 hours after ingestion of a small meal (30 ± 5 g carbo-
hydrate) using an MPC algorithm with a safety constraint Limitations/obstacles
that accounts for insulin on board.80
As an intermediate step to 24-hour fully automated Sensor delay/inaccuracy
closed-loop control, some groups have proposed a system At present, the most viable option for commercial launch
that provides day and night closed-loop control but relies of a closed-loop insulin delivery system utilizes CGMs that
on the user to input the timing, and in some cases carbohy- measure interstitial glucose as a marker of blood glucose.
drate content, of meals so that the algorithm can calculate Although these monitoring systems have a level of accuracy
appropriate bolus doses.81 One study found that peak post- that is safe, factors including calibration error and sensor
prandial glucose levels were lower when a premeal priming drift can mean that glucose values measured by interstitial
References 377
sensors are less accurate than that of finger-stick blood avoid glucose excursions, fully automated closed-loop
g lucose measurements.93 systems rely on changes in interstitial glucose to identify
Overestimation of glucose is of particular concern meal times and exercise before the algorithm can respond
because, if relied upon, these values would lead to excessive and change the insulin infusion rate. This creates an inher-
insulin doses and potentially hypoglycemia. ent delay. While improvements in algorithms and adjunct
Further, the time required for glucose molecules to move tools like heart rate monitors can be used to identify meals
from the blood to the interstitium creates a lag of approxi- and exercise early, without meal or exercise announcement,
mately 6 minutes between blood glucose and interstitial glu- automated closed-loop systems will always be acting reac-
cose.94 While some control algorithms account for sensor tively rather than proactively.
inaccuracy and the known delay between blood and intersti- Closed-loop systems that incorporate meal announce-
tial glucose, the closed-loop system ultimately relies on the ment or manual premeal boluses will likely provide the most
glucose readings obtained via CGMs to make calculations feasible way forward until the challenges of maintaining
and adjust insulin doses. good glucose control during meal times and exercise with-
out human intervention can be overcome. This is certainly
the case in pregnancy where gestational delays in glucose
Time to onset and peak of insulin activity disposal and slower insulin pharmacokinetics mean that
As with all automated control systems, there is a time lag human intervention with premeal boluses 15–30 minutes
between the point at which the algorithm makes a change before eating will be required until rapid-acting insulins are
and the time at which the effect of that change is realized. available.
Most algorithms currently under investigation take into
account the time required for delivery and absorption of
insulin and the time to peak insulin activity. However, it can Conclusion/future directions
be difficult to accurately predict these times because absorp-
tion and action times of insulin vary considerably between Closed-loop insulin delivery systems have the potential to
different individuals and within the same individual at dif- revolutionize the treatment of type 1 diabetes, resulting in
ferent times.89,95 stable glycemic control with minimal patient intervention.
To avoid accumulation of insulin and subsequent iatro- This technology could be particularly beneficial in preg-
genic hypoglycemia, closed-loop algorithms need to adjust nancy, when glucose levels can be difficult to control, the risk
glucose levels gradually and take into account insulin that of hypoglycemia is increased, and reducing exposure both
has already been administered but is yet to reach its full to hyperglycemia and to hypoglycemia could have lifelong
effect.96 Rapid-acting insulin analogues including aspart and implications for both mother and child.
lispro have reduced insulin-related time lag97,98; however, it In nonpregnant populations, closed-loop control can
can still take up to 90 minutes to reach a maximum glucose- increase the time spent with target glucose levels and reduce
lowering effect.99 frequency and severity of hypoglycemia. Trials are cur-
Preliminary studies have found that mechanisms includ- rently underway to explore the feasibility and effectiveness
ing site warming100 and coadministration of hyaluroni- of closed-loop systems for overnight glucose control in out-
dase101,102 can accelerate the pharmacokinetics of insulin, patient settings over longer periods of time for both pregnant
and attempts at development of more rapid-acting insulin and nonpregnant populations.
analogues are ongoing.
Algorithm, meals, and exercise

Acknowledgments
Algorithms for closed-loop glucose control are now able to ZAS is supported by the Gates Cambridge Trust and Jean
achieve near normoglycemia overnight under controlled Hailes for Women’s Health Australia. HRM is supported by
conditions. However, closed-loop insulin delivery after a National Institute for Health Research (NIHR) research
meals and exercise remains challenging. Where conventional fellowship (CDF-2013-06-035). The views expressed in this
management uses knowledge of food and activity to proac- publication are those of the author(s) and not necessarily
tively adjust carbohydrate and insulin doses to potentially those of the NHS, the NIHR, or the Department of Health.
REFERENCES
1. Casson IF, Clarke CA, Howard CV et al. Outcomes of preg- 3. Confidential Enquiry into Maternal and Child Health. Diabetes
nancy in insulin dependent diabetic women: Results of a in Pregnancy : Are We Providing the Best Care? Findings of
five year population cohort study. Br Med J 1997; 315(7103): a National Enquiry: England, Wales and Northern Ireland,
275–278. CEMACH: London, U.K., 2007.
2. Kitzmiller J, Gavin L, Gin G et al. Preconception care of diabe- 4. Middleton P, Crowther CA, Simmonds L. Different intensities of
tes: Glycemic control prevents congenital anomalies. J Am Med glycaemic control for pregnant women with pre-existing diabetes
Assoc 1991; 265: 731–736. (Review). Cochrane Database Syst Rev 2012; 8: CD008540.
5. Kinsley B. Achieving better outcomes in pregnancies compli- 26. Chico A, Saigi I, García-Patterson A et al. Glycemic control and
cated by type 1 and type 2 diabetes mellitus. Clin Ther 2007; 29 perinatal outcomes of pregnancies complicated by type 1 diabe-
(Suppl. D): S153–S160. tes: Influence of continuous subcutaneous insulin infusion and
6. García-Patterson A, Gich I, Amini SB et al. Insulin requirements lispro insulin. Diabetes Technol Ther 2010; 12(12): 937–945.
throughout pregnancy in women with type 1 diabetes mellitus: 27. Wender-Ozegowska E, Zawiejska A, Ozegowska K et al. Multiple
Three changes of direction. Diabetologia 2010; 53(3): 446–451. daily injections of insulin versus continuous subcutaneous insulin
7. Kerssen A, Evers IM, de Valk HW et al. Poor glucose control in infusion for pregnant women with type 1 diabetes. Aust N Z J
women with type 1 diabetes mellitus and “safe” hemoglobin A1c Obstet Gynaecol 2013; 53(2): 130–135.
values in the first trimester of pregnancy. J Matern Fetal Neonatal 28. Bruttomesso D, Bonomo M, Costa S et al. Type 1 diabetes con-
Med 2003; 13(5): 309–313. trol and pregnancy outcomes in women treated with continu-
8. Murphy HR, Rayman G, Duffield K et al. Changes in the glycemic ous subcutaneous insulin infusion (CSII) or with insulin glargine
profiles of women with type 1 and type 2 diabetes during preg- and multiple daily injections of rapid-acting insulin analogues
nancy. Diabetes Care 2007; 30(11): 2785–2791. (glargine-MDI). Diabetes Metab 2011; 37(5): 426–431.
9. Ringholm L, Pedersen-Bjergaard U, Thorsteinsson B et al. 29. Zoric S, Micic D, Kendereski A et al. Use of continuous subcu-
Hypoglycaemia during pregnancy in women with type 1 diabe- taneous insulin infusion by a portable insulin pump during preg-
tes. Diabet Med 2012; 29(5): 558–566. nancy in women with type 1 diabetes mellitus. Vojnosanit Pregl
10. Ter Braak EWMT, Evers IM, Erkelens DW et al. Maternal hypo- 2006; 63(7): 648–651.
glycemia during pregnancy in type 1 diabetes: Maternal and fetal 30. Farrar D, Tuffnell DJ, West J. Continuous subcutaneous insulin
consequences. Diabetes Metab Res Rev 2002; 18(2): 96–105. infusion versus multiple daily injections of insulin for pregnant
11. Evers IM, ter Braak EWMT, de Valk HW et al. Risk indicators women with diabetes. Cochrance Database Syst Rev 2007; (3):
predictive for severe hypoglycemia during the first trimester CD005542.
of type 1 diabetic pregnancy. Diabetes Care 2002; 25(March): 31. Chen R, Ben-Haroush A, Weismann-Brenner A et al. Level of
554–559. glycemic control and pregnancy outcome in type 1 diabetes: A
12. Combs C. Continuous glucose monitoring and insulin pump comparison between multiple daily insulin injections and contin-
therapy for diabetes in pregnancy. J Matern Fetal Neonatal Med uous subcutaneous insulin infusions. Am J Obstet Gynecol 2007;
2012; 25(10): 2025–2027. 197(4): 404.e1–404.e5.
13. Murphy HR, Elleri D, Allen JM et al. Closed-loop insulin delivery 32. Kallas-Koeman MM, Kong JM, Klinke JA et al. Insulin pump use in
during pregnancy complicated by type 1 diabetes. Diabetes Care pregnancy is associated with lower HbA1c without increasing the
2011; 34: 406–411. rate of severe hypoglycaemia or diabetic ketoacidosis in women
14. Murphy HR, Kumareswaran K, Elleri D et al. Safety and efficacy with type 1 diabetes. Diabetologia 2014; 57(4): 681–689.
of 24-h closed-loop insulin delivery in well-controlled pregnant 33. Kerssen A, de Valk HW, Visser GH. Day-to-day glucose vari-
women with type 1 diabetes: A randomised crossover case ability during pregnancy in women with type 1 diabetes mel-
series. Diabetes Care 2011; 34: 2527–2529. litus: Glucose profiles measured with the Continuous Glucose
15. Weissberg-Benchell J, Antisdel-Lomaglio J, Seshadri R. Insulin Monitoring System. BJOG 2004; 111(9): 919–924.
pump therapy: A meta-analysis. Diabetes Care 2003; 26(4): 34. Linkeschova R, Raoul M, Bott U et al. Less severe hypoglycaemia,
1079–1087. better metabolic control, and improved quality of life in type 1
16. Rodrigues IAS, Reid HA, Ismail K et al. Indications and efficacy of diabetes mellitus with continuous subcutaneous insulin infusion
continuous subcutaneous insulin infusion (CSII) therapy in Type (CSII) therapy; an observational study of 100 consecutive patients
1 diabetes mellitus: A clinical audit in a specialist service. Diabet followed for a mean of 2 years. Diabet Med 2002; 19(9): 746–751.
Med 2005; 22(7): 842–849. 35. Hoogma RPLM, Hammond PJ, Gomist R et al. Comparison of
17. Cypryk K, Kosiński M, Kamińska P et al. Diabetes control and the effects of continuous subcutaneous insulin infusion (CSII) and
pregnancy outcomes in women with type 1 diabetes treated dur- NPH-based multiple daily insulin injections (MDI) on glycaemic
ing pregnancy with continuous subcutaneous insulin infusion control and quality of life: Results of the 5-nations trial. Diabet
or multiple daily insulin injections. Pol Arch Med Wewn 2008; Med 2005; 23(2): 141–147.
118(6): 339–344. 36. d’Annunzio G, Minuto N, Emmanuele V et al. Use of continuous
18. Fresa R, Visalli N, Di Blasi V et al. Experiences of continuous subcutaneous insulin infusion since the first weeks of pregnancy
subcutaneous insulin infusion in pregnant women with type 1 in five women with type 1 diabetes mellitus. Diabetes Res Clin
diabetes during delivery from four Italian centers: A retrospective Pract 2007; 78(3): e11–e12.
observational study. Diabetes Technol Ther 2013; 15(4): 328–334. 37. Rosenn BM, Miodovnik M, Holcberg G et al. Hypoglycemia:
19. Giménez M, Conget I, Nicolau J et al. Outcome of pregnancy in The price of intensive insulin therapy for pregnant women with
women with type 1 diabetes intensively treated with continuous insulin-dependent diabetes mellitus. Obstet Gynecol 1995; 85(3):
subcutaneous insulin infusion or conventional therapy. A case- 417–422.
control study. Acta Diabetol 2007; 44(1): 34–37. 38. Kerssen A, de Valk HW, Visser GHA. Do HbA1c levels and the
20. Gonzalez-Romero S, Gonzalez-Molero I, Fernandez-Abellan self-monitoring of blood glucose levels adequately reflect glycae-
M et al. Continuous subcutaneous insulin infusion versus mul- mic control during pregnancy in women with type 1 diabetes
tiple daily injections in pregnant women with type 1 diabetes. mellitus? Diabetologia 2006; 49(1): 25–28.
Diabetes Technol Ther 2010; 12(4): 263–269. 39. Pickup JC, Freeman SC, Sutton AJ. Glycaemic control in type 1
21. Hiéronimus S, Cupelli C, Bongain A et al. Pregnancy in type 1 diabetes during real time continuous glucose monitoring com-
diabetes: Insulin pump versus intensified conventional therapy. pared with self monitoring of blood glucose: Meta-analysis of
Gynécol Obstet Fertil 2005; 33(6): 389–394. randomised controlled trials using individual patient data. Br
22. Lapolla A, Dalfrà MG, Masin M et al. Analysis of outcome of Med J 2011; 343: d3805.
pregnancy in type 1 diabetics treated with insulin pump or con- 40. Garg S, Zisser H, Schwartz S et al. Improvement in glycemic
ventional insulin therapy. Acta Diabetol 2003; 40(3): 143–149. excursions with a transcutaneous, real-time continuous glucose
23. Mukhopadhyay A, Farrell T, Fraser RB et al. Continuous subcu- sensor: A randomized controlled trial. Diabetes Care 2006; 29(1):
taneous insulin infusion vs intensive conventional insulin therapy 44–50.
in pregnant diabetic women: A systematic review and metaanaly- 41. Klonoff DC. Continuous Glucose Monitoring: Roadmap for 21st
sis of randomized, controlled trials. Am J Obstet Gynecol 2007; century diabetes therapy. Diabetes Care 2005; 28(5): 1231–1239.
197(5): 447–456. 42. Juvenile Diabetes Research Foundation Continuous Glucose
24. Cyganek K, Hebda-Szydlo A, Katra B et al. Glycemic control and Monitoring Study Group, Tamborlane WV, Beck RW et al.
selected pregnancy outcomes in type 1 diabetes women on con- Continuous glucose monitoring and intensive treatment of type
tinuous subcutaneous insulin infusion and multiple daily injec- 1 diabetes. N Engl J Med 2008; 359(14): 1464–1476.
tions: The significance of pregnancy planning. Diabetes Technol 43. Nørgaard K, Scaramuzza A, Bratina N et al. Routine sensor-
Ther 2010; 12(1): 41–47. augmented pump therapy in type 1 diabetes: The INTERPRET
25. Gabbe SG, Holing E, Temple P et al. Benefits, risks, costs, and study. Diabetes Technol Ther 2013; 15(4): 273–280.
patient satisfaction associated with insulin pump therapy for the 44. Bergenstal RM, Tamborlane WV, Ahmann A et al. Effectiveness
pregnancy complicated by type 1 diabetes mellitus. Am J Obstet of sensor-augmented insulin-pump therapy in type 1 diabetes.
Gynecol 2000; 182(6): 1283–1291. N Engl J Med 2010; 363: 311–320.
References 379
45. Hermanides J, Nørgaard K, Bruttomesso D et al. Sensor- 68. Ly TTT, Nicholas J, Retterath A et al. Effect of sensor-augmented
augmented pump therapy lowers HbA1c in suboptimally con- insulin pump therapy and automated insulin suspension vs stan-
trolled type 1 diabetes; a randomized controlled trial. Diabet dard insulin pump therapy on hypoglycemia in patients with type
Med 2011; 28(10): 1158–1167. 1 diabetes: A randomized clinical trial. J Am Med Assoc 2013;
46. Schmidt S, Nørgaard K. Sensor-augmented pump therapy at 310(12): 1240–1247.
36 months. Diabetes Technol Ther 2012; 14(12): 1174–1177. 69. Agrawal P, Welsh JB, Kannard B et al. Usage and effectiveness of
47. Battelino T, Conget I, Olsen B et al. The use and efficacy of the low glucose suspend feature of the Medtronic Paradigm Veo
continuous glucose monitoring in type 1 diabetes treated Insulin Pump. J Diabetes Sci Technol 2011; 5(5): 1137–1141.
with insulin pump therapy: A randomised controlled trial. 70. The Diabetes Control and Complications Trial Research Group.
Diabetologia 2012; 55(12): 3155–3162. Epidemiology of severe hypoglycemia in the diabetes control and
48. Langendam M, Luijf YM, Hooft L et al. Continuous glucose complications trial. Am J Med 1991; 90(4): 450–459.
monitoring systems for type 1 diabetes mellitus. Cochrane 71. Ringholm Nielsen L, Pedersen-Bjergaard U, Thorsteinsson B
Database Syst Rev 2012; (1): CD008101. et al. Hypoglycemia in pregnant women with type 1 diabetes:
49. Kerssen A, de Valk HW, Visser GH. The Continuous Glucose Predictors and role of metabolic control. Diabetes Care 2008; 31:
Monitoring System during pregnancy of women with type 1 9–14.
diabetes mellitus: Accuracy assessment. Diabetes Technol Ther 72. Kovatchev B, Renard E, Cobelli C et al. Feasibility of outpatient
2004; 6(5): 645–651. fully integrated closed-loop control: First studies of wearable arti-
50. Yogev Y, Chen R, Ben-Haroush A et al. Continuous glucose ficial pancreas. Diabetes Care 2013; 36: 1851–1858.
monitoring for the evaluation of gravid women with type 1 dia- 73. Kovatchev B, Cobelli C, Renard E et al. Multinational study of
betes mellitus. Obstet Gynecol 2003; 101(4): 633–638. subcutaneous model-predictive closed-loop control in type 1
51. Yogev Y, Ben-Haroush A, Chen R et al. Continuous glucose diabetes mellitus: Summary of the results. J Diabetes Sci Technol
monitoring for treatment adjustment in diabetic pregnancies— 2010; 4(6): 1374–1381.
A pilot study. Diabet Med 2003; 20(7): 558–562. 74. Kumareswaran K, Elleri D, Allen JM et al. Meta-analysis of over-
52. Murphy H, Rayman G, Lewis K. Effectiveness of continu- night closed-loop randomized studies in children and adults with
ous glucose monitoring in pregnant women with diabetes: type 1 diabetes: The Cambridge Cohort. J Diabetes Sci Technol
Randomised clinical trial. Br Med J 2008; 337: 1–8. 2011; 5(6): 1352–1362.
53. Kumareswaran K, Elleri D, Allen JM et al. Accuracy of con- 75. Hovorka R, Kumareswaran K, Harris J et al. Overnight closed
tinuous glucose monitoring during exercise in type 1 diabetes loop insulin delivery (artificial pancreas) in adults with type 1 dia-
pregnancy. Diabetes Technol Ther 2013; 15(3): 223–229. betes: Crossover randomised controlled studies. Br Med J 2011;
54. Steil GM, Panteleon AE, Rebrin K. Closed-loop insulin 342: 1855.
delivery—The path to physiological glucose control. Adv Drug 76. Steil GM, Rebrin K, Darwin C et al. Feasibility of automating insu-
Deliv Rev 2004; 56(2): 125–144. lin delivery for the treatment of type 1 diabetes. Diabetes 2006;
55. Hovorka R, Canonico V, Chassin LJ et al. Nonlinear model pre- 55(12): 3344–3350.
dictive control of glucose concentration in subjects with type 1 77. O’Grady MJ, Retterath AJ, Keenan DB et al. The use of an auto-
diabetes. Physiol Meas 2004; 25(4): 905–920. mated, portable glucose control system for overnight glucose
56. Atlas E, Nimri R, Miller S et al. MD-Logic Artificial Pancreas control in adolescents and young adults with type 1 diabetes.
System: A pilot study in adults with type 1 diabetes. Diabetes Diabetes Care 2012; 35(11): 2182–2187.
Care 2010; 33(5): 1072–1076. 78. Nimri R, Atlas E, Ajzensztejn M et al. Feasibility study of auto-
57. Kadish AH. Automation control of blood sugar. I. A servomech- mated overnight closed-loop glucose control under MD-logic
anism for glucose monitoring and control. Am J Med Electron artificial pancreas in patients with type 1 diabetes: The DREAM
1964; 39: 82–86. Project. Diabetes Technol Ther 2012; 14(8): 728–735.
58. Clemens AH, Chang PH, Myers RW. The development of 79. Ruiz JL, Sherr JL, Cengiz E et al. Effect of insulin feedback on
Biostator, a glucose-controlled insulin infusion system (GCIIS). closed-loop glucose control: A crossover study. J Diabetes Sci
Horm Metab Res Suppl 1977; 7: 23–33. Technol 2012; 6(5): 1123–1130.
59. Broekhuyse H, Nelson J, Zinman B et al. Comparison of algo- 80. Dassau E, Zisser H, Harvey R et al. Clinical evaluation of a per-
rithms for the closed-loop control of blood glucose using the sonalized artificial pancreas. Diabetes Care 2013; 36: 801–809.
artificial beta cell. IEEE Trans Biomed Eng 1981; 28: 678–687. 81. Lujif YM, DeVries JH, Zwinderman K et al. Day and night closed-
60. Pickup JC, Keen H. Continuous subcutaneous insulin infusion loop control in adults with type 1 diabetes: A comparison of two
at 25 years: Evidence base for the expanding use of insulin closed-loop algorithms driving continuous subcutaneous insulin
pump therapy in type 1 diabetes. Diabetes Care 2002; 25(3): infusion versus patient self-management. Diabetes Care 2013; 36:
593–598. 3882–3887.
61. Renard E, Costalat G, Chevassus H et al. Closed loop insulin deliv- 82. Weinzimer SA, Steil GM, Swan KL et al. Fully automated closed-
ery using implanted insulin pumps and sensors in type 1 diabetic loop insulin delivery versus semiautomated hybrid control in
patients. Diabetes Res Clin Pract 2006; 74(Suppl. 2): S173–S177. pediatric patients with type 1 diabetes using an artificial pan-
62. Renard E, Place J, Cantwell M et al. Closed-loop insulin deliv- creas. Diabetes Care 2008; 31: 934–939.
ery using a subcutaneous glucose sensor and intraperitoneal 83. Elleri D, Allen JM, Kumareswaran K et al. Closed-loop basal insu-
insulin delivery: Feasibility study testing a new model for the lin delivery over 36 hours in adolescents with type 1 diabetes.
artificial pancreas. Diabetes Care 2010; 33(1): 121–127. Diabetes Care 2013; 36: 838–844.
63. Garg S, Brazg RL, Bailey TS et al. Reduction in duration of 84. Murphy HR, Elleri D, Allen JM et al. Pathophysiology of post-
hypoglycemia by automatic suspension of insulin delivery: prandial hyperglycaemia in women with type 1 diabetes during
The in-clinic ASPIRE study. Diabetes Technol Ther 2012; 14(3): pregnancy. Diabetologia 2012; 55(2): 282–293.
205–209. 85. Goudie RJ, Lunn D, Hovorka R et al. Pharmacokinetics of insulin
64. Cengiz E, Swan KL, Tamborlane WV et al. Is an automatic Aspart in pregnant women with type 1 diabetes: Every day is dif-
pump suspension feature safe for children with type 1 diabetes? ferent. Diabetes Care 2014; 37(6): e121–e122.
An exploratory analysis with a closed-loop system. Diabetes 86. Cryer PE. Minireview: Glucagon in the pathogenesis of hypo-
Technol Ther 2009; 11(4): 207–210. glycemia and hyperglycemia in diabetes. Endocrinology 2012;
65. Elleri D, Allen JM, Nodale M et al. Suspended insulin infusion 153(3): 1039–1048.
during overnight closed-loop glucose control in children and 87. Castle JR, Engle JM, El Youssef J et al. Novel use of glucagon in
adolescents with type 1 diabetes. Diabet Med 2010; 27(4): a closed-loop system for prevention of hypoglycemia in type 1
480–484. diabetes. Diabetes Care 2010; 33(6): 1282–1287.
66. Buckingham B, Cobry E, Clinton P et al. Preventing hypogly- 88. Haidar A, Legault L, Dallaire M et al. Glucose-responsive insu-
cemia using predictive alarm algorithms and insulin pump sus- lin and glucagon delivery (dual-hormone artificial pancreas) in
pension. Diabetes Technol Ther 2009; 11(2): 93–97. adults with type 1 diabetes: A randomized crossover controlled
67. Danne T, Kordonouri O, Holder M et al. Prevention of hypo- trial. Can Med Assoc J 2013; 185(4): 297–305.
glycemia by using low glucose suspend function in sensor- 89. El-Khatib FH, Russell SJ, Nathan DM et al. A bihormonal closed-
augmented pump therapy. Diabetes Technol Ther 2011; 13(11): loop artificial pancreas for type 1 diabetes. Sci Transl Med 2010;
1129–1134. 2(27): 27ra27.
90. Russell SJ, El-Khatib FH, Nathan DM et al. Blood glucose control 97. Lindholm A, Jacobsen LV. Clinical pharmacokinetics and phar-
in type 1 diabetes with a bihormonal bionic endocrine pancreas. macodynamics of insulin Aspart. Clin Pharmacokinet 2001; 40(9):
Diabetes Care 2012; 35(11): 2148–2155. 641–659.
91. Van Bon AC, Jonker LD, Koebrugge R et al. Feasibility of a bihor- 98. Howey DC, Bowsher RR, Brunelle RL et al. [Lys(B28). Pro(B29)]-
monal closed-loop system to control postexercise and post- human insulin: A rapidly absorbed analogue of human insulin.
prandial glucose excursions. J Diabetes Sci Technol 2012; 6(5): Diabetes 1994; 43(3): 396–402.
1114–1122. 99. Hovorka R. Continuous glucose monitoring and closed-loop sys-
92. Bakhtiani PA, Zhao LM, El Youssef J et al. A review of artificial tems. Diabet Med 2006; 23(1): 1–12.
pancreas technologies with an emphasis on bi-hormonal therapy. 100. Raz I, Weiss R, Yegorchikov Y et al. Effect of a local heating
Diabetes Obes Metab 2013; 15(12): 1065–1070. device on insulin and glucose pharmacokinetic profiles in an
93. Castle JR, Ward WK. Amperometric glucose sensors: Sources of open-label, randomized, two-period, one-way crossover study
error and potential benefit of redundancy. J Diabetes Sci Technol in patients with type 1 diabetes using continuous subcutaneous
2010; 4(1): 221–225. insulin infusion. Clin Ther 2009; 31(5): 980–987.
94. Wentholt IME, Hart AAM, Hoekstra JBL et al. Relationship 101. Vaughn DE, Yocum RC, Muchmore DB et al. Accelerated phar-
between interstitial and blood glucose in type 1 diabetes patients: macokinetics and glucodynamics of prandial insulins injected
Delay and the push-pull phenomenon revisited. Diabetes Technol with recombinant human hyaluronidase. Diabetes Technol Ther
Ther 2007; 9(2): 169–175. 2009; 11(6): 345–352.
95. Heinemann L. Variability of insulin absorption and insulin action. 102. Hompesch M, Muchmore DB, Morrow L et al. Accelerated insu-
Diabetes Technol Ther 2002; 4(5): 673–682. lin pharmacokinetics and improved postprandial glycemic control
96. Hovorka R. Closed-loop insulin delivery: From bench to clinical in patients with type 1 diabetes after coadministration of prandial
practice. Nat Rev Endocrinol 2011; 7(7): 385–395. insulins with hyaluronidase. Diabetes Care 2011; 34(3): 666–668.
46 Noninvasive glucose monitoring
Itai Ben-David and Pierre Singer
The ideal system would be as follows:

Glucose monitoring
Diabetes mellitus has been recognized for thousands of ●● Accurate—The demand for this is more pronounced at
years, mainly described by its effects on urine. However, times of rapid glucose level changes, especially when in
it wasn’t until circa 1900 that real progress took place in the dangerous hypoglycemic regions.
understanding the physiological and pathological basis of ●● Precise—High reproducibility rate.
the disease.1 Once insulin was discovered, later extracted ●● Fast—Allowing point-of-care use.
from animals and then purified and injected into humans, ●● Physically comfortable—Without pain or other local side
a treatment became available for what had been a fatal effects.
disease, mainly type 1 diabetes mellitus (T1DM). Insulin ●● Easy to use—Should not impede on daily activities and
treatment obviously required monitoring body glucose lev- should not require multiple prolonged calibration phases;
els, not the least of which is the risk for hypoglycemia that its user’s instructions should be easy to follow with a com-
might be deadly. In the mid-1950s, tablets for urine glucose fortable interface.
level monitoring became widely available, soon replaced ●● Affordable—Keeping in mind that T2DM12,13 and GDM14
by strips. During the 1960s, blood glucose level monitor- are more prevalent among persons from low socioeco-
ing became available using color strips that reacted chemi- nomic states.
cally and changed their color according to glucose levels.
These had to be deciphered using a color chart, meaning
they offered estimated rather than exact measurements.2 Noninvasive and minimally invasive glucose
In 1971, the Ames reflectance meter invented by Anton monitoring: Principles
Clemens was the first meter to assess automatically the Given the demands from the optimal monitor, it should
color change produced by the reagents and offer a point-of- be as noninvasive as possible, meaning glucose levels should
care blood glucose level monitoring system. be measured externally. Attempts toward this objective have
Along with the evolution of glucose level monitoring, been in progress for decades.1 Obviously, we are not there
evidence started accumulating as to the effect of long-term quite yet; people are still pricking their fingers several times
blood glucose level control on related complications and a day. In order to understand why, we need to summarize the
mortality. These beneficial effects were observed in T1DM,3,4 methods that have been developed so far to measure glucose
type 2 diabetes mellitus (T2DM),5,6 and gestational diabetes levels noninvasively.
mellitus (GDM).7 The main site used for glucose sensing is the skin. The
Self-monitoring of blood glucose (SMBG) is still skin is comprised of the epidermis, which gets most glucose
the cornerstone of treating insulin-dependent diabetes by diffusion, and the dermis, which is supplied by blood
(types 1 and 2)8,9 as well as GDM.7 Its beneficial effects vessels.15 The dermis is the common site for glucose sens-
have been less consistently proven in patients with T2DM ing, whether from capillary blood or interstitial fluid. Blood
not treated with insulin.10 However, it is not without prob- flow to the skin changes with age and disease states. As light
lems. Adherence has been shown11 to be less than optimal passes through the skin, it undergoes absorption and scatter-
with the aforementioned regime that requires at least four ing. The epidermis is a light-absorbing layer for light closer
daily finger sticks. Poor adherence was related to pain and to the ultraviolet range, but most red and infrared light
T2DM as well as environmental and psychosocial vari- passes unchanged, independent of skin pigmentation. In the
ables. Obviously, this only underlines the need for better dermis, the majority of light scattering takes place, mostly
monitoring systems. due to the optical properties of collagen and elastic fibers.
381
Shorter wavelengths pass largely unaffected through these 500

fibers as well as the interstitial fluid and are absorbed by the 450 A
C
small fraction of tissue that is comprised of blood and cells.
This range is referred to as the “optical window” in which 400 A
red and near-infrared (NIR) light are less absorbed by the 350
other tissue components. Most truly noninvasive glucose E B
300
monitoring (NIGM) systems are optical in nature and follow
Device
these physiological concepts. As variations in skin condi- 250 B
tions, such as pigmentation, temperature, hydration, and 200
blood flow, affect the skin’s optical properties, glucose sens-
150
ing becomes challenging. D D
A different compartment used for glucose sensing is the 100
interstitial fluid.16 This is sometimes referred to as minimally 50
invasive glucose monitoring (MIGM), and its measurements C E
0
are transdermal. Glucose passes from the capillary wall to 0 100 200 300 400 500
the interstitial compartment by means of simple diffusion to Glucose reference (mg/dL)
be later taken up by the cells. This process is affected by many
factors, such as metabolic rate, vessel wall permeability, and Figure 46.1 An example of a Clarke error grid evaluation. In
local insulin levels.17 When blood glucose level changes, this example, 74.5% of the measurements fall within area A
there is a lag time, estimated at 2–45 minutes, between the and 20.7% within area B, giving an acceptable result of 95.5%
within these areas. Of the measurements, 1% falls within
blood and the interstitial fluids. The relationship between the area E, considered the most hazardous as it would lead to
two compartments, blood and interstitial fluid, constantly erroneous treatment.
changing represents one of the greatest challenges when
developing NIGM or MIGM.
Although the skin is the main site of measurement, there EGA.22 It was evaluated23 by Wentholt et al. in 2006 and was
are a few others. The oral mucosa was used18 with limited suc- found inferior in detecting differences compared to accuracy
cess relying on the same general principles as with the skin. assessment methods such as linear regression and correla-
Other sites used for measurement are the eye and the tym- tion. In general, a perfect accuracy analysis system does not
panic membrane. Yet other methods estimate glucose levels exist, and usually a combination of absolute difference, cor-
from bodily fluids—tears and exhaled breaths. These methods relation, linear regression, and Clarke EGA is employed24
will be discussed in further detail later. (shown in Figure 46.1).
Monitor accuracy assessment Noninvasive glucose monitoring

During the 1970s and 1980s, as SMBG became an impor- systems
tant tool for managing diabetes, many systems were used to
determine accuracy. Correlation coefficients, linear regres- The appeal of NIGM systems is obvious. NIGM allows
sion slopes, and linear correlation are limited when com- patients with diabetes to measure glucose levels without
paring sets of data approaching unity.19 Percent deviation pain, irritation, and risk of infection that are associated
is more useful when the difference is consistent across data, with other methods. They allow for tight control of glu-
but its applicability for glucose measurements is problematic cose levels, again with the possibility of acting as an affer-
as percentile differences of 20 mg/dL could be negligible in ent limb in a closed-loop system with an insulin pump as
the normoglycemic ranges but very meaningful in the hypo- the efferent limb. However, technology hasn’t reached that
glycemic range. In 1987, William L. Clarke et al. published20 far. Many billions of dollars have been spent trying to reach
their development of a novel system for comparison of glu- this goal but with very little to show. There are few nonin-
cose monitoring systems, the error grid analysis (EGA or vasive or minimally invasive systems in the market, and
Clarke EGA). The grid is divided into zones: results that fall their accuracy and reproducibility levels are generally low.
within zone A deviate less than 20% and are not in the hypo- Financial pressure and the urge by the diabetic population
glycemic range. Zone B represents a higher deviation but one sent some of these sensors to the market too early and with-
that is still acceptable as it should not affect treatment. Zones out a thorough clinical evaluation in different disease states.
C, D, and E represent areas in which results could lead to Dealing with such a technology- and industry-driven field
erroneous, potentially life-threatening decisions. This sys- of research, publication bias is expected to be high. Of the
tem was innovative for not only comparing reference and few sensors that were released to the global market, selling
study values but rather emphasizing the clinical implica- was stopped on some due to poor clinical performance plac-
tion of the readings. It underwent revision21 by Parkes et al. ing patients at risk. Glucose levels are rather estimated than
The Clarke EGA was designed at a time when continuous measured and the margins of error tend to be wide com-
measurements did not exist, and in 2004 a group led by pared to more invasive systems. We shall discuss many of
Kovatchev and Clarke proposed the continuous glucose these systems as it would be impossible to discuss all and
Transdermal devices 383
Table 46.1 A summary of transdermal techniques for noninvasive glucose monitoring
Technology Company/product Site Results, drawbacks, status

Reverse ionotophoresis GlucoWatch Biographer Wrist skin FDA approved in 2001.
(Cygnus Inc.) Poor accuracy, mainly in the hypoglycemic range.
Frequent finger sticks for calibration.
Discomfort associated with its use.
Withdrawn from market 2007.
Skin suction blister Experimental, University Volar Infection risk (low).
Hospital of Aarhus forearm Fair reliability but increased lag time in six type 1
diabetic patients in 1995.
Sonophoresis SonoPrep, Sontra Medical Skin In 2000, large variability between patients and sites.
Improved model with acceptable results in 2004.
Bioimpedance Pendra (Pendragon Medical Wrist skin Pendra was released in 2003 but soon withdrawn due
spectroscopy Ltd.) to poor reliability.
Requires a prolonged equilibration process.
Results affected by temperature and moisture.
Ultrasound with cymbal Experimental, Pennsylvania Skin Animal models only.
transducer array State University Fair results in rats in 2005.
Good results in pigs in 2009.
Electroporation BTX Applications Skin Animal models only.
Good results in rats in 2008.
Electromagnetic sensing Glucoband (Calisto Wrist skin Released after preclinical trials in 2005.
Medical) Accuracy affected by temperature and other blood
components.
Good safety profile.
as research is still very much going on. We will divide these Sonophoresis
into transdermal systems, which are generally minimally Skin permeability is increased using low-frequency ultra-
invasive, and optical systems, which are generally truly non- sound emitted from the SonoPrep™ (Sontra Medical) device.28
invasive (see Table 46.1). A low current is delivered while the return electrode is held
by the patient. A biosensor is coupled to the permeated skin
and converts glucose into hydrogen peroxide that in turn is
Transdermal devices oxidized. The current created is measured, and after a calibra-
tion phase, the level of glucose can be estimated. Results in
Reverse ionotophoresis a study29 from 2000 were mediocre with 92% falling within
The GlucoWatch® Biographer (Cygnus Inc.), up to its G2 areas A+B on the Clarke EGA. In a subsequent trial in 10 dia-
model, became Food and Drug Administration (FDA) betic patients,28 results improved with 95% falling within area
approved25 in 2001 as an adjunct to finger-stick testing to A+B with no report of pain or irritation by any of the patients.
improve therapeutic decisions. A hydrogel pad is placed over Since then, no significant advance has been reported.
the skin and left to warm up for 2 hours, a finger-stick test
is performed for calibration, and it starts measuring glu-
cose level every 10 minutes. The hydrogel pad is replaced Electroporation
every 13 hours. The device employs reverse ionotophoresis: Electroporation is achieved by applying short electrical
Application of a small current to the skin26 and subcutane- pulses to the skin, thus reversibly increasing its permeabil-
ous fluid passes through the skin to an electrode pair. This ity and allowing for transcutaneous sampling. Developed by
fluid contains glucose at a level proportional to the subcuta- BTX® Applications, it was tested30 on rats only, with all data
neous tissue. Oxidation of glucose is performed on the pad, points within areas A+B on the Clarke EGA. Clinical trials
allowing for level calculation. Its applicability in monitoring in humans have not been reported.
for hypoglycemic episodes was questioned as results were
disappointing showing very low sensitivity and specificity in
the hypoglycemic range.27 It was also deemed too expensive Impedance spectroscopy
when compared to SMBG and it did not alleviate the need A small current is passed across the skin, and impedance
for finger sticks. Perhaps the most unappealing effect was is recorded allowing for an indirect estimation of glucose
the skin irritation due to the electric current to the skin that levels. The Pendra (Pendragon Medical Ltd.) was released
afflicted most of all users. Its manufacture and sales were to the market after CE approval in 2003. Its use was cum-
halted in 2007. bersome as it required a prolonged equilibration phase, and
a year later, the company filed for bankruptcy.31 Its results32 passes through an array of cymbal transducers. These are
were very poor with only 78.4% of the results falling within specialized transducers with compact structures mak-
areas A+B and an unacceptable 4.3% within the most dan- ing them potentially comfortable as carried continuous
gerous area E. devices. In 2005, the technique was studied34 on rats with
good results. It was later studied35 on 200 lb pigs to show that
the technique can be used on human-sized animals as well,
Skin suction blister and results were promising. However, many of the operat-
This technique creates vacuum over a small area of the skin ing conditions of the cymbal array as well as the correlation
that leads to the formation of a blister. Fluid from the blister of glucose between the different components have not been
is collected and analyzed. The technique was studied33 on six sorted out. Clinical trials in humans are still awaited.
T1DM patients. Glucose levels in the fluid paralleled closely
with the venous levels but the lag time was unexpectedly
long in three of the patients. Since 1995, no new advances Electromagnetic sensing
have been reported. The Glucoband® (Calisto Medical) is placed on the wrist
skin where it applies a bioelectromagnetic wave specific for
glucose and measures resonance and impedance. Preclinical
Ultrasound with cymbal array trials quickly prompted its release to the market in 2005 but
Ultrasound waves permeate the skin and glucose is dif- since no other data have been published and its sales have
fused through. Glucose is sensed through a current that long ceased (Table 46.2).
Table 46.2 Summary of optical techniques for noninvasive glucose monitoring
Technology Company/product Site Results, drawbacks, status

Mid-infrared absorption Experimental, University Skin Signal influenced by other blood components.
spectroscopy of Cambridge Acceptable results in vitro only in 2003.
Near-infrared Diasensor (Biocontrol Skin, oral mucosa Highly sensitive but limited by physiological
spectroscopy Technology Inc.) variables.
TensorTip CoG (Cnoga Diasensor was rejected twice by the FDA due to
Ltd.) safety concerns.
TensorTip CoG combines invasive and
noninvasive apparatus.
Patented in 2011 and marketed in selected
countries for type 2 diabetics in 2012.
Raman spectroscopy C8 MediSensors Skin Promising results in 30 subjects in 2009.
CE approval in 2012.
Seems to have halted progress.
Occlusion red/near- OrSense NBM Base of finger skin Acceptable results in 23 diabetic patients in
infrared spectroscopy 2007.
Thermal emission Infratec Inc. Tympanic Good results in 31 insulin-dependent diabetic
spectroscopy membrane patients in 2002.
Photoacoustic Aprise (Glucon Inc.) Skin Acceptable results in 62 diabetic patients in
technique 2007.
Optical coherence Experimental, University Skin Good results in 15 healthy volunteers in 2002.
tomography of Texas Medical Accuracy limited by physiological changes as
Branch well as motion.
Fluorescence EyeSense Conjunctiva (long Good results in 28 diabetic patients in 2012.
term, surgically Problematic lag time.
placed) Hemorrhage and local reactions.
Fluorescence resonance Experimental, University Tears Discrepancy in tear glucose levels.
energy transfer of Western Ontario Tear collection takes 10 minutes, limited in situ
potential.
Requires external illumination.
Optical polarimetry Experimental, Texas A&M Eye Animal models only.
University Excellent results in three rabbits in 2012.
Mass spectroscopy Exhaled breath Mixed accuracy reports, mostly unsatisfactory.
Kromoscopy Experimental, University In vitro tissues only Good glucose and urea discrimination in vitro
of Iowa and Ohio in 2000. No in vivo data.
University
Optical devices 385
Optical devices
Absorption spectroscopy
Absorption spectroscopy utilizes light of variable frequencies
in order to measure changes in refraction (scattering), reflec-
tion, and absorbance within the tissue. These are affected by
changes in the concentrations of different molecules includ-
ing glucose. In general, these devices comprise of a light
source that delivers radiation at the set frequency and inten-
sity. After passing through, the tissue light waves are sensed
by the probe and then processed in the processing unit.
Mid-infrared spectroscopy
Absorption spectroscopy in the mid-infrared (MIR) range Figure 46.2 The OrSense NBM-200 device, using o cclusion
(2.5–50 µm) has been used to measure glucose levels accord- red/near-infrared spectroscopy. (By courtesy of the manufacturer.)
ing to the changes in the beam of light returned to the
device. However, many molecules other than glucose also
absorb this light.2 As light passage is slow, it is affected by the in 23 diabetic patients before and after a meal with 95.5% of
epidermal molecular structure. There are only in vitro data the measurements falling within areas A+B on the Clarke EGA
from a study36 on blood from 28 patients in which glucose compared to SMBG. However, since 2007 no further data have
level correlated well, but it seems this system is not practical been published on this device, and it is not commercially avail-
in vivo because of the confounding molecules. able for glucose monitoring (see Figure 46.2).
Near-infrared spectroscopy
In contrast, light in the NIR range (0.7–1.4 µm) largely passes Raman spectroscopy
through the epidermis unchanged and can therefore provide Light of a single wavelength (monochromatic) is passed
a molecular picture of the dermis. Different calculations are through a tissue. When it interacts with glucose, the energy
used to estimate glucose level. However, accuracy is ham- of the photons is shifted according to the vibrational and
pered by many factors such as other molecules (e.g., urea), rotational energy of glucose, thus providing selective
skin pigmentation, temperature, and hydration. A different information about the level of glucose. It may differentiate
measurement site is the oral mucosa that is richly vascular- between two similar molecules like glucose and galactose.39
ized but is impaired by the glucose contained in the saliva. In 2009, it was tested40 on 30 human subjects with 387 dif-
A few sensors have been released to the market. The ferent site measurements. Results were acceptable with 92%
Diasensor® (BioControl Technologies Inc.) failed twice to of data points falling within areas A+B on the Clarke EGA.
gain FDA approval, and in 2002 its CEO pleaded guilty to This device from C8 MediSensors has received CE approval
tax evasions and fraud practically ending a decade of invest- in 2012 but since progress seems to have halted.
ments of many millions of dollars in this product. The
TensorTip CoG® (Cnoga Ltd.) is a device combining invasive
and noninvasive measurements that became available in Photoacoustic technique
many countries since 2012. A new technology37 using single- This is a different spectroscopy technique. Light passed
walled carbon nanotubes is showing promise as a continuous through a tissue is converted into heat that emits ultrasonic
noninvasive monitor, but research is still at a preliminary waves caused by local expansion. These are interpreted by
level. As mentioned in the text ahead, NIR technology has the processor to calculate the glucose level. The Aprise™
also been researched as a screening tool for diabetes. (Glucon Inc.) device was examined41 on 62 diabetic patients
with 979 paired measurements. Results were acceptable with
94.6% of the points falling within areas A+B on a Clarke
Occlusion red/near-infrared spectroscopy EGA. Further research is awaited.
Occlusion spectroscopy is based on the fact that blood vessel
occlusion changes its physical and optical properties. As glu-
cose levels in the blood increase, it changes the scattering prop- Thermal emission spectroscopy
erties (smaller scattering coefficient). The device (OrSense®) The site of measurement for this device is the tympanic
employs an annular ring placed on the patient’s thumb that membrane. The human body emits electromagnetic radiation
contains a pneumatic cuff that inflates to an oversystolic pres- that stands in relation to spectral distribution. The handheld
sure for 2–3 seconds and then deflates causing erythrocyte apparatus analyzes light in the MIR range that is influenced
aggregation. The probe also contains a light source in the red/ by the body’s composition. In 2002, thermal infrared emis-
NIR spectrum and a light detector. The passage of photons to sion characteristics of different glucose concentrations were
the detector is affected by the glucose level as well as hemato- measured42 in 26 patients with insulin-dependent diabetes
crit concentration. It showed promising result in a pilot study38 (T1DM and T2DM). No side effects were reported. Out of
Glucose monitoring
Minimally invasive- Noninvasive-

Invasive
transdermal optical
Finger stick - self Ultrasound Photoacoustic Optical

Continuous glucose Spectroscopy Kromoscopy
blood glucose technique polarimetry
monitoring (CGM)
monitoring (SMBG)
Bioimpedance
spectroscopy Mid infrared (MIR)
Intravenous spectroscopy Fluorescence
Sonophoresis Near-infrared (NIR)

Subcutaneous
spectroscopy Fluorescence
Electromagnetic presonance energy
Microdialysis sensing Raman transfer (FRET)
spectroscopy
Skin suction Autofluorescence
blister Occlusion red/
near-infrared
spectroscopy Tear liquid
Reverse
Ionotophoresis fluorescence
Mass
spectroscopy
Electroporation
Thermal emission
spectroscopy
Figure 46.3 Existing and underdevelopment technologies for noninvasive glucose monitoring.
Other forms and applications of noninvasive glucose monitoring 387
432 paired data points, all fell within areas A+B (97.5% in a fluctuation in data points was observed as a result of
area A)—excellent results. Despite the good results, 12 years motion, limiting applicability, as well as being confounded
later, this is still the only published data in the field. by parameters such as temperature changes, tissue compo-
nents other than glucose, blood pressure, and heart rate (see
Figure 46.3).
Fluorescence and FRET technology
Different molecules in different states emit varying light
intensity.2 Using the molecule Concanavalin A (ConA) that Other forms and applications of
has four glucose binding sites,43 the level of glucose can be
calculated using a competitor such as fluorescein. This tech- noninvasive glucose monitoring
nique as a method for glucose sensing has been studied44 for Noninvasive glucose monitors as screening tools
over 30 years. Different enzymatic reactions have been used.
Apart from its place in managing glucose levels in diabetic
Most recently, EyeSense® released an implantable glucose
patients, NIGM is being increasingly investigated as a diag-
monitor that is placed between the eye and the conjunctiva.
nostic screening tool. The NSEEDS trial51 used skin fluores-
Studied45 on 28 diabetic patients, most patients suffered sub-
cence spectroscopy to evaluate the glycemic response to oral
conjunctival hemorrhage postoperatively but afterwards
glucose loading and was found to be similar to A1C and FPG
only complained of mild discomfort. Reading accuracy
for the detection of abnormal glucose tolerance. NIR was
was acceptable with more than 96% of the points falling
examined52 on 154 healthy and diabetic subjects, achieving
within areas A+B on the Clarke EGA; however, a shift from
77.3% sensitivity for the detection of diabetes with 70% spec-
area A to B was noticed. Lag time still poses a problem of
ificity, comparable with FPG measurements,53 while alleviat-
unknown significance.
ing the need for fasting.
A different technology is fluorescence resonance energy
Autofluorescence was evaluated54 for differentiating nor-
transfer (FRET). First studied46 in 1988, this technique is
mal pregnancies from GDM according to levels of advanced
based on the transfer of energy from donor molecules to
glycation end products; results were disappointing as it
acceptor molecules, thereby decreasing fluorescence levels
could not differentiate between the two groups at time of
allowing for miniscule changes in the level of molecules to be
diagnosis, nor was it able55 to diagnose gestational diabetic
detected. Studied in preclinical models,47 generally a contact
pregnant women at high risk for complications.
lens is used and boronic acid is usually the donor molecule.
A few devices have been patented, but many problems still
need to be addressed—the longer time it takes for tear col- Hypoglycemia monitoring
lection, the more the need for an external light source and As said earlier, one of the major pitfalls of most many NIGM
the lack of a predictable relationship between blood and tear systems is their failure to recognize hypoglycemic states
glucose levels. leading to delayed or even deleterious treatment. An alto-
gether different approach toward NIGM is not measuring
the glycemic level but rather monitoring for the signs of
Optical polarimetry hypoglycemia, mainly the autonomic reactions. This may
When polarized light hits certain molecules, it rotates. These help deal with nocturnal hypoglycemia, a prevalent56 path-
molecules are called optically active and they can be detected ological state that can cause seizures57 and might even be
using polarimetry. Using a dual-wavelength optical polar- fatal.58 The HypoMon® (AiMedics Pty. Ltd.)59 is a chest belt
imeter, this technique was studied48 in 2012 on nine rabbits worn during the night that monitors surface ECG and skin
employing a contact lens and an external light source, mea- bioimpedance. It also includes a radio-frequency transmitter
suring the glucose level within the aqueous humor. Accuracy that sends signals to a receiver that can be placed in the same
was very good with 100% of the results falling within areas room or an adjacent one. Tested on 52 patients with T1DM,
A+B on the Clarke EGA. These data are still preliminary, it showed a positive predictive value of only 38%, not yet con-
and many adaptations will be needed for human subjects as sidered acceptable. Another apparatus is the Gili Medical
an external light source must be placed directly on the eye- hypoglycemia noninvasive monitoring system (Gili Medical
lid, causing discomfort and also limiting the continuity of Ltd.), which contains four sensors that monitor for physi-
measurement. ologic effects of hypoglycemia—heart rate (tachycardia),
perspiration level (increased), skin temperature (peripheral
vasoconstriction), and body movement level (trembling).
Optical coherence tomography The monitor is placed on the hand and is connected to a
Introduced49 in 1991, this technique uses and detects light receiver. Tested60 on 10 patients with T1DM, it showed high
reflection from tissues combined with light from a reference sensitivity and specificity (100% and 85.7%, respectively).
arm. The device contains a photodetector and an interfer- This technology is still in development and its design must
ometer to achieve a high-resolution analysis of the tissue become less cumbersome. It should be noted that these mon-
scattering properties. Studied50 in 2002 on 15 healthy sub- itors are limited by many factors that hamper the autonomic
jects undergoing oral glucose tolerance testing, results were reaction to hypoglycemia such as hypoglycemia tolerance
in good correlation with blood glucose levels. However, and β-blockers.
Exhaled breath analysis The Minimed® (Medtronic Ltd.) is probably the best
A proposed alternative to finger-stick systems that collects studied system, 66–68 available in multiple models. It was
condensed humidity in exhaled breaths for glucose level studied in pregnant women as well,69 and its use was associ-
measurement has been developed.61 Within 10–20 minutes, ated with a better glycemic control, a lower birth weight, and
enough respiratory secretions (1–2 mL) can be collected for a reduced risk for macrosomia. It was prospectively studied
analysis.62 Respiratory fluids normally contain low levels of in 34 gravid patients with GDM70 and detected postprandial
glucose. In pathological conditions such as cystic fibrosis hyperglycemia as well as nocturnal hypoglycemia better than
and diabetes, these levels are increased.63 These secretions SMBG. Other widely available systems that have shown good
may serve as a surrogate for blood glucose measurement as and reliable results are the Dexcom Seven® (Dexcom Ltd.) that
blood glucose levels may be estimated by analyzing the glu- recently showed71 improved accuracy and the Freestyle® (vari-
cose level in the secretions, mainly in the pediatric popula- ous models, Abbott Diabetes Care). The latter was studied in
tion, but further research is pending. pregnant women suffering from T1DM72 and was found safe
but still not of optimal accuracy. In another study73 in pregnant
diabetics, it was found less accurate during moderate exercise.
Continuous glucose monitoring It should also be mentioned that some of the older models of
the Freestyle sensors were recalled in the beginning of 2014.
systems A different approach is microdialysis. A double-lumen
Similar to NIGM but more invasive are the continuous glu- catheter is placed within the subcutaneous tissue, and a
cose monitoring systems (usually referred to as CGM or glucose-free isotonic fluid is transfused continuously over a
CGMS). These are more invasive systems and have consis- dialysis membrane. Glucose permeates the membrane and
tently shown far better results. Because CGMS are widely its level is assayed. It can also be placed intravenously in
marketed, they are discussed in depth in a different chapter. hospitalized patients.74 This technique showed75 its ability to
However, they will be shortly mentioned here. Often, NIGM reduce HbA1C levels in patients with T1DM. The GlucoDay®
and MIGM are considered forms of CGM, but here they are (Menarini Diagnostics) is the main commercially available
reviewed separately. model, studied76 since the early 2000s.
The typical CGM device includes a sensor, sensing glu-
cose in the interstitial fluid or in the blood; a processing unit,
wireless or with wires; and a screen displaying the result. Conclusions
When the monitor is accurate and precise enough, allowing
a reliable and reproducible safety profile, it can be connected The place for NIGM is obvious—a painless system that can
to a fourth element, a continuous insulin pump. This is truly measure glucose levels and guide treatment could dramati-
an artificial pancreas (regarding the β-islet function). cally improve the quality of life of diabetic patients and reduce
In the subcutaneous or intravenous CGM devices, the sen- long-term organ damage. Results, however, have been gener-
sor is placed within the subcutis or a vein where it obtains a ally disappointing, especially when compared to the more
continuous reading of glucose level in the fluid. In the ear- invasive CGMS that have consistently shown good results
lier models, it was connected by a wire to the processing unit, and are becoming an integral part in the management of
but the more modern devices are completely implanted and diabetes, especially insulin dependent. This is certainly not
interface with the external controller wirelessly. The sensing for lack of investment. Billions of dollars77 have been spent
process is usually based on glucose oxygenation, and then an on research with very little to show. Devices released to the
assay by optical or electrochemical assay is performed. The market with great promises, especially in the lay press, were
reading is transmitted to the display unit (usually by radio later quietly withdrawn due to poor postmarketing results. Is
frequency). In earlier models, the data had to read retrospec- there a future for NIGM systems? In the authors’ opinion, it
tively but the newer models allow real-time CGM. An earlier remains unclear. It is clear, however, that devices must stand
systematic review64 couldn’t show significant improvement in more rigorous testing before being released into the market.
treatment targets. A more recent meta-analysis65 of six ran- Only time will tell whether advances in basic science could
domized controlled studies evaluating glycemic control in translate into accurate, precise, affordable, comfortable, and
T1DM using real-time CGM showed lower glycated hemoglo- easy-to-use systems that could become an integral part in
bin (HbA1c) and lower hypoglycemia rate in CGM users. the diagnosis and treatment of diabetes.
REFERENCES
1. American Diabetes Organization Official Site. Available from: 3. The Diabetes Control and Complications Trial Research Group.
http://www.diabetes.org, accessed on March 15, 2014. The effect of intensive treatment of diabetes on the development
2. Oliver NS, Toumazou C, Cass AEG. Glucose sensors: A review and progression of long-term complications in insulin-dependent
of current and emerging technology. Diabet Med 2009; 26: diabetes mellitus. N Engl J Med 1993; 329: 977–986.
197–210.
References 389
4. The Diabetes Control and Complications Trial/Epidemiology of 27. The Diabetes Research in Children Network (DirectNet) Study
Diabetes Interventions and Complications (DCCT/EDIC) Study Group. Accuracy of the GlucoWatch G2 Biographer and the
Research Group. Intensive diabetes treatment and cardiovascular continuous glucose monitoring system during hypoglycemia.
disease in patients with type 1 diabetes. N Engl J Med 2005; 353: Diabetes Care 2004; 27: 722–726.
2643–2653. 28. Chuang H, Taylor E, Davison TW. Clinical evaluation of a con-
5. UK Prospective Diabetes Study (UKPDS) Group. UK prospec- tinuous minimally invasive glucose flux sensor placed over
tive diabetes study VIII. Study design, progress and performance. ultrasonically permeated skin. Diabetes Technol Ther 2004;
Diabetologia 1991; 34: 877–890. 6(1): 21–30.
6. UK Prospective Diabetes Study (UKPDS) Group. Intensive blood- 29. Kost J, Mitragotri S, Gabbay RA et al. Transdermal monitoring
glucose control with sulphonylureas or insulin compared with of glucose and other analytes using ultrasound. Nat Med 2000;
conventional treatment and risk of complications in patients with 6: 347–350.
type 2 diabetes (UKPDS 33). Lancet 1998; 352: 837–853. 30. Murthy SS, Kiran VS, Mathur SK et al. Noninvasive transcuta-
7. Crowther CA, Hiller JE, Moss JR et al. For the Australian neous sampling of glucose by electroporation. J Diabetes Sci
Carbohydrate Intolerance Study in Pregnant Women (ACHOIS) Technol 2008; 2(2): 250–254.
Trial Group. Effect of treatment of gestational diabetes mellitus 31. DeVries JH, Wentholt I, Zwart A et al. Pendragoesdutch; lessons for
on pregnancy outcomes. N Engl J Med 2005; 352: 2477–2486. the CE mark in Europe. Diabetes Res Clin Pract 2006; 74: S93–S96.
8. Consensus Panel of the ADA, FDA and NIH. Consensus state- 32. Renard E. Implantable glucose sensors for diabetes monitoring.
ment on self-monitoring of blood glucose. Diabetes Care 1987; 1: Minim Invasive Ther Allied Technol 2004; 13: 78–86.
95–99. 33. Jensen BM, Bjerring P, Christiansen JS et al. Glucose content in
9. National Institute for Clinical Excellence (NICE). Type 2 diabetes: the human skin: Relationship with blood glucose levels. Scand J
The management of Type 2 diabetes (update). Guidance CG66, Clin Lab Invest 1995; 55(5): 427–432.
May 2008. 34. Lee S, Nayak V, Dodds J et al. Glucose measurements with
10. Welschen LM, Bloemendal E, Nijpels G et al. Self-monitoring of sensors and ultrasound. Ultrasound Med Biol 2005; 31(7):
blood glucose in patients with Type 2 diabetes who are not using 971–977.
insulin: A systematic review. Diabetes Care 2005; 28: 1510–1517. 35. Park EJ, Werner J, Beebe J et al. Noninvasive ultrasonic glucose
11. Vincze G, Barner CB, Lopez D. Factors associated with adher- sensing with large pigs (~200 pounds) using a lightweight cymbal
ence to self-monitoring of blood glucose among persons with transducer array and biosensors. J Diabetes Sci Technol 2009;
diabetes. Diabetes Educ 2004; 30(1): 112–125. 3(3): 617–623.
12. Hazuda HP, Haffner SM, Stern MP, Eifler CW. Effects of accul- 36. Shen YC, Davies AG, Linfield EH et al. The use of Fourier-
turation and socioeconomic status on obesity and diabetes transform infrared spectroscopy for the quantitative determina-
in Mexican Americans. The San Antonio Heart Study. Am J tion of glucose concentration in whole blood. Phys Med Biol
Epidemiol 1988; 128(6): 1289–1301. 2003; 48(13): 2023–2032.
13. Connolly V, Unwin N, Sherriff P et al. Diabetes prevalence 37. Yum K, McNicholas T, Mu B et al. Single-walled carbon nano-
and socioeconomic status: A population based study showing tube-based near-infrared optical glucose sensors toward in vivo
increased prevalence of type 2 diabetes mellitus in deprived continuous glucose monitoring. J Diabetes Sci Technol 2013; 7(1):
areas. J Epidemiol Community Health 2000; 54(3): 173–177. 72–87.
14. Phillipou G. Race/ethnicity and other risk factors for gestational 38. Amir O, Weinstein D, Zilberman S et al. Continuous noninvasive
diabetes. Am J Epidemiol 1992; 135: 965–973. glucose monitoring technology based on “occlusion spectros-
15. Sieg A, Guy RH, Delgado-Charro MB. Noninvasive and mini- copy.” J Diabetes Sci Technol 2007; 1(4): 463–469.
mally invasive methods for transdermal glucose monitoring. 39. Vaddiraju S, Burgess DJ, Tomazos I et al. Technologies for contin-
Diabetes Technol Ther 2005; 7(1): 174–197. uous glucose monitoring: Current problems and future promises.
16. Cengiz E, Tamborlane WV. A tale of two compartments: J Diabetes Sci Technol 2010; 4(6): 1540–1562.
Interstitial versus blood glucose monitoring. Diabetes Technol 40. Lipson J, Bernhardt J, Block U et al. Requirements for calibra-
Ther 2009; 11(1): S11–S16. tion in noninvasive glucose monitoring by Raman spectroscopy.
17. Koschinsky T, Heinemann L. Sensors for glucose monitoring: J Diabetes Sci Technol 2009; 3(2): 233–241.
Technical and clinical aspects. Diabetes Metab Res Rev 2001; 41. Weiss R, Yegorchikov Y, Shusterman A et al. Noninvasive con-
17: 113–123. tinuous glucose monitoring using photoacoustic technology-
18. Heise HM, Marbach R, Koschinsky TH et al. Noninvasive blood results from the first 62 subjects. Diabetes Technol Ther 2007;
glucose sensors based on near-infrared spectroscopy. Artif 9(1): 68–74.
Organs 1994; 18(6): 439–447. 42. Malchoff CD, Shoukri K, Landau JI et al. A novel noninvasive
19. Pohl S, Gonder-Frederick L, Cox D et al. Self-measurement of blood glucose monitor. Diabetes Care 2002; 25: 2268–2275.
blood glucose concentration: Clinical significance of patient- 43. Reeke GN, Becker JW, Edelman GM. The covalent and three-
generated measurements. Diabetes Care 1985; 8: 617–619. dimensional structure of concanavalin A. J Biol Chem 1975; 250:
20. Clarke WL, Cox D, Gonder-Frederick LA et al. Evaluating clini- 1525–1547.
cal accuracy of systems for self-monitoring of blood glucose. 44. Mansouri S, Schultz JS. A miniature optical glucose sensor based
Diabetes Care 1987; 10: 622–628. on affinity binding. Biotechnology 1984; 2: 885–890.
21. Parkes JL, Pardo S, Slatin SL et al. A new consensus error grid to 45. Hasslacher C, Auffarth G, Platten et al. Safety and accuracy of a
evaluate the clinical significance of inaccuracies in the measure- new long-term subconjunctival glucose sensor. J Diabetes 2012;
ment of blood glucose. Diabetes Care 2000; 23: 1143–1148. 4(3): 291–296.
22. Kovatchev B, Gonder-Fredrick LA, Cox DJ et al. Evaluating the 46. Meadows D, Schultz JS. Fibre-optic sensors based on fluores-
accuracy of continuous glucose monitoring sensors. Diabetes cence energy transfer. Talanta 1988; 35: 145–150.
Care 2004; 27: 1922–1928. 47. Zhang J, Hodge W, Hutnick C et al. Noninvasive diagnostic
23. Wentholt IM, Hoekstra JB, DeVries JH. A critical appraisal of the devices for diabetes through measuring tear glucose. J Diabetes
continuous glucose-error grid analysis. Diabetets Care 2006; 29: Sci Technol 2011; 5(1): 166–172.
1805–1811. 48. Pirnstill CW, Malik BH, Gresham VC. In vivo glucose monitoring
24. Brunner R, Kitzberger R, Miehsler W et al. Accuracy and reli- using dual-wavelength polarimetry to overcome corneal birefrin-
ability of a continuous glucose-monitoring system in critically ill gence in the presence of motion. Diabetes Technol Ther 2012;
patients. Crit Care Med 2011; 39(4): 659–664. 14(9): 819–827.
25. FDA, U.S. Food and Drug Administration. www.fda.gov/ 49. Huang D, Swanson EA, Lin CP et al. Optical coherence tomogra-
medicaldevices/productsandmedicalprocedures/deviceapprov- phy. Science 1991; 254: 1178–1181.
alsandclearances/recently-approveddevices/ucm083294.htm. 50. Larin KV, Eledrisi MS, Motamedi M et al. Noninvasive blood
26. Kurnik RT, Berner B, Tamada J et al. Design and simulation of a glucose monitoring with optical coherence tomography:
reverse iontophoretic glucose monitoring device. J Electrochem A pilot study in human subjects. Diabetes Care 2002; 25:
Soc 1998; 145: 4119–4125. 2263–2267.
51. Olson BP, Matter NI, Ediger MN et al. Noninvasive skin fluores- 64. Chetty VT, Almulla A, Odueyungbo A et al. The effect of con-
cence spectroscopy is comparable to hemoglobin A1c and fast- tinuous subcutaneous glucose monitoring (CGMS) versus inter-
ing plasma glucose for detection of abnormal glucose tolerance. mittent whole blood finger-stick glucose monitoring (SBGM) on
J Diabetes Sci Technol 2013; 7(4): 990–1000. hemoglobin A1c (HBA1c) levels in type I diabetic patients: A sys-
52. Brown CD, David HT, Ediger MN et al. Clinical assessment of tematic review. Diabetes Res Clin Pract 2008; 81(1): 79–87.
near-infrared spectroscopy for noninvasive diabetes screening. 65. Pickup JC, Freeman SC, Sutton AJ. Glycaemic control in type 1
Diabetes Technol Ther 2005; 7(3): 456–466. diabetes during real time continuous glucose monitoring com-
53. NHANES. Reference manuals, National Center for Health pared with self monitoring of blood glucose: Meta-analysis of
Statistics, Centers for Disease Control. Available from: http:// randomized controlled trials using individual patient data. Br
www.cdc.gov/nchs/about/ major/nhanes/NHANESIII_Reference_ Med J 2011; 343: d3805.
Manuals.htm, accessed on February 10, 2014. 66. Gross TM, Bode BW, Einhorn D et al. Performance evaluation
54. Ranitz-Greven WLD, Bos DC, Poucki WK et al. Advanced gly- of the MiniMed continuous glucose monitoring system during
cation end products, measured as skin autofluorescence, at patient home use. Diabetes Technol Ther 2000; 2: 49–56.
diagnosis in gestational diabetes mellitus compared with normal 67. Bergenstal RM, Tamborlane WV, Ahmann A et al. Effectiveness
pregnancy. Diabetes Technol Ther 2012; 14(1): 43–49. of sensor-augmented insulin-pump therapy in type 1 diabetes. N
55. Ranitz-Greven WLD, Kaasenbrood L, Poucki WK et al. Engl J Med 2010; 363: 311–320.
Advanced glycation end products, measured as skin autofluo- 68. Battelino T, Conget I, Olsen B et al. The use and efficacy of con-
rescence, during normal pregnancy and pregnancy compli- tinuous glucose monitoring in type 1 diabetes treated with insulin
cated by diabetes mellitus. Diabetes Technol Ther 2012; 14(12): pump therapy: A randomized controlled trial. Diabetologia 2012;
1134–1139. 55: 3155–3162.
56. Porter PA, Keating B, Byrne G et al. Incidence and predictive 69. Murphy HR, Rayman G, Lewis K et al. Effectiveness of con-
criteria of nocturnal hypoglycaemia in young children with tinuous glucose monitoring in pregnant women with diabetes:
insulin-dependent diabetes mellitus. J Pediatr 1997; 130(3): Randomized clinical trial. Br Med J 2008; 337: a1680.
366–372. 70. Yogev Y, Chen R, Ben-Haroush et al. Continuous glucose moni-
57. Buckingham B, Wilson DM, Lecher T et al. Duration of noctur- toring for the evaluation of gravid women with type 1 diabetes
nal hypoglycaemia before seizures. Diabetes Care 2008; 31(11): mellitus. Obstet Gynecol 2003; 101: 633–638.
2110–2112. 71. Garcia A, Rack-Gomer AL, Bhavaraju NC et al. Dexcom G4AP:
58. Sovik O, Thordarson H. Dead-in-bed syndrome in young dia- An advanced continuous glucose monitor for the artificial pan-
betic patients. Diabetes Care 1999; 22(Suppl. 2): B40–B42. creas. Diabetes Technol Ther 2013; 7(6): 1436–1445.
59. Skladnev VN, Ghevondian N, Tarnavski S et al. Clinical evalua- 72. Murphy HR, Elleri D, Allen JM et al. Closed-loop insulin delivery
tion of a noninvasive alarm system for nocturnal hypoglycemia. during pregnancy complicated by type 1 diabetes. Diabetes Care
J Diabetes Sci Technol 2010; 4(1): 67–74. 2011; 34: 406–411.
60. Schechter A, Eyal O, Zuckerman-Levin N et al. A prototype of 73. Kumareswaran K, Elleri D, Allen JM et al. Accuracy of continuous
a new noninvasive device to detect nocturnal hypoglycemia glucose monitoring during exercise in type 1 diabetes pregnancy.
in adolescents with type 1 diabetes—A pilot study. Diabetes Diabetes Technol Ther 2013; 15(3): 223–229.
Technol Ther 2012; 14(8): 683–689. 74. Rostami E, Bellander BM. Monitoring of glucose in brain, adipose
61. Roberts K, Jaffe A, Verge C et al. Noninvasive monitoring of glu- tissue, and peripheral blood in patients with traumatic brain injury:
cose levels: Is exhaled breath the answer? J Diabetes Sci Technol A microdialysis study. J Diabetes Sci Technol 2011; 5(3): 596–604.
2012; 6(3): 659–664. 75. Bolinder J, Ungerstedt U, Arner P. Long-term continuous glucose
62. Horváth I, Hunt J, Barnes PJ; ATS/ERS Task Force on Exhaled monitoring with microdialysis in ambulatory insulin-dependent
Breath Condensate. Exhaled breath condensate: Methodological diabetic patients. Lancet 1993; 342(8879): 1080–1085.
recommendations and unresolved questions. Eur Respir J 2005; 76. Maran A, Crepaldi C, TiengoA et al. Continuous subcutaneous
26(3): 523–548. glucose monitoring in diabetic patients: A multicenter analysis.
63. Baker EH, Clark N, Brennan AL et al. Hyperglycemia and cystic Diabetes Care 2002; 25(2): 347–352.
fibrosis alter respiratory fluid glucose concentrations estimated 77. Heinemann L. Noninvasive glucose monitoring systems: Will we
by breath condensate analysis. J Appl Physiol 2007; 102(5): ever have such sensors for practical use? J Diabetes Sci Technol
1969–1975. 2007; 1(6): 936–939.
47 Reproduction and its impact on
health and disease
Sara Ornaghi and Michael J. Paidas
With the exception of the stomach, there is no organ that

holds such numerous ramifications of sympathy with
Implications of pregnancy history on
other organs as the womb. subsequent women’s health
J.M. Good (1825) CVD risk
Globally, one of three women dies from CVD.4 In response
to the growing appreciation that many preventive efforts
start too late to be effective, there has been a call for “pri-
Introduction mordial prevention”—prevention of the major CVD risk fac-
tors themselves.5,6 In this context, pregnancy complications
Women’s reproductive health is intimately linked with their
have the potential to be effective CVD risk “stress tests” to
overall physical and mental health. The vitality and perfor-
identify women who would most benefit from primordial or
mance of women’s reproductive axis—as revealed in patterns
primary prevention efforts to reduce CVD risk.
of menstruation, fertility, and pregnancy—yield important
In a large retrospective cohort study in more than one
information about subclinical disease trajectories. Many of
million women in Canada, the risk of premature CVD has
the factors that contribute to coronary heart disease, stroke,
been shown to be strongly related to the severity of metabolic
and cancers—inflammation, vasculopathy, angiogenesis,
thrombosis, insulin resistance, and altered endocrine disturbances that have occurred during pregnancy.7
profiles—also underlie aberrant menstrual patterns, female A key factor underlying CVD is the metabolic syndrome.
infertility, pregnancy complications, and adverse pregnancy The metabolic syndrome is a spectrum of metabolic abnor-
outcomes such as preterm delivery (PTD) or fetal growth malities associated with insulin resistance, which is manifest
restriction.1 as relative hyperglycemia, hyperlipidemia, and disturbance
It has long been understood that pregnancy complica- of coagulation. The normal physiological response to preg-
tions are important for the lifelong health of offspring in nancy represents a transient excursion into a metabolic syn-
the context of fetal programming but much less appreci- drome, with a relative degree of insulin resistance, definite
ated that these complications also have key implications hyperlipidemia, and an increase in coagulation factors.8,9
for the long-term health of the mother.1 Indeed, pregnancy Normal pregnancy also involves upregulation of the inflam-
complications may be the first warning that a woman is at matory cascade.10 Such upregulation in nonpregnant women
an elevated risk for future cardiovascular disease (CVD).2 has recently been recognized as an independent predictor of
An accumulating body of research has shown that com- cardiovascular events and DM.11 Such metabolic responses,
mon pregnancy complications, including PTD, fetal which are the result of the pregnancy-related hormonal
growth restriction, preeclampsia, and gestational diabetes changes,12 could be considered as “stress” tests of maternal
mellitus (GDM), predict the future risk of chronic diseases carbohydrate and lipid pathways and vascular function. In
in women, including CVD, diabetes mellitus (DM), and this way, adverse pregnancy outcomes may be indicators of
breast cancer.1,3 Of note, a high proportion of women will, an increased risk of metabolic and vascular diseases in later
in the course of their reproductive career, have at least one life (Figure 47.1).
of these pregnancy adverse outcomes. Indeed, the preva- Recently, a review analyzing the evidence for asso-
lence of any one of these conditions in any given pregnancy ciations of parity and common pregnancy complications
ranges from 2% to greater than 12%. 3 Because of the actual (low birth weight, PTD, hypertensive disorders of preg-
magnitude of this problem, an in-depth insight appears nancy, and GDM) with future CVD risk has been pub-
mandatory. lished. Physiological mechanisms that might underlie these
391
392 Reproduction and its impact on health and disease
Population with complicated pregnancies the result of constitutional determinants and external factors
Population with physiological pregnancies that influence fetal growth throughout pregnancy, whereas
Threshold for vascular/metabolic diseases the ponderal index may reflect restriction or acceleration of
Vascular risk factors
nutritional status in the last part of pregnancy.23

Offspring birth weight has been reported to strongly pre-
dict maternal life span.24–28 One meta-analysis has calculated
that, for every standard deviation (roughly 500 g) higher
birth weight of the firstborn child, maternal CVD mortality
is decreased by 25%.24
A twofold increase of the maternal CVD incidence and
mortality is reported for pregnancies complicated by hav-
Neonatal life Pregnancies Middle age ing low-birth-weight infants (<2500 g), whose incidence is
Age
nearly 8%.24–26,28 Further, adjustment for cigarette smoking
decreased this risk only modestly, and no correcting effect
Figure 47.1 Vascular risk factors and metabolic syndrome was identified for prepregnancy BMI.29,30 The presence of
of pregnancy. (Adapted from Sattar, N. and Greer, I.A., Br. insulin resistance, hypertension, and inflammation likely
Med. J., 325, 157, 2002.) underlies this increased risk.27 However, as previously men-
tioned, the basic birth weight is not so accurate as standard-
associations and implications for future research and for ized birth weight and ponderal index in evaluating actual
healthcare design and policy have been also investigated.3 fetal growth. Therefore, Lykke et al. investigated the rela-
Risk factors for vascular disease are identifiable during tion between the standardized birth weight by gestational
excursions into the metabolic syndrome of pregnancy. age and gender and the ponderal index and the mother’s
subsequent mortality and cardiovascular morbidity.31 The
Parity authors showed that the risk profile for the standardized
The association between parity and later CVD has been birth weight and subsequent maternal death progressively
investigated in several studies.13–16 increased with decreasing fetal growth, peaking at < −3
To date, the largest study published included 1.3 m
illion SD (HR 2.75; 95% CI, 2.37–3.19) compared to the median.
women (included in the Swedish registry data) with a median Further, the risk profile for subsequent DM by standard-
follow-up time of 9.5 years.16 Parity was shown to be associ- ized ponderal index appeared to rise with increasing fetal
ated with CVD in a J-shaped fashion, with two births rep- growth and to peak at > +3 SD (HR 17.8; 95% CI, 15.0–21.0).
resenting the nadir of risk. Specifically, women with 0 and The authors concluded that fetal growth is to be considered
≥5 births had multivariable-adjusted hazard ratios (HRs) a pivotal marker of subsequent risk for premature death,
of 1.11 (95% confidence interval [CI], 1.09–1.14) and 1.57 CVD, and DM in the mother. Specifically, since higher
(95% CI, 1.52–1.64), respectively, as compared with women estimates were identified for standardized birth weight
with 2 births. compared with ponderal index, the authors hypothesized
Investigation of the association between number of chil- that the fetal growth degree could be superior to the fetal
dren and paternal CVD risk has helped in gaining a bet- growth mode and that maternal constitutional factors such
ter understanding of the parity associated with maternal as genetic growth potential could be more important than
CVD risk. Indeed, similar associations for mothers and fetal nutritional deprivation to indicate long-term mater-
fathers would suggest that the association between parity nal health. This hypothesis appears in accordance with “the
and maternal CVD is not causal but is more likely a result thrifty genotype hypothesis.”32
of confounding by socioeconomic position and/or behaviors Birth weight predicts maternal CVD risk,24,29,31 as does
related to child-rearing and shared by both parents. Three gestational length (see the section “Preterm delivery”). The
reports have endorsed this hypothesis.17–20 coincidence of restricted fetal growth and prematurity shows
an additive effect, yielding a more than threefold increased
Birth weight CVD risk.29
Birth weight is the conventional measurement used to indi- Conflicting results are reported in literature about the
cate how well fetal growth has succeeded. However, it is association between large birth weight (macrosomia) and
dependent on gestational age and to a lesser extent on gender. maternal CVD risk, recognizing both a decreased and a
In this respect, the fetal growth degree can be defined as birth higher risk.24,25,28,29,31,33 However, given the strong associa-
weight standardized for gestational age and gender, with small tions of macrosomia with GDM and later type 2 DM,31,34
for gestational age (SGA) and large for gestational age being the presence and magnitude of the association of large birth
the two extremes.21,22 Another measurement of fetal growth weight with future CVD risk may depend on the popula-
is the physical shape of the newborn, the ponderal index, tion prevalence of GDM and chronic DM during preg-
which is an indicator of infant leanness or obesity, parallel nancy. Indeed, the association of macrosomia with CVD
to body mass index (BMI). This can be construed to repre- risk is attenuated by adjustment for GDM,29 indicating that
sent the fetal growth mode, which is also dependent on gesta- a substantial portion of the association of macrosomia and
tional age. The standardized birth weight may be regarded as CVD is explained by underlying maternal metabolic risk.
Implications of pregnancy history on subsequent women’s health 393
Similarly, the association of low birth weight with maternal Hypertensive disorders
CVD risk may be driven by endothelial dysfunction and Hypertensive disorders of pregnancy are common obstet-
other pathologies associated with restricted fetal growth ric complications that presage CVD. Preeclampsia affects
and preterm birth.3 approximately 2%–5% of pregnancies, 52,53 and it still rep-
resents a major cause of perinatal morbidity and mortality.54
Preterm delivery In turn, incidence of gestational hypertension varies from
PTD (<37 weeks of gestation) accounts for 6%–12% of deliv- 3% to 14%.52,53,55
eries in the developed world.35 The HRs for CVD associated Although a link between preeclampsia and future chronic
with total PTD are on the order of 1.3–2.6 as compared with hypertension has already been suggested almost 50 years
term births.3,28–31,36–42 However, different values of relative risk ago,56 it was only recently that preeclampsia has been
(RR) are reported when considering the distinct phenotypes accepted to predispose also to premature CVD.57
of PTD, that is, spontaneous (spontaneous labor, preterm Case–control and cohort studies consistently reported
premature rupture of membranes) or iatrogenic (medically that preeclampsia is predictive of future CVD and cere-
induced labor, cesarean section). Spontaneous preterm birth brovascular disease. This risk has been summarized in
is generally associated with uteroplacental inflammation, two systematic reviews of controlled studies that evalu-
ischemia, and hemorrhage.43 Further, women with a history ated the risk of late cardiovascular events in women with
of spontaneous PTD show higher inflammatory levels, larger and without a history of preeclampsia. 58,59 Compared
waist circumference, and dyslipidemia years after deliv- with women with no history of the disease, preeclamptic
ery.27 In turn, the main reasons for the medically indicated patients were at increased risk of developing hyperten-
preterm deliveries include preeclampsia and fetal growth sion (RR 3.70; 95% CI, 2.70–5.05 at mean follow-up of
restriction, well-known risk factors for subsequent mater- 14 years), ischemic heart disease (RR 2.16; 95% CI, 1.86–
nal CVD development. Consequently, hypertensive preterm 2.52 at mean follow-up of 11.7 years), stroke (RR 1.81; 95%
deliveries show a stronger association with maternal CVD CI, 1.45–2.27 at mean follow-up of 10.4 years), and venous
outcomes as compared with normotensive preterm deliver- thromboembolism (RR 1.79; 95% CI, 1.37–2.33 at mean
ies, though the latter are still associated with a 1.2–3-fold follow-up of 4.7 years). The absolute risk that a woman
increased risk compared with term deliveries.36,40,44,45 This with or without a history of preeclampsia would develop
can be likely related to the presence of common antecedents one of these cardiovascular events at age 50–59 years was
in preterm birth and CVD.45,46 As mentioned earlier, the pos- estimated to be 17.8% and 8.3%, respectively. 59 In addi-
sible causes of spontaneous preterm birth include infection tion, a graded relationship was observed between severity
and inflammation involving pro-inflammatory cytokines, of preeclampsia and risk of future cardiac disease (mild
the prostaglandin cascade, and matrix metalloproteinases.47 preeclampsia [RR 2.00; 95% CI, 1.83–2.19], moderate pre-
The process of atherosclerosis is also largely inflammatory eclampsia [RR 2.99; 95% CI, 2.51–3.58], severe preeclamp-
in its origin. Further, biochemical markers of inflammation sia [RR 5.36; 95% CI, 3.96–7.27]), as well as correlation
are related to the risk of plaque rupture and endothelial dys- between preeclampsia and future peripheral artery dis-
function and, subsequently, to the CVD risk but also to the ease (RR 1.87; 95% CI, 1.94–3.73). 58
risk of preterm birth.48,49 In addition, thrombin activation Similar findings have been reported by prospective
by vascular lesions can be detected in the placentas of spon- cohort studies published afterward.52,60–62 Further, the Child
taneous preterm deliveries. It has been reported to be a key Health and Development Study in California, providing the
component in the atherosclerotic process.50,51 longest follow-up (37 years), has confirmed that women with
A 39% higher risk in women delivering moderately pre- a history of preeclampsia in any pregnancy have a twofold
term infants and a more than doubled risk in women deliv- increase in the risk of CVD death (adjusted HR 2.14; 95%
ering very (28–32 weeks) and extremely (<28 weeks) preterm CI, 1.29–3.57).
infants has been reported, as compared with women who Women with early-onset/severe preeclampsia and PTD
had a first term delivery, even after adjustment for potential are at the highest risk of CVD later in life, including during
confounders, including preeclampsia and smoking.29 The the premenopausal period (Table 47.1).36 In two large stud-
impact of the severity of PTD on the risk of specific car- ies, these women had an eight- to ninefold increased risk of
diovascular events (hypertension, ischemic heart disease, death from cardiovascular causes compared with women
thromboembolism) and type 2 DM in the mother has also without a history of preeclampsia.36,62 In contrast, mild pre-
been investigated in a registry-based retrospective cohort eclampsia occurring late in gestation does not appear to be
study conducted by Lykke et al.39 Included in the analysis associated with a high risk of remote CVD.63 The stronger
were 782,287 women with a first delivery. After adjustments association between CVD and preterm preeclampsia sug-
for maternal age, year of delivery, hypertensive pregnancy gests that the pathogenesis of early versus late preeclampsia
disorders, fetal growth deviation, placental abruption, and may be different.
stillbirth, the authors identified that women delivering Endothelial dysfunction is a hallmark of preeclampsia.
between 32 and 36 weeks of gestation were at higher risk for Several studies have demonstrated that women with a history
all the cardiovascular events analyzed, as compared with of preeclampsia or severe early-onset fetal growth restric-
term controls. Further, the adjusted risk of subsequent type tion exhibit impaired endothelial function and peripheral
2 DM increased 1.89-fold. vasodilation remote from pregnancy, thus suggesting that
hypertension in the absence of GDM were associated with

Table 47.1 Deaths from cardiovascular causes
a twofold increase in the incidence of type 2 DM during
16.5 years of postdelivery follow-up, after controlling for
Relative hazard rate several confounding variables. Unfortunately, analysis was
Population (95% CI) not adjusted for prepregnancy BMI and obesity, well-known
Nonpreeclamptic, term delivery 1 risk factors for type 2 DM development.88 In women who had
Nonpreeclamptic, preterm delivery 2.95 (2.12–4.11) preeclampsia or gestational hypertension and GDM, the risk
Preeclamptic, term delivery 1.65 (1.01–2.70) of future type 2 DM resulted to have a 16–18-fold increase,
Preeclamptic, preterm delivery 8.12 (4.31–15.33)
which is above the already elevated 13-fold increase in risk
associated with GDM alone. These findings and those from
Source: Adapted from Irgens, H.U. et al., Br. Med. J., 323, 1213, previous reports89–91 suggest that clinicians should inform
2001.
women with a history of preeclampsia or gestational hyper-
tension that they may be at increased risk of developing
DM and, likely, cardio- or cerebrovascular events. However,
persistence of endothelial dysfunction postpartum may the available evidence did not have supported changes in
contribute to the increased CVD risk identified in this standard screening guidelines so far.
patient population.64–67 Mangos et al. tested, 2–12 years after Finally, since the severity of the hypertensive disorders
delivery, women who have had preeclampsia or gestational seems to predispose to thromboembolic events,59 caution is
hypertension during their pregnancy, in order to investigate mandatory when prescribing oral contraceptives for these
endothelial dysfunction, sympathetic (over)activity and/or women.92
renin–angiotensin system activation, and abnormalities of
renal function.68 Women with previous hypertensive disor- Gestational diabetes mellitus
ders of pregnancy showed a greater BMI, more insulin resis- Pregnancy is characterized by progressive insulin resistance
tance, and a higher awake and sleep blood pressure values, as from midgestation onward. The severe insulin resistance of
compared with patients who have had normal pregnancies. late pregnancy can be looked upon as a physiological test of
However, none of the physiological stress tests of the sym- beta-cell compensatory capacity. An optimal response to this
pathetic nervous system and endothelial function differed test would be the maintenance of normal glucose tolerance
among the groups, as well as the renal function analysis. In in pregnancy. In turn, any abnormality of antepartum glu-
contrast, others have reported a reduced peripheral vasodi- cose homeostasis can be inferred to reflect some degree of
lation in women with a history of preeclampsia, when using underlying beta-cell dysfunction, with GDM representing the
different methods of evaluation.64–66,69–72 most severe case thereof.93 Beta-cell function has been shown
Preeclampsia and CVD share many risk conditions such to progressively decrease, as glucose tolerance status worsens
as systemic and endothelial inflammation, insulin resistance, from normal to hyperglycemia to GDM.94 The long-term sig-
DM, obesity, and preexistent hypertension. Elevated fasting nificance of this observation is underscored by the central role
insulin and glucose levels have been demonstrated during of beta-cell dysfunction in the pathogenesis of type 2 DM.95
glucose tolerance tests in women with preeclampsia.73 Also, GDM, defined as glucose intolerance that is first rec-
women with a history of preeclampsia have higher serum ognized during pregnancy, confers a fourfold to seven-
concentrations of fasting lipids, insulin, and coagulation fold greater risk of the incidence of type 2 DM.34,96,97 A
factors postpartum than do controls matched for BMI.74–76 population-based study reported that the incidence of type 2
Data from some epidemiological studies suggest that the DM in women with previous GDM is 3.7% 9 months postpar-
increased risk of late cardiovascular morbidity in previously tum, 4.9% 15 months postpartum, 13.1% 5 years postpartum,
preeclamptic women reflects an underlying predisposition and 18.9% 9 years postpartum (versus 2% in controls with-
in these women for both disorders (genetic factors, shared out previous GDM).98 It has been estimated that 3%–70% of
risk factors), but it is also possible that preeclampsia results women with a history of GDM will develop type 2 DM within
in permanent arterial changes leading to late CVD.77–80 Some three decades of the pregnancy.99,100 Waist circumference and
investigators have hypothesized that increased insulin resis- BMI are the strongest anthropometric measures associated
tance, sympathetic overactivity, pro-inflammatory status, with the development of type 2 DM in women with GDM, as
endothelial dysfunction, and the abnormal lipid profile in they are in women without GDM.98,101 Indeed, type 2 DM has
preeclamptic women constitute an early manifestation of been shown to develop in 50%–70% of obese women with a
the metabolic syndrome and that these changes persist after history of GDM, versus fewer than 25% of women with GDM
pregnancy, thereby putting affected women at increased risk who achieve normal body weight after delivery.102–105 Other
of CVD.81–85 One group estimated that lifestyle interventions major risk factors include gestational requirement for insulin
after preeclampsia would decrease CVD risk by 4%–13%.86 and early gestational age at time of diagnosis.101
However, it is still incompletely clarified whether preeclamp- In addition to type 2 DM, this patient population is at
sia itself, underlying risk conditions, or a combination of increased risk of several other cardiometabolic abnormali-
both contributes to future hypertension and CVD.87 ties, including obesity, hypertension, and dyslipidemia.106–111
In a population-based retrospective cohort study includ- Given this constellation of traditional cardiovascular risk
ing over one million women, preeclampsia or gestational factors, it is not surprising that women with a history of
Implications of pregnancy history on subsequent women’s health 395
previous GDM exhibit high rates of the metabolic syndrome population-based interventional study including deliver-
in the first decade following the index pregnancy.106,110,111 ies occurred during a 25-year period with a mean follow-up
Further, these women have been linked to a variety of non- duration of 11.2 years, the authors recognized higher rates of
traditional cardiometabolic risk factors including increased total renal morbidity in women with previous GDM (OR 2.3,
insulin resistance, low circulating levels of adiponectin, and 95% CI 1.4–3.7; p < 0.001) and a linear correlation between
upregulation of inflammatory proteins.112 Moreover, these the number of GDM episodes and the risk for future renal
patients display evidence of vascular dysfunction, including morbidity.
increased vessel stiffness, and impaired cardiac autonomic Altogether these data further emphasize the importance
function.113–116 Finally, they have been shown to be at risk for of GDM screening to maternal future morbidity risk assess-
the early onset of subclinical atherosclerosis, as indicated ment and may initiate the discussion regarding the introduc-
by increased carotid artery intima-media thickness (IMT) tion of specific screening programs after GDM pregnancies
(the CARDIA Study—NCT00005130).87 The carotid artery in order to early detect vascular, metabolic, and renal dis-
IMT has been shown to strongly predict heart disease and eases, thus using the benefits of secondary prevention.
stroke, particularly in women. Therefore, on the basis of all
of these findings, it has long been suspected that the diagno- First-trimester bleeding and miscarriage
sis of GDM likely identifies a population of young women at First-trimester bleeding is common in pregnancy, with
increased risk of CVD, in addition to type 2 DM.117,118 an incidence up to 20%–27%, and for about half of these
Studies have reported a 66%–85% higher risk of CVD, women, the pregnancy will end in miscarriage.132,133 In these
including coronary artery disease, myocardial infarc- patients, a higher risk of ischemic heart disease later in life
tion, and/or stroke.99,109,119–121 Results of the aforementioned has already been reported.38,134 In turn, for women with an
CARDIA study demonstrated that a history of GDM signi- ongoing pregnancy, evidence has mounted in the last few
fies increased CVD risk apart from greater prepregnancy years for an increased risk of later pregnancy complications
obesity, race, parity, and age. Weight gain, insulin resistance, such as PTD, placental abruption, and preeclampsia.135,136
and blood pressure partially accounted for the GDM-IMT As reported herein, these obstetric complications, as well as
association, thus leading the authors to suggest that body size, fetal growth restriction, are associated with increased mater-
blood pressure control, and insulin resistance could be impor- nal mortality and cardiovascular morbidity.
tant modifiable risk factors that may influence progression of Lykke et al. conducted a registry-based cohort study, aim-
atherosclerosis during the midlife in women with a history of ing to investigate the association of first-trimester bleeding
GDM. Two cohort studies reported a RR of CVD of 1.71 (95% without miscarriage with both prepregnancy cardiovascu-
CI 1.08–2.69, with a median follow-up of 12 years)99 and 1.58 lar morbidity and maternal long-term cardiovascular out-
(95% CI 1.00–2.49, with a median follow-up of 30 years).109 come.137 The results showed that women with prepregnancy
However, it is still unclear whether a history of GDM increases CVD were at increased risk of first-trimester bleeding and,
CVD risk independent of subsequent metabolic disorders, vice versa, that women experiencing first-trimester bleeding
mostly due to the lack of biochemical screening for metabolic were at higher risk of subsequent CVD (Table 47.2). The lat-
diseases before and after pregnancy in the previous studies.122 ter association persisted after adjustment for later pregnancy
Not only GDM but also lesser degrees of antepartum complications, which are associated with both first-trimester
hyperglycemia have been associated with an elevated risk of bleeding136 and subsequent maternal cardiovascular morbid-
subsequent type 2 DM and CVD.93,120,123–125 These data sug- ity,31,38,39,89 and thus being potential confounders. According
gest that further study of the potential relationship between to the intrinsic model of early vaginal bleeding, impaired
mild glucose intolerance in pregnancy and vascular disease cytotrophoblast invasion and spiral artery remodeling may
is needed, as it is possible that current antepartum GDM underlie this obstetric complication. Interestingly, an altered
screening may provide unrecognized insight into the future placentation also explains the observed link to later adverse
cardiovascular risk potential in young women.112 pregnancy outcomes,4,15,135,138 thus indicating the potential
Since the maternal cardiovascular risk is associated with existence of a common etiology to first-trimester bleeding,
GDM development, the ACOG,126 the American Diabetes adverse pregnancy outcomes, and subsequent CVD.
Association (ADA),127 and the Fifth International Workshop
Conference on Gestational Diabetes128 have recommended
long-term follow-up of these patients. All of them need to Recurrent pregnancy complications and maternal
undergo an oral glucose tolerance test 6–12 weeks after deliv- CVD risk
ery, using a 2-hour 75 g oral glucose tolerance test. In case of There is evidence that recurrent preeclampsia89 and PTD39,40
a negative result, women should be counseled regarding their are associated with a greater risk of CVD than a single com-
risk of developing GDM in subsequent pregnancies and type 2 plicated pregnancy in multiparous women. Specifically,
DM and CVD in the future. Since lifestyle interventions, such Lykke et al. demonstrated that PTD is associated with sub-
as weight loss and exercise, are clearly beneficial for reducing sequent overt type 2 DM and that the recurrence of PTD
the incidence of these disorders,129,130 they should be always causes a twofold increase in this risk.39
performed by clinicians. One-third to two-thirds of women with GDM in their first
Recently, a significant correlation between GDM and pregnancy will have GDM in a subsequent pregnancy.139–142
long-term renal morbidity has been also reported.131 In a In a study including over 65,000 pregnancies, the risk of GDM
Table 47.2 First-trimester bleeding without miscarriage and subsequent maternal cardiovascular
morbidity
Controls Cases
n = 763,976 n = 18,311 Adjusteda
Long-term outcome n % n % HR 95% CI P
Hypertension 18,101 2.4 450 2.5 1.20 1.09–1.32 <0.001
Ischemic heart disease 8,264 1.1 252 1.4 1.58 1.39–1.79 <0.001
Stroke 8,738 1.1 249 1.4 1.41 1.25–1.60 <0.001
Thromboembolic event 3,760 0.5 121 0.7 1.61 1.34–1.93 <0.001
Diabetes mellitus 6,520 0.9 200 1.1 1.51 1.31–1.74 <0.001
a Adjusted for the preterm delivery, primary rupture of membranes, hypertensive disorders, small for gestational age/large for
gestational age, placental abruption, and stillbirth.
in the second pregnancy among women with and without by their large, healthy placentas,150,151 thus hypothesizing that
previous GDM was 41% and 4%, respectively.139 Although the endocrine, angiogenic, and thrombotic processes might explain
association of recurrent GDM with CVD risk has not been the links between offspring size and breast cancer risk.3
studied, after a first GDM pregnancy, each subsequent GDM
pregnancy has been associated with a modestly increased Preterm delivery
risk of type 2 DM (adjusted HR = 1.16; 95% CI, 1.01–1.34)
and, likely, of cardio- or cerebrovascular events later in life.143 Several studies have reported that PTD, especially in a first
pregnancy, is associated with an increased risk of future breast
Additionally, having preeclampsia,144 PTD,44 or GDM143
cancer36,152–155 and, perhaps, ovarian cancer.156 However, data
in the last pregnancy appears to be associated with higher
in literature are still controversial on this topic.151,157,158
risk of future CVD in multiparous mothers, thus suggesting
that pregnancy complications severe enough to contraindi-
Hypertensive disorders
cate or discourage a subsequent pregnancy may be particu-
larly potent predictors of future CVD risk.3 No significant association between hypertensive disorders of
pregnancy and later development of cancer has been iden-
tified.59,159 Observational studies from the United State,
Non-CVD risk Sweden, and Norway reported that women with preeclamp-
Breast cancer risk sia were at reduced risk or had no excess risk of cancer when
Parity followed 13–19 years postpartum.36,150,154,157,160–162 Further, in
While high parity has shown a protective effect for hormone- the registry-based retrospective cohort study of Lykke et al.,
receptor positive breast cancers,145 an increased risk of estrogen women experiencing hypertensive disorders of pregnancy
receptor (ER)- and progesterone receptor (PR)-negative (ER- were not identified to be at increased risk of subsequent death
PR-) tumors, known to be associated with poorer prognosis from noncardiovascular causes in general. Interestingly, mild
compared with other cancer subtypes, has been identified.146 preeclampsia was associated with a slightly reduced risk of
However, emerging evidence suggests that breastfeeding may death from this kind of causes.163 In turn, a study from Israel
reduce this risk.146 Using cases and controls taken from the reported an increased risk of cancer in such women (HR 1.27;
Breast Cancer Family Registry, Work et al. examined reproduc- 95% CI, 1.03–1.57) with a median follow-up of 29 years.164,165
tive risk factors in relation to ER and PR status of breast cancers. The discordant results may be explained by several factors,
High parity (≥3 live births) without breastfeeding was identi- including differences in patient populations, the absence of or
fied as positively associated with ER-PR-tumors (OR 1.57; 95% insufficient adjustment for confounders, differences in length
CI, 1.10–2.24). In turn, no association with parity was found in of follow-up, and incomplete ascertainment.
women who breastfed for a total duration of ≥12 months (OR
0.93; 95% CI, 0.71–1.22).147 Hypothesized potential protective Gestational diabetes mellitus
mechanisms include the removal of estrogens via breast fluid, Although a handful of studies have indicated links between
excretion of carcinogenic agents through breast milk, delay in type 2 DM and breast cancer,166,167 only a few studies have
ovulation associated with breastfeeding, and induction of ter- investigated whether GDM increases the risk of cancer. Thus
minal differentiation of breast epithelial cells.148 far, reports include increased,166,168 decreased,169 and null
associations of GDM with breast cancer risk.158
Birth weight
Offspring birth weight has been positively associated with risk Renal cancer risk
of breast cancer.24,28,149 Intriguingly, any increased breast cancer Clinical and experimental findings suggest that female hor-
risk associated with bearing large infants seems to be explained monal and reproductive factors could influence kidney cancer
Conclusion 397
Table 47.3 Combinations of preterm delivery, small for gestational age, and preeclampsia
and the risk of subsequent death from cardiovascular causes
All women Death from cardiovascular causes

Pregnancy outcomes n % n % HR 95% CI P
None 643,935 82.3 824 0.1 1.00 Reference <0.001
PTD 31,132 4.0 70 0.2 1.90 1.49–2.43 <0.001
SGA 33,311 4.3 136 0.4 2.51 2.10–3.01 <0.001
PTD + SGA 5,206 0.7 27 0.5 3.30 2.25–4.84 <0.001
Preeclampsia 25,184 3.2 72 0.3 2.08 1.63–2.64 <0.001
Preeclampsia + PTD 2,662 0.3 6 0.2 2.41 1.08–5.37 <0.001
Preeclampsia + SGA 2,374 0.3 17 0.7 4.67 2.89–7.55 <0.001
Preeclampsia + PTD + SGA 2,218 0.3 10 0.5 3.85 2.07–7.19 <0.001
Missing values 36,265 4.6 148 0.4 1.25 1.04–1.50 <0.001
Source: Reproduced from Lykke, J.A. et al., Paediatr. Perinat. Epidemiol., 24, 323, 2010. With permission.
Abbreviations: PTD, preterm delivery; SGA, small-for-gestational-age offspring.
development.170 Interestingly, it has been reported that ana- with decreasing gestational age. Within each stratum of
tomical changes to the kidney during pregnancy might gestational age, mothers to very-small-for-gestational-age
make nephrons more vulnerable to inflammation and oxi- infants consistently had higher risks of CVD than mothers
dative stress.171 to non-SGA or moderately SGA infants, who appeared to
Since most studies examining the association of female have similar risks. Compared with women giving birth to a
hormonal and reproductive factors with kidney cancer risk non-SGA infant at term, women giving birth to a very-small-
have been based on small case numbers, with limited statis- for-gestational-age infant very preterm had a tripled risk for
tical power, Karami et al. decided to conduct analyses in two CVD during follow-up. Further, HRs for CVD remained
large, prospective cohort studies to more comprehensively essentially unchanged after additional adjustment for pre-
investigate this association.172 No clear evidence of an asso- pregnancy BMI.
ciation with parity and number of live births was found.
Combined pregnancy complications Physiological mechanisms linking

Recently, Lykke et al. have reported the results of a registry- pregnancy complications to maternal
based retrospective cohort study of 782,287 women, aiming
to elucidate the impact of PTD, delivery of an SGA offspring, CVD risk
hypertensive disorders of pregnancy, placental abruption, Pathways that link pregnancy complications to later-life
and stillbirth in a first delivery on subsequent rates of death CVD are still incompletely defined. Considerable evidence
from cardiovascular and noncardiovascular causes.163 supports the existence of common predisposing factors for
The cardiovascular mortality was increased in women both pregnancy complications and CVD risk. However,
delivering preterm (HR 1.90; CI 95%, 1.49–2.43), having an almost no studies have investigated the alternative that
SGA newborn (HR 2.51; CI 95%, 2.10–3.01) or having prepregnancy complications might cause increased CVD risk.
eclampsia only (HR 2.08; CI 95%, 1.63–2.64). A significant Rich-Edwards and coworkers have analyzed evidence com-
additive effect was identified when experiencing combi- paring CVD risk before, during, and after pregnancies with
nations of these (Table 47.3). In turn, only PTD of an SGA and without complications, trying to address this issue.3 The
newborn related to an increased risk of death from noncar- authors conclude that the associations of pregnancy com-
diovascular causes (HR 2.12; CI 95%, 1.72–2.62). plications with future CVD events in women are explained
An interaction between gestational age and fetal growth only in part by their associations with classic CVD risk fac-
with respect to mothers’ CVD risk has also been reported tors of hypertension, dyslipidemia, type 2 DM, inflamma-
by Bonamy et al. in their nationwide registry-based study. tion, and thrombosis, which are evident before, during, and
Accounting for the possible influence of maternal charac- after such complicated pregnancies.
teristics (e.g., smoking) and pregnancy complications (e.g.,
hypertensive diseases), factors associated with fetal growth
restriction, preterm birth, and mother’s long-term risks of Conclusion
CVD,36,45,173,174 the authors identified a significant multipli-
cative interaction between gestational age and birth weight The reproductive health of women provides glimpses into
standardized for gestational age. In each stratum of stan- subclinical processes and their future risk of chronic dis-
dardized birth weight, incidence rates of CVD increased ease. Specifically, the pregnancy complications may serve
as an early indication that a woman is on a high CVD risk public health, given the magnitude of the associations, the
trajectory, before the classic CVD risk factors (hypertension, prevalence of the pregnancy complications, and the impor-
dyslipidemia, type 2 DM, and endothelial dysfunction) are tance of CVD in women. Further, pregnancy complications
clinically detected. However, whether pregnancy compli- occur early enough in a woman’s life course to offer a signifi-
cations are either mere markers of a woman’s subsequent cant preventive intervention by lifestyle changes and drug
disease susceptibility, or important modifiers of pathologi- administration (i.e., statins and antihypertensive drugs).
cal processes, e.g., via accelerated systemic atherosclerosis Accordingly to what reported herein, histories of preeclamp-
leading to cardiovascular morbidity,175 is still incompletely sia and GDM have been recently included as part of CVD
clarified. In some cases, pregnancy complications are pre- risk assessment by both the American Heart Association and
ceded by manifestations of the metabolic syndrome, which the European Society of Cardiology.180,181
will predispose to later CVD.77,176,177 Further, a possible
relationship between the individual’s genotype, pregnancy
complications, and subsequent CVD and death has yet to be Summary points
discovered.178
Further work to elucidate the endocrine, metabolic, vas- ●● Women with a history of adverse pregnancy outcome
cular, immune, and inflammatory factors shared by repro- appear to be at increased risk of metabolic and vascular
ductive disorders and chronic disease is needed. Indeed, diseases in later life.
many of the reproductive disorders overlap in definition or ●● Pregnancy complications and CVD and cerebrovascular
etiology, as in the predominance of preterm infants among diseases may have common pathogenic mechanisms.
the low birth weight. ●● Pregnancy may be viewed as a cardiovascular stress test,
Because of the progressive increase in maternal age at the and having a severe pregnancy complication might be the
first pregnancy and in the use of assisted reproduction, and first sign of CVD susceptibility in a woman.
the recent advances in obstetric care, the number of women ●● Women with a history of GDM should be screened for
having a history of pregnancy complications will continue to type 2 DM and be given counseling and appropriate life-
rise. Pregnancy may be viewed as a cardiovascular stress test, style advice.
and a preterm and/or SGA delivery might be the first sign of ●● Women who have had a very-low-birth-weight baby
CVD susceptibility in a woman.46,179 or combined complications seem to be at severalfold
Information about reproductive health might be used as increased risk of mortality from cardiovascular causes
a form of “personalized medicine” to tailor health advice to and should be screened for vascular risk factors.
women, from early in their lives, in order to change disease ●● Since the severity of the hypertensive disorders seems to
trajectories before they emerge as chronic disease at meno- predispose to thromboembolic events, caution is man-
pause. The use of pregnancy complication history to screen datory when prescribing oral contraceptives for these
women for targeted CVD prevention has potential to improve women.
REFERENCES
1. Rich-Edwards JW. Reproductive health as a sentinel of chronic 10. Sacks GP, Studena K, Sargent K, Redman CW. Normal preg-
disease in women. Womens Health 2009; 5: 101–105. nancy and preeclampsia both produce inflammatory changes in
2. Sattar N, Greer IA. Pregnancy complications and maternal car- peripheral blood leukocytes akin to those of sepsis. Am J Obstet
diovascular risk: Opportunities for intervention and screening? Br Gynecol 1998; 179: 80–86.
Med J 2002; 325: 157–160. 11. Haffner SM. Do interventions to reduce coronary heart disease
3. Rich-Edwards JW, Fraser A, Lawlor DA, Catov JM. Pregnancy reduce the incidence of type 2 diabetes? A possible role for
characteristics and women’s future cardiovascular health: An inflammatory factors. Circulation 2001; 103: 346–347.
underused opportunity to improve women’s health? Epidemiol 12. Sattar N, Gaw A, Packard CJ, Greer IA. Potential pathogenic roles
Rev 2014; 36: 57–70. of aberrant lipoprotein and fatty acid metabolism in pre-eclamp-
4. Shah RU, Klein L, Lloyd-Jones DM. Heart failure in women: sia. Br J Obstet Gynaecol 1996; 103: 614–620.
Epidemiology, biology and treatment. Womens Health 2009; 5: 13. Green A, Beral V, Moser K. Mortality in women in relation to
517–527. their childbearing history. Br Med J 1988; 297: 391–395.
5. Labarthe DR. Prevention of cardiovascular risk factors in the first 14. Steenland K, Lally C, Thun M. Parity and coronary heart disease
place. Prev Med 1999; 29: S72–S78. among women in the American Cancer Society CPS II popula-
6. Weintraub WS, Daniels SR, Burke LE et al. Value of primordial tion. Epidemiology 1996; 7: 641–643.
and primary prevention for cardiovascular disease: A policy 15. Ness RB, Harris T, Cobb J, Flegal KM, Kelsey JL, Balanger A,
statement from the American Heart Association. Circulation Stunkard AJ, D’Agostino RB. Number of pregnancies and the sub-
2011; 124: 967–990. sequent risk of cardiovascular disease. N Engl J Med 1993; 328:
7. Ray JG, Vermeulen M, Schull MJ, Redelmeier DA. Cardiovascular 1528–1533.
health after maternal placental syndromes (CHAMPS): 16. Parikh NI, Cnattingius S, Dickman PW, Mittleman MA,
Population-based retrospective cohort study. Lancet 2005; 366: Ludvigsson JF, Ingelsson E. Parity and risk of later-life maternal
1797–1803. cardiovascular disease. Am Heart J 2010; 159: 215–221.
8. Greer IA. Thrombosis in pregnancy: Maternal and fetal issues. 17. Dekker JM, Schouten E. Number of pregnancies and risk of car-
Lancet 1999; 353: 1258–1265. diovascular disease. N Engl J Med 1993; 329: 1893–1894.
9. Martin U, Davies C, Hayavi S, Hartland A, Dunne F. Is normal 18. Grinblatt JA. Number of pregnancies and risk of cardiovascular
pregnancy atherogenic? Clin Sci 1999; 96: 421–425. disease. N Engl J Med 1993; 329: 1894–1895.
References 399
19. Ness RB, Cobb J, Harris T, D’Agostino RB. Does number of 41. Pell JP, Smith GC, Walsh D. Pregnancy complications and sub-
children increase the rate of coronary heart disease in men? sequent maternal cerebrovascular events: A retrospective cohort
Epidemiology 1995; 6: 442–445. study of 119,668 births. Am J Epidemiol 2004; 159: 336–342.
20. Lawlor DA, Emberson JR, Ebrahim S, Whincup PH, Wannamethee 42. Nardi O, Zureik M, Courbon D, Ducimetiere P, Clavel-Chapelon
SG, Walker M, Smith GD. Is the association between parity and F. Preterm delivery of a first child and subsequent mothers’ risk
coronary heart disease due to biological effects of pregnancy of ischaemic heart disease: A nested case-control study. Eur J
or adverse lifestyle risk factors associated with child-rearing? Cardiovasc Prev Rehabil 2006; 13: 281–283.
Findings from the British Women’s Heart and Health Study 43. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology
and the British Regional Heart Study. Circulation 2003; 107: and causes of preterm birth. Lancet 2008; 371: 75–84.
1260–1264. 44. Abstracts of the 45th Annual SER (Society for Epidemiologic
21. Steer P. The management of large and small for gestational age Research) Meeting. Minneapolis, Minnesota, USA. June 27–30,
fetuses. Semin Perinatol 2004; 28: 59–66. 2012. Am J Epidemiol 2012; 175(Suppl. 11): S1–S145.
22. Goldenberg RL, Cliver SP. Small for gestational age and intra- 45. Hastie CE, Smith GC, Mackay DF, Pell JP. Maternal risk of isch-
uterine growth restriction: Definitions and standards. Clin Obstet aemic heart disease following elective and spontaneous pre-term
Gynecol 1997; 40: 704–714. delivery: Retrospective cohort study of 750 350 singleton preg-
23. Walther FJ, Ramaekers L. The ponderal index as a measure of nancies. Int J Epidemiol 2011; 40: 914–919.
the nutritional status at birth and its relation to some aspects of 46. Satter N. Do pregnancy complications and CVD share common
neonatal morbidity. J Perinat Med 1982; 10: 42–47. antecedents? Atheroscler Suppl 2004; 5: 3–7.
24. Davey Smith G, Hyppönen E, Power C, Lawlor DA. Offspring 47. Menon R, Bhat G, Saade GR, Spratt H. Multivariate adaptive
birth weight and parental mortality: Prospective observa- regression splines (MARS) analysis to predict biomarkers of spon-
tional study and meta-analysis. Am J Epidemiol 2007; 166: taneous preterm birth. Acta Obstet Gynecol Scand 2014; 93(4):
160–169. 382–391.
25. Davey Smith G, Hart C, Ferrell C, Upton M, Hole D, Hawthorne 48. Blake GJ, Ridker PM. Novel clinical markers of vascular wall
V, Watt G. Birth weight of offspring and mortality in the Renfrew inflammation. Circ Res 2001; 89: 763–771.
and Paisley study: Prospective observational study. Br Med J 49. Pitiphat W, Gillman M, Joshipura K, Williams PL, Douglass CW,
1997; 315: 1189–1193. Rich-Edwards JW. Plasma C-reactive protein in early pregnancy
26. Smith GD, Harding S, Rosato M. Relation between infants’ birth and preterm delivery. Am J Epidemiol 2005; 162: 1108–1113.
weight and mothers’ mortality: Prospective observational study. 50. Tracy RP. Thrombin, inflammation, and cardiovascular disease:
Br Med J 2000; 320: 839–840. An epidemiologic perspective. Chest 2003; 124: 49S–57S.
27. Catov JM, Newman A, Roberts JM et al. Association between 51. Vidaeff AC, Monga M, Ramin SM, Saade G, Sangi-Haghpeykar H.
infant birth weight and maternal cardiovascular risk factors in the Is thrombin activation predictive of subsequent preterm delivery?
health, aging, and body composition study. Ann Epidemiol 2007; Am J Obstet Gynecol 2013; 208: 037.
17: 36–43. 52. Fraser A, Nelson S, Macdonald-Wallis C, Cherry L, Butler E,
28. Smith GD, Whitley E, Gissler M, Hemminki E. Birth dimensions of Sattar N, Lawlor DA. Associations of pregnancy complications
offspring, premature birth, and the mortality of mothers. Lancet with calculated cardiovascular disease risk and cardiovascular
2000; 356: 2066–2067. risk factors in middle age: The Avon Longitudinal Study of Parents
29. Bonamy AK, Parikh N, Cnattingius S, Ludvigsson JF, Ingelsson and Children. Circulation 2012; 125: 1367–1380.
E. Birth characteristics and subsequent risks of maternal car- 53. Wallis AB, Saftlas A, Hsia J, Hani K, Atrash HK. Secular trends
diovascular disease: Effects of gestational age and fetal growth. in the rates of preeclampsia, eclampsia, and gestational hyper-
Circulation 2011; 124: 2839–2846. tension, United States, 1987–2004. Am J Hypertens 2008; 21:
30. Smith GD, Sterne J, Tynelius P, Lawlor DA, Rasmussen F. Birth 521–526.
weight of offspring and subsequent cardiovascular mortality of 54. Hernandez-Diaz S, Toh S, Cnattingius S. Risk of pre-eclampsia
the parents. Epidemiology 2005; 16: 563–569. in first and subsequent pregnancies: Prospective cohort study. Br
31. Lykke JA, Paidas M, Triche EW, Langhoff-Roos J. Fetal growth and Med J 2009; 338: b2255.
later maternal death, cardiovascular disease and diabetes. Acta 55. Roberts CL, Algert C, Morris JM, Ford JB, Henderson-Smart DJ.
Obstet Gynecol Scand 2012; 91: 503–510. Hypertensive disorders in pregnancy: A population-based study.
32. Neel JV. Diabetes mellitus: A “thrifty” genotype rendered detri- Med J Aust 2005; 182: 332–335.
mental by “progress”? Am J Hum Genet 1962; 14: 353–362. 56. Adams EM, Macgillivray I. Long-term effect of preeclampsia on
33. Friedlander Y, Paltiel O, Manor O, Deutsch L, Yanetz R, Calderon- blood-pressure. Lancet 1961; 2: 1373–1375.
Margalit R, Siscovick DS, Harlap S. Birthweight of offspring and 57. Carty DM, Delles C, Dominiczak AF. Preeclampsia and future
mortality of parents: The Jerusalem perinatal study cohort. Ann maternal health. J Hypertens 2010; 28: 1349–1355.
Epidemiol 2007; 17: 914–922. 58. McDonald SD, Malinowski A, Zhou Q, Yusuf S, Devereaux
34. Metzger BE, Cho NH, Roston SM, Radvany R. Prepregnancy PJ. Cardiovascular sequelae of preeclampsia/eclampsia: A
weight and antepartum insulin secretion predict glucose toler- systematic review and meta-analyses. Am Heart J 2008; 156:
ance five years after gestational diabetes mellitus. Diabetes Care 918–930.
1993; 16: 1598–1605. 59. Bellamy L, Casas J, Hingorani AD, Williams DJ. Pre-eclampsia
35. Beck S, Wojdyla D, Say L, Betran AP, Merialdi M, Requejo JH. and risk of cardiovascular disease and cancer in later life:
The worldwide incidence of preterm birth: A systematic review of Systematic review and meta-analysis. Br Med J 2007; 335: 1.
maternal mortality and morbidity. Bull World Health Organ 2010; 60. Hermes W, Franx A, van Pampus MG et al. Cardiovascular risk
88: 31–38. factors in women who had hypertensive disorders late in preg-
36. Irgens HU, Reisaeter L, Irgens LM, Lie RT. Long term mortality nancy: A cohort study. Am J Obstet Gynecol 2013; 208: 474.
of mothers and fathers after pre-eclampsia: Population based e1–474.e8.
cohort study. Br Med J 2001; 323: 1213–1217. 61. Smith GN, Pudwell J, Walker M, Wen SW. Ten-year, thirty-year,
37. Wikstrom AK, Haglund B, Olovsson M, Lindeberg SN. The risk of and lifetime cardiovascular disease risk estimates following a
maternal ischaemic heart disease after gestational hypertensive pregnancy complicated by preeclampsia. J Obstet Gynaecol Can
disease. Br J Obstet Gynaecol 2005; 112: 1486–1491. 2012; 34: 830–835.
38. Smith GC, Pell JP, Walsh D. Pregnancy complications and mater- 62. Mongraw-Chaffin ML, Cirillo P, Cohn BA. Preeclampsia and car-
nal risk of ischaemic heart disease: A retrospective cohort study diovascular disease death: Prospective evidence from the child
of 129,290 births. Lancet 2001; 357: 2002–2006. health and development studies cohort. Hypertension 2010; 56:
39. Lykke JA, Paidas M, Damm P, Triche EW, Kuczynski E, Langhoff- 166–171.
Roos J. Preterm delivery and risk of subsequent cardiovascu- 63. Chesley SC, Annitto J, Cosgrove RA. The remote prognosis of
lar morbidity and type-II diabetes in the mother. Br J Obstet eclamptic women. Sixth periodic report. Am J Obstet Gynecol
Gynaecol 2010; 117: 274–281. 1976; 124: 446–459.
40. Catov JM, Wu CS, Olsen J, Sutton-Tyrrell K, Li J, Nohr EA. Early or 64. Chambers JC, Fusi L, Malik IS, De Swiet M, Kooner JS. Association
recurrent preterm birth and maternal cardiovascular disease risk. of maternal endothelial dysfunction with preeclampsia. J Am
Ann Epidemiol 2010; 20: 604–609. Med Assoc 2001; 285: 1607–1612.
65. Agatisa PK, Ness RB, Roberts JM, Costantino JP, Kuller LH, 88. Feig DS, Shah B, Lipscombe LL, Wu CF, Ray JG, Lowe J, Hwee J,
McLaughlin MK. Impairment of endothelial function in women Booth GL. Preeclampsia as a risk factor for diabetes: A popula-
with a history of preeclampsia: An indicator of cardiovascular tion-based cohort study. PLoS Med 2013; 10: e1001425.
risk. Am J Physiol Heart Circ Physiol 2004; 286: H1389–H1393. 89. Lykke JA, Langhoff-Roos J, Sibai BM, Funai EF, Triche EW, Paidas
66. Lampinen KH, Rönnback M, Kaaja RJ, Groop PH. Impaired MJ. Hypertensive pregnancy disorders and subsequent cardio-
vascular dilatation in women with a history of pre-eclampsia. vascular morbidity and type 2 diabetes mellitus in the mother.
J Hypertens 2006; 24: 751–756. Hypertension 2009; 53: 944–951.
67. Yinon Y, Kingdom JC, Odutayo A, Moineddin R, Drewlo S, Lai V, 90. Engeland A, Bjørge T, Daltveit AK, Skurtveit S, Vangen S, Vollset
Cherney DZ, Hladunewich MA. Vascular dysfunction in women SE, Furu K. Risk of diabetes after gestational diabetes and pre-
with a history of preeclampsia and intrauterine growth restric- eclampsia. A registry-based study of 230,000 women in Norway.
tion: Insights into future vascular risk. Circulation 2010; 122: Eur J Epidemiol 2011; 26: 157–163.
1846–1853. 91. Callaway LK, Lawlor DA, O’Callaghan M, Williams GM,
68. Mangos GJ, Spaan J, Pirabhahar S, Brown MA. Markers of cardio- Najman JM, McIntyre HD. Diabetes mellitus in the 21 years
vascular disease risk after hypertension in pregnancy. J Hypertens after a pregnancy that was complicated by hypertension:
2012; 30: 351–358. Findings from a prospective cohort study. Am J Obstet Gynecol
69. Germain AM, Romanik M, Guerra I, Solari S, Reyes MS, Johnson 2007; 197: e1–e7.
RJ, Price K, Karumanchi SA, Valdés G. Endothelial dysfunction: 92. WHO Collaborative Study of Cardiovascular Disease and Steroid
A link among preeclampsia, recurrent pregnancy loss, and future Hormone Contraception. Acute myocardial infarction and com-
cardiovascular events? Hypertension 2007; 49: 90–95. bined oral contraceptives: Results of an international multicentre
70. Paez O, Alfie J, Gorosito M, Puleio P, de Maria M, Prieto N, Majul case-control study. Lancet 1997; 349: 1202–1209.
C. Parallel decrease in arterial distensibility and in endothelium- 93. Retnakaran R, Qi Y, Sermer M, Connelly PW, Hanley AJ, Zinman
dependent dilatation in young women with a history of pre- B. Glucose intolerance in pregnancy and future risk of pre-diabe-
eclampsia. Clin Exp Hypertens 2009; 31: 544–552. tes or diabetes. Diabetes Care 2008; 31: 2026–2031.
71. McDonald SD, Han Z, Walsh MW, Gerstein HC, Devereaux 94. Retnakaran R, Qi Y, Sermer M, Connelly PW, Zinman B, Hanley
PJ. Kidney disease after preeclampsia: A systematic review and AJ. Pre-gravid physical activity and reduced risk of glucose
meta-analysis. Am J Kidney Dis 2010; 55: 1026–1039. intolerance in pregnancy: The role of insulin sensitivity. Clin
72. Hamad RR, Eriksson M, Silveira A, Hamsten A, Bremme K. Endocrinol 2009; 70: 615–622.
Decreased flow-mediated dilation is present 1 year after a pre- 95. Kahn SE. The relative contributions of insulin resistance and
eclamptic pregnancy. J Hypertens 2007; 25: 2301–2307. beta-cell dysfunction to the pathophysiology of Type 2 diabetes.
73. Kaaja R, Laivuori H, Laakso M, Tikkanen MJ, Ylikorkala O. Diabetologia 2003; 46: 3–19.
Evidence of a state of increased insulin resistance in preeclamp- 96. Bellamy L, Casas J, Hingorani AD, Williams D. Type 2 diabetes
sia. Metabolism 1999; 48: 892–896. mellitus after gestational diabetes: A systematic review and meta-
74. Girouard J, Giguère Y, Moutquin JM, Forest JC. Previous hyper- analysis. Lancet 2009; 373: 1773–1779.
tensive disease of pregnancy is associated with alterations of 97. Gunderson EP, Lewis CE, Tsai A, Chiang V, Carnethon M,
markers of insulin resistance. Hypertension 2007; 49: 1056–1062. Quesenberry CP Jr, Sidney S. A 20-year prospective study of
75. Wolf M, Hubel C, Lam C et al. Preeclampsia and future cardio- childbearing and incidence of diabetes in young women, con-
vascular disease: Potential role of altered angiogenesis and insu- trolling for glycemia before conception: The Coronary Artery Risk
lin resistance. J Clin Endocrinol Metab 2004; 89: 6239–6243. Development in Young Adults (CARDIA) Study. Diabetes 2007;
76. Sattar N, Ramsay J, Crawford L, Cheyne H, Greer IA. Classic 56: 2990–2996.
and novel risk factor parameters in women with a history of pre- 98. Feig DS, Zinman B, Wang X, Hux JE. Risk of development of dia-
eclampsia. Hypertension 2003; 42: 39–42. betes mellitus after diagnosis of gestational diabetes. Can Med
77. Magnussen EB, Vatten L, Lund-Nilsen TI, Salvesen KA, Davey Assoc J 2008; 179: 229–234.
Smith G, Romundstad PR. Prepregnancy cardiovascular risk fac- 99. Shah BR, Retnakaran R, Booth GL. Increased risk of cardiovascu-
tors as predictors of pre-eclampsia: Population based cohort lar disease in young women following gestational diabetes mel-
study. Br Med J 2007; 335: 1. litus. Diabetes Care 2008; 31: 1668–1669.
78. Bytautiene E, Bulayeva N, Bhat G, Li L, Rosenblatt KP, Saade GR. 100. Kim C, Newton K, Knopp RH. Gestational diabetes and the inci-
Long-term alterations in maternal plasma proteome after sFlt1- dence of type 2 diabetes: A systematic review. Diabetes Care
induced preeclampsia in mice. Am J Obstet Gynecol 2013; 208: 13. 2002; 25: 1862–1868.
79. Magnussen EB, Vatten L, Smith GD, Romundstad PR. Hypertensive 101. Nicholson WK, Wilson LM, Witkop CT et al. Therapeutic man-
disorders in pregnancy and subsequently measured cardiovascu- agement, delivery, and postpartum risk assessment and screening
lar risk factors. Obstet Gynecol 2009; 114: 961–970. in gestational diabetes. Evid Rep Technol Assess 2008; 162: 1–96.
80. Romundstad PR, Magnussen E, Smith GD, Vatten LJ. Hypertension 102. Dornhorst A, Bailey P, Anyaoku V, Elkeles RS, Johnston DG,
in pregnancy and later cardiovascular risk: Common anteced- Beard RW. Abnormalities of glucose tolerance following gesta-
ents? Circulation 2010; 122: 579–584. tional diabetes. Q J Med 1990; 77: 1219–1228.
81. van Rijn BB, Nijdam M, Bruinse HW, Roest M, Uiterwaal CS, 103. O’Sullivan JB. Body weight and subsequent diabetes mellitus.
Grobbee DE, Bots ML, Franx A. Cardiovascular disease risk fac- J Am Med Assoc 1982; 248: 949–952.
tors in women with a history of early-onset preeclampsia. Obstet 104. O’Sullivan JB. Diabetes mellitus after GDM. Diabetes 1991; 2:
Gynecol 2013; 121: 1040–1048. 131–135.
82. Stekkinger E, Zandstra M, Peeters LL, Spaanderman ME. Early- 105. Baptiste-Roberts K, Barone BB, Gary TL, Golden SH, Wilson LM,
onset preeclampsia and the prevalence of postpartum metabolic Bass EB, Nicholson WK. Risk factors for type 2 diabetes among
syndrome. Obstet Gynecol 2009; 114: 1076–1084. women with gestational diabetes: A systematic review. Am J Med
83. Zandstra M, Stekkinger E, van der Vlugt MJ, van Dijk AP, 2009; 122: 207–214.
Lotgering FK, Spaanderman ME. Cardiac diastolic dysfunction 106. Verma A, Boney CM, Tucker R, Vohr BR. Insulin resistance syn-
and metabolic syndrome in young women after placental syndrome in women with prior history of gestational diabetes mel-
drome. Obstet Gynecol 2010; 115: 101–108. litus. J Clin Endocrinol Metab 2002; 87: 3227–3235.
84. Kaaja RJ, Greer I. Manifestations of chronic disease during preg- 107. Caballero AE. Metabolic and vascular abnormalities in subjects
nancy. J Am Med Assoc 2005; 294: 2751–2757. at risk for type 2 diabetes: The early start of a dangerous situation.
85. Kaaja RJ, Pöyhönen-Alho MK. Insulin resistance and sympathetic Arch Med Res 2005; 36: 241–249.
overactivity in women. J Hypertens 2006; 24: 131–41. 108. Meyers-Seifer CH, Vohr BR. Lipid levels in former gestational dia-
86. Berks D, Hoedjes M, Raat H, Duvekot JJ, Steegers EA, Habbema betic mothers. Diabetes Care 1996; 19: 1351–1356.
JD. Risk of cardiovascular disease after pre-eclampsia and the 109. Carr DB, Utzschneider K, Hull RL et al. Gestational diabetes
effect of lifestyle interventions: A literature-based study. Br J mellitus increases the risk of cardiovascular disease in women
Obstet Gynaecol 2013; 120: 924–931. with a family history of type 2 diabetes. Diabetes Care 2006; 29:
87. Ghossein-Doha C, Spaanderman M, van Kuijk SM, Kroon AA, 2078–2083.
Delhaas T, Peeters L. Long-term risk to develop hypertension in 110. Kousta E, Efstathiadou Z, Lawrence NJ et al. The impact of ethnic-
women with former preeclampsia: A longitudinal pilot study. ity on glucose regulation and the metabolic syndrome following
Reprod Sci 2014; 21(7): 846–853. gestational diabetes. Diabetologia 2006; 49: 36–40.
References 401
111. Lauenborg J, Mathiesen E, Hansen T, Glumer C, Jorgensen T, Borch- 132. Everett C. Incidence and outcome of bleeding before the 20th
Johnsen K, Hornnes P, Pedersen O, Damm P. The prevalence of the week of pregnancy: Prospective study from general practice.
metabolic syndrome in a Danish population of women with pre- Br Med J 1997; 315: 32–34.
vious gestational diabetes mellitus is three-fold higher than in the 133. Hasan R, Baird D, Herring AH, Olshan AF, Jonsson Funk ML,
general population. J Clin Endocrinol Metab 2005; 90: 4004–4010. Hartmann KE. Association between first-trimester vaginal bleed-
112. Retnakaran R. Glucose tolerance status in pregnancy: A window ing and miscarriage. Obstet Gynecol 2009; 114: 860–867.
to the future risk of diabetes and cardiovascular disease in young 134. Kharazmi E, Dossus L, Rohrmann S, Kaaks R. Pregnancy loss and
women. Curr Diabetes Rev 2009; 5: 239–244. risk of cardiovascular disease: A prospective population-based
113. Anastasiou E, Lekakis JP, Alevizaki M, Papamichael CM, Megas cohort study (EPIC-Heidelberg). Heart 2011; 97: 49–54.
J, Souvatzoglou A, Stamatelopoulos SF. Impaired endothelium- 135. Lykke JA, Dideriksen K, Lidegaard O, Langhoff-Roos J. First-
dependent vasodilatation in women with previous gestational trimester vaginal bleeding and complications later in pregnancy.
diabetes. Diabetes Care 1998; 21: 2111–2115. Obstet Gynecol 2010; 115: 935–944.
114. Hannemann MM, Liddell WG, Shore AC, Clark PM, Tooke JE. 136. van Oppenraaij RH, Jauniaux E, Christiansen OB, Horcajadas JA,
Vascular function in women with previous gestational diabetes Farquharson RG, Exalto N. Predicting adverse obstetric outcome
mellitus. J Vasc Res 2002; 39: 311–319. after early pregnancy events and complications: A review. Hum
115. Gasic S, Winzer C, Bayerle-Eder M, Roden A, Pacini G, Kautzky- Reprod Update 2009; 15: 409–421.
Willer A. Impaired cardiac autonomic function in women with prior 137. Lykke JA, Langhoff-Roos J. First trimester bleeding and mater-
gestational diabetes mellitus. Eur J Clin Invest 2007; 37: 42–47. nal cardiovascular morbidity. Eur J Obstet Gynecol Reprod Biol
116. Heitritter SM, Solomon CG, Mitchell GF, Skali-Ounis N, Seely 2012; 164: 138–141.
EW. Subclinical inflammation and vascular dysfunction in women 138. Johns J, Hyett J, Jauniaux E. Obstetric outcome after threatened
with previous gestational diabetes mellitus. J Clin Endocrinol miscarriage with and without a hematoma on ultrasound. Obstet
Metab 2005; 90: 3983–3988. Gynecol 2003; 102: 483–487.
117. Volpe L, Di Cianni G, Lencioni C, Cuccuru I, Benzi L, Del Prato 139. Getahun D, Fassett MJ, Jacobsen SJ. Gestational diabetes: Risk
S. Gestational diabetes, inflammation, and late vascular disease. of recurrence in subsequent pregnancies. Am J Obstet Gynecol
J Endocrinol Invest 2007; 30: 873–879. 2010; 203: 13.
118. Banerjee M, Cruickshank J. Pregnancy as the prodrome to vascu- 140. Philipson EH, Super DM. Gestational diabetes mellitus: Does it
lar dysfunction and cardiovascular risk. Nat Clin Pract Cardiovasc recur in subsequent pregnancy? Am J Obstet Gynecol 1989; 160:
Med 2006; 3: 596–603. 1324–1329.
119. Sarwar N, Gao P, Seshasai SR, Gobin R, Kaptoge S, Di 141. Moses RG. The recurrence rate of gestational diabetes in subse-
Angelantonio E. Diabetes mellitus, fasting blood glucose concen- quent pregnancies. Diabetes Care 1996; 19: 1348–1350.
tration, and risk of vascular disease: A collaborative meta-analysis 142. MacNeill S, Dodds L, Hamilton DC, Armson BA, Vanden Hof
of 102 prospective studies. Lancet 2010; 375: 2215–2222. M. Rates and risk factors for recurrence of gestational diabetes.
120. Retnakaran R, Shah BR. Mild glucose intolerance in pregnancy Diabetes Care 2001; 24: 659–662.
and risk of cardiovascular disease: A population-based cohort 143. Retnakaran R, Austin P, Shah BR. Effect of subsequent pregnan-
study. Can Med Assoc J 2009; 181: 371–376. cies on the risk of developing diabetes following a first pregnancy
121. Carpenter MW. Gestational diabetes, pregnancy hypertension, complicated by gestational diabetes: A population-based study.
and late vascular disease. Diabetes Care 2007; 30: S246–S250. Diabet Med 2011; 28: 287–292.
122. Gunderson EP, Chiang V, Pletcher MJ, Jacobs DR, Quesenberry 144. Skjaerven R, Wilcox A, Klungsoyr K, Irgens LM, Vikse BE, Vatten
CP, Sidney S, Lewis CE. History of gestational diabetes mellitus LJ, Lie RT. Cardiovascular mortality after pre-eclampsia in one
and future risk of atherosclerosis in mid-life: The Coronary Artery child mothers: Prospective, population based cohort study. Br
Risk Development in Young Adults Study. J Am Heart Assoc Med J 2012; 345: e7677.
2014; 3: 000490. 145. Palmer JR, Boggs D, Wise LA, Ambrosone CB, Adams-Campbell
123. Retnakaran R, Shah BR. Abnormal screening glucose challenge LL, Rosenberg L. Parity and lactation in relation to estrogen
test in pregnancy and future risk of diabetes in young women. receptor negative breast cancer in African American women.
Diabet Med 2009; 26: 474–477. Cancer Epidemiol Biomarkers Prev 2011; 20: 1883–1891.
124. Retnakaran R, Qi Y, Sermer M, Connelly PW, Hanley AJ, 146. Bao PP, Shu XO, Gao YT et al. Association of hormone-related
Zinman B. An abnormal screening glucose challenge test in characteristics and breast cancer risk by estrogen receptor/pro-
pregnancy predicts postpartum metabolic dysfunction, even gesterone receptor status in the shanghai breast cancer study. Am
when the antepartum oral glucose tolerance test is normal. Clin J Epidemiol 2011; 174: 661–671.
Endocrinol 2009; 71: 208–214. 147. Work ME, John EM, Andrulis IL et al. Reproductive risk factors
125. Retnakaran R, Qi Y, Sermer M, Connelly PW, Zinman B, Hanley and oestrogen/progesterone receptor-negative breast cancer
AJ. Isolated hyperglycemia at 1 hour on oral glucose tolerance in the Breast Cancer Family Registry. Br J Cancer 2014; 110:
test in pregnancy resembles gestational diabetes mellitus in pre- 1367–1377.
dicting postpartum metabolic dysfunction. Diabetes Care 2008; 148. Lipworth L, Bailey L, Trichopoulos D. History of breast-feeding
31: 1275–1281. in relation to breast cancer risk: A review of the epidemiologic
126. The American College of Obstetricians and Gynecologysts, literature. J Natl Cancer Inst 2000; 92: 302–312.
Committee on Practice Bulletins—Obstetrics. Practice Bulletin 149. Wohlfahrt J, Melbye M. Maternal risk of breast cancer and birth
No. 137: Gestational diabetes mellitus. Obstet Gynecol 2013; characteristics of offspring by time since birth. Epidemiology
122: 406–416. 1999; 10: 441–444.
127. American Diabetes Association. Standards of medical care in 150. Cohn BA, Cirillo P, Christianson RE, van den Berg BJ, Siiteri PK.
diabetes—2014. Diabetes Care 2014; 37: S14–S80. Placental characteristics and reduced risk of maternal breast can-
128. Metzger BE, Buchanan T, Coustan DR et al. Summary and rec- cer. J Natl Cancer Inst 2001; 93: 1133–1140.
ommendations of the Fifth International Workshop-Conference 151. Cnattingius S, Torrång A, Ekbom A, Granath F, Petersson G,
on Gestational Diabetes Mellitus. Diabetes Care 2007; 30: Lambe M. Pregnancy characteristics and maternal risk of breast
S251–S260. cancer. J Am Med Assoc 2005; 294: 2474–2480.
129. Ratner RE, Christophi CA, Metzger BE, Dabelea D, Bennett 152. Hsieh CC, Wuu J, Lambe M, Trichopoulos D, Adami Ho, Ekbom
PH, Pi-Sunyer X, Fowler S, Kahn SE. Prevention of diabetes in A. Delivery of premature newborns and maternal breast-cancer
women with a history of gestational diabetes: Effects of metfor- risk. Lancet 1999; 353: 1239.
min and lifestyle interventions. J Clin Endocrinol Metab 2008; 93: 153. Innes KE, Byers T. First pregnancy characteristics and subsequent
4774–4779. breast cancer risk among young women. Int J Cancer 2004; 112:
130. Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin 306–311.
JM, Walker EA, Nathan DM. Reduction in the incidence of type 154. Vatten LJ, Romundstad P, Trichopoulos D, Skjaerven R. Pre-
2 diabetes with lifestyle intervention or metformin. N Engl J Med eclampsia in pregnancy and subsequent risk for breast cancer.
2002; 346: 393–403. Br J Cancer 2002; 87: 971–973.
131. Beharier O, Shoham-Vardi I, Pariente G et al. Gestational diabe- 155. Melbye M, Wohlfahrt J, Andersen AM, Westergaard T, Andersen
tes mellitus is a significant risk factor for long-term maternal renal PK. Preterm delivery and risk of breast cancer. Br J Cancer 1999;
disease. J Clin Endocr Metab 2015; 100: 1412–1416. 80: 609–613.
156. Mucci LA, Dickman P, Lambe M, Trichopoulos D, Riman T, 170. Cavalieri EL, Kumar S, Todorovic R, Higginbotham S, Badawi
Hsieh CC, Cnattingius S. Gestational age and fetal growth in rela- AF, Rogan EG. Imbalance of estrogen homeostasis in kidney and
tion to maternal ovarian cancer risk in a Swedish cohort. Cancer liver of hamsters treated with estradiol: Implications for estrogen-
Epidemiol Biomarkers Prev 2007; 16: 1828–1832. induced initiation of renal tumors. Chem Res Toxicol 2001; 14:
157. Polednak AP, Janerich D. Characteristics of first pregnancy in 1041–1050.
relation to early breast cancer. A case-control study. J Reprod 171. Ramin SM, Vidaeff AC, Yeomans ER, Gilstrap LC 3rd. Chronic
Med 1983; 28: 314–318. renal disease in pregnancy. Obstet Gynecol 2006; 108:
158. Troisi R, Weiss H, Hoover RN et al. Pregnancy characteristics and 1531–1539.
maternal risk of breast cancer. Epidemiology 1998; 9: 641–647. 172. Karami S, Daugherty S, Schonfeld SJ, Park Y, Hollenbeck AR,
159. Thadhani R, Solomon CG. Preeclampsia—A glimpse into the Grubb RL 3rd, Hofmann JN, Chow WH, Purdue MP. Reproductive
future? N Engl J Med 2008; 359: 858–860. factors and kidney cancer risk in 2 US cohort studies, 1993–2010.
160. Arnadottir GA, Geirsson R, Arngrimsson R, Jonsdottir LS, Am J Epidemiol 2013; 177: 1368–1377.
Olafsson O. Cardiovascular death in women who had hyperten- 173. Kyrklund-Blomberg NB, Cnattingius S. Preterm birth and mater-
sion in pregnancy: A case-control study. Br J Obstet Gynaecol nal smoking: Risks related to gestational age and onset of deliv-
2005; 112: 286–292. ery. Am J Obstet Gynecol 1998; 179: 1051–1055.
161. Mogren I, Stenlund H, Hogberg U. Long-term impact of repro- 174. Rasmussen S, Irgens LM. The effects of smoking and hyperten-
ductive factors on the risk of cervical, endometrial, ovarian and sive disorders on fetal growth. BMC Pregnancy Childbirth 2006;
breast cancer. Acta Oncol 2001; 40: 849–854. 6: 16.
162. Aagaard-Tillery KM, Stoddard G, Holmgren C, Lacoursiere DY, 175. Rodie VA, Freeman D, Sattar N, Greer IA. Pre-eclampsia and
Fraser A, Mineau GP, Varner MW. Preeclampsia and subsequent cardiovascular disease: Metabolic syndrome of pregnancy?
risk of cancer in Utah. Am J Obstet Gynecol 2006; 195: 691–699. Atherosclerosis 2004; 175: 189–202.
163. Lykke JA, Langhoff-Roos J, Lockwood CJ, Triche EW, Paidas MJ. 176. Ray JG, Diamond P, Singh G, Bell CM. Brief overview of mater-
Mortality of mothers from cardiovascular and non-cardiovascular nal triglycerides as a risk factor for pre-eclampsia. Br J Obstet
causes following pregnancy complications in first delivery. Gynaecol 2006; 113: 379–386.
Paediatr Perinat Epidemiol 2010; 24: 323–330. 177. Catov JM, Ness R, Kip KE, Olsen J. Risk of early or severe pre-
164. Paltiel O, Friedlander Y, Tiram E, Barchana M, Xue X, Harlap S. eclampsia related to pre-existing conditions. Int J Epidemiol 2007;
Cancer after pre-eclampsia: Follow up of the Jerusalem perinatal 36: 412–419.
study cohort. Br Med J 2004; 328: 5. 178. Lykke JA, Langhoff-Roos J, Young B, Paidas MJ. Population-based
165. Calderon-Margalit R, Friedlander Y, Yanetz R, Deutsch L, Perrin investigations to study the association of cardiovascular polymor-
MC, Kleinhaus K, Tiram E, Harlap S, Paltiel O. Preeclampsia and phisms and adverse pregnancy outcome. Semin Perinatol 2007;
subsequent risk of cancer: Update from the Jerusalem Perinatal 31: 219–222.
Study. Am J Obstet Gynecol 2009; 200: 25. 179. Williams D. Pregnancy: A stress test for life. Curr Opin Obstet
166. Lawlor DA, Smith GD, Ebrahim S. Hyperinsulinaemia and Gynecol 2003; 15: 465–471.
increased risk of breast cancer: Findings from the British 180. Mosca L, Benjamin E, Berra K et al. Effectiveness-based guide-
Women’s Heart and Health Study. Cancer Causes Control 2004; lines for the prevention of cardiovascular disease in women—
15: 267–275. 2011 update: A guideline from the American heart association.
167. Wolf I, Sadetzki S, Catane R, Karasik A, Kaufman B. Diabetes Circulation 2011; 123: 1243–1262.
mellitus and breast cancer. Lancet Oncol 2005; 6: 103–111. 181. Regitz-Zagrosek V, Blomstrom Lundqvist C, Borghi C et al. ESC
168. Perrin MC, Terry MB, Kleinhaus K et al. Gestational diabetes as Guidelines on the management of cardiovascular diseases during
a risk factor for pancreatic cancer: A prospective cohort study. pregnancy: The Task Force on the Management of Cardiovascular
BMC Med 2007; 5: 25. Diseases during Pregnancy of the European Society of Cardiology
169. Rollison DE, Giuliano A, Sellers TA, Laronga C, Sweeney C, (ESC). Eur Heart J 2011; 32: 3147–3197.
Risendal B, Baumgartner KB, Byers T, Slattery ML. Population-
based case-control study of diabetes and breast cancer risk in
Hispanic and non-Hispanic white women living in US southwestern
states. Am J Epidemiol 2008; 167: 447–456.
48 Diabetes, pregnancy, and the
developmental origins of health
and disease
Gerard H.A. Visser and Mark A. Hanson
disease such as diabetes in the offspring. However, it is now

Introduction recognized that diabetes can also be passed from one gen-
Multiple fields of inquiry support the “developmental eration to a subsequent generation, or even to more than
origins of health and disease” (DOHaD) concept, which one generation.4 Thus, for example, populations in which
suggests that the risk of developing some chronic, non- the prevalence of diabetes has increased following economic
communicable diseases (NCDs) in adulthood is influenced improvements leading to mismatch now suffer greater
not only by genetic and adult lifestyle factors but also by incidence in subsequent generations as “diabetes begets
environmental factors acting in early life.1 Thus, the risk of diabetes.”
disease as an adult depends on a range of influences acting In this chapter, the impact of pregnancy in women with
across the life course, starting with preconceptional attri- preexisting or gestational diabetes mellitus (GDM) on the
butes of the parents, such as nutrition and health behav- long-term outcome of their offspring is discussed, in the
iors, the intrauterine environment, birth phenotype, and context of the altered intrauterine environment, with mater-
aspects of the postnatal environment. These factors do not nal obesity as an increasing comorbidity. Prevention of
initiate disease in the developing embryo, fetus, or infant excessive fetal growth is likely to be an important preventive
but have substantial influence on the ways in which the measure to prevent long-term morbidity in these offspring.
individual will respond to the challenges of their later life- Early diagnosis of those at risk will be crucial if interven-
style. Research in a range of species has shown that signals tions are to be instituted in good time.
such as maternal nutritional state, stress levels, or body
composition “prime” her offspring to respond to challenges
such as living in an obesogenic environment.2 Thus, the Diabetes in pregnancy and long-term
emerging epidemic of obesity with its resulting effects on outcome on offspring
risk of adult diabetes and metabolic abnormalities, cardio-
vascular disease, and some forms of cancer is not only due Animal studies
to the adoption of a Western lifestyle but also in part due First data on the effects of altered intrauterine glucose
to the intrauterine environment and other environmental homeostasis on the next generation came from studies in
factors early in life. rats. Mild diabetes during pregnancy induced by strepto-
The rise in the prevalence of NCDs with changes in zotocin resulted in impaired glucose tolerance and GDM
lifestyle in high-income countries is well documented. Of in offspring; the effects on glucose tolerance extended to
perhaps greater concern is the faster rise in low- to middle- the third generation. 5 A large number of studies followed
income countries, especially those undergoing economic in many species, showing that such effects on glucose/
transition, with higher numbers of the population living in insulin homeostasis could also be produced by unbal-
urban settings, the availability of fast foods of poor nutri- anced maternal nutrition (over- or undernutrition) by
tional quality, and adoption of more sedentary lifestyle. administration of glucocorticoids or exposure to endo-
In addition, global improvements in public health have crine disruptor chemicals in pregnancy.6–8 The processes
reduced mortality in early life and increased longevity, lead- by which these effects occur have been shown to involve
ing to greater incidence of chronic disease. The concept of developmental plasticity, in particular epigenetic effects
“mismatch”3 explains how the transition from a poor to a that permit multiple phenotypes to be induced from a
rich lifestyle within a generation leads to a greater risk of genotype, dependent on the developmental environment.9
403
404 Diabetes, pregnancy, and the developmental origins of health and disease
Most recently, such effects on the offspring of several gen- 16-year-old adolescents was related to maternal obesity,
erations have also been shown to operate via the paternal and especially to the combination of maternal obesity and
line.10,11 GDM, but not to GDM alone.16 A meta-analysis has sug-
A set of identified, independent risk factors impacting gested that there is an association between all types of dia-
outcome in offspring of women with diabetes is shown in betes in pregnancy and childhood obesity; however, only
the following: three studies in the meta-analysis corrected for maternal
BMI, and in all three of those studies, the direct relation-
●● Genetic predisposition ship between maternal diabetes and childhood obesity was
●● (Gestational) diabetes lost following correction for BMI.17 Similarly, a study of
●● Maternal body mass index (BMI) metabolic syndrome in infants at the age of 7–11 years has
●● Weight gain in pregnancy shown that maternal obesity or being large for gestational
●● Macrosomia at birth age at birth were independently associated with metabolic
●● Cesarean delivery (CD) syndrome but not maternal GDM.18 In a recent review by
●● Excessive weight gain >2 years of age Donovan and Cundy,19 it was concluded that exposure to
●● Socioeconomic circumstances hyperglycemia in utero has minimal direct effects on the
later risk of obesity and type 2 diabetes. The increased risk
Maternal diabetes and diabetes in offspring of obesity in the offspring of women with type 2 diabetes
or GDM can be explained by confounding factors, such
The impact of hyperglycemia during fetal development
as parental obesity. In women with type 1 diabetes, it has
on the offspring has been studied in Pima Indians in the
been shown that childhood obesity at the age of 7 and
United States, a population poorly adapted for Westernized
increased insulin resistance were independently related to
diets and sedentary lifestyles common today, and thus with
maternal BMI and being large for gestational age at deliv-
a high propensity toward the development of diabetes. In
ery. 20 Thus, there is doubt as to the independent impact of
Pima Indian women, non-insulin-dependent diabetes mel-
maternal preexisting diabetes or GDM on obesity in the
litus (NIDDM) during pregnancy results in offspring who,
offspring, although data corrected for maternal BMI are
at ages 20–24, themselves have a much higher prevalence of
limited. The effect of maternal diabetes on outcome may
NIDDM (45%) than do the offspring of nondiabetic Pima
be related to an increased and proportionally abnormal
Indian women (1.4%) or the offspring of Pima Indian moth-
fetal growth resulting in macrosomia at birth. The impact
ers who developed diabetes after the pregnancy (8.6%).12
of maternal obesity, irrespective of the presence of diabe-
These differences remained statistically significant after tak-
tes or otherwise, has also been shown in a large popula-
ing into account paternal diabetes, age at onset of diabetes
tion study in Scotland, in which an excess of premature
in the parents, and the offspring’s weight relative to height.
mortality from cardiovascular events was found in adult
Thus, in this diabetes-prone population, there is a dramatic
offspring. 21 In this study, data were adjusted for maternal
increase in the development of diabetes among those off-
age at delivery, socioeconomic status, birth weight, and
spring who gestated in an abnormal intrauterine environ-
gestational age at delivery. Maternal BMI was not taken
ment defined by diabetes-driven hyperglycemia. In contrast,
into account.
in Europe, women with either GDM (compared with preg-
nant women at genetic high risk for, but without, GDM) or
type 1 diabetes (compared to the general population) tended Weight gain during pregnancy
to have offspring that are approximately 8% more likely to An increased weight gain during pregnancy has been shown
develop T2D or impaired glucose tolerance by the age of 2713; to be related to fetal macrosomia at birth in women with
in this European population less prone to develop diabetes, type 1 diabetes.22 High maternal weight gain is also associ-
the effects of intrauterine hyperglycemia are less pronounced ated with increased body weight in a 12-year-old offspring,
but still detectable. It, therefore, appears that both genetic and this effect appeared only partly mediated through a
and epigenetic predispositions may differ between popula- higher birth weight.23 Similarly, associations have been
tions and play a role in determining risk of adult diabetes. found between gestational weight gain with offspring BMI
Unfortunately, maternal BMI was not taken into account in and blood pressure at 21 years of age.24
these studies.
Fetal macrosomia
Maternal BMI and obesity in offspring Being large for gestational age at birth is a risk factor for
GDM in the mother is associated with a range of effects childhood obesity both in offspring of women with diabe-
on the offspring.14,15 Fetuses of women with GDM develop tes16,20 and in nondiabetic populations.16 This may be due
insulin resistance in utero, but mainly in the case of to epigenetic changes induced by an abnormal intrauter-
maternal obesity. In fact, data suggest that obesity is the ine environment, by genetic differences (in relation to a
more important factor impacting long-term weight out- high maternal BMI) or altered nutrition and activity pat-
comes for offspring, not GDM. A recent Finnish study tern during childhood. The latter may be due to impaired
demonstrated that overweight and abdominal obesity in socioeconomic circumstances and related unhealthy
Discussion 405
lifestyle. Animal data have shown that “cafeteria style” potential value in ensuring that children over the age of 2 do
diet during the newborn period will result in obesity not grow too quickly or gain weight disproportionately to
and glucose intolerance during pregnancy and in glucose their gains in height, which may induce epigenetic changes
intolerance in the following generation. 25 There is evidence that impair glucose tolerance and set the stage for diabetes.
that the highest incidence of childhood obesity occurs in Similarly, accelerated growth has been found in infants of
offspring of women with type 2 diabetes in whom abnor- women with type 1 and type 2 diabetes.20,26 Concentrated
mal intrauterine environment, increased maternal BMI, public health efforts focused on adequate diet and sufficient
and suboptimal socioeconomic circumstance may come exercise early in childhood and more study on interventions
together. 26 In offspring of women with type 1 diabetes, such as folic acid that may potentially minimize epigenetic
type 2 diabetes, and GDM, childhood obesity is most insults during pregnancy37 are both high priorities to help
common in infants being large for dates at birth, with stem the ongoing epidemic of diabetes across the globe, an
the highest incidence of obesity in large-for-dates infants epidemic that threatens to become a worldwide pandemic in
of obese women with type 2 diabetes (Hammoud et al., the years ahead, with particular burden placed on the devel-
unpublished data). oping world.38
Cesarean delivery
Women with diabetes deliver generally two to three times as
much by cesarean section than control women. CDs result
Discussion
in a 20%–30% higher incidence of autoimmune disease and The data presented in this chapter indicate that many fac-
obesity in the offspring. This has been shown most convinc- tors may play a role in the outcome of offspring of women
ingly in meta-analyses of observational studies regarding with diabetes. The strongest relationship with diabetes and/
childhood diabetes and asthma.27,28 Although these data did or obesity in offspring concerns most likely the preexisting
not derive from randomized controlled trials, subanalysis maternal BMI and fetal macrosomia at birth. The latter is
did not reveal effects of low-birth-weight infants, breast- affected by maternal diabetes and concomitant abnormal
feeding, and passive smoking. Moreover, funnel plots did intrauterine environment and maternal weight gain during
not show evidence of publication bias.27 The effects of CDs pregnancy. There is evidence that macrosomia in fetuses of
on the immune response may be due to the absence of the women with type 1 and type 2 diabetes is increasing in the
stress of the normal labor process, resulting in maladaptive Western world, despite improvements in glucose control
immune activation. Other possibilities include altered epi- during pregnancy.39–42 This unexpected finding is likely to
genetic regulation of gene expression and perturbed bacterial be due to an increase in maternal BMI, reduction in mater-
colonization of the gastrointestinal tract.29 After a CD, there nal vascular complications, better glucose control in early
is a delayed and altered bacterial colonization with reduced pregnancy resulting in a better placentation, increase in
variation.30,31 This may result in delayed developmental bal- weight gain during pregnancy, and poorer control (espe-
ance between TH-1 and TH-2 immune responses.32 A meta- cially postprandially) near term since women are not admit-
analysis of 15 studies with a combined population of over ted routinely to the hospital anymore.22,43,44 Reduction of
160,000 individuals has shown a 22% increase in obesity in macrosomia at birth seems difficult, with reducing of
children, adolescents, and adults after having been born by weight gain during pregnancy as one of the few modifiable
cesarean section.33 The microflora in the gut may stimulate options. Prevention of excessive growth during childhood
fat deposition and promote obesity through several mecha- might be a secondary tool to prevent negative long-term
nisms, by improving energy yield from food; by regulating effects of fetal overgrowth. Achievement of normoglycemia
gut permeability, low-grade inflammation, and immune is the most important factor but appears difficult at present.
balance; or by modulating metabolism and/or genes directly Continuous glucose measurements have shown that glucose
in the liver.34 All in all, it can be concluded that microflora values of women with type 1 diabetes are on average almost
during early life play a significant role in health and dis- 2 mmol/L higher than in control pregnancies.45 A first ran-
ease.35 In women with diabetes, a CD may have an effect domized trial on the use of a real-time continuous glucose
on the incidence of type 1 diabetes and/or obesity in their sensor, five times during pregnancy for 4–6 days, failed to
offspring. show an improved outcome in women with type-1/2 dia-
betes, maybe because this device facilitates enhanced food
Accelerated growth in offspring >2 years of age intake that may promote fetal growth.46 However, a recent
Finally, persuasive evidence of the impact of early childhood study from China has shown that continuous glucose mea-
growth on the development of T2D has been presented by surements and intensive dietary counseling in women with
Eriksson et al. in their longitudinal archival study of 8760 GDM were associated with better glycemic control and a
subjects born in Helsinki from 1934 to 1944.36 Subjects significantly lower incidence of large-for-dates newborn
were divided into those weighing ≤3.5 kg at birth, and those infants.47 Intensified glucose monitoring in combination
weighing >3.5 kg at birth. In both groups of subject, rapid with regular dietary counseling may therefore be able to
gain in BMI after the age of 2 years was associated with an reduce macrosomia at birth and hopefully reduce long-term
increased risk of T2D in later life. These data suggest the impairments.
406 Diabetes, pregnancy, and the developmental origins of health and disease
REFERENCES
1. Gluckman PD, Hanson MA. Developmental origins of disease 23. Ludwig DS, Rouse HL, Currie J. Pregnancy weight gain and
paradigm: A mechanistic and evolutionary perspective. Pediatr childhood body weight: A within-family comparison. PLoS
Res 2004; 56: 311–317. Med 2010; 10: e1001521.
2. Hanson MA, Gluckman PD. Developmental processes and 24. Mamun AA, O’Callaghan M, Callaway L, Williams G, Najman
the induction of cardiovascular function: Conceptual aspects. J, Lawlor DA. Associations of gestational weight gain with
Physiol Rev 2005; 565: 27–34. offspring body mass index and blood pressure at 21 years of
3. Gluckman PD, Hanson MA. Mismatch, Oxford University age: Evidence from a birth cohort study. Circulation 2009; 119:
Press: Oxford, U.K., 2006, pp. 1–285. 1720–1727.
4. Ma RCW, Chan JCN, Tam WH, Hanson MA, Gluckman PD. 25. Holemans K, Caluwaerts S, Poston L, Van Assche FA. Diet-
Gestational diabetes, maternal obesity and the NCD burden. induced obesity in the rat: A model for gestational diabetes
Clin Obstet Gynecol 2013; 56: 633–641. mellitus. Am J Obstet Gynecol 2004; 190: 858–865.
5. Aerts L, Van Assche FA. Is gestational diabetes an acquired 26. De Valk HW, van Nieuwaal NH, Visser GHA. Pregnancy out-
condition? J Dev Physiol 1979; 1: 219–225. come in type 2 diabetes mellitus: A retrospective analysis from
6. Ozanne SE. Metabolic programming in animal type 2 diabetes. the Netherlands. Rev Diabet Stud 2006; 3: 134–142.
Br Med Bull 2001; 60: 143–152. 27. Thavagnamam S, Fleming J, Bromley A, Shields MD, Cardwell
7. Bertram CE, Hanson MA. Animal models and programming of CR. A meta-analysis of the association between cesarean
the metabolic syndrome Type 2 diabetes. Br Med Bull 2001; section and childhood asthma. Clin Exp Allergy 2008; 38:
60: 103–121. 629–633.
8. Alonso-Magdalena P, Quesada I, Nadal A. Endocrine disruptors 28. Cardwell CR, Stene LC, Joner G et al. Cesarean section is
in the etiology of type 2 diabetes mellitus. Nat Rev Endocrinol associated with an increased risk of childhood-onset type 1
2011; 7: 346–353. diabetes mellitus: A meta-analysis of observational studies.
9. Godfrey KM, Lillycrop KA, Burdge GC, Gluckman PD, Hanson Diabetologia 2008; 51: 726–735.
MA. Epigenetic mechanisms and the mismatch concept of the 29. Cho CE, Norman M. Cesarean section and development of the
developmental origins of health and disease. Pediatr Res 2007; immune system in offspring. Am J Obstet Gynecol 2013; 208:
61: 5R–10R. 249–254.
10. Hughes V. Epigenetics: The sins of the father. Nature 2014; 507: 30. Gronlund MM, Lehtonen OP, Eerola P, Kero P. Fecal micro-
22–24. flora in healthy infants born by different methods of delivery:
11. Radford EJ, Ito M, Shi H et al. In utero undernourishment per- Permanent changes in intestinal flora after cesarean delivery.
turbs the adult sperm methylome and intergenerational metab- J Pediatr Gastroenterol Nutr 1999; 28: 19–25.
olism. Science 2014; 345(6198): 1255903. 31. Azad MB, Konya T, Maughan H et al. Gut microbiota of healthy
12. Pettit DJ, Aleck KA, Baird HR, Carraher MJ, Bennett PH, Canadian infants: Profiles by mode of delivery and infant diet at
Knowler WC. Congenital susceptibility to NIDDM. Role of 4 months. Can Med Assoc J 2013; 185(5): 385–394.
intrauterine environment. Diabetes 1988; 37: 622–628. 32. Strachan DP. Hay fever, hygiene and household size. Br Med J
13. Clausen TD, Mathiesen ER, Hansen T et al. High prevalence of 1989; 299: 1259–1260.
type 2 diabetes and pre-diabetes in adult offspring of women 33. Darmasseelane K, Hyde MJ, Santhakumaran SM, Gale C, Modi
with gestational diabetes mellitus or type 1 diabetes the role of N. Mode of delivery and offspring body mass index, over-
intrauterine hyperglycemia. Diabetes Care 2008; 31: 340–346. weight and obesity in adult life: A systematic review and meta-
14. Catalano PM, Ehrenberg HM. The short- and long-term impli- analysis. PLOS ONE 2014; 9: e87896.
cations of maternal obesity on the mother and her offspring. Br 34. Friedman JE. Thirteenth Annual Meeting of the Diabetes in
J Obstet Gynaecol 2006; 113: 1126–1113. Pregnancy Study Group of North America, Washington, DC,
15. Catalano PM, Presley L, Minium J, Hauguel-de Mouzon S. October 31–November 2, 2013.
Fetuses of obese mothers develop insulin resistance in utero. 35. Flint HJ. The role of the gut microbiota in nutrition and health.
Diabetes Care 2009; 32: 1076–1080. Nat Rev Gastroenterol Hepatol 2012; 9: 577–589.
16. Pirkola J, Pouta A, Bloigu A et al. Risks of overweight and 36. Eriksson JG, Forsen TJ, Osmond C, Barker DJ. Pathways of
abdominal obesity at age 16 years associated with prenatal infant and childhood growth that lead to type 2 diabetes.
exposures to maternal prepregnancy overweight and gesta- Diabetes Care 2003; 26: 3006–3010.
tional diabetes mellitus. Diabetes Care 2010; 33: 1115–1121. 37. Lillycrop KA, Phillips ES, Jackson AA, Hanson MA, Burdge GC.
17. Philipps LH, Santhakumaran S, Gale C et al. The diabetic preg- Dietary protein restriction of pregnant rats induces and folic acid
nancy and offspring BMI in childhood: A systematic review and supplementation prevents epigenetic modification of hepatic
meta-analysis. Diabetologia 2011; 54: 1957–1966. gene expression in the offspring. J Nutr 2005; 135: 1382–1386.
18. Boney CM, Verma A, Tucker R, Vohr BR. Metabolic syndrome 38. Hossain P, Kawar B, El Nahas M. Obesity and diabetes in the
in childhood: Association with birth weight, maternal obe- developing world—A growing challenge. N Engl J Med 2007;
sity, and gestational diabetes mellitus. Pediatrics 2005; 115: 356: 213–215.
e290–e296. 39. Evers IM, de Valk HW, Visser GHA. Risk of complications of
19. Donovan LE, Cundy T. Does exposure to hyperglycaemia in pregnancy in women with type 1 diabetes: Nationwide prospec-
utero increase the risk of obesity and diabetes in the offspring? tive study in the Netherlands. Br Med J 2004; 328(7445): 915.
A critical reappraisal. Diabet Med 2015; 32: 295–304. 40. Jensen DM, Damm P, Moelsted-Pedersen L et al. Outcomes in
20. Rijpert M, Evers IM, de Vroede MA, de Valk HW, Heijnen CJ, type 1 diabetic pregnancies: A nationwide, population-based
Visser GHA. Risk factors for childhood overweight in offspring study. Diabetes Care 2004; 27: 2819–2823.
of type 1 diabetic women with adequate glycemic control dur- 41. Penney GC, Mair G, Pearson DW. Outcomes of pregnan-
ing pregnancy nationwide follow-up study in the Netherlands. cies in women with type 1 diabetes in Scotland: A national
Diabetes Care 2009; 32: 2099–2104. population-based study. Br J Obstet Gynaecol 2003; 110(3):
21. Reynolds RM, Allan KM, Raja EA et al. Maternal obesity during 315–318.
pregnancy and premature mortality from cardiovascular event 42. Persson M, Norman M, Hanson U. Obstetric and perinatal
in adult offspring: Follow-up of 1,323,275 person years. Br Med outcomes in type 1 diabetic pregnancies: A large, population-
J 2013; 34: f4539. based study. Diabetes Care 2009; 32: 2005–2009.
22. Seifres CM, Feghali MN, Althouse AD, Caritis SN, Catov JM. 43. Kuc S, Wortelboer EJ, Koster MP, de Valk HW, Schielen PC,
Effect of excess gestational weight gain on pregnancy out- Visser GHA. Prediction of macrosomia at birth in type-1 and
comes in women with type-1 diabetes. Obstet Gynecol 2014; 2 diabetic pregnancies with biomarkers of early placentation.
123: 1295–1302. BJOG 2011; 118: 748–754.
References 407
44. Glinianaia SV, Tennant PW, Bilous RW, Rankin J, Bell R. 46. Secher A, Ringholm L, Andersen HU, Damm P, Mathiesen ER.
HbA(1c) and birthweight in women with pre-conception type The effect of real-time glucose monitoring in pregnant women
1 and type 2 diabetes: A population-based cohort study. with diabetes: A randomized controlled-trial. Diabetes Care
Diabetologia 2012; 55: 3193–3203. 2013; 36: 1877–1883.
45. Kerssen A, de Valk HW, Visser GHA. Increased second tri- 47. Yu F, Lv L, Liang Z et al. Continuous glucose monitoring effects
mester maternal glucose levels are related to extremely large- on maternal; glycemic control and pregnancy outcomes in
for-gestational-age infants in women with type 1 diabetes. patients with gestational diabetes mellitus: A prospective
Diabetes Care 2007; 30: 1069–1074. cohort study. J Clin Endocrinol Metab 2014; 99: 4674–4682.
49 Interventions to improve
pregnancy outcome in obese
pregnancy: Implications for
mother and child
Rahat Maitland and Lucilla Poston
The majority of adverse complications of obese pregnancies on GWG may be less appropriate for obese women than for
for both mother and child are strongly associated with pre- those with a lower BMI, although achievement of the Institute
pregnancy BMI1; therefore, it can be reasonably assumed that of Medicine (IOM) recommendation of 5–9 kg (for women
preconception is the optimal time to introduce strategies to with a BMI > 30 kg/m2) would inevitably reduce postpartum
encourage weight loss so that women become pregnant with weight retention and thereby decrease the risk of adverse out-
a healthy BMI. However, as a large proportion of pregnan- comes, including delivery of an large for gestational age (LGA)
cies are unplanned, this approach will only be appropriate infant in a subsequent pregnancy.14 Recent meta-analyses have,
for a motivated subgroup, until such time when public health however, identified many of these studies to be poorly designed
strategies to reduce weight in adolescents and young adults and most to be underpowered for clinical outcomes.15–19 The
have met with success. Most intervention strategies therefore results have generally been disappointing with very modest
focus on the pregnant woman, and in recent years, as the reductions in GWG, e.g., 2.2 kg, in one meta-analysis17 and
global epidemic of obesity has spiralled, so have the num- with relatively few women achieving the IOM recommended
ber of studies attempting to reduce the burden of obesity on limits of GWG. Moreover, evaluation of an intervention on
the health of the mother and her child. While interventions the woman’s diet or PA, in order to determine compliance,
to obese subjects often meet with failure in the general pub- has not always been undertaken. Of those that have evaluated
lic, encouragement can be taken from the knowledge that the women’s diet, some have shown the desired effect, with
pregnancy is often a time of high motivation for women, to significant changes including reduced calorie and saturated
address and maximize not only their own health but that fat intake and an increase in polyunsaturated fats or improve-
of their unborn child. Given the increasing evidence that ments in fiber consumption.3,9,20 Less frequently, an increase
a child nurtured in utero by an obese woman may have an in PA has been reported.13,21 Compliance to the intervention
increased risk of obesity and cardiometabolic disease in later is important to assess, as failure to change GWG could arise
life,2 interventions in pregnancy may also have the benefit of either from poor compliance to the intervention or because
impacting favorably upon the health of the next generation. of the lack of effectiveness. Some interventions have been
underpinned with theoretical strategies designed to support
behavioral change and others have not. One meta-analysis to
determine the most effective intervention strategy for limiting
Interventions to limit gestational GWG in women of all BMIs concluded that those that did not
weight gain address theories of behavioral change were less effective than
those that did22 but an updated systematic review suggests that
Intervention strategies designed to improve outcomes for either approach is equally effective.23 These authors concluded
mother and child in overweight and obese pregnant women that motivational interviewing and behavioral self-monitoring
have included dietary advice3–5 or physical activity (PA)6–8 or, are likely to be key factors in achieving limitation of GWG,
more frequently, a combination of the two.9–13 To date, the especially if coupled with dietary interventions.
outcome of choice has predominantly been gestational weight The evidence for any clinical benefit of a modest limi-
gain (GWG). Since there is a strong association between mater- tation of GWG achieved through intervention studies
nal prepregnancy BMI and clinical outcomes, the emphasis in overweight and obese pregnant women is less than
408
Physical activity 409
encouraging, as indicated by several meta-analyses. Dodd The DALI study (ISRCTN70595832), a collaborative
et al. reported no significant effect on birth weight, mac- European randomized controlled trial (RCT) of a vitamin
rosomia, cesarean section rate, gestational diabetes melli- D and lifestyle intervention in different combinations for
tus (GDM), or any other clinical outcome.15 More recently, obese pregnant women, is currently underway and may pro-
Oteng-Ntim et al. reported a significant reduction in GWG vide more insights into the effects of restricting GWG (target
of 2.21 kg in obese and overweight women with an associ- sample size 880) and the safety in terms of clinical outcomes.
ated lower incidence of GDM (odd ratio 0.80, [95% confi- Women in the lifestyle intervention arms (dietary, PA, or a
dence interval [CI] 0.58–1.10]), but the strength of evidence combination of both) are advised to limit GWG to <5 kg.
was considered low.17 In a subgroup of obese women, part Primary outcomes are GWG, insulin sensitivity (homeo-
of a large meta-analysis of interventions on GWG in women static model assessment of insulin resistance [HOMA-IR]),
of all BMIs, Thangaratinam et al. also arrived at the same and fasting blood glucose (FBG) measured following 75 g
conclusion.18,19 One of the few studies in obese pregnant oral glucose tolerance test (OGTT) at 24–28 and 35–37 weeks
women to have shown improvement of relevant biochemi- gestation. Prenatal ultrasound scanning will enable detailed
cal outcomes was a rigorous, intensive dietary intervention intrauterine dimensions to be made, and within 48 of birth,
in 50 obese pregnant women, designed to restrict GWG neonates will have full anthropometric assessment.28
to 6–7 kg and proposing adherence to the official Danish
dietary recommendations (fat 30E%, protein 15E%–20E%,
and carbohydrate 50E%–55E% [E% = % of total energy]). Intervention studies aimed to improve glucose
Compared to other studies, a considerable reduction in metabolism and reduce GDM incidence
GWG was achieved, being reduced by half (6.6 vs. 13.3 kg, Since the increased risk of GDM is a consequence of insu-
[95% CI 2.6–10.8], p = 0.002). This was also associated with lin resistance (IR) and because IR underpins the majority of
a reduction in the fasting plasma insulin concentration adverse outcomes of pregnancy in obese women, approaches
from 15 weeks onward. At 27 weeks gestation, a 20% reduc- that focus on IR or GDM as primary outcomes in high-risk
tion in plasma insulin was observed with a further 23% women may be more profitable than a focus on restriction
reduction at 36 weeks. 3 The intervention, however, had no of GWG. However, despite the strong evidence to support
effect on the incidence of GDM, but the study was likely to lifestyle changes in reducing the risk of type 2 diabetes mel-
have been underpowered for this outcome. litus (T2DM),29–31 results of meta-analyses regarding inter-
These negative data from interventions primarily tar- ventions to reduce GDM in women of all BMIs have been
geting GWG are perhaps unsurprising as evidence sug- contradictory, leaving many unanswered questions.32,33
gests that prepregnancy BMI may contribute as much to
adverse outcomes as excessive GWG.1 Approaches that focus
on limitation of weight gain in obese pregnant women may Physical activity
therefore be misplaced.
PA is an established method of improving insulin sensi-
tivity and overall glycemic control in addition to reducing
Safety considerations of limiting GWG cardiovascular risk by influencing multiple clinical and bio-
Thus far, no adverse effects of limiting GWG within rec- chemical variables associated with adverse events including
ommended limits in obese pregnant women by dietary or glycated hemoglobin (HbA1c), systolic blood pressure, adi-
exercise interventions have been reported.4,11,24–26 However, posity, and triglycerides in subjects with T2DM.34 The acute
caution should be heeded, as observational studies have effects of exercise can increase insulin-mediated uptake by
highlighted deleterious consequences for those who gain the skeletal muscle by up to 40% for 16 hours, with long-term
less than 5 kg (the lowest gain advised by the IOM for BMI effects from sustained activity thought to be mediated via
≥ 30 kg/m2), including adverse effects on fetal growth. The increased glucose transporter type 4 (GLUT) 4 production.35
most comprehensive review to address this issue is a retro- In the absence of any weight loss, improvements in periph-
spective analysis of 1241 overweight and obese women who eral insulin sensitivity and activity of lipoprotein lipase, an
had participated in prospective observational studies.27 Of enzyme involved in the catabolism of triglyceride-rich lipo-
the total sample, 15.2% (n = 188) gained or lost less than 5 kg proteins and strongly associated with IR, have been reported
and for these women, the incidence of small for gestational in sedentary adults following a 6-month walking program.35
age (SGA) and low birth weight was significantly greater The progression to overt diabetes in high-risk individuals
than for those whose GWG exceeded 5 kg. Their infants had including those who are obese with known glucose intoler-
significantly less total (403 ± 175 vs. 471 ± 193 g, p < 0.0001) ance can be limited, but the optimal mode, frequency, inten-
and lean fat mass (2855 ± 321 vs. 2995 ± 347 g, p < 0.0001) sity, and duration of exercise is yet to be identified.29,36
with smaller head circumference (34.2 ± 1.7 vs. 34.5 ± 1.7, In pregnancy, among women of normal BMI, increasing
p = 0.02) and reduced length (49.3 ± 2.3 vs. 50.0 ± 2.8, p = PA also improves glucose tolerance.37 However, advancing
0.001), reflecting reduced skeletal growth. The risk of SGA gestation and physical restriction coupled with maternal
was independent of maternal glucose status, and the exclu- anxiety is often considered a barrier to exercise in pregnancy
sion of all women with GDM from secondary analysis did in women of all BMIs. Most exercise-only interventions cen-
not alter the results. ter around supervised exercise classes two to three times a
410 Interventions to improve pregnancy outcome in obese pregnancy
week, adapted for pregnancy with a qualified trainer, with or The Metformin in Gestational Diabetes (MiG) trial, the
without a home-based element.6,21,38–40 largest RCT of 751 subjects comparing metformin therapy
While one meta-analysis showed strong inverse associa- vs. insulin for women diagnosed with GDM from 20 to
tions (25%–55% reduced risk) between PA before or dur- 33 weeks gestation (mean BMI for metformin and insulin
ing early pregnancy and GDM,32 a more recent report of groups 35.1 ± 8.3 kg/m2 and 34.6 ± 7.2 kg/m2, respectively),
six RCTs of PA in of a total of 1278 subjects33 suggested that found no significant difference in the composite primary out-
there was insufficient evidence to support the use of PA to come of neonatal complications.48 Acceptability and compli-
prevent GDM (risk ratio 0.91 [95% CI 0.57–1.44], p = 0.68). ance with metformin treatment were high, and women in
Following the Hyperglycemia and Adverse Pregnancy this group experienced less GWG and greater postpartum
Outcomes study,41 new guidelines for the diagnosis of GDM weight reduction at 6–8 weeks.
were agreed by the International Association of Diabetes and Metformin has also been used to induce ovulation in
Pregnancy Study Groups (IADPSG) advocating lower glu- women with subfertility and polycystic ovary syndrome, a
cose thresholds than those used previously (fasting glucose condition also characterized by IR, with no adverse preg-
≥5.1 mmol/L, or 1-hour glucose ≥10.0 mmol/L, or 2-hour nancy outcomes reported. Thus, it might appear reasonable
glucose ≥8.5 mmol/L [75 g OGTT]).42 In 2013 these were also to use metformin as an intervention in nondiabetic obese
endorsed by the World Health Organization.43 Using these pregnant women from early pregnancy onward. Two multi-
new diagnostic criteria, a recent RCT of PA and GDM in center RCTs of metformin vs. placebo with no other lifestyle
510 healthy Spanish pregnant women of normal BMI found or dietary restrictions are in progress. The “Metformin in
no reduction in GDM between exercising and nonexercising Obese Nondiabetic Pregnant Women” (MOP: NCT01273584)
groups.26 This concurs with a Norwegian study of lean women and “Efficacy of Metformin in Pregnant Obese Women”
(n = 702) that adopted a less intense intervention.25 However, (EMPOWaR: ISRCTN51279843) studies are recruiting
neither study measured compliance to the intervention, women from 12 to 16 weeks gestation with slightly different
i.e., whether PA changed in the intervention arm of the study. entry BMI values (BMI ≥ 35 kg/m2 for MOP and ≥30 kg/m2
There has been limited success in PA interventions in for EMPOWaR). Exclusion and inclusion criteria are other-
overweight and obese pregnant women. In a feasibility study, wise similar but non-Caucasian women are excluded from
Callaway et al. reported a modest increase in PA among EMPOWaR.
obese pregnant women who received an individualized exer- Theoretically, the metabolic profile of infants exposed to
cise program, but did not consider this would be adequate to metformin could be favorably altered and accompanied by
reduce GDM.21 The FitFor 2 program found no significant changes in adipose tissue distribution. This has the poten-
differences in maternal FBG, fasting insulin, HbA1c, BMI, or tial to reduce the heightened lifetime risk of cardiovascu-
birth weight following a PA intervention (60 minutes aerobic lar morbidity and premature death reported in children of
exercise in sessions twice a week) in women at high risk of obese mothers.2,49 The MiG TOFU (the offspring follow-up
GDM (BMI ≥ 30 or >25 kg/m2 plus one other risk factor).7 study) follow-on study has, however, identified some
Lack of statistical power may have confounded the results; potentially concerning differences in body composition in
however, compliance to the sessions was poor, with only children at 2 years using a combination of body composi-
16.3% of women attending half of the total number of sessions tion measures in a subgroup of 318 children from the MiG
(two perweek), and compliance decreased with progression study.50 Overall body fat and estimates of visceral fat mea-
of pregnancy. Several studies have addressed the barriers to sured by dual energy X-ray absorptiometry and bioimped-
changing PA behavior among obese pregnant women, which ance studies were equivalent but children in the metformin
have included physical discomfort, concerns about safety for arm had greater measures of subcutaneous fat stores at three
the baby, and social and cognitive factors.44,45 The high inci- sites: subscapular (6.3 ± 1.9 vs. 6.0 ± 1.7 mm, p = 0.02) and
dence of depression among obese pregnant women may also biceps skinfold (6.03 ± 1.9 vs. 5.6 ± 1.7 mm, p = 0.04) and
contribute to a lack of enthusiasm.46 mid-upper arm circumference (17.2 ± 1.5 vs. 16.7 ± 1.5 cm,
p = 0.002). The original hypothesis that in utero metformin
exposure would be associated with less visceral fat and thus
Pharmacological interventions less IR in the offspring was not met. Nonetheless, the authors
concluded that increases in subcutaneous fat stores may be
Metformin exerts its glucose-lowering effects by reducing a preferential and healthier response to metformin with an
hepatic gluconeogenesis and improving peripheral insulin ultimate reduction of visceral fat in the long term. Longer-
sensitivity, by mechanisms that are not fully understood. The duration follow-up studies, ideally from large RCTs such as
resultant reduction of blood glucose concentration without MOP and EMPOWaR, are required.
corresponding hypoglycemia or weight gain, an unwanted
side effect of many other agents used for T2DM, makes it
a desirable option for treatment of GDM since they both Targeted postprandial glucose interventions and the GI
share a common metabolic pathway of IR and inflammation. Studies from more than 20 years ago showed that postpran-
Metformin crosses the placenta and therefore its use in preg- dial glucose has a stronger association with macrosomia than
nancy has until recently been limited because of concern of FBG therefore interventions specifically aimed at reduc-
its potential effects on the mother and fetus.47 ing the postprandial glycemic response seem warranted to
Low glycemic index (LGI) diet 411
improve pregnancy outcome in women who have GDM or LGI and obesity
who are obese and insulin resistant.51 A Cochrane review of the evidence for the use of LGI diets
Dietary advice remains the cornerstone of treatment for in obese subjects (in nonpregnant subjects) included six
GDM, with increasing use of second-line therapies, particu- eligible RCTs with a total of 202 subjects. Obese and over-
larly insulin and metformin for women who are unable to weight subjects who followed the LGI diet had significantly
achieve adequate control with dietary changes alone.52,53 Two greater weight loss (weighted mean difference −1.1 kg [95%
seminal RCTs have unequivocally demonstrated that treat- CI −2.0 to −0.2]) and reductions in total cholesterol and LDL
ment of mild glucose intolerance in pregnancy improves neo- concentrations compared to subjects on standard or high-
natal and maternal outcomes.54,55 In one study, the Australian GI diets (WMD [weighted mean difference] −0.22 mmol/L,
Carbohydrate Intolerance Study in Pregnant Women, women [95% CI −0.43 to −0.02] and WMD −0.24 mmol/L, [95% CI
with mild glucose intolerance, randomized to the intervention −0.44 to −0.05] for standard and high GI, respectively).59
group, received dietary advice and only commenced insulin Subsequently, Liu et al. demonstrated a significant reduction
when fasting and/or postprandial blood glucose targets were in serum glucose and insulin AUC (42% and 29%, respec-
exceeded on more than two occasions. A significant reduc- tively) when comparing diets of LGI/low carbohydrate to
tion in serious perinatal complications was achieved in the high GI/high carbohydrate in overweight and obese people.60
intervention arm, and although 20% of women in this group
were commenced on insulin to optimize glucose control, it is
important to highlight that the majority of women were suc- LGI diet and diabetes
cessfully managed with dietary advice alone.54 A Cochrane review published in 2009 and an earlier meta-
analysis both concluded that a modest, yet clinically useful,
reduction in HbA1c of 0.5% can be achieved with a LGI diet
Low glycemic index (LGI) diet in the management of diabetes, equivalent to the changes
observed with single-agent pharmacotherapy.61,62 Similar
Otto and Niklas first observed the different glycemic limitations of small sample size and variation in duration of
response to various foods in 1980 with the subsequent con- diet were reported and both reviews included participants
cept of the glycemic index (GI) food classification system with T1 and T2DM.
developed by Jenkins in Toronto shortly thereafter.56 Initial
studies compared a 50 g portion of carbohydrate calculated
from published food tables to a standard of 50 g glucose. LGI and pregnancy
Following consumption of the carbohydrate, serial venous Two systematic reviews are inconclusive to support univer-
blood samples were taken up to 2 hours and glucose val- sal recommendation of LGI diets to prevent or treat GDM in
ues plotted to determine the area under the curve (AUC). pregnant women.63,64
Similar methods are adopted today however the use of cap- The Camden study, a prospective analysis of dietary GI
illary blood glucose is more frequently utilized against a calculated from three 24-hour recall questionnaires over the
standard of white bread. The AUC for each food is expressed course of pregnancy, in 1082 women reported a positive and
as a percentage of the mean AUC for the standard, from a significant relationship between GI, maternal HbA1c, and
sample of 10 subjects, and used to calculate the GI.57 The GI plasma glucose 1 hour following 50 g glucose challenge test
predicts the ranking of the glycemic response of foods in (GCT) at 24–28 weeks gestation.65 Similarly, GI was posi-
an individual; foods that are digested and absorbed slowly tively related to fetal growth.
raise blood glucose concentrations gradually and are thus Louie et al. conducted a systematic review of LGI diets in
given low GI (LGI) values. The division of GI values into pregnancy for women of all BMI categories including three
three simple categories has facilitated a useable system for studies complicated by GDM.66 Of the eight studies included,
individuals and generated an additional tool for the delivery half showed positive influence of a LGI diet on pregnancy
of dietary advice for healthcare professionals; high GI ≥70, outcomes, with a reduction in LGA reported in two but a
medium GI 56–69, and LGI <55. moderate increase in the delivery of an SGA infant in one.
Inherent and important limitations of GI classification In general, LGI diets have been shown to be safe in preg-
exist when applying the principles to the wider population and nancy with no adverse maternal effects reported, but issues
when interpreting individual responses to foods of different regarding restricted fetal growth have been identified.65
categories. Aberrations in methodology are known to affect These concerns have not been reproduced in subsequent
the GI including food portion size and preparation, the choice studies and will continue to be addressed by ongoing larger
of standard, timing of test, frequency of blood sampling, and intervention trials.
the formula used to determine the AUC.57 Subject character-
istics such as age, sex, ethnicity, degree of glucose intolerance,
and coexisting disease that may influence glucose absorption Randomized controlled trials of LGI interventions in
are also confounders but attempts to standardize that these over weight and obese pregnant women
factors would serve to limit the variability in glucose response The three largest RCTs aimed at lowering the GI of the diet,
and maintain consistencies across food groups tested, there- in conjunction with supporting a healthy lifestyle in high-risk
fore minimizing the effect on overall GI.58 pregnant women, have completed recruitment,12,67,68 and two
have published the outcomes.12,67 While the recently reported UPBEAT (ISRCTN 89971375)
Pregnancy and Glycemic Index Outcomes study did not spe- The UPBEAT (U.K. Pregnancies Better Eating and Activity
cifically recruit obese women (mean BMI 24.5 kg/m2) and Trial), which has recently completed recruitment and has
included those with normal glucose tolerance, the results yet to report outcomes, is a multicenter RCT of a complex
nonetheless merit mention.69 In an RCT, LGI dietary advice dietary and PA intervention aimed at improving glucose
was compared to routine healthy eating with both arms receiv- homeostasis in obese pregnant women (n = 1546). Following
ing equivalent support and contact from the research team and randomization at 15+0–17+6 weeks gestation to intervention
dietitian from 20 weeks onward. In contrast to an earlier but or standard antenatal care, the program is delivered over
smaller study from the same group, no significant differences 8 weekly group sessions with a health trainer until OGTT
were observed for the primary outcomes of birth weight, birth at 27+0–28+6 weeks gestation. The objective of the dietary
centile, or ponderal index.70 The failure to replicate the find- recommendation is to reduce the GI and glycemic load
ings in the larger cohort with women of similar characteristics (GL = GI × carbohydrates [g]/100) using food swaps and to
is perhaps unsurprising since the advice given to the control exchange saturated fatty acids (SFAs) for monounsaturated
arm was a healthy diet, reinforced with dietary counseling, in and polyunsaturated fatty acids. Women maintain contact
contrast to standard antenatal care in the earlier study, where with the research team with further three structured vis-
women were reported to consume a high GI diet reflective of its until delivery where dietary data, blood samples, and
habitual diet. anthropometry are obtained. Within 72 hours of birth, a
detailed anthropometric assessment of the baby is made that
is repeated at 6 months’ postnatal review.
Low glycemic index diet in pregnancy to prevent The study endpoints include maternal and neonatal pri-
macrosomia (ROLO study): Randomised control trial mary outcomes of, respectively, GDM diagnosed by 75 g
The ROLO study hypothesized that an LGI diet would OGTT using the recommendations of the IADPSG42 and
reduce recurrence of macrosomia in subsequent pregnancies LGA delivery, defined as adjusted birth weight >90th centile
for women who had previously delivered an infant >4000 g for gestational age using customized centiles, where adjust-
vs. standard antenatal care with no dietary advice (n = 781, ment is made for maternal height, weight, ethnicity, parity,
mean BMI 26.8 kg/m2).71 No difference in the primary out- and infant sex.73
come was observed, but positive findings were noted for A pilot trial for UPBEAT was undertaken in 183 women
those following the LGI diet. At 40 weeks gestation, women in four urban U.K. hospitals, to evaluate changes in dietary
in the intervention arm had gained 1.5 kg less weight than and PA behaviors in response to the intervention, to trial
controls (mean difference −1.3, [95% CI −2.4 to −0.2], p = all aspects of the protocol, and to undertake process evalu-
0.01), and fewer exceeded the 2009 IOM guidelines for GWG ation.74 Baseline habitual diets for all women were similar
(38% vs. 48% for LGI and control, respectively). Both FPG with no differences in the following: total energy intake,
and 1 hour glucose following 50 g GCTs were significantly GI, GL, and macronutrient composition. At 28 weeks gesta-
lower in the LGI group, although the incidence of GDM fol- tion, there was a significant reduction in total energy intake,
lowing 100 g glucose load was no different.67 GL, total fat, and SFA with an increase in fiber consump-
tion (nonstarch polysaccharides) for the intervention group.
Women entered the study with a medium GI diet (58), but
Antenatal lifestyle advice for women who are after 8 weeks of adherence to the program, a borderline sig-
overweight or obese: LIMIT randomised trial nificant 7-point reduction in the GI was achieved, and the
LIMIT was the first adequately powered lifestyle interven- diet of the intervention group was reclassified as LGI (differ-
tion RCT for obese and overweight pregnant women with ence −7, [95% CI −15 to 0], p = 0.054). On completion of the
2212 participants randomized to a combined program of study, analysis will determine whether these positive dietary
dietary and exercise advice (in conjunction with behavioral changes are maintained throughout pregnancy and beyond
strategies) or to standard antenatal care. The intervention and whether they translate into favorable clinical outcomes.
was delivered over three visits with a research dietitian and In summary, for all the pregnancy interventions dis-
reinforced by three additional telephone consultations with cussed, the additional commitment over and above the time
trained assistants.12 Dietary advice was not principally cen- required for routine antenatal care is a limitation particu-
tered on LGI principles but incorporated food exchanges, larly for women in full-time employment or with childcare
increased fiber and fruit consumption, and reduction of responsibilities. Poor compliance and high dropout rates are
refined sugars to maintain a eucaloric diet with a global greatest in women from ethnic minorities and those who have
effect of lowering overall GI.72 The primary outcome of a not received higher education.7,11,21 Authors have attempted
reduction in LGA incidence (birth weight ≥90th centile to facilitate participation by various means including local
for gestational age and sex) compared to control was not delivery, providing transport to and from research centers,
met. Pregnancy outcomes including the diagnosis of GDM and regular phone, postal, and email contact, but greater
and GWG were the same; however, there was a reduction in understanding of barriers to behavioral change, particularly
the risk for the secondary outcome of macrosomia in the to increased PA, is needed.
intervention group (15% vs. 19%, relative risk 0.83, [95% CI Adequately powered RCTs are required to determine
0.68–0.99], p = 0.04). whether the interventions described are successful in
References 413
improving maternal, obstetric, and offspring outcomes and study designs that encourage and support pregnant women
whether they have the real translational potential to be a and that can adapt to meet the cultural requirements of the
therapeutic option for obese pregnant women. Pragmatic local population, are essential.
REFERENCES
1. Nohr EA, Vaeth M et al. Combined associations of prepregnancy 21. Callaway LK, Colditz PB et al. Prevention of gestational diabetes:
body mass index and gestational weight gain with the outcome Feasibility issues for an exercise intervention in obese pregnant
of pregnancy. Am J Clin Nutr 2008; 87(6): 1750–1759. women. Diabetes Care 2010; 33(7): 1457–1459.
2. Poston L. Maternal obesity, gestational weight gain and diet as 22. Gardner B, Wardle J et al. Changing diet and physical activity to
determinants of offspring long term health. Best Pract Res Clin reduce gestational weight gain: A meta-analysis. Obes Rev 2011;
Endocrinol Metab 2012; 26(5): 627–639. 12(7): e602–e620.
3. Wolff S, Legarth J et al. A randomized trial of the effects of dietary 23. Hill B, Skouteris H et al. Interventions designed to limit ges-
counseling on gestational weight gain and glucose metabolism in tational weight gain: A systematic review of theory and meta-
obese pregnant women. Int J Obes 2008; 32(3): 495–501. analysis of intervention components. Obes Rev 2013; 14(6):
4. Thornton YS, Smarkola C et al. Perinatal outcomes in nutritionally 435–450.
monitored obese pregnant women: A randomized clinical trial. J 24. Barakat R, Stirling JR et al. Does exercise training during preg-
Natl Med Assoc 2009; 101(6): 569–577. nancy affect gestational age? A randomised controlled trial. Br J
5. Quinlivan JA, Lam LT et al. A randomised trial of a four-step mul- Sports Med 2008; 42(8): 674–678.
tidisciplinary approach to the antenatal care of obese pregnant 25. Stafne SN, Salvesen KA et al. Regular exercise during pregnancy
women. Aust N Z J Obstet Gynaecol 2011; 51(2): 141–146. to prevent gestational diabetes: A randomized controlled trial.
6. Barakat R, Cordero Y et al. Exercise during pregnancy improves Obstet Gynecol 2012; 119(1): 29–36.
maternal glucose screen at 24–28 weeks: A randomised con- 26. Barakat R, Pelaez M et al. Exercise during pregnancy and gesta-
trolled trial. Br J Sports Med 2012; 46(9): 656–661. tional diabetes-related adverse effects: A randomised controlled
7. Oostdam N, van Poppel M et al. No effect of the FitFor2 exercise trial. Br J Sports Med 2013; 47(10): 630–636.
programme on blood glucose, insulin sensitivity, and birthweight 27. Catalano PM, Mele L et al. Inadequate weight gain in overweight
in pregnant women who were overweight and at risk for ges- and obese pregnant women: What is the effect on fetal growth?
tational diabetes: Results of a randomised controlled trial. Br J Am J Obstet Gynecol 2014; 211(2): 137.e1–137.e7.
Obstet Gynaecol 2012; 119: 1098–1107. 28. Jelsma JG, van Poppel MN et al. DALI: Vitamin D and lifestyle
8. Sui Z, Grivell RM et al. Antenatal exercise to improve outcomes intervention for gestational diabetes mellitus (GDM) prevention:
in overweight or obese women: A systematic review. Acta Obstet An European multicentre, randomised trial—Study protocol.
Gynecol Scand 2012; 91(5): 538–545. BMC Pregnancy Childbirth 2013; 13: 142.
9. Guelinckx I, Devlieger R et al. Effect of lifestyle intervention on 29. Tuomilehto J, Lindstrom J et al. Prevention of type 2 diabetes mel-
dietary habits, physical activity, and gestational weight gain in litus by changes in lifestyle among subjects with impaired glu-
obese pregnant women: A randomized controlled trial. Am J Clin cose tolerance. N Engl J Med 2001; 344(18): 1343–1350.
Nutr 2010; 91(2): 373–380. 30. Knowler WC, Barrett-Connor E et al. Reduction in the incidence
10. Phelan S, Phipps MG et al. Randomized trial of a behavioral of type 2 diabetes with lifestyle intervention or metformin. N Engl
intervention to prevent excessive gestational weight gain: The Fit J Med 2002; 346(6): 393–403.
for Delivery Study. Am J Clin Nutr 2011; 93(4): 772–779. 31. Lindström J, Louheranta A et al. The Finnish Diabetes
11. Vinter CA, Jensen DM et al. The LiP (Lifestyle in Pregnancy) Prevention Study (DPS): Lifestyle intervention and 3-year
study: A randomized controlled trial of lifestyle intervention results on diet and physical activity. Diabetes Care 2003;
in 360 obese pregnant women. Diabetes Care 2011; 34(12): 26(12): 3230–3236.
2502–2507. 32. Tobias DK, Zhang C et al. Physical activity before and during
12. Dodd JM, Turnbull D et al. Antenatal lifestyle advice for women pregnancy and risk of gestational diabetes mellitus: A meta-
who are overweight or obese: LIMIT randomised trial. Br Med J analysis. Diabetes Care 2011; 34(1): 223–229.
2014; 348: 1285. 33. Yin YN, Li XL et al. Physical activity during pregnancy and the
13. Renault KM, Norgaard K et al. The Treatment of Obese Pregnant risk of gestational diabetes mellitus: A systematic review and
Women (TOP) study: A randomized controlled trial of the effect meta-analysis of randomised controlled trials. Br J Sports Med
of physical activity intervention assessed by pedometer with or 2014 Feb; 48(4): 290–295.
without dietary intervention in obese pregnant women. Am J 34. Chudyk A, Petrella RJ. Effects of exercise on cardiovascular risk
Obstet Gynecol 2014; 210(2): 134.e1–134.e9. factors in type 2 diabetes: A meta-analysis. Diabetes Care 2011;
14. Villamor E, Cnattingius S. Interpregnancy weight change and 34(5): 1228–1237.
risk of adverse pregnancy outcomes: A population-based study. 35. Duncan GE, Perri MG et al. Exercise training, without weight
Lancet 2006; 368(9542): 1164–1170. loss, increases insulin sensitivity and postheparin plasma lipase
15. Dodd JM, Grivell RM et al. Antenatal interventions for overweight activity in previously sedentary adults. Diabetes Care 2003;
or obese pregnant women: A systematic review of randomised 26(3): 557–562.
trials. Br J Obstet Gynaecol 2010; 117(11): 1316–1326. 36. Eriksson KF, Lindgärde F. Prevention of type 2 (non-insulin-
16. Muktabhant B, Lumbiganon P et al. Interventions for preventing dependent) diabetes mellitus by diet and physical exercise. The
excessive weight gain during pregnancy. Cochrane Database Syst 6-year Malmö feasibility study. Diabtologia 1991; 34(12): 891–898.
Rev 2012; 4: CD007145. 37. Clapp JF. Effects of diet and exercise on insulin resistance during
17. Oteng-Ntim E, Varma R et al. Lifestyle interventions for overweight pregnancy. Metab Syndr Relat Disord 2006; 4(2): 84–90.
and obese pregnant women to improve pregnancy outcome: 38. Ong MJ, Guelfi KJ et al. Supervised home-based exercise may
Systematic review and meta-analysis. BMC Med 2012; 10: 47. attenuate the decline of glucose tolerance in obese pregnant
18. Thangaratinam S, Rogozińska E et al. Effects of interventions in women. Diabetes Metab 2009; 35(5): 418–421.
pregnancy on maternal weight and obstetric outcomes: Meta- 39. Oostdam N, van Poppel MN et al. Design of FitFor2 study: The
analysis of randomised evidence. Br Med J 2012; 344: e2088. effects of an exercise program on insulin sensitivity and plasma
19. Thangaratinam S, Rogozinska E et al. Interventions to reduce or glucose levels in pregnant women at high risk for gestational dia-
prevent obesity in pregnant women: A systematic review. Health betes. BMC Pregnancy Childbirth 2009; 9: 1.
Technol Assess 2012; 16(31): iii–iv, 1–191. 40. Nascimento SL, Surita FG et al. The effect of an antenatal physi-
20. Luoto R, Kinnunen TI et al. Primary prevention of gestational dia- cal exercise programme on maternal/perinatal outcomes and
betes mellitus and large-for-gestational-age newborns by lifestyle quality of life in overweight and obese pregnant women: A
counseling: A cluster-randomized controlled trial. PLOS Med randomised clinical trial. Br J Obstet Gynaecol 2011; 118(12):
2011; 8(5): e1001036. 1455–1463.
41. The HAPO Study Cooperative Research Group. Hyperglycemia 60. Liu AG, Most MM et al. Reducing the glycemic index or carbo-
and adverse pregnancy outcomes. N Engl J Med 2008; 358(19): hydrate content of mixed meals reduces postprandial glycemia
1991–2002. and insulinemia over the entire day but does not affect satiety.
42. IADPSG. International association of diabetes and pregnancy Diabetes Care 2012; 35(8): 1633–1637.
study groups recommendations on the diagnosis and classifica- 61. Brand-Miller J, Hayne S et al. Low-glycemic index diets in the
tion of hyperglycemia in pregnancy. Diabetes Care 2010; 33(3): management of diabetes: A meta-analysis of randomized con-
676–682. trolled trials. Diabetes Care 2003; 26(8): 2261–2267.
43. WHO. Diagnostic criteria and classification of hyperglycaemia 62. Thomas D, Elliott EJ. Low glycaemic index, or low glycaemic
first detected in pregnancy. World Health Organization: Geneva, load, diets for diabetes mellitus. Cochrane Database Syst Rev
Switzerland, 2013. 2009; (1): CD006296.
44. Foxcroft KF, Rowlands IJ et al. Exercise in obese pregnant women: 63. Tieu J, Crowther CA et al. Dietary advice in pregnancy for pre-
The role of social factors, lifestyle and pregnancy symptoms. venting gestational diabetes mellitus. Cochrane Database Syst
BMC Pregnancy Childbirth 2011; 11: 4. Rev 2008; (2): CD006674.
45. Sui Z, Dodd JM. Exercise in obese pregnant women: Positive 64. Han S, Crowther CA et al. Different types of dietary advice for
impacts and current perceptions. Int J Womens Health 2013; 5: women with gestational diabetes mellitus. Cochrane Database
389–398. Syst Rev 2013; 3: CD009275.
46. Molyneaux E, Poston L et al. Obesity and mental disorders during 65. Scholl TO, Chen X et al. The dietary glycemic index during
pregnancy and postpartum: A systematic review and meta-analysis. pregnancy: Influence on infant birth weight, fetal growth, and
Obstet Gynecol 2014; 123(4): 857–867. biomarkers of carbohydrate metabolism. Am J Epidemiol 2004;
47. Charles B, Norris R et al. Population pharmacokinetics of metfor- 159(5): 467–474.
min in late pregnancy. Ther Drug Monit 2006; 28(1): 67–72. 66. Louie JC, Brand-Miller JC et al. Glycemic index and pregnancy: A
48. Rowan JA, Hague WM et al. Metformin versus insulin for the systematic literature review. J Nutr Metab 2010; 2010: 282464.
treatment of gestational diabetes. N Engl J Med 2008; 358(19): 67. Walsh JM, McGowan CA et al. Low glycaemic index diet in preg-
2003–2015. nancy to prevent macrosomia (ROLO study): Randomised control
49. Reynolds RM, Allan KM et al. Maternal obesity during pregnancy trial. Br Med J 2012; 345: e5605.
and premature mortality from cardiovascular event in adult off- 68. Briley AL, Barr S et al. A complex intervention to improve preg-
spring: Follow-up of 1 323 275 person years. Br Med J 2013; 347: nancy outcome in obese women; the UPBEAT randomised con-
f4539. trolled trial. BMC Pregnancy Childbirth 2014; 14(1): 74.
50. Rowan JA, Rush EC et al. Metformin in gestational diabetes: The 69. Moses RG, Casey SA et al. Pregnancy and glycemic index out-
offspring follow-up (MiG TOFU). Diabetes Care 2011; 34(10): comes study: Effects of low glycemic index compared with con-
2279–2284. ventional dietary advice on selected pregnancy outcomes. Am J
51. Combs CA, Gunderson E et al. Relationship of fetal macroso- Clin Nutr 2014; 99(3): 517–523.
mia to maternal postprandial glucose control during pregnancy. 70. Moses RG, Luebcke M et al. Effect of a low-glycemic-index diet
Diabetes Care 1992; 15(10): 1251–1257. during pregnancy on obstetric outcomes. Am J Clin Nutr 2006;
52. Dornhorst A, Frost G. The principles of dietary management of 84(4): 807–812.
gestational diabetes: Reflection on current evidence. J Hum Nutr 71. Walsh J, Mahony R et al. A randomised control trial of low gly-
Diet 2002; 15(2): 145–156. caemic index carbohydrate diet versus no dietary intervention
53. NICE (National Institute for Health and Clinical Excellence). in the prevention of recurrence of macrosomia. BMC Pregnancy
Diabetes in Pregnancy: Management of Diabetes and Its Childbirth 2010; 10: 16.
Complications from Pre-Conception to the Postnatal Period 72. Dodd JM, Turnbull DA et al. Limiting weight gain in overweight
[CG63]. RCOG Press: London, U.K., 2008. and obese women during pregnancy to improve health out-
54. Crowther CA, Hiller JE et al. Effect of treatment of gestational comes: The LIMIT randomised controlled trial. BMC Pregnancy
diabetes mellitus on pregnancy outcomes. N Engl J Med 2005; Childbirth 2011; 11: 79.
352(24): 2477–2486. 73. Gardosi J, Figueras F et al. The customised growth potential: An
55. Landon MB, Spong CY et al. A multicenter, randomized trial international research tool to study the epidemiology of fetal
of treatment for mild gestational diabetes. N Engl J Med 2009; growth. Paediatr Perinat Epidemiol 2011; 25(1): 2–10.
361(14): 1339–1348. 74. Poston L, Briley AL et al. Developing a complex intervention for
56. Jenkins DJ, Wolever TM et al. Glycemic index of foods: A physi- diet and activity behaviour change in obese pregnant women (the
ological basis for carbohydrate exchange. Am J Clin Nutr 1981; UPBEAT trial); assessment of behavioural change and process
34(3): 362–366. evaluation in a pilot randomised controlled trial. BMC Pregnancy
57. Wolever TM, Jenkins DJ et al. The glycemic index: Methodology Childbirth 2013; 13(1): 148.
and clinical implications. Am J Clin Nutr 1991; 54(5): 846–854. 75. Otto H, Niklas L. Differences d’action sur la glycemie d’ailments
58. Pi-Sunyer FX. Glycemic index and disease. Am J Clin Nutr 2002; contenant des hydrates de carbone: consequences pour le trait-
76(1): 290S–298S. ment dietetique du diabete sucre. (The different actions on gly-
59. Thomas DE, Elliott EJ et al. Low glycaemic index or low glycae- cemia of foods containing carbohydrates: Consequences for the
mic load diets for overweight and obesity. Cochrane Database nutritional treatment of diabetes mellitus.) Med Hyg 1980; 38:
Syst Rev 2007; (3): CD005105. 3424–3429 (in French).
50 Lifestyle interventions to
reduce risk of diabetes
among high-risk pregnant
and postpartum women
Lisa Chasan-Taber
lifestyle interventions designed to reduce the risk of diabe-

Background tes or diabetes risk factors among women with a history of
Gestational diabetes mellitus (GDM) is one of the most com- GDM. The chapter first describes the impact of the lifestyle
mon complications of pregnancy with a prevalence rate vary- interventions on the incidence of type 2 diabetes and bio-
ing from 1% to 20% depending on the population studied markers of insulin resistance, weight change, and healthy
and diagnostic criteria applied.1,2 With the recent adoption behaviors such as physical activity, diet, and breastfeeding.
of the International Association of Diabetes and Pregnancy The chapter then goes on to describe the study design and
Study Groups Consensus Panel diagnostic criteria, it is esti- methods of new randomized trials that have been recently
mated that 18% of pregnant women will be diagnosed with launched. The chapter concludes with recommendations for
GDM.1 Obesity is strongly associated with risk of GDM,3 and future research and practice.
it is expected that the incidence of GDM among women of
reproductive age will further increase as the prevalence of
obesity continues to rise among this age group. Methods
GDM is related to short- and long-term adverse health
outcomes for the mother. Compared with women with We conducted a literature search of PubMed for English-
healthy pregnancies, women with histories of GDM have language studies of randomized controlled trials of lifestyle
elevated CVD risk factors including higher blood pressure, interventions among women with a history of GDM. Only
triglyceride levels, and lower HDL.4 Consistent with these published peer-reviewed journal articles of original research
findings, a meta-analysis found that GDM confers a seven- in the English language were included. Keyword searches
fold risk for future type 2 diabetes5 and up to one-third of included lifestyle intervention, randomized controlled trial,
women with type 2 diabetes have previously been diagnosed type 2 diabetes, prevention, diet, physical activity, postpar-
with GDM.6 tum, pregnancy, weight retention, gestational diabetes melli-
According to a systematic review, the highest-risk period tus, and health behaviors. Additional relevant articles cited in
for the development of type 2 diabetes is within the first the reference lists of identified papers were retrieved manually.
5 years after a GDM pregnancy.7 Therefore, both GDM and In total, nine randomized controlled trials of lifestyle
milder glucose intolerance in pregnancy identify women interventions conducted among women with a history of
who are at high risk for subsequent glucose intolerance and GDM were identified to fulfill the eligibility criteria (Tables
type 2 diabetes and offer the opportunity to implement inter- 50.1 and 50.2).13–21 Of these studies, two examined the
ventions to decrease such risk.7,8 However, while lifestyle impact of the lifestyle intervention on subsequent incidence
interventions involving exercise and a healthy diet in high- of diabetes,18,21 and four examined the impact on postpar-
risk adults have been found to reduce progression to type 2 tum biomarkers of insulin resistance.15–17,20 In terms of other
diabetes by more than 50%,9–12 little attention has been given risk factors for diabetes, all with the exception of one study21
to the potential benefits of such strategies in women with a examined the impact of the lifestyle intervention on weight
history of GDM. change and physical activity. Four of the trials examined the
Therefore, the goal of this chapter is to provide a com- impact on diet.14,15,19,20 Only one study to date reported the
prehensive overview of the randomized controlled trials of impact on breastfeeding.14
415
Table 50.1 Randomized trials of lifestyle interventions to reduce risk of type 2 diabetes among women with gestational diabetes mellitus: Study designs
Source Name Design Pilot Population Intervention Mode Goals

Cheung RCT; FU: Pilot 43 women with previous Exercise intervention Individualized in PA: 150 minute/week or 10,000
et al.13 12 months GDM <4 years. vs. usual care person; telephone; steps/day for 5 days/week
previously; Australia control mailings
Ferrara Diet, Exercise, RCT; FU: Pilot 197 women with current Lifestyle intervention Individualized in Weight, return to prepregnancy
et al.14 and 12 months GDM; California (diet, exercise, person; telephone weight if normal, lose 5% of
Breastfeeding breastfeeding) vs. prepregnancy weight if
Intervention usual care control overweight; PA, 150 minute/
week of moderate intensity or
harder; diet, ≤25% calories
from fat; breastfeeding,
exclusively for 6 months
Hu et al.15 Tianjin RCT; 1 year Preliminary 404 women with previous Lifestyle intervention Individualized in Weight, no weight loss if normal
Gestational results GDM from 2005 to 2009; (diet and exercise) person weight, lose 5%–10% of
Diabetes China vs. usual care prepregnancy weight if BMI ≥
Mellitus control 24 kg/m2 through reduction of
Prevention ≥10% total calories; PA, 150 of
Program moderate intensity or harder;
diet, TF <30%, fiber 20–30 g/
day, % calories from fat
Kim RCT; FU: Pilot 49 women with previous Exercise intervention Web based Steps: up to 10,000 steps/day
et al.16 13 weeks GDM within past 3 years; vs. usual care
Michigan control
McIntyre RCT; FU: Pilot 28 women with previous Exercise intervention Individualized in PA: 150 minute/week
et al.17 12 weeks GDM 6 weeks vs. usual care person; telephone
postpartum; Australia control
Ratner Diabetes RCT; FU: 350 women with previous Lifestyle intervention Individualized in Weight, reduction ≥7% of initial
et al.18 Prevention 2.8 years GDM and current elevated (diet and exercise) person; group body weight; diet, low calorie,
Program glucose levels from the vs. placebo sessions low fat; PA, moderate intensity
DPP; United States ≥150 minute/week
Reinhardt RCT; FU: Pilot 38 women following GDM Lifestyle intervention Telephone; mailings Healthy eating and PA
et al.19 6 months diagnosis; Australia (diet and exercise)
vs. usual care control
Shyam RCT; FU: 77 women with previous Low GI diet vs. usual In person; text Weight, 5%–7% reduction in body
et al.20 6 months GDM within 2 months; care control messaging, e-mails weight if BMI > 23 and
Malaysia maintenance if normal; PA,
moderate intensity for 30
minutes/day five times per week
416 Lifestyle interventions to reduce risk of diabetes among high-risk pregnant and postpartum women
Wein RCT; FU: 796 200 women with previous Diet intervention vs. Telephone; mailings Diet, healthy eating; PA, exercise
et al.21 person-years GDM from 1989 to 1991 control (30 minutes three times per
(median and subsequent IGT week)
51 months)
Abbreviations: RCT, randomized clinical trial; FU, follow-up; GDM, gestational diabetes mellitus; IGT, impaired glucose tolerance; PA, physical activity; FG, fasting glucose; GI, glycemic index; BMI,
body mass index.
Table 50.2 Randomized trials of lifestyle interventions to reduce risk of type 2 diabetes among women with gestational diabetes mellitus: Findings
Impact on
type 2 Impact on biomarkers Impact on
Author Name diabetes of insulin resistance Impact on weight Impact on PA Impact on diet breastfeeding
Cheung NA NA BMI (kg/m2): 28 (95% Steps (% achieving NA
et al.13 CI: 23.9, 34.3) vs. goal), 30.8 vs.
25.5 (95% CI: 22.5, 17.6 p = 0.34; PA
28.7), p = 0.14 (% achieving
goal), 70.0 vs.
57.9, p = 0.51
Ferrara Diet, Exercise, NA NA Weight (% achieving PA (difference in Fat (% difference in mean Breastfeeding
et al.14 and goal): 37.5% vs. mean change in change) −3.6, p = 0.002 (difference in
Breastfeeding 21.4%, p = 0.07 minute/week): % partially or
Intervention 25.3, p = 0.91 exclusively
breastfeeding):
15%, p = 0.09
Hu et al.15 Tianjin NA FG (change in Weight change, −1.4 ± LTPA (% increased): Fat (% decrease), 77.1 vs. NA
Gestational mmol/L), −0.09 ± 3.44 kg vs. −0.21 ± 59.4% vs. 26.9%, 68.9, p = 0.064; fiber
Diabetes 0.52 vs. −0.09 ± 3.52 kg (0.3%), p = p < 0.001 (% increase), 59.5 vs.
Mellitus 0.6, p = 0.97; 0.001; BMI change, 47.4, p = 0.012
Prevention insulin (change in −0.50 ± 1.41 kg/m2
Program pmol/L), −11.8 ± vs. −0.09 ±
27.4 vs. −3.2 ± 1.37 kg/m2, p = 0.004
31.2, p = 0.004
Kim NA FG (change in Weight (change in kg): PA (% moderate NA NA
et al.16 mmol/L), −0.046 vs. −0.14 kg vs. −1.5 kg, intensity): 58 vs.
0.038, p = 0.65; p = 0.13 47, p = 0.51
2- hour glucose on
75 g OGTT (change
in mmol/l), −0.48
vs. −0.42, p = 0.91
McIntyre NA FG (change in Change in weight (kg): PA (median [range] NA NA
et al.17 mmol/L), 0.25 ± .56 0.97 + 3.7 vs. 0.22 + increase in
vs. 0.12 + 0.42, NS; 4.2 NS planned PA
insulin (change in minutes/week),
nU/mL), 1.49 + 4.23 60 (0–540) vs. 0
vs. 0.06 + 3.89, NS (0–580); p =
0.234; walking,
NS
(Continued)
Methods 417
Table 50.2 (Continued ) Randomized trials of lifestyle interventions to reduce risk of type 2 diabetes among women with gestational diabetes mellitus: Findings
Impact on
type 2 Impact on biomarkers Impact on
Author Name diabetes of insulin resistance Impact on weight Impact on PA Impact on diet breastfeeding
Ratner Diabetes Diabetes: NA Weight (change in kg): PA (change in hour/ NA NA
et al.18 Prevention 53% risk −5.13 ± 0.43 vs. week): 1.5 hour/
Program reduction approx. 0 in placebo week vs. NA in
vs. placebo, at 6 months; −1.6 ± year 1; <0.5 vs.
p = 0.002 0.80 vs. approx. 0 in NA in year 3
placebo at 3 years
Reinhardt NA NA BMI (difference in LTPA (change in Total fat (change in g/ NA
et al.19 change in kg/m2), minutes/day): 11 day), −19 (95%CI: −37,
−1.5 (95% CI: −2.8, (95% CI: 1, 22) −1); GL (unit change),
− 0.1); weight −26 (95% CI −48, −4)
(difference in change
in kg), −4.0 (95% CI:
−7.6, −0.5)
Shyam NA Glucose, 2 hours post Weight (% achieving PA (median MET- Fat (g), 58 ± 18 vs. 53 ±
et al.20 75 g OGTT (median goal): 33% vs. 8%, minute/week, 16, p = 0.695 for
mmol/L, IQR), −0.2 p = 0.01 IQR): 933 (1403) difference in change;
(2.8) vs. 0.8 (2.0), vs. 965 (857), fiber (g), 17 ± 4 vs. 13
p = 0.025; insulin p = 0.908 ± 4, p = 0.02 for
(<2 μU/L), 61.5% difference in change;
vs. 52.6%, p = GI, 57 ± 5 vs. 64 ± 6,
0.228 p = 0.033 for
difference in change
Wein Diabetes NA NA NA NA
et al.21 (annual
incidence
rate): 6.1%
vs. 7.3%
(incidence
rate ratio =
0.83, 95%
CI: 0.47,
1.48)
Abbreviations: RCT, randomized clinical trial; FU, follow-up; GDM, gestational diabetes mellitus; IGT, impaired glucose tolerance; PA, physical activity; FG, fasting glucose; GI, glycemic index; LTPA,
leisure time physical activity.
Results 419
Results insulin levels (−11.8 ± 27.4 pmol/L) as compared to those in

the control arm (−3.2 ± 31.2 pmol/L, p = 0.004). The authors
Type 2 diabetes and biomarkers of insulin resistance also observed the suggestion of a positive impact of the
Among the nine randomized controlled trials conducted intervention on fasting glucose levels.
among women with GDM, two evaluated the impact of a life- Similar findings were observed by Shyam et al., who
style intervention on subsequent incidence of diabetes,18,21 randomized 77 Asian women who had been diagnosed
and four examined the impact on postpartum biomarkers with GDM in the prior 2 months to a low-glycemic-index
of insulin resistance (Table 50.2).15–17,20 The most success- dietary intervention or to a usual care control.20 Goals of
ful intervention to date in terms of impact on subsequent the intervention were a 5%–7% reduction in body weight
diabetes was observed by the Diabetes Prevention Program if prepregnancy BMI was greater than 23 kg/m 2 and mod-
(DPP),9 a multicenter randomized trial of an intensive life- erate-intensity physical activity of 30 minutes at least five
style intervention conducted among a population of adults times per week. At 6 months of follow-up, the intervention
who had elevated fasting and postload plasma glucose con- group had significantly greater decreases in 2-hour post-
centrations. Goals of the lifestyle intervention were at least load blood glucose after 75 g oral glucose tolerance test
a 7% reduction in enrollment weight; a low-calorie, low-fat (median [IQR, inter-quartile range]: −0.2 [2.8] mmol/L) as
diet; and at least 150 minutes/week of moderate-intensity compared to the control arm (0.8 [2.0] mmol/L), p = 0.025).
physical activity. In a subset of this population limited to The suggestion of a beneficial impact on plasma insulin was
women with a self-reported history of GDM (n = 350/2190), not statistically significant.
Ratner et al.18 found that the incidence of type 2 diabetes in
those randomized to the lifestyle intervention was 7.4/100
person-years, compared with 15.2/100 person-years in the Weight change
placebo group, for a 53% reduction in incidence (p = 0.002). With the exception of one study,21 all of the intervention
However, this study involved an intensive intervention not studies conducted among women with prior GDM exam-
easily administered in a clinical setting and was conducted ined the impact on weight (Table 50.2). It is important to note
an average of 12 years after GDM diagnosis such that inter- here that four of these studies14,17,19,20 had conducted among
vening lifestyle factors and subsequent pregnancies may women with current GDM or very recent (e.g., within the
have modified findings. For example, women with early past 2 months) GDM, and therefore, focused on postpartum
postpartum conversion to diabetes, and therefore at high- weight loss and return to prepregnancy weight. The major-
est risk, were not eligible. ity observed a statistically significant positive impact of the
In the second trial evaluating the impact of the inter- intervention.15,18–20
vention on subsequent diabetes, Wein et al.21 randomized In a feasibility study, Ferrara et al. randomized women
200 women with previous GDM and subsequent impaired with a current diagnosis of GDM in late pregnancy to a
glucose tolerance to an intensive dietary intervention tar- lifestyle intervention that continued for 12 months post-
geting healthy eating and regular exercise (i.e., 30 min- partum or to a usual care control arm.14 Goals of the
utes for three times per week) or to routine dietary advice. intervention were to return to prepregnancy weight, if
Follow-up continued for a median of 51 months. Women it was normal, or achieve a 5% reduction from prepreg-
randomized to the dietary intervention had an annual inci- nancy weight if overweight. The proportion of women who
dence rate of type 2 diabetes of 6.1% as compared to 7.3% in reached the postpartum weight goal was higher, although
the control arm for an incident rate ratio of 0.83 (95% con- not statistically significantly, in the lifestyle intervention
fidence interval (CI) 0.47–1.48), which was not statistically arm as compared to the usual care arm (37.5% vs. 21.4%,
significant. As with the study by Ratner et al., this trial was absolute difference 16.1%, p = 0.07). However, in the sub-
conducted years after GDM diagnosis and therefore faced group of women not exceeding gestational weight gain
the same limitations described earlier. guidelines, the lifestyle intervention was more effective
Of the four studies that evaluated the impact of the life- (difference in the proportion of women meeting the weight
style intervention on biomarkers of insulin resistance,15–17,20 goals: 22.5%, p = 0.04).
two observed a statistically significant positive impact on Findings of similar magnitude were observed by Hu
blood glucose measures.15,20 For example, Hu et al. random- et al. in the TGDMPP.15 Specifically, the authors found a
ized 404 participants in the Tianjin Gestational Diabetes 1.4 kg (2.1%) weight loss in the intervention arm at 1 year
Mellitus Prevention Program (TGDMPP) diagnosed with of follow-up compared to a 0.21 kg (0.3%) weight loss in
GDM from 2005 to 2009 to either a lifestyle intervention the control arm (p = 0.001), as well as a positive impact
or a control group.15 The goals of the intervention included on BMI, body fat, and waist circumference. Ratner et al.
weight loss of 5%–10% of pregnancy weight if overweight in the DPP observed a weight loss of −5.13 ± 0.43 kg in
through reduction of at least 10% of total calories, partici- the lifestyle arm after 6 months that decreased to a weight
pation in 150 minutes/week of moderate-intensity activ- loss of −1.6 ± 0.80 kg at year 3 of follow-up.18 Shyam et al.
ity or harder, and reduction of total fat to less than 30% found that 33% of women in the intervention arm achieved
of calories. Interim 1-year results indicate that women in the prespecified weight loss goal as compared to 8% in the
the intervention group had a greater reduction in plasma control arm (p = 0.01). 20
Physical activity Berry et al.22 are conducting a randomized controlled

All of the trials with the exception of one21 examined trial among 100 African-American, non-Hispanic white,
the impact of a lifestyle intervention on physical activity and bilingual Hispanic women between 22 and 36 weeks of
among women diagnosed with GDM (Table 50.2). Of these, pregnancy who are diagnosed with GDM in North Carolina.
three15,18,19 observed a statistically significant impact on Women are randomized in late pregnancy (22–36 weeks ges-
one or more measures of activity, while the others tended tation) to a 14-week lifestyle intervention including diet and
to observe the suggestion of a beneficial effect. For example, physical activity or to a wait-listed control group. Follow-up
Reinhardt et al. randomized 38 women in rural Australia will continue to 10 months postpartum. Primary outcomes
following their diagnosis of GDM into an exercise interven- will include fasting blood glucose and BMI from baseline to
tion or a control arm.19 At 6 months follow-up, the interven- 10 months postpartum. Secondary maternal outcomes will
tion group increased leisure physical activity compared to include clinical, adiposity, health behaviors, and self-efficacy
the control group by 11 minutes/day (95% CI: 1, 22); however, outcomes.
changes in total physical activity levels were not statistically Chasan-Taber et al. are conducting Estudio Parto
significantly different between groups. (Project Aiming to Reduce Type twO diabetes), a random-
Hu et al. in the TGDMPP observed a similar posi- ized controlled trial in western Massachusetts.23 A total of
tive impact on leisure time physical activity, with a percent 300 Hispanic prenatal care patients who screen positive for
increase of 59.4% in the intervention arm as compared to a GDM at 24–28 weeks gestation are randomized to a cul-
26.9% increase (p < 0.001) in the control arm, but no signifi- turally and linguistically modified, individually tailored
cant impact on walking.15 Ratner et al. in the DPP observed lifestyle intervention or to a health and wellness compari-
an increase of approximately 1.5 hours/week in moderate- son control group. Follow-up will continue to 12 months
intensity physical activity after 1 year, which diminished to an postpartum. The intervention is delivered via three in-per-
increase of less than 30 minutes of physical activity by year 3.18 son sessions, telephone booster calls, and mailed materi-
als. Targets of the intervention are (1) postpartum weight
reduction to prepregnancy weight if prepregnancy BMI
Dietary factors was in the normal range or a 5% reduction from prepreg-
Four of the trials examined the impact of the lifestyle inter- nancy weight if prepregnancy BMI was overweight/obese,
vention on diet,14,15,19,20 and all observed a statistically signif- (2) at least 150 minute/week of moderate-intensity physi-
icant beneficial impact on one or more dietary components. cal activity, and (3) reduction in postpartum total caloric
For example, in the pilot feasibility study by Ferrara et al., intake via reduced consumption of popular calorie dense
the authors observe a percent difference in mean change foods, reduced portion size, modifications in ethnic reci-
in fat of −3.55% between the lifestyle arm and the control pes, and higher fruit and vegetable intake. Primary out-
arm (p = 0.002).14 Hu et al. in the TGDMPP observed that comes will include postpartum weight loss, biomarkers
the lifestyle arm had 77.1% decrease in energy from fat associated with insulin resistance, other cardiovascular
as compared to a 68.9% decrease in the control arm (p = risk factors, and the adoption and maintenance of healthy
0.064).15 Reinhardt et al.19 in their 6-month pilot study in physical activity and dietary behaviors.
Australian women observed a change in total fat between Ferrara et al. are conducting the Gestational Diabetes’
arms of −19 g/day (95% CI: −37, −1). A beneficial impact of Effects on Moms study,24 a cluster randomized clinical trial of
the lifestyle intervention on fiber was observed by Hu et al. 44 medical facilities at Kaiser Permanente Northern California.
and Shyam et al. A total of 2320 women with a GDM diagnosis between March
2011 and March 2012 are randomized to either the inter-
vention or usual care conditions. Follow-up will continue to
Breastfeeding 12 months postpartum. The intervention is a DPP-derived
Only one study to date examined the impact of the lifestyle print/telephone lifestyle intervention. Primary outcomes will
intervention on breastfeeding among women with prior include the achievement of postpartum weight goals and total
GDM.14 Specifically, Ferrara et al. found that the difference weight change. Secondary outcomes include postpartum gly-
in percent of women partially or exclusively breastfeeding cemia, blood pressure, depression, percent of calories from fat,
between the intervention and control arms was 15% (p = 0.09) total caloric intake, and physical activity levels.
Infanti et al.25 are conducting Croi MyAction, a two-
group, parallel randomized controlled trial for women with
Newly launched randomized trials prior GDM. A total of 54 women with a history of GDM
and persistent postpartum glucose dysfunction are ran-
of lifestyle interventions to prevent domly assigned to an intervention group or to a control arm.
diabetes among women with GDM Follow-up will continue to 1 year postintervention. The life-
style intervention is delivered via in-person sessions includ-
The following section provides an overview of the recently ing one-on-one sessions, group exercise, and education
launched lifestyle intervention studies among women with programs. Primary outcomes include fasting plasma glucose
GDM. Five trials are currently ongoing (Table 50.3).22–24 levels on a 75 g oral glucose tolerance test.
Table 50.3 Newly launched randomized trials of lifestyle interventions to reduce risk of type 2 diabetes among women with gestational diabetes mellitus
Author Name Design Population Intervention Mode Goals Outcomes

Berry RCT; FU: 100 ethnically/racially Lifestyle Individualized Healthy behaviors Primary, FPG, BMI;
et al.22 10 months diverse women with intervention in person; secondary, clinical,
current GDM; North (diet and group adiposity, behaviors,
Carolina exercise) vs. sessions; text self-efficacy
control messaging
Chasan- Estudio Parto RCT; FU: 300 Hispanic women Lifestyle Individualized Weight, return to Primary, postpartum weight
Taber 12 months with current GDM; intervention in person; prepregnancy weight retention, biomarkers of
et al.23 Massachusetts (diet and telephone; if normal weight, lose insulin resistance, and
exercise) vs. mailings 5% of prepregnancy CVD; secondary, exercise,
health and weight if overweight; diet
wellness PA, 150 minute/week
control of moderate intensity
or harder; diet, ≤25%
calories from fat
Ferrara Gestational RCT; FU 2320 women with a Lifestyle (diet, Telephone; Weight, return to Primary, postpartum weight;
et al.14 Diabetes’ Effects 12 months GDM diagnosis exercise) vs. mailings prepregnancy weight secondary, postpartum
on Moms study between March usual care if normal weight, lose glycemia, blood pressure,
2011 and March control 5% of prepregnancy depression, % calories
2012; California weight if overweight from fat, total caloric
or obese; PA and diet, intake; PA levels
individualized goals
Infanti Croi MyAction RCT; FU: 54 women with Lifestyle Individualized Healthy lifestyle Primary (FPG), secondary
et al.25 12 months previous GDM and intervention in person; (insulin resistance, diet
persistent (diet and group adherence, weight, and
postpartum glucose exercise) vs. BMI), PA and fitness, lipid
dysfunction; Ireland control profile, psychological
factors
Shih Mothers After RCT; FU: 574 women with Lifestyle Individualized Weight, reduce >5%; Primary, diabetes risk;
et al.26 Gestational 12 months GDM in most recent intervention in person; PA ≥30 minute/day in secondary, psychosocial,
Diabetes in pregnancy; Australia (diet and group moderate intensity; QOL, CVD risk
Australia Diabetes exercise) vs. sessions; diet, fat intake <30%; Factors
Prevention control telephone fiber >15 g/1000 kcal
Program
Abbreviations: RCT, randomized clinical trial; FU, follow-up; GDM, gestational diabetes mellitus; IGT, impaired glucose tolerance; PA, physical activity; FG, fasting glucose; GI, glycemic index; LTPA,
leisure time physical activity; BMI, body mass index; kcal, kilocalories; QOL, quality of life; CVD, cardiovascular disease.
Newly launched randomized trials of lifestyle interventions to prevent diabetes among women with GDM 421
Shih et al. are conducting the Mothers After Gestational low-risk perceptions for future type 2 diabetes and subopti-
Diabetes in Australia Diabetes Prevention Program, a ran- mal levels of physical activity and fruit and vegetable intake.
domized controlled trial among 574 women with a diagnosis The majority of studies reveal a distinct knowledge–behavior
of GDM in their most recent pregnancy.26 Women are ran- gap among this population as well as a lack of knowledge
domized to 12-month diabetes prevention program or to a regarding necessary lifestyle modifications.30
usual care control group. The intervention is delivered via Promising strategies to address these challenges can be
in-person and group sessions and telephone. Follow-up confound in recent studies that have found a number of enablers
tinues for 12 months. Primary outcomes include incidence to postpartum lifestyle changes including social support.31
of diabetes, and secondary outcomes include cardiovascular Dasgupta et al.32 conducted focus groups among women
risk factors and psychosocial and quality of life factors. within 5 years of a GDM diagnosis. Participants stated that
their participation in a diabetes prevention program would
be enhanced by face-to-face interactions with professionals
Discussion and peers, provision of childcare support, and inclusion of
spouses/partners. Therefore, interventions that integrate the
Postpartum lifestyle interventions are critical in light of entire family and influence family members, along with the
recent findings from long-term follow-up studies that a sig- participant, to adopt health promoting behaviors may be
nificant proportion of women with GDM go on to develop particularly successful.
type 2 diabetes, especially during the first decade after the In this vein and to address transportation barriers, home-
index pregnancy.5 With the growing rates of diabetes and based interventions conducted via mail, telephone, the
obesity in US women, evidence regarding the effective- Internet/e-mail, and text messaging or involving home vis-
ness of lifestyle modification for the prevention of diabetes its by health educators may be more feasible and acceptable
in women with GDM is critical. Therefore, GDM offers an to women in the postpartum period. For example, several
important opportunity for the development, testing, and recent trials that relied largely on the Internet, mail, or tele-
implementation of clinical strategies for the prevention of phone have observed promising results.16,19,21 Furthermore,
subsequent type 2 diabetes.27 Such protocols can capital- there is evidence that Internet-based lifestyle interventions
ize on the teachable moment of pregnancy28 and empower can increase exercise in a general postpartum population.33
women to make postpartum lifestyle changes. However, at the same time, Internet access may be a barrier
To date, the majority of randomized controlled trials of as women with GDM tend to have lower socioeconomic sta-
lifestyle interventions in women with GDM designed to tus than women without GDM.34
prevent type 2 diabetes have been limited to pilot or feasibil- Weight loss interventions that begin during pregnancy
ity studies. However, preliminary findings suggest that such may be more effective than those initiated only in the post-
interventions can improve postpartum biomarkers of insu- partum period given the strong association between GWG
lin resistance and other diabetes risk factors in women with and postpartum weight retention and the fact that it may
a history of GDM. Specifically, the trials conducted to date be difficult to reduce postpartum weight retention with-
have observed favorable impacts on fasting glucose, insulin, out first preventing excessive gestational weight gain dur-
postpartum weight, leisure time physical activity, and intake ing pregnancy.35 Two of the published trials began in late
of total fat, fiber, and glycemic load. Only one study to date pregnancy14,19 and suggest that such protocols can have a
examined the impact on breastfeeding and found the sug- beneficial impact on gestational weight gain as well as pre-
gestion of a beneficial impact.14 pare women for postpartum changes. Consideration should
While evidence is rapidly accumulating that behavior in also be given to translating such programs to clinical care.
the postpartum period may be critical in the prevention of In their newly launched study,24 Ferrara et al. are evaluat-
longer-term progression toward diabetes, lifestyle changes ing whether delivering a diabetes prevention program at the
can be difficult to implement in these critical years after health system level is able to successfully reach women with
delivery. Postpartum women may be faced with the pres- prior GDM.
sures of caring for a new baby in addition to their existing Breastfeeding has been associated with reduced blood
household and caregiving responsibilities. Recent qualita- glucose levels and a reduced incidence of type 2 diabetes
tive data show that having young children is a major bar- among women with a history of GDM.36 However, only one
rier to an active lifestyle in the first 12 months postpartum.29 of the nine published trials included breastfeeding as one of
Other qualitative and quantitative studies indicate a number their goals.14 Future trials should focus on promoting a com-
of barriers to physical activity during postpartum, includ- bination of breastfeeding, diet, and physical activity.
ing physical discomfort, parenting duties, tiredness, lack Finally, rates of progression to type 2 diabetes vary by
of time, not prioritizing their health over other competing ethnicity, with, for example, Asian and Hispanic women
responsibilities, and lack of spousal/partner support.3 having higher rates of progression than non-Hispanic white
A second challenge to implementing behavior change women.37 At the same time, Hispanic women have higher
in the postpartum period is the relatively low perceived rates of being overweight/obesity when entering pregnancy
risk of future diabetes among women with recent GDM. A and are more likely to be sedentary than non-Hispanic white
review30 of studies examined the risk perceptions and health women.37 Future studies should focus on these high-risk
behaviors of women with previous GDM. The authors found groups. One newly launched study23 is being conducted in
References 423
Hispanic women only and will evaluate whether a culturally preventing subsequent progression to type 2 diabetes among
modified intervention will be effective in reducing diabetes women with GDM. Such interventions that focus on the
risk in Latinas with GDM. acquisition of healthy lifestyle skills in late pregnancy and
In summary, trials of diabetes prevention programs postpartum have a high potential for preventing the inter-
among high-risk pregnant and postpartum women are generational cycle of diabetes in this high-risk population.
sparse. Collectively, the prior body of evidence suggests that
such lifestyle interventions, if delivered to women with a
history of GDM, would have the potential to delay or pre- Acknowledgment
vent one-sixth of type 2 diabetes cases in the female popu-
lation.38 Larger, well-designed controlled randomized trials This work was funded by NIH/NIDDK 2R01DK064902 and
are needed to assess the effects of lifestyle interventions on DK097011.
REFERENCES
1. Metzger B, Gabbe S, Persson B et al. International association 18. Ratner RE, Christophi CA, Metzger BE et al. Prevention of dia-
of diabetes and pregnancy study groups recommendations on betes in women with a history of gestational diabetes: Effects of
the diagnosis and classification of hyperglycemia in pregnancy. metformin and lifestyle interventions. J Clin Endocrinol Metab
Diabetes Care 2010; 33(3): 676–682. 2008; 93(12): 4774–4779.
2. Anonymous. Standards of medical care in diabetes—2012. 19. Reinhardt JA, van der Ploeg HP, Grzegrzulka R, Timperley JG.
Diabetes Care 2012; 35 (Suppl. 1): S11–S63. Implementing lifestyle change through phone-based motiva-
3. Downs D, Chasan-Taber L, Evenson K. Obesity and physical tional interviewing in rural-based women with previous gesta-
activity during pregnancy and postpartum: Evidence, guide- tional diabetes mellitus. Health Promot J Austr 2012; 23(1): 5–9.
lines, and recommendations. In: Nicholson W, Baptiste-Roberts 20. Shyam S, Arshad F, Abdul Ghani R, Wahab NA, Safii NS, Nisak
K, eds., Obesity during Pregnancy in Clinical Practice, Springer: MY Chinna K, Kamaruddin NA. Low glycaemic index diets
London, U.K., 2014, pp. 183–227. improve glucose tolerance and body weight in women with pre-
4. Kim C. Maternal outcomes and follow-up after gestational diabe- vious history of gestational diabetes: A six months randomized
tes mellitus. Diabet Med 2014; 31(3): 292–301. trial. Nutr J 2013; 12: 68.
5. Bellamy L, Casas J, Hingorani A, Williams D. Type 2 diabetes 21. Wein P, Beischer N, Harris C, Permezel M. A trial of simple ver-
mellitus after gestational diabetes: A systematic review and meta- sus intensified dietary modification for prevention of progression
analysis. Lancet 2009; 373(9677): 1773–1779. to diabetes mellitus in women with impaired glucose tolerance.
6. Cheung NW, Byth K. Population health significance of gesta- Aust N Z J Obstet Gynaecol 1999; 39(2): 162–166.
tional diabetes. Diabetes Care 2003; 26(7): 2005–2009. 22. Berry DC, Neal M, Hall EG et al. Rationale, design, and meth-
7. Kim C, Newton KM, Knopp RH. Gestational diabetes and the odology for the optimizing outcomes in women with gesta-
incidence of type 2 diabetes: A systematic review. Diabetes Care tional diabetes mellitus and their infants study. BMC Pregnancy
2002; 25(10): 1862–1868. Childbirth 2013; 13: 184.
8. Retnakaran R. Glucose tolerance status in pregnancy: A window 23. Chasan-Taber L, Marcus BH, Rosal MC et al. Estudio Parto:
to the future risk of diabetes and cardiovascular disease in young Postpartum diabetes prevention program for hispanic women with
women. Curr Diabetes Rev 2009; 5(4): 239–244. abnormal glucose tolerance in pregnancy: A randomised controlled
9. Knowler WC, Barrett-Connor E, Fowler SE et al. Reduction in the trial—Study protocol. BMC Pregnancy Childbirth 2014; 14(1): 100.
incidence of type 2 diabetes with lifestyle intervention or metfor- 24. Ferrara A, Hedderson MM, Albright CL et al. A pragmatic clus-
min. N Engl J Med 2002; 346(6): 393–403. ter randomized clinical trial of diabetes prevention strategies
10. Tuomilehto J, Lindstrom J, Eriksson JG et al. Prevention of type for women with gestational diabetes: Design and rationale of
2 diabetes mellitus by changes in lifestyle among subjects the Gestational Diabetes’ Effects on Moms (GEM) study. BMC
with impaired glucose tolerance. N Engl J Med 2001; 344(18): Pregnancy Childbirth 2014; 14: 21.
1343–1350. 25. Infanti JJ, Dunne FP, O’Dea A, Gillespie P, Gibson I, Glynn LG,
11. Pan XR, Li GW, Hu YH et al. Effects of diet and exercise in pre- Noctor E, Newell J, McGuire BE. An evaluation of Croi MyAction
venting NIDDM in people with impaired glucose tolerance. The community lifestyle modification programme compared to
Da Qing IGT and Diabetes Study. Diabetes Care 1997; 20(4): standard care to reduce progression to diabetes/pre-diabetes in
537–544. women with prior gestational diabetes mellitus (GDM): Study
12. Chiasson JL, Josse RG, Gomis R et al. Acarbose for prevention protocol for a randomised controlled trial. Trials 2013; 14(1): 121.
of type 2 diabetes mellitus: The STOP-NIDDM randomised trial. 26. Shih ST, Davis-Lameloise N, Janus ED et al. Mothers after ges-
Lancet 2002; 359(9323): 2072–2077. tational diabetes in Australia Diabetes Prevention Program
13. Cheung NW, Smith BJ, van der Ploeg HP, Cinnadaio N, Bauman (MAGDA-DPP) post-natal intervention: Study protocol for a ran-
A. A pilot structured behavioural intervention trial to increase domized controlled trial. Trials 2013; 14: 339.
physical activity among women with recent gestational diabetes. 27. Buchanan TA. What is gestational diabetes? Diabetes Care 2007;
Diabetes Res Clin Pract 2011; 92(1): e27–e29. 30 (Suppl. 2): S105.
14. Ferrara A, Hedderson M, Albright C et al. A pregnancy and post- 28. Phelan S. Pregnancy: A “teachable moment” for weight control and
partum lifestyle intervention in women with gestational diabetes obesity prevention. Obstet Gynecol 2010; 202(2): 135.e1–135.e8.
mellitus reduces diabetes risk factors: A feasibility randomized 29. Nicklas JM, Zera CA, Seely EW, Abdul-Rahim ZS, Rudloff ND,
control trial. Diabetes Care 2011; 34(7): 1519–1525. Levkoff SE. Identifying postpartum intervention approaches to
15. Hu G, Tian H, Zhang F et al. Tianjin Gestational Diabetes prevent type 2 diabetes in women with a history of gestational
Mellitus Prevention Program: Study design, methods, and 1-year diabetes. BMC Pregnancy Childbirth 2011; 11: 23.
interim report on the feasibility of lifestyle intervention program. 30. Jones EJ, Roche CC, Appel SJ. A review of the health beliefs and
Diabetes Res Clin Pract 2012; 98(3): 508–517. lifestyle behaviors of women with previous gestational diabetes.
16. Kim C, Draska M, Hess ML, Wilson EJ, Richardson CR. A web- J Obstet Gynecol Neonatal Nurs 2009; 38(5): 516–526.
based pedometer programme in women with a recent history of 31. Evenson K, Aytur S, Borodulin K. Physical activity beliefs, barri-
gestational diabetes. Diabet Med 2012; 29(2): 278–283. ers, and enablers among postpartum women. J Womens Health
17. McIntyre HD, Peacock A, Miller YD, vette Koh D, Marshall AL. 2009; 18(12): 1925–1934.
Pilot study of an individualised early postpartum intervention to 32. Dasgupta K, Da Costa D, Pillay S et al. Strategies to optimize par-
increase physical activity in women with previous gestational ticipation in diabetes prevention programs following gestational
diabetes. Int J Endocrinol 2012; 2012: 892019. diabetes: A focus group study. PLOS ONE 2013; 8(7): e67878.
33. Lewis BA, Martinson BC, Sherwood NE, Avery MD. A pilot study 36. Kjos SL, Henry O, Lee RM, Buchanan TA, Mishell DR. The
evaluating a telephone-based exercise intervention for pregnant effect of lactation on glucose and lipid metabolism in women
and postpartum women. J Midwifery Womens Health 2011; with recent gestational diabetes. Obstet Gynecol 1993; 82(3):
56(2): 127–131. 451–455.
34. Kim C, Sinco B, Kieffer EA. Racial and ethnic variation in access 37. Chasan-Taber L. Physical activity and dietary behaviors associ-
to health care, provision of health care services, and ratings of ated with weight gain and impaired glucose tolerance among
health among women with histories of gestational diabetes mel- pregnant Latinas. Adv Nutr 2012; 3(1): 108–118.
litus. Diabetes Care 2007; 30(6): 1459–1465. 38. Bentley-Lewis R, Levkoff S, Stuebe A, Seely EW. Gestational dia-
35. Ferrara AE, Ehrlich SF. Strategies for diabetes prevention before betes mellitus: Postpartum opportunities for the diagnosis and
and after pregnancy in women with GDM. Curr Diabetes Rev prevention of type 2 diabetes mellitus. Nat Clin Pract Endocrinol
2011; 7(2): 75–83. Metab 2008; 4(10): 552–558.
51 Can fetal macrosomia be
predicted and prevented?
Maria Farren and Michael Turner
such as cesarean section, shoulder dystocia, brachial plexus

Introduction injuries, and neonatal morbidity. Based on the finding of
There is no consensus about the definition of macroso- this study, a grading system for macrosomia was proposed.
mia. Some authors prefer large for gestational age (LGA) Grade I macrosomia was defined as 4000–4499 g, grade II
(>90th%) and others birth weights (BW) of ≥4.0 kg, ≥4.1 kg, macrosomia as 4500–4999 g, and grade III macrosomia as
or ≥4.5 kg.1–3 The American College of Obstetricians and ≥5000 g. Labor complications, birth injuries, and newborn
Gynecologists (ACOG) defines macrosomia as an infant morbidity increased with increasing gradation. It was found
with a BW of ≥4.5 kg irrespective of gestational age or other that perinatal mortality was increased in neonates ≥5000 g.6
demographic characteristics.4 The use of centiles depends on
the pregnancy being accurately dated by ultrasound, and this
may not be possible in resource-poor settings. Macrosomia Incidence
based on BW >4.0 kg in our hospital would include 12%–13%
of neonates.5 Adverse clinical consequences in this group The incidence of macrosomia is widely reported to be
are uncommon. This definition makes little sense because increasing.16,17 However, some of this evidence is more than
adverse consequences for a BW of 4.0–4.5 kg may occur, and 20 years old. The incidence of fetal macrosomia in our unit
any interventions have not been shown to improve clinical has decreased over 20 years whether the definition used is
outcomes. either ≥4.0 or ≥4.5 kg (Figure 51.1).
For the purpose of this chapter, we are defining fetal In the Republic of Ireland, the incidence of macrosomia
macrosomia as a BW of ≥4.5 kg. This is based on studies has not increased over the last 10 years. The rate of the mac-
showing increased fetal and maternal trauma at the delivery rosomia (infants ≥ 4.5 kg) has been 2.5%–2.9% (Figure 51.2).
of a baby weighing ≥4.5 kg.6 The definition of macrosomia as The mean BW for singleton births in the Irish population
≥4.5 kg represents 2%–3% of our neonatal population. The from 1990 to 2012 shows little variation (Figure 51.3).
4.5 kg limit is only appropriate for term babies and can be Similarly in the United States, the incidence of macroso-
criticized for being arbitrary as the incidence of shoulder mia across all subgroups is decreasing (Figure 51.4), despite
dystocia rises between BWs of 4.0 and 4.25 kg. It could be rising obesity rates.18
argued that 4.25 kg be used as the cutoff value.7,8 However, The incidence of macrosomia in Scandinavia may be
the ≥4.5 kg limit defines unequivocally and unambiguously increasing. This is reflected in studies from Sweden and
a high-risk group that may benefit from accurate diagnosis Denmark. In Sweden in 2001, a large retrospective study
and management over the peripartum period.8 reviewed all birth records from 1992 and 2001. This included
The macrosomic fetus is at risk of perinatal complications 874,163 live, term, singleton births. They found that mean
such as shoulder dystocia, brachial plexus injury, clavicu- BW increased from 3596 to 3631 g and the percentage of
lar fracture, and meconium aspiration.3,9–12 In the neona- infants weighing >4500 g or more increased from 3.7% to
tal period, macrosomic infants are at risk of hypoglycemia, 4.6%.19 This study cites an increasing average maternal BMI
hyperbilirubinemia, and hypomagnesemia.13 The mother of and decreasing smoking rate over the same time period as
a macrosomic infant is at increased risk of prolonged labor, contributing factors to this increase. However, the increase
operative vaginal delivery, perineal trauma, and cesarean in mean BMI is hardly significant clinically and may reflect
section.9,12,14,15 the timing of obstetric interventions at term, e.g., cesarean
The effect of BW on neonatal risk, morbidity, and mortal- section or induction of labor.
ity was studied in a retrospective review in Birmingham, AL, In Denmark a retrospective study in 2001 reviewed all
between 1995 and 1997.6 From this review, it was concluded term singleton deliveries from 1990 to 1999. This included 26,
that babies ≥4.5 kg were at greater risk of labor complications 392 infants. Similarly, they found that mean BW increased,
425
426 Can fetal macrosomia be predicted and prevented?
16
14
Percentage of neonates
12
10
6 ≥4.5 g 4.0–4.5 kg
0
1993 1995 1997 1999 2001 2003 2005 2007 2009 2011
Year
Figure 51.1 Birth weights in 1992–2012, Coombe Women and Infants University Hospital.
16
14
12
10
Percent
8 4000–4499 kg ≥4500 kg
0
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Year
Figure 51.2 Percentage of babies born in the Republic of Ireland in 2002–2012, ESRI/NPRS Perinatal Statistics Report 2003–2012.
3530
3520
3510
3500
Birth weight (kg)
3490
3480
3470
3460
3450
3440 Singletons Total
3430
1990 1993 2000 2002 2004 2006 2008 2010 2012
Year
Figure 51.3 Mean birth weight in Ireland, ESRI/NPRS Perinatal Statistics Report.
as did the proportion of infants weighing 4000–4249 g, Overall, it is observed that the incidence of macrosomia
4250–4499 g, and 4500 g or more.20 However, it is worth has remained stable over the last 10–15 years. Anecdotally
noting that within their study, only 60.2% had their preg- babies are getting bigger. However, with the possible excep-
nancies dated by early pregnancy ultrasound. The remaining tion of Scandinavia, this is not the case as is reflected in
subjects had their pregnancies dated on the basis of the last our analysis at local, national, and international level.
menstrual period or on the gestational age recorded as book- Commentary in the literature is fraught with discrepancies,
ing. This may explain the apparent increase in BW. assumptions, and inaccurate measurements of gestation at
Risk factors 427
10 For a variety of reasons, the rate of induction of labor

9 at term is increasing in both primiparous and multiparous
women.26,27 Induction before term will result in the delivery
8
of an infant much smaller than if that infant had been left to
7 deliver at term or beyond, in the postdates period.
6 Indeed, the suspicion of macrosomia may be given as a
reason for induction of labor. This diagnosis, often made on
Percent
5 clinical examination or ultrasound, is fraught with error

4000–4499 kg
4 and interobserver variation. However, it can drive clinicians
4500–4999 kg
3
to plan an induction before term, again resulting in the
5000 kg delivery of a lighter infant than would have been delivered
2 after term.
1
0
1996 1998 2000 2002 2009 2012 Risk factors
Year
The risk factors for macrosomia are either modifiable or
Figure 51.4 Percentage of singleton births in all macroso- unmodifiable (Table 51.1). Unmodifiable risk factors include
mic groups in the United States in 1996–2012, National Vital maternal age, parity, ethnicity, parental height, gender of the
Statistics, CDC.
fetus, and a history of a previous LGA infant.
The effect of maternal age was demonstrated in a retro-
recruitment and delivery. These discrepancies illustrate the spective American review in 1985.15 A total of 574 infants
importance of accurate dating of a pregnancy so that any weighing <4500 kg were compared to a control sample of
diagnosis of macrosomia is not erroneous. 18,739 infants weighing 2500–3499 kg. This study found
The reason for any decreasing trends in fetal macrosomia that women delivering macrosomic infants were older and
is multifactorial. The increasing rate of assisted reproduction of higher parity.
has led to an increased number of twin and multiple preg- A prospective observational study in Philadelphia
nancies.21 Such an increase of multiple pregnancies leads to between 1991 and 1994 explored the effect of ethnicity on
a decrease in the overall BW as multiple pregnancies tend macrosomia.28 Macrosomia was defined as a BW >90th cen-
to be delivered earlier and overall weigh less than their sin- tile. Within the cohort, 103 American and 36 Latino subjects
gleton counterparts. were followed prospectively. All subjects had been diag-
The diagnosis and management of gestational diabe- nosed with GDM. They found ethnicity is an independent
tes mellitus (GDM) also may account for the plateau in risk factor for fetal macrosomia. However, ethnic variation
the BWs. The criteria for screening remain controversial, in the United States is difficult to interpret as there is such
but even with selective screening, more women are being wide ethnic intermingling.
diagnosed with GDM.22 The Hyperglycemia and Adverse The effect of parental height was extrapolated from the
Pregnancy Outcome (HAPO) trial revealed the effects of Millennium Cohort Study in the United Kingdom between
mild hyperglycemia on fetal size and has subsequently led 2000 and 2002.29 There were 8053 infants studied. It showed
to changes in the parameters once considered for the diag- maternal weight was more influential than paternal weight
nosis of GDM.23 on infant BW.
Larger numbers of women being screened for GDM may Women with a history of fetal macrosomia are at risk of
lead to earlier intervention and this improves glycemic con- delivering another macrosomic infant. A retrospective study
trol. The control of blood glucose levels can be achieved in Dublin between 1998 and 1999 reviewed 14,461 pregnan-
through diet, oral hypoglycemics, and subcutaneous insulin. cies.30 From this cohort, 529 infants (3.7%) were macrosomic,
The overall effect of treatment modalities is stricter glycemic with the incidence higher in parous women (4.6%) compared
control and subsequent reduction in BW.24 This may also
explain the plateau in the number of macrosomic infants. In
the United States, there is a trend toward universal screen-
Table 51.1 Risk factors of fetal macrosomia
ing for GDM. Therefore, even greater numbers of women are
diagnosed with GDM, and their pregnancies are managed
ideally within strict glycemic control.24 Unmodifiable Modifiable
In addition to the regulation of hyperglycemia, the preg- Maternal age Hyperglycemia
nancy complicated by GDM is unlikely to progress past 38 Parity Hypertriglyceridemia
weeks gestation especially in the case of insulin-dependent Ethnicity Gestational age at birth
GDM.25 As a result, these fetuses are not being given the Parental height Maternal weight
opportunity for a growth surge in the latter part of preg- Gender of the fetus Gestational weight gain
nancy and in the postdates period. This may impact on the History of LGA infant Dysfunctional lifestyle
overall percentage of macrosomic infants.
with nulliparas (2.4%, p < 0.0001). In the following 5 years, to the accuracy of ultrasound estimation. However, this
164 women went on to deliver a macrosomic fetus. They study considered those women in the general obstetric pop-
found that women with a history of one macrosomic fetus ulation. The accuracy of maternal estimates of BW among
are at increased risk of another macrosomic fetus in a sub- women with GDM has not been studied.
sequent pregnancy (odds ratio [OR] 15.8, 95% confidence Ultrasound has been extensively studied for sensitivity
interval (CI) 11.45–21.9, p < 0.001). For women with two or and specificity in detecting fetal macrosomia. Twenty-two
more macrosomic fetuses, the risk is greater (OR 47.4, 95% articles were reviewed in 2005.36 The inclusion criteria were
CI 19.9–112.9, p < 0.001). any study that considered the sensitivity and specificity of
While all the previous factors are factored in, there are ultrasound-estimated fetal weight to correctly identify a
influences on BW that can be modified. These are of great macrosomic fetus. The authors considered macrosomia as
interest to the clinician, for it is these that we can focus on in fetal weight ≥4000 g. The majority of the reports that met
an effort to reduce the adverse maternal and fetal outcomes the inclusion criteria were from the United States. Of the
associated with the macrosomic infant. Modifiable risk fac- 22 reports, 14 were in the general obstetric population, 4
tors may include maternal hyperglycemia, increased gesta- included pregnancies complicated by GDM, and 4 included
tional age at delivery, increased maternal weight, increased postdate pregnancies.36
gestational weight gain, and an unhealthy lifestyle. The incidence of macrosomia among the general obstetric
The link between maternal and fetal hyperglycemia was population (defined as BW >4000 g) ranged from 3% to 55%.
first described in Copenhagen in 1952.31 Maternal hypergly- The sensitivity and specificity in the detection of macrosomia
cemia leads to fetal hyperglycemia as glucose crosses the plain this group was wide ranging. The sensitivity ranged from
centa and insulin does not.32 Glucose is the main substrate 70% to 99%, while the specificity ranged from 12% to 79%.
for growth in the fetus.33 In babies of mothers with diabe- The time span of the articles included was 10 years (1993–
tes, the effect of hyperglycemia is apparent as their BWs are 2003). The review considered the possibility that a variation
increased. A prospective study in the northern region of the in the regression equation used by each study to predict BW
United Kingdom in 1994 enrolled 113 women with preex- influenced the results. Hadlock’s proposed formula was used
isting diabetes mellitus. The study found that 35% of those by 57% of the reports.38 The post-test probability using this
delivered (36 out of 104 deliveries) had a BW >95th centile.34 equation ranged from 17% to 76%. Of the 14 studies, 3 used
In 2008, the HAPO study found that continuous exposure the equation proposed by Shepard with a post-test probabil-
by the fetus to glucose, at levels below those diagnostic of ity ranging from 16% to 32%, and 1 study used various equa-
diabetes, had a variety of adverse effects on the pregnancy tions with a post-test probability ranging from 27% to 47%.39
including BW >90th centile. It concluded that the equation chosen was not a factor in the
Gestational age at delivery is an important risk factor for inconsistent results.
macrosomia. An observational study in California in 1997 The review found that despite advances in equipment and
studied the BW of 326 singleton fetuses born between 37 and expertise, our ability to predict the macrosomic infant has
42 weeks gestation. They ensured the pregnancies were accu- not improved over the decade. The expertise of the person-
rately dated and controlled for obesity, ethnicity, and life- nel conducting the ultrasound examination also does not
style factors, e.g., smoking. The study found that BW was a influence the accuracy of prediction of fetal macrosomia.40
linear function of gestational age between 37 and 42 weeks.35 A retrospective review in California recruited 365 women
Despite this, however, studies have shown no benefit to over a 7-month period. The prevalence of macrosomia (BW
induction of labor or cesarean section in nondiabetic moth- ≥ 4000 g) was 12%. The prediction of BW in the growth-
ers with suspected fetal macrosomia.2 restricted fetus and the appropriately grown fetus was more
accurate among sonographers. However, both sonographers
and fetomaternal specialists had similar accuracy in detect-
Prediction ing the macrosomic fetus. The post-test probability to detect
the macrosomic fetus was 53% and 56%, respectively.
There has long been an interest in accurate methods or mod- The ability to predict macrosomia in the general obstetric
els for the prediction of fetal macrosomia. Clinical assessment population is fraught with inaccuracy. However, evidence
has been the longest standing method. This involves subjec- suggests it is feasible to predict macrosomia in pregnancies
tive clinical palpation of the abdomen and measurement of complicated by diabetes mellitus and in postdate pregnan-
the symphysiofundal height (SFH). These methods are fraught cies. It is most likely that the reason for greater detection is
with inaccuracies as interobserver variation occurs. The mea- the higher prevalence of macrosomia among these groups.
surement of SFH does not allow for variables such as polyhy- The post-test probability of detection of a macrosomic fetus
dramnios, uterine anomalies, or m ultiple pregnancies.36 ranged from 71% to 81% in those with GDM and 61%–63%
Some studies suggest that maternal prediction of mac- in postdate pregnancies.41–44 These studies defined macroso-
rosomia is accurate. A report in 1995 asked 70 postdates mia as ≥4000 g.
women to predict the fetal weight of their baby.37 Among the Neonatal morbidity due to birth trauma is more likely to
cohort, the sensitivity was 56% and the specificity was 94%. occur when the BW is ≥4500 g.8 Therefore, it is prudent to
The pre-test probability of macrosomia was taken to be 20%, consider if macrosomia can be predicted—either clinically
and therefore, the post-test probability was 70%, comparable or sonographically—in this select group of patients. In the
Prevention 429
literature, reviews suggest a post-test probability of detection to reduce and also weight gain in the current pregnancy,
of macrosomia in infants ≥4500 g varies from 22% to 37% if BMI is to be optimized.
on ultrasound and 12% to 36% on clinical examination.2,45–47 Maternal obesity may influence fetal macrosomia.
Present evidence indicates that it remains difficult to predict However, possibly of more importance is the concept of
infants who weigh ≥4500 g either clinically or on ultrasound. maternal body composition. Considering the overall body
Serum biomarkers that are sensitive and specific could composition and not just BMI may be a more appropriate
potentially predict fetal macrosomia. This area remains in its focus for both the prediction and prevention of fetal mac-
infancy, and the results to date show only a relatively small rosomia. In a prospective observational study, our research
improvement in overall prediction of macrosomia.48 Low team studied the correlation between fetal BW and mater-
pregnancy-associated plasma protein A and increased beta nal body composition.53 The study included 2618 subjects of
human chorionic gonadotropin levels have been shown to be whom 16.5% were obese. We concluded there was no rela-
predictive in pregnancies complicated by growth restriction. tionship between maternal fat mass and fetal macrosomia.
However, the reverse has not been proven true when it comes However, fat-free mass was a strong predictor of BW ≥4.0 kg.
to the prediction of macrosomia. The use of biomarkers for Women who were in the highest fat-free mass quartile had
predicting macrosomia, from as early as the first trimester, an odds ratio of 3.6 for a BW >4000 g compared to those in
has been investigated but with only marginal increase in the lowest quartile (Figure 51.5).
detection.48 These markers were used in combination with The impact of body composition on BW is further dem-
maternal risk profile. onstrated by work carried out in our unit. This prospective
study recruited 368 women across all BMI categories. The
BW of babies was recorded and compared against maternal
body composition. This shows the range of fetal BW across
Prevention all BMI categories.
The root of macrosomia is multifactorial. Therefore, there Obstetric interventions are made in a bid to reduce the
are many aspects that can be considered when it comes to incidence and impact of fetal macrosomia. It has previously
preventing this obstetric challenge. been postulated that elective cesarean section or induc-
The effect of maternal hyperglycemia on fetal macroso- tion of labor may avoid the adverse outcomes of macroso-
mia is established. Glucose crosses the placenta, while insulin mia. However, neither approach has shown clear benefit in
does not. This results in fetal hyperglycemia and a subsequent women with otherwise uncomplicated pregnancies.
hyperinsulinemia and fetal macrosomia.49 In a pregnancy A retrospective analysis in 1983 reviewed the outcomes
complicated by diabetes, there is central deposition of subcuta- for infants with a BW > 4000 g.54 They concluded that mac-
neous fat in the abdominal and interscapular areas. However, rosomia was rare at 37 weeks gestation and increasingly
the skeletal growth is unaffected.50 A prospective study of 479 more common thereafter. They advocated ultrasound analy-
healthy, nondiabetic mother and baby pairs concluded that sis between 36 and 38 weeks gestation and subsequent induc-
the effect of maternal hyperglycemia on fetal growth is mainly tion of labor for suspected macrosomia.
related to fat deposition.30 Therefore, it stands to reason that It must also be considered that induction of labor in itself
control of hyperglycemia is essential to prevent macrosomia. increases the rate of cesarean section. Eleven studies, both
There is some evidence also that altering the carbohydrate observational and randomized, which compared expectant
type consumed from high to low glycemic index changes glu-
cose and insulin response, in turn altering fetal weight gain.51
The impact of maternal obesity on fetal BW may be a factor 5.00
to be considered in preventing macrosomia. A retrospective
observational study in Cleveland reviewed 12,950 deliveries 4.50
from 1997 to 2001 to evaluate whether an abnormal body
habitus contributed to the birth of a macrosomic infant.52 4.00
Birth weight
The subjects were classified as underweight, normal, over-

weight, or obese (defined as >30 kg/m2). Obesity was not 3.50
further classified into subgroups. Macrosomia was defined
as BW >90th centile. They concluded that obese women had 3.00
an increased risk of macrosomia (16.8% vs. 10.5%, p < 0.001)
as were overweight women (12.3% vs. 10.5%, p < 0.01). The 2.50
rate of abnormal body habitus is steadily rising, particularly
in the obesity classes. This may impact on the rate of macro- 2.00
somia and, in doing so, will increase the morbidity associ-
Underweight Normal Overweight Obese 1 Obese 283
ated with macrosomia for both mother and baby. There are
also uncertainties about whether any epidemiological asso- BMI category
ciations between obesity and macrosomia may be genetic or Figure 51.5 The relationship between BMI and birth weight.
may be due to concomitant GDM. Ideally, it is prepregnancy (From Higgins, O. et al., Ir. Med. J., 107, 10, November 2014.
or interpregnancy weight gain that we should be aiming With permission.)
management versus induction of labor when fetal macroso- contentious. Some advise a cutoff value of >4000 g, others
mia was suspected were reviewed.55 These 11 studies included advise >4250 g, and the ACOG recommends >5000 g.56,58,59
3751 subjects of which 2700 were managed expectantly and It remains reasonable that cesarean section should be
1051 underwent labor induction. Analysis showed that, com- discussed in the context of GDM and suspected fetal mac-
pared with those whose labor was induced, women who expe- rosomia >4500 g.
rienced spontaneous onset of labor had a lower incidence of
cesarean section (OR 0.39, 95% CI 0.30, 0.50) and higher rates
of spontaneous vaginal delivery (OR 2.07, 95% CI 1.34, 3.19). Conclusion
No differences were noted in rates of operative vaginal deliv-
ery, shoulder dystocia, or abnormal Apgar scores in the analy- Macrosomia remains a contentious and controversial topic
ses of the observational or randomized studies.55 in obstetric practice. There is little doubt that a pregnancy
Pregnancy complicated by GDM is an independent risk complicated by macrosomia leads to a greater risk of mater-
factor for neonatal morbidity. Shoulder dystocia was more nal and neonatal morbidity. It is, therefore, a pregnancy
common in diabetic women.56 Within this study, it was con- complication that we strive to both predict and prevent.
cluded that the combination of diabetes and a macrosomic However, the effects are wide reaching. Our prediction tools
infant, with macrosomia defined as ≥4000 g, accounted for at present are ineffective.
73% of shoulder dystocia cases among women with GDM. The suspicion of macrosomia heightens both maternal
Therefore, it was concluded that cesarean section was indi- and clinician anxiety and often leads to more investigation
cated in those with GDM and an estimated fetal weight and more intervention than in the pregnancy where fetal
≥4000 g.56 A similar increased risk of brachial plexus injury growth is considered to be normal. The challenges facing us
associated with GDM has also been demonstrated.57 when it comes to macrosomia are wide ranging. It is hoped
The complications associated with macrosomia in those that through greater research and an agreed definition of
with GDM are recognized and cesarean delivery in these macrosomia, the tools for prediction and the strategies for
cases has been widely recommended.56–58 However, the prevention will become more precise and widely used among
cutoff value at which cesarean section is advised remains clinicians worldwide.
REFERENCES
1. Johnstone FD, Prescott RJ, Steel JM et al. Clinical and ultrasound 16. Power C. National trends in birth weight: Implications for future
prediction of macrosomia in diabetic pregnancy. Br J Obstet adult disease. Br Med J 1994; 308: 1270–1271.
Gynaecol 1996; 103: 747–754. 17. Boneille SR. Why are babies getting heavier? Comparison of
2. Gonen O, Rosen DJ, Dolfin Z. Induction of labour versus expect- Scottish births from 1980 to 1992. Br Med J 1997; 315: 1205.
ant management in macrosomia: A randomized study. Obstet 18. Finucane MM, Stevens GA, Cowan MJ et al. National, regional,
Gynecol 1997; 89: 913–917. and global trends in body-mass index since 1980: Systematic
3. Golditch IM, Kirkham K. The large fetus, management and out- analysis of health examination surveys and epidemiological stud-
come. Obstet Gynecol 1978; 52: 26–30. ies with 960 country-years and 9·1 million participants. Lancet
4. American College of Obstetricians and Gynecologists. Fetal mac- 2011; 377: 557–567.
rosomia. Technical Bulletin 159, ACOG: Washington, DC, 1991. 19. Surkan PJ, Hsieh CC, Johansson AL et al. Reasons for increasing
5. Perinatal Statistics Report, 2012. Health Research and Information trends in large for gestational age births. Obstet Gynecol 2004;
Division, ESRI, November 2013. 104: 720–726.
6. Boulet SL, Alexander GR, Salihu HH et al. Macrosomic births in 20. Orskou J, Ulrk K, Henriksen TB et al. An increasing proportion
the United States: Determinants, outcomes and proposed grades of infants weigh more than 4000 grams at birth. Acta Obstet
of risk. Am J Obstet Gynecol 2003; 188: 1372–1378. Gynecol Scand 2001; 80: 931–936.
7. Nesbitt TS, Gilbert WM, Herrchen B et al. Shoulder dystocia 21. Blondel B, Kogan MD, Alexander GR et al. The impact of the
and associated risk factors with macrosomic infants born in increasing number of multiple births on the rates of preterm birth
California. Am J Obstet Gynecol 1998; 179(2): 476–480. and low birthweight. Am J Public Health 2002; 92(8): 1323–1330.
8. Campbell S. Fetal macrosomia: A problem in need of a policy. 22. Ali FM, Farah N, O’Dwyer V et al. The impact of the new national
Ultrasound Obstet Gynecol 2014; 43: 3–10. guidelines on screening for gestational diabetes mellitus. Ir Med J
9. Walker D. The dangers of fetal macrosomia. J Obstet Gynecol 2014; 107: 3.
1985; 66: 158–161. 23. HAPO Study Cooperative Research Group; Metzger BE, Lowe LP,
10. Rossi C, Mullin P, Prefumo F et al. Prevention, management and Dyer AR et al. Hyperglycemia and adverse pregnancy outcomes.
outcomes of macrosomia: A systematic review of literature and N Engl J Med 2008; 358: 1991–2002.
meta-analysis. Obstet Gynecol Surv 2013; 68: 702–709. 24. Cosson E, Benchimol M, Carbillon L et al. Universal rather than
11. Walsh JM, Kandamany N, Ni Shuibne N et al. Neonatal brachial selective screening for gestational diabetes mellitus may improve
plexus injury: Comparison of incidence and antecedents between fetal outcomes. Diabetes Metab 2006; 32: 140–146.
2 decades. Am J Obstet Gynecol 2011; 204: 324.e1–324.e6. 25. Casey B, Lucas M, McIntire DD et al. Pregnancy outcomes in
12. Modanlou HD, Dorchester WL, Thorosian A et al. Macrosomia- women with gestational diabetes compared with the general
maternal, fetal and neonatal implications. Obstet Gynecol 1980; obstetric population. Obstet Gynecol 1997; 90: 869–873.
55: 420–424. 26. Coombe Women and Infants University Hospital Annual Clinical
13. Nold JL, Georgieff MK. Infants of diabetic mothers. Prediatr Clin Report 2012.
North Am 2004; 51: 619–637. 27. Rayburn WF, Zhang J. Rising rates of labor induction: Present
14. Turner MJ, Rasmussen MJ, Turner JE et al. The influence of birth concerns and future strategies. Obstet Gynecol 2002; 100:
weight on labour in nulliparas. Obstet Gynecol 1990; 76: 79–83. 164–167.
15. Spellacy WN, Miller S, Winegar A et al. Macrosomia-maternal 28. Homko C, Sivan E, Nyirjesy P et al. The interrelationship between
characteristics and infant complications. Obstet Gynecol 1985; ethnicity and gestational diabetes in fetal macrosomia. Diabetes
66: 158–161. Care 1995; 18: 1442–1445.
References 431
29. Griffiths L, Dezateux C, Cole TJ. Differential parental weight and 44. Pollock RN, Hauer-Pollack G, Divon HV et al. Macrosomia in
height contributions to offspring birthweight and weight gain in postdates pregnancies: The accuracy of routine ultrasonographic
infancy. Int J Epidemiol 2007; 36: 104–107. screening. Am J Obstet Gynecol 1992; 167: 7–11.
30. Walsh CA, Mahony RT, Foley ME et al. Recurrence of fetal 45. Chauhan SP, Hendrix NW, Magann EF et al. Limitation of clinical
macrosomia in non-diabetic pregnancies. J Obstet Gynaecol
and sonographic estimate of birth weight: Experience with 1034
2007; 27: 374–378. parturients. Obstet Gynecol 1998; 91: 72–77.
31. Pedersen J. Diabetes and pregnancy blood sugar of new- 46. Smith GC, Smith MF, Mc Nay MB et al. The relation between
born infants. PhD thesis, Danish Science Press: Copenhagen, fetal abdominal circumference and birth weight: Findings in 3512
Denmark, 1952. pregnancies. Br J Obstet Gynaecol 1997; 104: 186–190.
32. Boskovic R, Feig DS, Derewlany L et al. Transfer of insulin lispro 47. O Reilly-Green CP, Divon MY. Receiver operating characteris-
across the human placenta: In vitro perfusion studies. Diabetes tic curves of sonographic estimated fetal weight for prediction
Care 2003; 26: 1390–1394. of macrosomia in prolonged pregnancies. Ultrasound Obstet
33. Moses RG, Luebcke M, Davis WS et al. Effect of a low glycemic- Gynecol 1998; 91: 72–77.
index diet during pregnancy on obstetric outcomes. Am J Clin 48. Nanda S, Akoledar R, Mulhaisen M et al. Maternal serum adipo-
Nutr 2006; 84: 807–812. nectin at 11 to 13 weeks of gestation in the prediction of macro-
34. Hawthorne G, Robson S, Ryall EA et al. Prospective popula- somia. Prenat Diagn 2011; 31: 479–483.
tion based survey of outcome of pregnancy in diabetic women: 49. Hellerstrom C, Swenne I. Functional maturation and proliferation
Results of the Northern Diabetic Pregnancy Audit, 1994. Br Med of fetal pancreatic beta cells. Diabetes 1991; 40: 89–93.
J 1997; 315(7103): 279–281. 50. McFarland MB, Trylovich CG, Langer O et al. Anthropometric
35. Nahum GG, Stanislaw H, Huffaker BJ et al. Fetal weight gain at differences in macrosomic infants of diabetic and nondiabetic
term: Linear with minimal dependence on maternal obesity. Am mothers. J Matern Fetal Neonatal Med 1998; 7: 292–295.
J Obstet Gynecol 1995; 172: 1387–1394. 51. Clapp 3rd JF. Maternal carbohydrate intake and pregnancy out-
36. Chauhan SO, Grobman WA, Gherman RA et al. Suspicion come. Proc Nutr Soc 2002; 61: 45–50.
and treatment of the macrosomic fetus: A review. Am J Obstet 52. Ehrenberg H, Mercer B, Catalano PM et al. The influence of obe-
Gynecol 2005; 193: 332–346. sity and diabetes on the prevalence of macrosomia. Am J Obstet
37. Chauhan SP, Sullivan CA, Sharman RS et al. Parous patients’ esti- Gynecol 2004; 191: 964–996.
mate of birth weight in postterm pregnancy. J Perinatol 1995; 15: 53. Kent E, O’Dwyer V, Fattah C et al. Correlation between birth
192–194. weight and maternal body composition. Obstet Gynecol 2013;
38. Hadlock FP, Harrist RB, Lutton TC et al. Estimation of fetal weight 121: 46–50.
with the use of head, body and femur measurements: A prospec- 54. Boyd ME, Usher RH, McLean FH et al. Fetal macrosomia:
tive study. Am J Obstet Gynecol 1985; 155: 333–337. Prediction, risks, proposed management. Obstet Gynecol 1983;
39. Nahum GG, Stanislaw H, Huffaker BJ et al. Accurate prediction 61: 715–722.
of term birth weight from prospectively measurable maternal 55. Sanchez-Ramos L, Bernstein S, Kaunitz AM et al. Expectant man-
characteristics. J Reprod Med 1999; 44: 705–712. agement versus labor induction for suspected fetal macrosomia:
40. Humphries J, Reynolds D, Bell Scarbrough L et al. Sonographic A systematic review. Obstet Gynecol 2002; 100: 997–1002.
estimate of birth weight: Relative accuracy of sonographers ver- 56. Acker DB, Sachs BP, Friedman EA et al. Risk factors for shoulder
sus maternal-fetal medicine specialists. J Matern Fetal Neonatal dystocia. Obstet Gynecol 1985; 66: 762–768.
Med 2002; 11: 108–112. 57. Ecker JL, Greenberg JA, Norwitz ER et al. Birth weight as a
41. Benson CB, Doubilet PM, Saltzman DH et al. Sonographic deter- predictor of brachial plexus injury. Obstet Gynecol 1997; 89:
mination of fetal weights in diabetic pregnancies. Am J Obstet 643–647.
Gynecol 1987; 156: 441–444. 58. Langer O, Berkus MK, Huff RN et al. Shoulder dystocia: Should
42. McLaren RA, Puckett JL, Chauhan SP et al. Estimators of birth the fetus weighing ≥4000 grams be delivered by caesarean sec-
weight in pregnant women requiring insulin: A comparison tion? Am J Obstet Gynecol 1991; 165: 831–837.
of seven sonographic models. Obstet Gynecol 1995; 85: 59. American College of Obstetricians and Gynaecologists. Shoulder
565–569. dystocia. Practice Bulletin No. 40, The College: Washington, DC,
43. Best G, Pressman EK et al. Ultrasonographic prediction of birth 2002.
weight in diabetic pregnancies. Obstet Gynecol 2002; 99: 60. O’Higgins A, Doolan A, McCartan T et al. Ir Med J Nov 2014;
740–744. 107: 10.
52 Hypoglycemia in diabetic
pregnancy
Graziano Di Cianni, Cristina Lencioni, Emilia Lacaria, and Laura Russo
type 1 diabetes. Nevertheless, the prevalence rates rise

Introduction up to 70%–80% in clinical trials using insulin to achieve
Hypoglycemia is a major factor preventing patients with both good metabolic control. 5
type 1 and type 2 diabetes from achieving near-normal gly-
cemia. The risk of hypoglycemia in diabetic patients is due to
both the imperfect pharmacokinetics of current therapy that Symptoms
produces inappropriately high insulin concentrations and a A single threshold value for plasma glucose concentration
failure in the physiological protective mechanism that limits that defines hypoglycemia in diabetes cannot be assigned. In
the decrease in blood glucose concentrations. Hypoglycemia fact, glycemic threshold for symptoms of hypoglycemia shift
occurs as a result of good metabolic control and is a limit- from lower plasma glucose concentration in patients with
ing factor in achieving the highest possible value in the daily more frequent episodes of hypoglycemia to higher plasma
glycemic profile, in every condition in which it is required. glucose concentrations in those with poorly controlled
Maternal blood glucose control in pregnancy, as close to diabetes. However, hypoglycemia in diabetes usually has a
that of normal pregnancy, is essential in minimizing the neg- plasma glucose concentration below 70 mg/dL (3.9 mmol/L).6
ative effects on the mother and the fetus. Strategies to achieve Symptoms of hypoglycemia, which are the same also dur-
and maintain normal glycemia in pregnancy are onerous but ing pregnancy, are divided into two categories: neurogenic
not negotiable, and hypoglycemia becomes nearly inevitable and neuroglycopenic (Figure 52.1). Neurogenic (autonomic)
using available strategies. In addition, pregnancy itself may symptoms, activated by the autonomic nervous system and
be associated with impaired counterregulation system and mediated by norepinephrine, epinephrine, and acetylcho-
hypoglycemia unawareness. This could result in an increase line, include anxiety, palpitations, sweating, shakiness, dry
of severe hypoglycemic episodes. mouth, pallor, pupil dilatation, diaphoresis, hunger, and
The incidence, risk factors, and possible pathophysi- paresthesia.
ological factors of maternal hypoglycemia, its effects on the The insufficient contribution of glucose to the brain
mother and on the fetus, and the management of hypoglyce- (neuroglycopenia) with an associated cerebral reduction
mia during pregnancy are discussed in detail in this chapter. of oxygen causes neuroglycopenic symptoms. The cogni-
tive function becomes impaired when plasma glucose con-
centration falls below 50–55 mg/dL (2.7–3 mmol/L).7 These
Hypoglycemia in diabetes symptoms include abnormal mentation, irritability, dif-
ficult speaking, ataxia, paresthesia, headache, stupor, and
Hypoglycemia is the most frequent acute complication of eventually (if untreated) seizure, coma, and even death.
type 1 diabetes mellitus therapy. It has been reported that Neuroglycopenic symptoms also include transient focal
diabetic people live about 10% of their lives with glycemic neurological deficits (e.g., diplopia, hemiparesis). These
values lower than 50 mg/dL (2.8 mmol/L) and that on aver- symptoms usually represent the alarm signal of a more
age once a week they present an episode of symptomatic serious hypoglycemic attack, and they usually precede the
hypoglycemia and one episode of severe hypoglycemia per alterations of the cortical function. In this phase, the per-
year.1,2 About every 4–5 years, a case of this can lead to coma son is completely vigilant, but he/she can completely prevent
with the need of assistance and admission to hospital.3 In the attack from occurring by consuming foods rich in fast-
2%–4% of the cases, hypoglycemia causes death for people absorbing carbohydrates.
with type 1 diabetes mellitus.4 Severe hypoglycemia appears when glucose levels con-
Hypoglycemia is a common complication also for tinue to decrease. In this case, the person is incapable of
people with type 2 diabetes. The rate of severe hypoglycemia starting a treatment for the resolution of the episode and he
in type 2 diabetes is roughly one-third to that of or she needs a rapid and adequate assistance.
432
Hypoglycemia in diabetes 433
5.0 result of a complex integration within the brain, these sig-

nals are responsible for the autonomic response (sympathetic
4.6 mmol/L: Inhibition of insulin secretion and parasympathetic) organized within the hypothala-
4.5
mus.9 Glycemic reduction in the physiological area (about
4.4 mmol/L) causes a decline in insulin secretion that results
4.0 3.8 mmol/L: Glucagon release in an increased production of glucose from the liver. This rep-
resents the initial defense against decreasing glucose levels.10
3.7 mmol/L: Epinephrine and growth hormone release Glucagon stimulates hepatic glycogenolysis and it also
3.5
favors gluconeogenesis, but it also seems to determine auto-
nomic inputs. Epinephrine has a similar action to that of
3.2 mmol/L: Cortisol release/autonomic symptoms
3.0 glucagon, but it becomes of critical importance when there
2.8 mmol/L: Neuroglycopenic symptoms
is insufficient glucagon secretion. These two hormones are
2.6 mmol/L: Cognitive dysfunction therefore considered the ones mainly involved in the coun-
2.5
terregulation response; in fact, the development or the
1.7 mmol/L: Hepatic autoregulation accentuation of the hypoglycemic crisis does not take place
1.5 1.5 mmol/L: Convulsions, coma without these two hormones.
Other hormones, however, such as the growth hormone
and cortisol, are involved in counterregulation, especially
1.0
in prolonged hypoglycemia. Their effects are carried on
Figure 52.1 Symptoms of hypoglycemia. by limiting the glucose consumption and by favoring their
production.11
The defense mechanisms against hypoglycemia are
Pathophysiology clearly impaired in patients with type 1 diabetes and advanced
Glucose, in physiological conditions, provides 90% of the type 2 diabetes, which became unable to decrease insulin
energy necessary for brain functioning, and the cerebral levels and to increase glucagon levels in response to hypo-
functions are totally dependent on the level of the circulating glycemia. The loss of β-cell function determines that insulin
glucose. The decrease of the glycemic values results in a series circulating levels are only a function of exogenous insulin
of redundant neuroendocrine responses, aiming to bring the clearance and there is a reduced signaling of functional
glycemic levels to the physiological range8 (Figure 52.2). alpha cells to secrete glucagon as glucose concentration lev-
Decreasing plasma glucose concentrations are sensed els decrease. Under these conditions, epinephrine remains
in the hypothalamus and other regions of the brain as well critical for counterregulation because the two physiological
as in the hepatic portal vein and the carotid body. Glucose defense mechanisms against hypoglycemia are impaired.7
sensing in visceral sites send their information to the brain However, many patients with type 1 diabetes, especially
via the cranial nerve (especially the vagus nerve), and as a after recurrent hypoglycemia and in the presence of extended
Central
Glucose
nervous system
Adrenal
Increased sympathoadrenal outflow
Pancreas medulla
Beta cell Alpha cell
Epinephrine Epinephrine Increased

(palpitation, acetylcholine
tremor, arousal) (sweating, hunger)
Insulin Glucagon
Increased Increased
glycogenolysis neurogenic
and increased Liver
symptoms
gluconeogenesis
Increased
ingestion of
carbohydrates
Increased glucose
production Glucose
Figure 52.2 Physiological mechanism against hypoglycemia. The decreased in glucose concentration in blood includes a
decreased in insulin secretion, an increase in glucagon secretion, and an increase in epinephrine secretion.
434 Hypoglycemia in diabetic pregnancy
duration diabetes (10–20 years), suffer from reduced responses epinephrine during hypoglycemia. The attenuated epineph-
of epinephrine. Therefore, patients with type 1 diabetes with rine response to hypoglycemia is a marker of an attenuated
combined defect of glucagon and epinephrine are at about autonomic and sympathetic response.7 It causes the initial
sixfold increased risk for severe iatrogenic hypoglycemia loss of the warning symptoms of hypoglycemia and, over
during intensive insulin therapy.12 time, through a downward vicious spiral including pro-
Hyperinsulinemia is the rule in diabetes mellitus, both gressively impaired physiological responses and increasing
type 1 and type 2, because of the therapeutic delivery of clusters of hypoglycemic episodes, the development of hypo-
insulin into the peripheral rather than portal circulation glycemia unawareness.18
and because of the empirical algorithms used to administer
insulin. Absolute hyperinsulinemia, due to excessive levels
of circulating insulin (excess of dosage or irregularity of Cardiovascular, cognitive impact, and mortality
absorption), causes hypoglycemia more frequently during associated with hypoglycemia
the hours preceding meals or in the first hours in the morn- Severe and prolonged hypoglycemia can cause death: recent
ing. Relative hyperinsulinemia, due to other conditions such studies have demonstrated that up to 10% of patients with
as delayed or inadequate diet (especially as far as the range of type 1 and sulfonylurea-treated type 2 diabetes mellitus died
carbohydrates is concerned), physical exercise, renal failure, of hypoglycemia. Potential mechanisms include brain death
excessive alcohol consumption, and delayed gastric empty- and cardiac arrhythmias (including torsade de pointes and
ing, usually causes hypoglycemia after meals.13 atrial fibrillation).19
Type 2 diabetes is characterized by insulin resistance and Acute prolonged hypoglycemia causes a marked increase
persistent beta cell function, which allows insulin secretion in the workload of the heart, compromises endothelial func-
to decrease as blood glucose decreases, and apparently intact tion, and increases blood coagulability and viscosity.20
counterregulation. Consequently, as reported in the litera- A recent retrospective study reported a 79% higher risk in
ture,9 the rate of severe hypoglycemia results is lower than in cardiovascular events compared to patients without hypo-
type 1 diabetes. However, the United Kingdom Prospective glycemic events in patients with type 2 diabetes.21 Three large
Diabetes Study (UKPDS) showed that the frequency of hypo- trials, Action to Control Cardiovascular Risk in Diabetes
glycemia increases over the years with a prolonged duration (ACCORD),22 Action in Diabetes and Vascular Disease:
of insulin treatment.14 Patients with advanced type 2 diabetes Preterax and Diamicron Modified Release Controlled
and long duration of insulin treatment have reduced both glu- Evaluation (ADVANCE),23 and Veterans Affairs Diabetes Trial
cagon and sympathoadrenal response to hypoglycemia.9 (VADT),24 examined the effects of glucose lowering the cardio-
In people with type 2 diabetes, hypoglycemia is often a vascular events in patients with type 2 diabetes: a total of 24,000
consequence of the pharmacological treatments with hypo- patients with high cardiovascular risk were assigned to either
glycemic agents that stimulate pancreatic insulin secretion intensive control or standard therapy. Subjects assigned to the
(sulfonylureas, glinides). intensive arm experienced more episodes of hypoglycemia
than those randomly assigned to standard treatment arm. In
the ACCORD trial, subjects assigned to more intensive treat-
Hypoglycemia unawareness ment also experienced a 20% increase in mortality. In contrast
Hypoglycemia unawareness is a common long-term compli- in the ADVANCE or VADT study, recent severe hypoglycemia
cation of diabetes. Recent evidences suggest that 10%–15% of was a strong predictor of negative cardiovascular outcome. The
individuals with type 1 diabetes mellitus exhibit hypoglyce- results are still conflicting and a matter of debate.
mia unawareness, whereas impaired awareness (patients who Data from a mean 18 years of follow-up in the Diabetes
usually, but not always, experience symptoms of hypoglyce- Control and Complications Trial (DCCT) show no asso-
mia during hypoglycemic episodes) is seen in 40%–50%.1,15 ciation between hypoglycemia and a decline in any cogni-
In insulin-treated type 2 diabetes mellitus, hypoglycemia tive domain.25 Recent findings show that adults with type 1
unawareness has been reported to be approximately 10%.16 diabetes require a greater volume of the brain to perform a
The exact mechanism of hypoglycemia unawareness is working memory task during hypoglycemia.26
not completely understood. In type 2 diabetes mellitus, current evidence suggests that
It is thought to be due to an alteration in the relationship recurrent hypoglycemia does not cause cognitive impair-
between glycemic thresholds of the various components ment in young middle-aged patients. Data are conflicting
of the physiological response to hypoglycemia and cogni- regarding the impact of severe hypoglycemia on the risk
tive functions.17 Patients with hypoglycemia unawareness of dementia among elderly patients with type 2 diabetes
demonstrate sympathoadrenal activation at lower glucose mellitus.27,28
concentrations than those who have awareness and, impor-
tantly, at a lower level than that for cognitive impairment.
Strong evidence indicates that defective loss of symptoms Hypoglycemia in diabetic pregnancy
of hypoglycemia in patients with type 1 diabetes is largely
the result of antecedent, frequent hypoglycemia. It seems In pregnant women with diabetes (type 1 and type 2), a good
likely that antecedent hypoglycemic episodes contribute metabolic control before and during pregnancy is essential
to defective glucagon response and impaired secretion of to reduce the maternal–fetal morbidity and mortality.29 This
Hypoglycemia in diabetic pregnancy 435
is true for women with pregestational type 1 or type 2 dia- Pathophysiology

betes as well as for those with gestational diabetes mellitus. Impairment of glucose counterregulation
To obtain and maintain an optimal daily glycemic profile Diabetic pregnancy is characterized per se by an impairment
and HbA1c levels to near-normal values during pregnancy, of the counterregulatory response. As observed by experi-
women with diabetes are at increased risk of severe hypogly- mental studies performed in diabetic dogs, epinephrine and
cemic episodes. Moreover, in nondiabetic pregnancy, nor- glucagon responses to hypoglycemia result in an impaired
mal blood glucose levels are 20% lower than in nonpregnant glucose counterregulation during pregnancy.38 Moreover,
women,30,31 making the definition and the detection of hypo- these data have also been confirmed in human pregnancy by
glycemia more challenging. using the hypoglycemic clamp technique.16,39
Around 45%–71%32–34 of women with type 1 diabe- In pregnant women with intensively treated type 1 diabe-
tes experienced severe hypoglycemia during pregnancy. tes, consistent lower epinephrine and glucagon responses were
Furthermore, in front of similar glycemic control, almost found with respect to nondiabetic nonpregnant women. In
60% of severe hypoglycemic episodes are experienced by addition, the glycemic thresholds for epinephrine and growth
only 10% of pregnant women with type 1 diabetes.35 hormone secretion decreased. This may be the consequence, at
Severe hypoglycemia occurs three to five times more fre- least in part, of the intensive insulin treatment of the patients.
quently in the first trimester and at a lower rate in the third An impairment of secretion of both hormones, especially that
trimester.35 This is despite the fact that, in the third trimester, of glucagon, has also been demonstrated in nondiabetic preg-
a more stringent metabolic control is followed by the major- nant women. The exact mechanisms involved in the reduction
ity of women by administering much higher insulin doses of glucagon and epinephrine secretion during pregnancy are
as compared with early pregnancy. The first event usually not completely clear. Nevertheless, as for glucagon, one possi-
occurs before 20 weeks, and 80% of all events are recorded bility is that the placental hormones exert a suppressive effect
before 20 weeks32 (Figure 52.3). on it. In this context, it has been reported that the arginine-
Despite the higher incidence of severe hypoglycemia in induced secretion of glucagon is decreased in women using
pregnancy, the proportion of nocturnal severe hypoglyce- hormonal contraception.16 These data imply a suppression of
mia events is similar before and during pregnancy and at a these hormones by the high levels of estrogen and progester-
similar level to that reported in the DCCT among nonpreg- one that characterize pregnancy. Moreover, the results of a
nant patients with type 1 diabetes.36 A pre-bedtime glucose more recent experimental study suggest that the blunting of
value below 6.0 mmol/L predicts nocturnal biochemical glucagon secretion during insulin-induced hypoglycemia in
hypoglycemia in the first trimester among pregnant women pregnancy is related to a generalized impairment of a number
with type 1 diabetes.35 Interestingly, the vast majority of of different signals, including parasympathetic and sympa-
these hypoglycemia episodes are asymptomatic, suggesting thoadrenal stimuli and altered sensing of circulating and/or
that the nocturnal severe hypoglycemia events recorded by intraislet insulin.40
patients only represent the tip of the iceberg of all nocturnal Also, growth hormone response during pregnancy is pro-
hypoglycemic periods. Nocturnal asymptomatic hypoglyce- gressively reduced in pregnant women both with and with-
mia may be associated with the development of hypoglyce- out diabetes, and this can be determined by the progressive
mia unawareness.37
16
14
12
Number of events
10
0
4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38
Gestational week
Figure 52.3 Rate of severe hypoglycemic events per week during pregnancy in 108 women with type 1 diabetes. (Modified from
Ringholm, L. et al., Diabet. Med., 29, 558, 2012.)
increase of placental hormones.16 Another study using hyper-

Table 52.1 Risk factors for severe hypoglycemia
insulinemic hypoglycemic clamp in pregnant women with
in pregnant diabetic women
type 1 diabetes in the third trimester of pregnancy showed
a significant increase of placental growth hormone during
History of severe hypoglycemia in the preceding year
hypoglycemia. These observations indicate that the placenta
is an endocrine organ that can play an active role in acute Long duration of diabetes
metabolic processes such as the hormonal counterregula- Impaired hypoglycemia awareness
tion of hypoglycemia.41 In fact, placental growth hormone Low HbA1c in early pregnancy
has well-established hyperglycemic and insulin-antagonistic Fluctuating plasma glucose levels
properties and may exert its glycemic effects directly or Excessive use of supplementary insulin between meals
indirectly by insulin-like growth factor-I (IGF-I).42–44 More Nausea and vomiting
recent studies showed that, in normal pregnancy, maternal
placental growth hormone levels increase and gradually
replace the pulsatile pituitary growth hormone produc-
tion45; moreover, from 20 weeks, pituitary growth hormone Finally, nausea and vomiting, common in the first trimes-
levels decrease and remain at minimal levels in the third ter, may contribute to hypoglycemia by reducing the inges-
trimester.44 tion of carbohydrate (Table 52.1).
In pregnant women with type 1 diabetes, low IGF-I levels
are present in the first half of pregnancy and tend to increase Symptoms
until late pregnancy.43,46 In particular, women with type 1 According to the entity of the clinical presentation, both
diabetes experiencing repeated severe hypoglycemia dur- in pregnancy and outside pregnancy, hypoglycemia can be
ing pregnancy showed lower IGF-I levels as compared with classified as severe (an event requiring assistance of another
women without repeated severe hypoglycemia.46 Thus, the person) documented symptomatic hypoglycemia (an event
additive influence of pituitary growth hormone and placen- with typical symptoms that are accompanied by a measured
tal growth hormone on peripheral metabolism may be at plasma glucose concentration ≤70 mg/dL), asymptomatic
its lowest in early pregnancy where the incidence of severe hypoglycemia (an event not accompanied by typical symp-
hypoglycemia is highest. toms but with a measured plasma glucose concentration
Pregnancy induces substantial changes in the renin– ≤70 mg/dL), probable symptomatic hypoglycemia (an event
angiotensin system. During the early stages of pregnancy, with typical symptoms not accompanied by a plasma glucose
activation of the local and systemic renin–angiotensin concentration), and pseudohypoglycemia (an event with
system has been demonstrated and may be important for typical symptoms of hypoglycemia with a measured plasma
implantation and embryonic growth and development. glucose concentrations >70 mg/dL).
Angiotensin-converting enzyme activity may be higher
in pregnant women with type 1 diabetes compared with Fetal consequences
healthy pregnant women. However, only a weak association A series of experimental studies in animals have demon-
between angiotensin-converting enzyme activity and severe strated a teratogenic effect of maternal hypoglycemia occur-
hypoglycemia has been detected during pregnancy.35,47 ring during the embryogenesis. The exposure, in vitro, of
mouse embryos to hypoglycemic milieu even for short periods
Risk factors (2 hours) in the first stages of embryogenesis resulted in mal-
Hypoglycemia during diabetic pregnancy is certainly linked formations. This could be explained by the fact that younger
to the intensified insulin treatment useful to reach and embryos are very sensitive to hypoglycemia because they are
maintain the targets of glycemic control recommended in dependent on uninterrupted glycolysis and have not yet devel-
these pregnant women. Moreover, the impairment of glu- oped the ability for aerobic glucose metabolism.50,51
cose counterregulation42,48 and other features such as an Studies regarding a potential teratogenic effect of hypo-
increase in insulin sensitivity during the first trimester glycemia in pregnant women are few. Impastato et al.,52 more
of gestation49 are important in determining an increase in than 40 years ago, reported that 6 out of 19 normal pregnant
hypoglycemic episodes during pregnancy. Diabetic preg- women who received insulin coma therapy, for psychiatric
nancies are often characterized by hypoglycemia unaware- disorders, during the early phase of pregnancy, had mal-
ness; the mechanisms operating during pregnancy are not formed fetus. The following clinical studies evaluating the
different from those outside pregnancy. In particular, the potential teratogenic effect of maternal hypoglycemia in dia-
exposure to low blood glucose levels requested to obtain an betic patients, as reported in Table 52.2, showed no increase
ideal daily glycemic profile is the main factor involved in the in the frequency of fetal malformations in women who expe-
reduced h ypoglycemia awareness in pregnancy. Risk factors rienced frequent episodes of hypoglycemia.35,52–64
for severe hypoglycemia in pregnancy include a history of However, due to the limited data reported in literature
severe hypoglycemia in the preceding years, long duration of and to the fact that no prospective studies have been per-
diabetes, low HbA1c in early pregnancy, fluctuating plasma formed on this topic, the possibility that severe episodes of
glucose levels, and excessive use of supplementary insulin hypoglycemia may contribute to congenital malformations
between meals and impaired hypoglycemia awareness. cannot be completely excluded.
Treatment 437
Table 52.2 Maternal hypoglycemia and embryogenesis in pregnancies with type 1 diabetes
Study author Time of exposure glycemic levels Pregnancy outcome

Impastato52 <10 gestational week (gw) (19 nondiabetic Abortion,2 stillbirth,2 multiple malformations,1
women) Hirschsprung’s disease1
Molsted-Pedersen53 First trimester Lower frequency of maternal hypoglycemia in
malformed infants
Bergman54 Whole pregnancy (BG < 3.3mmol/L) No relationship with malformations
Rayburn55 Whole pregnancy (altered consciousness requiring No relationship with malformations
assistance by another person)
Mills56 5–12 gw No relationship with malformations
BG < 2.8 mmol/L or symptoms
Steel57 <9 gw No relationship with malformations
Severe hypoglycemia
Kitzmiller58 2–8 gw No relationship with malformations
BG < 2.7 mmol/L or symptoms
Kimmerle35 Not known No relationship with malformations
Hypoglycemic coma
Rosenn59 7–13 gw No relationship with malformations
Severe hypoglycemia
DCCT60 First trimester No relationship with malformations
Severe hypoglycemia
Palomar-Morales61 Whole pregnancy No relationship with malformation
Taylor62 Whole pregnancy No relationship with malformation
Mitanchez63 Whole pregnancy No relationship with malformation
Talaviya et al.64 Whole pregnancy No relationship with malformation
Abbreviation: BG, blood glucose.
Studies aimed to evaluate fetal heart rate (FHR) during that are developmentally adaptive for fetal survival, but later
hypoglycemia in diabetic pregnant women are conflicting. in life might predispose offspring to adult-onset diabetes.70
While some studies showed fetal tachycardia during mater- The degree of glycemic control during pregnancy affects
nal hypoglycemic coma,65 others showed bradycardia and perinatal outcome; nevertheless, data of pregnancy out-
reduced fetal movements during maternal hypoglycemia.66 come in diabetic women with repeated hypoglycemia are
An increased FHR has been demonstrated during hypogly- few. Perinatal outcome seems not to be affected by severe
cemic clamp studies, probably as a consequence of mater- episodes of hypoglycemia.35,54 Finally, studies in children
nal increased sympatico adrenal activity.67 On the contrary, of mothers who experienced frequent severe hypoglycemic
other studies have reported no changes in FHR and fetal episodes during pregnancy showed no relation between
movements, suggesting that the fetus could use alternative intellectual performance, psychomotor development, and
substrate during episodes of maternal hypoglycemia.68 With hypoglycemia. Rizzo et al. studied children of mothers with
correction of maternal blood glucose levels, FHR and fetal hypoglycemia during the second and third trimester of
activity returned to normal. pregnancy. They found normal intelligence scores at 2 and
Although data regarding congenital malformations 5 years of age.71 Maternal hypoglycemia in the first or third
and functional cardiac anomalies are not conclusive, it is trimester has not been shown to predict cognitive function
accepted that low levels of maternal glucose during preg- in adult offspring of women with type 1 diabetes.72
nancy may cause fetal growth retardation and small-for-
gestational-age infants.
Langer has shown that in gestational diabetic pregnant Treatment
women, the frequency of small-for-gestational-age infants
was 20% in women with low mean plasma glucose levels Diabetic people should be educated to identify correctly
(<4.8 mmol/L), significantly higher with respect to normal the alarm symptoms of hypoglycemia and to treat them-
pregnant women (11%).69 selves promptly to prevent progression to neuroglycopenia.
In addition, maternal hypoglycemia may also determine All hypoglycemic episodes, even in the absence of symp-
impaired fetal beta cell function. In fact, animal experimen- toms, require treatment. The goal of the treatment is to
tal studies suggest that fetus with intrauterine growth restric- increase the blood glucose to a safe level to avoid clinical
tion caused by chronic placental insufficiency have impaired consequences.
insulin secretion, insulin biosynthesis, and insulin content. Relatives, friends, teachers, and coworkers should be
The fetal beta cell dysfunction might indicate mechanisms taught to recognize symptoms of hypoglycemia, and in case
of doubt, they must treat immediately the person that may Effective approaches known to decrease the risk of iatro-
have hypoglycemia. genic hypoglycemia include patient education, dietary and
Mild to moderate hypoglycemia is usually treated exercise modifications, medication adjustment, careful glu-
with food, oral glucose tablets, or sucrose solutions. It cose monitoring by the patient, and conscientious surveil-
is sufficient to administer 15–20 g of glucose as glucose lance by the clinician.6
tablets (3–5 g/10 kg body weight) or sports drinks con- Optimizing insulin therapy is essential to reduce the risk
taining pure glucose to raise blood glucose by 65 mg/dL of hypoglycemia in diabetic pregnant women; the use of both
(3 mmol/L). Fruit juice is absorbed more slowly than glu- short- and long-acting insulin analogues during pregnancy
cose and is not as effective in raising the blood glucose can reduce the risk of severe hypoglycemia.73–75 The data that
levels as tablets or liquid glucose. Likewise, honey is less insulin pumps can help with the reduction of severe hypo-
efficient than glucose because it contains only 40%–50% glycemia during pregnancy are not well established.76 A case
glucose and the same amount of fructose. Sucrose solu- report suggested that insulin pump therapy combined with
tion as granulated sugar in orange juice or milk does real-time continuous glucose monitoring, with careful mon-
not provide a quick rise in blood glucose levels for rapid itoring of glucose values, is useful in preventing severe hypo-
symptomatic relief and may require the administration glycemia in pregnant women with impaired hypoglycemia
of a larger volume. awareness.35,77
Moderate hypoglycemia may require a second adminis-
tration of glucose.
In case of loss of consciousness, the treatment is 10–25 g Conclusions
(20–50 mL of 33% glucose solution injected intravenously
over 1–3 minutes). The intravenous infusion of glucose solu- Diabetic pregnancy represents the most important challenge
tions (20–50 mL at 33%) is the advised treatment for the hos- of insulin therapy, requiring a continuous effort to maintain
pitalized patient. euglycemia and avoid severe hypoglycemic episodes. The fre-
The administration of glucose solutions at 33% must be quent exposure to low glucose levels determines an impaired
carefully controlled to avoid phlebitis and pain in case of glucose counterregulation and hypoglycemia unawareness.
breaking of the vein. Hypoglycemia may be dangerous for the mother, but it is
Glucagon is ideal to use at home and the standard dose unclear whether and to which extent it determines embry-
is 1 mg given subcutaneously. The maximal blood glu- opathy. Finally, the worry about hypoglycemic episodes can
cose response occurs after 15 minutes, with a peak at 20 have a negative impact on the experience of pregnancy for the
minutes to 1 hour. When glucagon is used, it is important mother and her partner. Women should avoid hypoglycemic
to remember that postglucagon nausea and vomiting are episodes and be able to prevent them. Therefore, women have
common and that careful monitoring of blood glucose to be adequately instructed to perform self-monitoring of
levels should be continued until the patient is able to eat blood glucose and observe the diet especially of the intermeal
normally. snacks and insulin plan. The new technology for monitoring
Glucagon does not cross the placenta and its use in preg- the glucose profile daily will be a useful tool to reduce hypo-
nancy is not dangerous.58 glycemic episodes in pregnancy with diabetes in the next years.
REFERENCES
1. Cryer PE. Hypoglycemia in type 1 diabetes mellitus. Endocrinol 9. Segel SA, Paramore DS, Cryer PE. Hypoglycemia-associated auto-
Metab Clin North Am 2010; 39(3): 641–654. nomic failure in advanced type 2 diabetes. Diabetes 2002; 51:
2. American Diabetes Association. Defining and reporting hypo- 724–733.
glycaemia in diabetes: A report from the American Diabetes 10. Mitrakou A, Ryan C, Venemen T et al. Hierarchy of glycemic
Association Workgroup on Hypoglycemia. Diabetes Care 2005; thresholds for activation of counterregulatory hormone secretion,
28: 1245–1249. initiation of symptoms and onset of cerebral dysfunction in nor-
3. The DCCT Research Group. Hypoglycemia in the diabetic con- mal human. Am J Physiol 1991; 260: E67–E74.
trol and complications trial. Diabetes 1997; 37: 271–286. 11. Heller SR. Hypoglycemia and diabetes. In: Pickup JC, Williams
4. Laing SP, Swerdlow AJ, Slater SD et al. The British Diabetic G, eds., Textbook of Diabetes 1, 3rd ed., Blackwell Publishing:
Association Cohort Study. Cause-specific mortality in patients with Oxford, U.K., 2003.
insulin treated diabetes mellitus. Diabet Med 1999; 16: 466–471. 12. Fanelli C, Pampanelli S, Lalli C et al: Long term intensive therapy
5. The U.K. Prospective Diabetes Study Group. Intensive blood- of IDDM diabetic patients with clinically overt autonomic neu-
glucose control with sulfonylureas or insulin compared with con- ropathy on awareness and counterregulation to hypoglycaemia.
ventional treatment and risk of complication in patients with type Diabetes 1997; 46: 1172–1181.
2 diabetes. Lancet 1998; 352: 837–853. 13. Fanelli C, Porcellati F, Pampanelli S, Bolli G. Insulin therapy and
6. Seaquist ER, Anderson J, Childs B et al. Hypoglycemia and hypoglycaemia: The size of the problem. Diabetes Metab Res
diabetes: A report of a Workgroup of the American Diabetes Rev 2004; 20: S32–S42.
Association and the Endocrine Society. Diabetes Care 2013; 14. The UKPDS Research Group. Overview of 6 years of therapy
36(5): 1384–1395. of type 2 diabetes: A progressive disease. Diabetes 1995; 44:
7. Cryer PE. Mechanism of hypoglycemia-associated autonomic 1249–1258.
failure in diabetes. Review article. N Engl J Med 2013; 369: 15. Pedersen-Bjergaard U, Pramming S, Thorsteinsson B. Recall
362–372. for severe hypoglycemia and self-estimated state of aware-
8. Cryer PE, Davis S, Shamoon H. Hypoglycaemia in diabetes. ness in type 1 diabetes. Diabetes Metab Res Rev 2003; 19:
Diabetes Care 2003; 26: 1902–1912. 232–240.
References 439
16. Schopman JE, Geddes J, Frier BM. Prevalence of impaired aware- 40. Canniff KM, Smith MS, Lacy DB, Williams PE, Moore MC.
ness of hypoglycemia and frequency of hypoglycemia in insulin- Glucagon secretion and autonomic signaling during hypoglyce-
treated type 2 diabetes. Diabetes Res Clin Pract 2010; 87: 64–68. mia in late pregnancy. Am J Physiol Regul Integr Comp Physiol
17. Heller SR. How should hypoglycaemia unawareness be man- 2006; 223: 4–28.
aged? In: Gill G, Williams G, Pickup J, eds., Difficult Diabetes: 41. Bjorklund AO, Adauson UK, Calstrom KA, Hennnem G, Igant
Current Management Challenges. Blackwell Science Ltd: Oxford, A, Lins PE, Wastgrenn LM. Placental hormones during induced
U.K., 2001, pp. 168–187. hypoglycemia in pregnant women with insulin dependent diabe-
18. Cryer PE. Hypoglycaemia-associated autonomic failure in diabetes mellitus: Evidence of an active role for placenta in hormonal
tes. Am J Physiol 2001; 281: E115–E121. counter-regulation. Br J Obstet Gynecol 1998, 105(6): 649–655.
19. Oyer DS. The science of hypoglycemia in patients with diabetes. 42. Chellakooty M, Vangsgaard K, Larsen T et al. A longitudinal
Curr Diabetes Rev 2013; 9: 195–208. study of intrauterine growth and the placental growth hor-
20. Frier BM, Schernathaner G, Heller SR. Hypoglycemia and cardio- mone (GH)-insulin-like growth factor I axis in maternal circula-
vascular risk. Diabetes Care 2011; 34(Suppl. 2): S132–S137. tion: Association between placental GH and fetal growth. J Clin
21. Johnston SS, Conner C, Aagren M, Smith DM, Bouchard J, Brett Endocrinol Metab 2004; 89(1): 384–391.
J. Evidence linking hypoglycemic events in patients with type 2 43. Fuglsang J, Lauszus F, Flyvbjerg A, Ovesen P. Human placental
diabetes. Diabetes Care 2011; 34: 1164–1170. growth hormone, insulin-like growth factor I and -II, and insu-
22. Gerstein HC, Miller ME, Bygton RP et al.; Action to Control lin requirements during pregnancy in type 1 diabetes. J Clin
Cardiovascular Risk in Diabetes Study Group. Effects of intensive Endocrinol Metab 2003; 88(10): 4355–4361.
glucose lowering in type 2 diabetes. N Engl J Med 2008; 358: 44. Caufriez A, Frankenne F, Hennen G, Copinschi G. Regulation of
2545–2559. maternal IGF-I by placental GH in normal and abnormal human
23. Patel A, MacMahon S, Chalmers J et al.; ADVANCE Collaborative pregnancies. Am J Physiol 1993; 265(4 Pt. 1): E572–E577.
Group. Intensive blood glucose control and vascular outcomes 45. Lønberg U, Damm P, Andersson AM et al. Increase in maternal
in patients with type 2 diabetes. N Engl J Med 2008; 358: placental growth hormone during pregnancy and disappearance
2560–2572. during parturition in normal and growth hormone-deficient preg-
24. Duckworth W, Abraira C, Moritz T et al.; VADT Investigators. nancies. Am J Obstet Gynecol 2003; 188(1): 247–251.
Glucose control and vascular complications in veterans with type 46. Ringholm Nielsen L, Juul A, Pedersen-Bjergaard U et al. Lower
2 diabetes. N Engl J Med 2009; 360: 129–139. levels of circulating IGF-I in type 1 diabetic women with frequent
25. Jacobson AM, Musen G, Ryan CM et al. Long-term effect of dia- severe hypoglycaemia during pregnancy. Diabet Med 2008;
betes and its treatment on cognitive function. N Engl J Med 2007; 25(7): 826–833.
356: 1842–1852. 47. Nielsen LR, Pedersen-Bjergaard U, Thorsteinsson B et al. Severe
26. Bolo NR, Musen G, Jacobson AM et al. Brain activation during hypoglycaemia during pregnancy in women with type 1 diabe-
working memory is altered in patients with type 1 diabetes dur- tes: Possible role of renin-angiotensin system activity? Diabetes
ing hypoglycemia. Diabetes 2011; 60: 3256–3264. Res Clin Pract 2009; 84(1): 61–67.
27. Bruce DG, Davis WA, Casey GP et al. Severe hypoglycemia 48. Rosenn BM, Miodovnik M, Khoury JC, Siddiqi TA. Counter
and cognitive impairment in older patients with diabetes: The regulatory hormonal responses to hypoglycemia during preg-
Fremantle Diabetes Study. Diabetologia 2009; 52: 385–390. nancy. Obstet Gynecol 1996; 87: 568–574.
28. Whitmer RA, Karter AJ, Yaffe K, Quesenberry CP, Selby JV. 49. Di Cianni G, Miccoli R, Volpe L, Lencioni C, Del Prato S.
Hypoglycemic episodes and risk of dementia in older patients with Intermediate metabolism in normal pregnancy and in gestational
type 2 diabetes mellitus. J Am Med Assoc 2009; 301: 1565–1572. diabetes. Diab Met Res Rev 2003; 19: 259–270.
29. Metzger BE, Lowe LP, Dyer AR et al.; HAPO Study Cooperative 50. Illsley NP. Glucose transporters in human placenta. Placenta
Research Group. Hyperglycemia and adverse pregnancy out- 2000; 21: 14–22.
comes. N Engl J Med 2008; 358: 1991–2002. 51. Smoak IW, Sadler TW. Embryopathic effects of short-term expo-
30. Parretti E, Mecacci F, Papini M et al. Third-trimester maternal sure to hypoglycemia in mouse embryo in vitro. Am J Obstet
glucose levels from diurnal profiles in nondiabetic pregnan- Gynecol 1990; 163: 619–624.
cies: Correlation with sonographic parameters of fetal growth. 52. Impastato DJ, Gabriel AR, Lardaro EH. Electric and insulin shock
Diabetes Care 2001; 24(8): 1319–1323. therapy during pregnancy. Dis Nerv Syst 1964; 23: 542–546.
31. Yogev Y, Ben-Haroush A, Chen R et al. Diurnal glycemic profile 53. Molsted-Pedersen L, Tygstrup I, Pedersen J. Congenital malforma-
in obese and normal weight nondiabetic pregnant women. Am J tions in newborn infants of diabetic women. Correlation with mater-
Obstet Gynecol 2004; 191(3): 949–953. nal diabetic vascular complications. Lancet 1964; 23: 1124–1126.
32. Nielsen LR, Pedersen-Bjergaard U, Thorsteinsson B, Johansen 54. Bergman M, Seaton TB, Auerhahn CC, Aaron Young C, Glasser
N, Damm P, Mathiensen ER. Hypoglycemia in pregnant women M, Shapiro LR. The incidence of gestational hypoglycaemia in
with type 1 diabetes: Predictors and role of metabolic control. insulin-dependent and non-insulin-dependent diabetic women.
Diabetes Care 2008; 31: 9–14. N Y State J Med 1986; 86: 174–177.
33. Evers IM. Risk indicators predictive for severe hypoglycemia dur- 55. Rayburn W, Piehl E, Sanfield J, Compton A. Reversing severe
ing the first trimester of type 1 diabetic pregnancy. Diabetes Care hypoglycemia during pregnancy with glucagon therapy. Am J
2002; 25: 554–555. Perinatol 1987; 4(3): 259–261.
34. Kimmerle R, Heinemann L, Delecki A, Berger M. Severe hypo- 56. Mills JL, Knopp RH, Simpson JL et al. Lack of relation of increased
glycemia incidence and predisposing factors in 85 pregnancies malformation rates in infants of diabetic mothers to glycemic
of type 1 diabetic women. Diabetes Care 1992; 15: 1034–1037. control during organogenesis. N Engl J Med 1988; 318: 671–676.
35. Ringholm L, Pedersen-Bjergaard U, Thorsteinsson B, Damm P, 57. Steel JM, Johnstone FD, Hepburn DA, Smith AF. Can pregnancy
Mathiesen ER. Hypoglycaemia during pregnancy in women with care of diabetic women reduce the risk of abnormal babies? Br
type 1 diabetes. Diabet Med 2012; 29: 558–566. Med J 1990; 301: 1070–1074.
36. The DCCT Research Group. Epidemiology of severe hypoglyce- 58. Kitzmiller JL, Gavin LA, Gin GD, Jovanovic Peterson L, Main EK,
mia in diabetic control and complications trial. Am J Med 1991; Zigrang WD. Preconception care of diabetes. Glycemic con-
90: 450–459. trol prevents congenital anomalies. J Am Med Assoc 1991; 265:
37. Schopman JE, Geddes J, Frier BM. Frequency of symptomatic 731–736.
and asymptomatic hypoglycaemia in type 1 diabetes: Effect of 59. Rosenn B, Siddiqi TA, Miodovnik M. Normalization of blood
impaired awareness of hypoglycaemia. Diabet Med 2011; 28(3): glucose in insulin-dependent diabetic pregnancies and the risks
352–355. of hypoglycemia: A therapeutic dilemma. Obstet Gynecol Surv
38. Connolly CC, Aglione LN, Smith MS, Lacy B, Moore MC. 1995; 50: 56–61.
Pregnancy impairs the counterregulatory response to insulin- 60. The Diabetes Control and Complications Trial (DCCT) Group.
induced hypoglycemia in the dog. Am J Physiol Metab 2004; 287: Pregnancy outcome in the Diabetes Control and Complications
480–488. Trial. Am J Obstet Gynaecol 1996; 174: 1343–1353.
39. Diamond MP, Reece EA, Caprio S, Jones TW, Amiel S, De 61. Palomar-Morales M, Méndez JD. Hypoglycemia as a risk
Gennaro N. Impairment of counterregulatory hormone responses factor in embryonic development in maternal diabetes mel-
to hypoglycemia in pregnant women with insulin-dependent dia- litus. Review of the literature. Ginecol Obstet Mex 1999; 67:
betes mellitus. Am J Obstet Gynecol 1992; 166: 70–77. 425–432.
62. Taylor R, Lee C, Kyne-Grzebalski D, Marshall SM, Davison JM. 70. Limesand SW, Rozane PJ, Zerbe GO, Hutton JC, Hav WW Jr.
Clinical outcomes of pregnancy in women with type 1 diabe- Attenuated insulin release and storage in fetal sheep pancreatic
tes. Obstet Gynecol 2002; 99(4): 537–541. islet with intrauterine growth restriction. Endocrinology 2006;
63. Mitanchez D. Fetal and neonatal complications of gestational 147(3): 1488–1497.
diabetes: Perinatal mortality, congenital malformations, mac- 71. Rizzo T, Metzger BE, Burns WJ, Burns K. Correlations between
rosomia, shoulder dystocia, birth injuries, neonatal outcomes. antepartum maternal metabolism and child intelligence. N Engl J
J Gynecol Obstet Biol Reprod 2010; 39(8 Suppl. 2): S189–S199. Med 1991; 325: 911–916.
64. Talaviya PA, Saboo BD, Joshi SR, Padhiyar JN, Chandarana HK, 72. Clausen TD, Mortensen EL, Schmidt L et al. Cognitive function in
Shah SJ, Vyas CK, Shah AN. Pregnancy outcome and glyce- adult offspring of women with type 1 diabetes. Diabet Med July
mic control in women with type 1 diabetes: A retrospective 2011; 28(7): 838–844.
comparison between CSII and MDI treatment. Diabetes Metab 73. Mathiesen ER, Kinsley B, Amiel SA, Heller S, McCance D, Duran
Syndr 2013; 7(2): 68–71. S, Bellaire S, Raben A; Insulin Aspart Pregnancy Study Group.
65. Confino E, Ismajovic B, David MP, Gleicher N. Fetal heart rate Maternal glycemic control and hypoglycemia in type 1 diabetic
in maternal hypoglicaemic coma. Int J Gynaecol Obstet 1985; pregnancy: A randomized trial of insulin aspart versus human
23: 59–60. insulin in 322 pregnant women. Diabetes Care 2007; 30(4):
66. Langer O, Cohen WR. Persistent fetal bradycardia during 771–776.
maternal hypoglycaemia. Am J Obstet Gynaecol 1984; 149: 74. Heller S, Damm P, Mersebach H, Skjøth TV, Kaaja R, Hod M,
688–690. Durán-García S, McCance D, Mathiesen ER. Hypoglycemia in
67. Bjorlund AO, Adamson UK, Almstrom NH. Effects of hypo- type 1 diabetic pregnancy: Role of preconception insulin aspart
glycaemia on fetal heart activity and umbilical artery doppler treatment in a randomized study. Diabetes Care 2010; 33(3):
velocity waveforms in pregnant women with insulin dependent 473–478.
diabetes mellitus. Br J Obstet Gynaecol 1996; 103: 413–420. 75. Lambert K, Holt RI. The use of insulin analogues in pregnancy.
68. Lapidot A, Haber S. Effect of acute insulin-induced hypoglyce- Diabetes Obes Metab 2013; 15(10): 888–900.
mia on fetal versus adult brain fuel utilization assessed by 13C 76. Bode BW, Steed RD, Davidson PC. Reduction in severe hypogly-
MRS isotopomer analysis of (U-13 C) glucose metabolism. Dev cemia with long-term continuous subcutaneous insulin infusion
Neurosci 2000; 22: 444–455. in type I diabetes. Diabetes Care April 1996; 19(4): 324–327.
69. Langer O, Levy J, Brustman L, Anyaegbunam A, Merkatz R, 77. Worm D, Nielsen LR, Mathiesen ER, Nørgaard K. Continuous
Divon M. Glycemic control in gestational diabetes mellitus— glucose monitoring system with an alarm: A tool to reduce hypo-
How tight is tight enough: Small for gestational age versus large glycemic episodes in pregnancy with diabetes. Diabetes Care
for gestational age? Am J Obstet Gynaecol 1989; 161: 646–653. 2006; 29(12): 2759–2760.
53 Hypertensive disorders and
diabetic pregnancy
Jacob Bar, Moshe Hod, and Michal Kovo
studies reported an association of low birth weight with

Metabolic syndrome: When high risk of later ischemic heart disease7–12 and stroke13,14 or
hypertension and diabetes meet impaired glucose tolerance and DM.15,16 It was also found to
be associated with high blood pressure (BP) in childhood17,18
The striking increase in the prevalence of obesity, diabetes and adult life.19 The evidence was strongest for BP and glu-
mellitus (DM), hypertension, and cardiovascular disease in cose tolerance,20 which could be measured earlier in life and
the last two decades1 has led to the concept of the metabolic for which more data, and sometimes also prospective data,
syndrome.2 Also termed syndrome X,3 insulin resistance were available.19,21 The evidence was weaker, though still
syndrome,4 and the deadly quartet,5 metabolic syndrome convincing for heart disease, for which data were sparse and
is characterized by a constellation of well-documented risk often confined to men. The findings in the few studies on
factors for cardiovascular disease, namely, glucose intol- stroke, particularly the hemorrhagic type, were consistent.14
erance, insulin resistance, central obesity, dyslipidemia, In another study, Barker et al.22 observed that the effects
and hypertension, that co-occur in individuals at a higher of impaired fetal growth are modified by subsequent growth.
rate than expected by chance. Extensive research has still As such, individuals who were small at birth but became over-
not completely elucidated the precise cause of the syn- weight in adulthood were at the highest risk of heart disease
drome, although some strong positions have been taken. and type 2 DM (a physiological resistance to insulin action).
Nevertheless, it is widely recognized that a combination of This finding led to the second part of the hypothesis, the thrifty
genetic predisposition and environmental factors, particu- phenotype (Figure 53.1). The authors proposed that the pro-
larly those associated with socioeconomic status, is involved. cess of adaptation to undernutrition in fetal life leads to per-
The environmental factors include both postnatal life hab- manent metabolic and endocrine changes. These are beneficial
its and nutrition and—no less important—intrauterine if the undernutrition persists after birth, but may predispose
conditions. Indeed, there is plentiful evidence linking low the individual to obesity and impaired glucose tolerance if it
birth weight due to fetal growth restriction (FGR) with an does not. The most unfavorable growth pattern is smallness
increased risk of vascular disease in later adult life.6 and thinness at birth, continued slow growth in early child-
hood, followed by an acceleration of growth so that height and
weight approach the population means, with a continued rise
Intrauterine factors in metabolic in body mass index above the mean. The growth pattern dif-
fers by sex6,23 and ponderal index.6 However, as birth weight
syndrome: The fetal origin of adult and ponderal index, as well as body mass index, are only crude
disease measures of the manner in which fetal nutrition affects body
composition and the balance of lean body mass to fat, the true
Barker6 pioneered the idea that the epidemic of coronary impact of fetal growth on later disease remains unclear. Be that
heart disease in Western countries in the twentieth cen- as it may, there is no doubt that low birth weight and high body
tury, which paradoxically coincided with improved stan- mass index interact and that their effects on BP and impaired
dards of living and nutrition, originated in fetal life. He glucose tolerance are multiplicative.6
postulated that the low birth weight and impaired fetal The thrifty phenotype paradigm has stimulated a wealth
growth that were characteristic of deprived regions in the of animal and human research on FGR and its sequelae.
1900s may have predisposed the survivors to heart disease The hypothesis predicts that a population undergoing a
in later life. Support was provided by studies conducted in transition from poor to better nutrition will be character-
Hertfordshire, England, showing a higher rate of cardiovas- ized by more heart disease and impaired glucose tolerance.
cular mortality in men who had been small at birth and at This is epitomized by the rapidly rising incidence of type
1 year of age.6 Thereafter, at least seven retrospective cohort 2 DM, ischemic heart disease, and obesity in increasingly
441
442 Hypertensive disorders and diabetic pregnancy
Maternal body composition and diet
Placenta
Fetal undernutrition
Adapted liver Changes in structure Re-set hypothalamic- Changes in pancreas

metabolism of heart, blood vessels pituitary-adrenal and muscle
and kidney and growth
hormone-IGF axes
Increased circulating LDL Hypertension and left Non-insulin-dependent

cholesterol and fibrinogen ventricular hypertrophy diabetes
Coronary heart disease
Figure 53.1 Possible mechanisms linking fetal undernutrition and coronary artery disease. IGF, insulin growth factor.
urbanized India.24–26 Indian infants are exceptionally small, of cell types, gene expression, or both. Some researchers
with a mean birth weight of 2700 g, and their mothers tend have accused the authors who formulated the hypothesis
to be short and underweight. The infants also have low mus- of incorrect statistical interpretations because of chance,
cle mass, small viscera, and a relative excess of fat—a body artifacts, or confounding factors in later life, but these have
composition particularly likely to lead to insulin resistance.27 been resolved.34 Nevertheless, it should be emphasized that
In a cohort study in Indian children, Yajnik et al.28 showed support for the hypotheses comes mainly from studies in
that lower weight at birth and higher body mass index in rodents35 that cannot rule out environmental causes, partic-
childhood were associated with impaired glucose tolerance. ularly those associated with socioeconomic status,36,37genetic
Although improving the growth and nutrition of the mother predisposition to low birth weight or hypertension and
before pregnancy would seem to be the ideal strategy to hypertension-related diseases, and postnatal factors.38,39
improve fetal growth, animal studies have shown that more Unfortunately, testing these parameters in humans is neither
than one generation of improved maternal nutrition may be ethical nor practical.
necessary for an optimal outcome.29,30 Thus, in India, where In an attempt to separate genetic from extrauterine envi-
women begin childbearing already in their teens, before they ronmental influences, some researchers have studied multiple
are fully grown, postponing marriage might be a good first pregnancies. For example, the Tasmanian Infant Health
step.31 There is only limited evidence that nutritional supple- Survey of a cohort of monozygotic, dizygotic, and singleton
ments in pregnancy improve fetal growth in undernourished pregnancies reported a stronger association between birth
mothers.32 Furthermore, the effect of supplements varies weight and BP in children from multiple pregnancies.40 The
according to the stage of pregnancy: giving them early in association also held true within the monozygotic pairs,
pregnancy may even worsen fetal growth.6 suggesting that a genetic predisposition may need to be
Stene et al.,33 in a large population-based cohort study, combined with specific mechanisms within the fetoplacen-
noted a relatively weak but significant and nearly linear asso- tal unit.40 A study of 492 pairs of female twins showing an
ciation between birth weight and risk of type 1 DM. The ratio inverse relationship of birth weight and adult BP41 proved
of children with a birth weight of 4500 g or more to children further corroboration for the assumption that restricted
with a birth weight of less than 2000 g was 2.21. This find- intrauterine growth is due to placental dysfunction rather
ing raised the possibility that perinatal factors influence the than inadequate maternal nutrition or genetic factors.
risk of type 1 DM. The underlying mechanisms of this asso- Two other studies stress the importance of primary pre-
ciation are unknown, but they probably differ from those vention of high BP and cardiovascular disease and the con-
responsible for the association between low birth weight and troversy still surrounding Barker’s fetal origin hypothesis. In
later onset of type 2 DM. the first, school children with a history of low birth weight
There may also be factors other than nutrition that play a were found to have impaired endothelial function and a
role in the casual pathway leading to high BP, cardiovascular trend toward carotid stiffness, which may represent early
disease, or type 2 DM. expressions of vascular compromise.42 However, another
The fetal origin of adult disease (FOAD) hypothesis group of investigators showed no difference in flow-mediated
assumes that a poor nutrient supply during a critical period endothelial-dependent vasodilatation (early stage in the

of in utero life may “program” a permanent structural or development of atherosclerosis) between adolescents who
functional change in the fetus, altering the distribution had a low birth weight and controls.43
Intrauterine factors in metabolic syndrome: The fetal origin of adult disease 443
Maternal or fetal Smoking or High or low Certain drugs Hypo- or hyper-

Preeclampsia
malnutrition alcoholism salt intake (i.e., steroids, CSA) vitaminosis
Adverse conditions during pregnancy
Epigenetic modification of DNA
Maternal
Impaired fetal development
diabetes/
(birth weight or , low nephron number,
paternal hyper-
arterial remodeling)
lipidemia
Cardiovascular disease Metabolic disease

Renal disease
(i.e., hypertension, (i.e., insulin resistance
(i.e., glomerulonephritis)
coronary heart disease) dyslipidemia)
Figure 53.2 Factors that directly or indirectly affect fetal development and may thus favor programming diseases that occur in
later life. (From Koleganove, N. et al., Nephrol. Dial. Transplant., 27, 3003, 2012. With permission.)
An additional factor that may underlie the proposed con- with or without FGR, FGR without hypertension, intra-
tribution of the intrauterine environment to adult disease is uterine fetal death, placental abruption, and even preterm
the reduced nephron mass documented in infants in disad- labor.49 Defective deep placentation is dependent upon
vantaged populations with intrauterine growth restriction genetic and environmental factors, as well as time of onset,
and exposure to maternal diabetes and vitamin A deficiency. duration, and extent of the ischemic insult. Moreover, evo-
A lower nephron mass may impair nephrogenesis, thereby lutionary pressures derived from the potential conflict rela-
increasing the susceptibility of the infant to later kidney tionship between the fetus and the mother that include their
damage from diseases such as hypertension and diabetes or response to insults also play a role in determining the phe-
diabetic kidney, which also commonly affect disadvantaged notypic expression of disorders of the placental bed.49 FGR
people.44,45 is caused by placental dysfunction, resulting in impaired
In summary, the FOAD hypothesis is not without criti- transfer of oxygen and nutrients to the fetus that may lead
cism, implying that the association between anthropomet- to fetal hypoxia and early growth impairment. To assess the
ric measures at birth and outcomes in adulthood is also vascular function of the placenta in FGR and to estimate
influenced by confounders, such as socioeconomic status the hemodynamic condition of the growth-restricted fetus,
and lifestyle (diet, cigarette smoking, physical exercise).46 Doppler ultrasonography is widely used. Hemodynamic
Nevertheless, it is likely to remain an important perspec- deterioration in cases of severe placental insufficiency is
tive and better position as part of a broader life course well recognized through the increased impedance to blood
perspective rather than as an independent hypothesis.47 flow in the umbilical artery (UA)50 and the compensatory
Recently, the World Health Organization included low mechanisms that maintain cerebral oxygen supply.51 There is
birth weight as a risk factor for cardiovascular disease a significant association between pathological Doppler flow
(Figure 53.2). pattern and placental vascular compromise. Indeed, the pla-
cental component has been studied thoroughly in FGR,52,53
with and without pathological UA Doppler indices.54,55
Vicious cycle of pregnancy: The contributing factors to Stereological studies observed substantial changes in placen-
adult disease tal morphology demonstrating poor growth of villi and fetal
Intrauterine environment component vasculature of FGR placentas.56 Most of the studies included
Successful pregnancy requires normal placental function patients with hypertensive disorders.57
from adequate placentation early in pregnancy to sustained The fetus, in response to placental undernutrition and in
placental function in the second half of pregnancy. Abnormal order to maintain sufficient vascular supply to the brain, by
placentation, which is defined as defective remodeling of the using mechanisms such as arterial redistribution, decreases
uteroplacental arteries, results in inadequate uteroplacental resistance to blood flow in the middle cerebral artery, a phe-
blood flow,48 leading to unsuccessful pregnancy outcome. nomenon that has been termed as a “head sparing effect.”
In recent years, it has become clear that poor placentation Simultaneously, Doppler studies demonstrate that because
occurs in a broader range of pregnancy complications than of the limited blood supply to the fetus, the arterial redis-
initially thought. These complications include preeclampsia tribution process is accompanied by an increased resistance
to flow to other fetal vital organs, such as the kidneys, liver, Postnatal period component
and pancreas. These independent Doppler studies may shed When undernutrition during early development is followed
light on the thrifty phenotype paradigm and on the asso- by improved nutrition later in development, many mam-
ciation between poor intrauterine environment and disease mals retain some capacity to compensate by increasing their
at adulthood. The arterial redistribution as a compensation growth rate. Such compensatory changes will carry costs.
mechanism may explain why individuals exposed to isch- This may explain why rapid childhood growth, especially in
emic changes in utero develop dyslipidemia, lower nephron people who were born small or were thin in infancy, appears
number, and impaired glucose tolerance, all contributing to to have deleterious effects on later health.63,64
metabolic syndrome later in life.
Combined component: The epigenesis
Maternal component: Genetics plus environment Developmental plasticity is defined as the ability of an organism
The association between pregnancy complications in gen- to develop in various ways, depending on the particular envi-
eral and preeclampsia in particular to future metabolic ronment or setting.65 Developmental plasticity requires stable
syndrome is well established in recent years.58 By the envi- modulation of gene expression, and this appears to be medi-
ronmental changes, either metabolic or hormonal, preg- ated, at least in part, by epigenetic processes such as DNA meth-
nancy complications, especially preeclampsia and GDM, ylation and histone modification. Thus, both the genome and
occur in women with underlying vascular compromise.59 the epigenome interactively influence the mature phenotype
The same women who develop pregnancy complications and determine sensitivity to later environmental factors and
are more obese, have more familial history of diabetes and the subsequent risk of disease.66 The human placenta executes a
hypertension, and have the same risk factors to develop later complex of specific functions influenced by maternal, fetal, and
atherosclerotic cardiovascular disease. Indeed, our group60 environmental signals. It is highly possible that placental epig-
and another investigator clearly demonstrated an associa- enome is linked to its development and function, in general, and
tion between early thrombo-occlusive disorders before the between its imprinting status and fetal growth, in particular.67
age of 50 and history of pregnancy complications.61 One of In summary, pregnancy is a key point in the FOAD
the most common complications in their pregnancy was hypothesis. The maternal background that leads to preg-
FGR (odds ratio [OR] of 8.4).60 Another investigator demon- nancy complications and placental compromise is expressed
strated that preeclamptic women have a 4.3 times increased through various mechanisms including genetic and epigen-
odds of having a fetus with growth restriction compared to etic mechanisms in the intrauterine and postnatal environ-
controls.62 ment, all contributing factors to adult disease (Figure 53.3).
Potential intervention
Suboptimal Potential
Environmental exposure intrauterine monitoring
(e.g., diet) environment (maternal blood)
Placenta
Disruption in epigenetic Changes in
Stochastic factors profile DNA methylation, gene expression
Genetic sex-specific miRNA
effects Pregnancy disease
Aberrant placenta (e.g. PE, IUGR)
development/function
Metabolic/endocrine Modified tissue

disruption function/development
Fetal programming/
Maladaption?
? ?
Adverse birth outcome
including low birth weight
Predisposition to
adult-onset disease (e.g. T2D)
Figure 53.3 Influences and consequences of placental epigenetic disruption. Several factors are known to influence placental
epigenetic profile, including environmental exposures, stochastic events, and genetic and sex effects. Disruption of DNA methylation
patterns can lead to aberrant gene expression and placental development and function. In turn, placental dysfunction can lead to
pregnancy-associated disease as preeclampsia or fetal growth restriction. Disruptions in normal fetal development can lead to predis-
position to adult-onset disease and health problems. (From Novakovic, B. and Saffery, R., Placenta, 33, 959, 2012. With permission.)
Insulin resistance and hypertension in the nonpregnant state 445
Mechanisms underlying metabolic Cardiovascular disease

Peripheral vascular disease
syndrome Nephropathy
Stroke
Metabolic syndrome is characterized by a cluster of clinically
recognizable physiological abnormalities: glucose intoler-
Irreversible complications
ance, high BP, and unfavorable lipid profile—all alterations
induced by the compensatory hyperinsulinemia. It also
involves biochemical abnormalities.68 Upregulation of the Hypertension Diabetes
inflammatory cascade has been recognized as an additional
Dyslipidemia Hyperinsulinemia
risk factor for the impaired cardiovascular component of the
syndrome.69
Insulin resistance now appears to be the epidemiological Clinical threshold
link between high BP and obesity. Insulin resistance induces
hypertension via mechanisms at the cellular, circulatory, Increasing insulin resistance
and neurological levels as well as via possible polygenic fac-
tors. Acquired or transient insulin resistance is associated
with certain physical conditions, such as pregnancy, obesity, Aging, obesity, physiological stress, corticosteroid use
oral contraceptive use, and severe distress. Type 2 DM is a
state of increased insulin secretion caused by the physiologi- Figure 53.4 Factors influencing the generation of insulin
resistance and its clinical correlates in the nonpregnant state.
cal resistance of insulin action and a lower-than-normal
beta-cell reserve. Diabetes in pregnancy or gestational DM
(GDM) may precede the clinical expression of type 2 DM in Dyslipidemia component
the nonpregnant state, even by several years. Preeclampsia Several other metabolic disturbances, such as elevated
and other hypertensive disorders, which are known to have a levels of triglycerides, decreased levels of high-density
higher incidence in GDM, can be linked to increased insulin lipoproteins, high cholesterol level, glucose intolerance,
resistance.70 and hyperuricemia, have also been related to hyperinsu-
linemia.76 The metabolic consequences of these disturbances
include changes in the lipid profile, resulting in atheroscle-
rosis, increased deposition of body fat, and proliferation of
Insulin resistance and hypertension in vascular smooth muscle cells, which place the hypertensive,
the nonpregnant state hyperinsulinemic individual at increased risk of cardiac
complications and stroke.77 Studies of the evolution of the
To understand the association between insulin resistance clinical and biological disturbances in women with a poly-
and hypertensive disorders in pregnancy, we first need cystic ovary (PCO) support the view that insulin resistance,
to elucidate the role of insulin resistance in hypertensive dyslipidemia, and hypertension are all manifestations of a
disorders in the nonpregnant state. The pathogenesis of single syndrome. Often obese, these women have hyperin-
essential hypertension is multifactorial, involving complex sulinemia that disrupts sex hormone production,78 resulting
interactions between endocrine, metabolic, and genetic in androgenization and clinical manifestations of hirsutism
factors. and infertility. During pregnancy, they have more glucose
intolerance79 and pregnancy-induced hypertension (PIH).80
Later in life, women with PCO acquire a male-pattern risk
Obesity component profile for coronary artery disease, including dyslipidemia
The worldwide obesity epidemic has been a major driving and hypertension.81
force in the recognition of metabolic syndrome.68 Several
of the definitions proposed for metabolic syndrome include
increased waist circumference.70,71 This factor is known to Inflammatory component
be associated with a relative predominance of visceral over The association of the metabolic syndrome with inflamma-
subcutaneous adipose tissue,72,73 which results in a higher tion is well documented.82 The increase in proinflammatory
rate of flux of adipose-tissue-derived free fatty acids to the cytokines, including interleukin-6, resistin, tumor necrosis
liver through the splanchnic circulation, thereby effecting factor alpha (TNF-α), and C-reactive protein,83 reflects an
glucose production, lipid synthesis, and prothrombotic overproduction by monocyte-derived macrophages and pos-
protein secretion—all features of metabolic syndrome.73 sibly other cells within the expanded adipose tissue mass.84–86
Obesity, aging, and diabetes can amplify genetic tendencies
toward the clinical expression of the disorder (Figure 53.4).
Familial clustering of DM and hypertension has been Mechanisms of action
reported by several investigators, who also observed a Physiological studies suggest that insulin resistance occurs
close association of insulin resistance with obesity-related primarily in the peripheral muscles and is mediated through
hypertension.74,75 the nonoxidative intracellular pathways of glucose disposal.
Insulin modulates BP through several pathways, including Gestational Pregnancy-induced

stimulation of sympathetic neural activity, direct vasculo- diabetes hypertension Pre-eclampsia
pathic actions, changes in cellular ion flux, and promotion
of sodium retention.87–89
Clinical threshold
polygenic influences, hormonal
variation, obesity, pre-existing disease
Clinical consequences of insulin resistance
Insulin resistance impairs glucose tolerance while pro-
moting dyslipidemia, obesity, hypertension, and athero- Dyslipidemia Insulin Hyperinsulinemia Pre-existing
sclerosis. Patients with three or more of these risk factors resistance hypertension
have an increased incidence of stroke, nephropathy, isch-
emic heart disease, and peripheral vascular disease.90
Long-term diabetic complications are the most common Pregnancy
cause of blindness, renal failure, and limb amputation in Figure 53.5 Risk factors for vascular disease and metabolic
the United States today. Meticulous glycemic control has syndrome in pregnancy.
been shown to decrease the incidence of eye disease among
diabetic patients. Antihypertensive therapy, specifically
with angiotensin-converting enzyme inhibitors (ACE-Is), primary fuel.97 A steady supply of glucose is essential for
is effective in reducing the rate of progression of diabetic the growing fetomaternal unit; normally, pregnant women
kidney disease. To prevent the peripheral vascular remod- are able to increase their insulin secretion to three times
eling that results in stroke, limb loss, and heart disease, that of nonpregnant women.98 In GDM, however, there
the underlying pathophysiological mechanism needs to be is no increase in maternal insulin secretion in reaction
reversed. This has become possible with the introduction to the increasing insulin resistance.99 Some investigators
of metformin91 and troglitazone,92 which are prototypes believe this effect is due to a metabolically limited beta-cell
of the new classes of insulin-sensitivity-enhancing agents. reserve.100,101 In most women with GDM, insulin sensitivity
These drugs are an important addition to the weight loss, is restored after pregnancy. However, some may later develop
exercise, and diet modification programs used to date, life- type 2 DM. The reported cumulative incidence rate of type
style habits that are effective but rarely adhered to for more 2 DM after GDM is approximately 50% after 5 years.102,103 It
than a few years. is even higher in women with excessive weight gain or with
repeated pregnancies, who continue to experience insulin
Insulin resistance and type 1 DM resistance.104 Thus, the strong association between insulin
The recent acceptance of the role of insulin resistance in resistance, hypertension, obesity, and dyslipidemia, as part
type 1 DM was supported by the concurrent rise in the of the metabolic syndrome or sharing a common pathway in
incidence of the disease with a steady increase over the intrauterine life, may explain the higher incidence of hyper-
last 20–30 years in overweight and sedentary habits in tensive complications in diabetic pregnancy. Researchers
children and adolescents in many populations. Medical reported that metabolic syndrome may also involve other
evidence of the link between type 1 DM and insulin resis- metabolic abnormalities besides hyperglycemia, hyperin-
tance/metabolic syndrome continues to grow. The insulin sulinemia, and dyslipidemia, namely, increased concentra-
resistance associated with the rising prevalence of weight tions of plasminogen activator inhibitor-1,105 leptin,106 and
gain may reflect a more aggressive form of autoimmune TNF-α.107 Although these markers are surrogate measures
disease mediated by the same immuno-inflammatory fac- of insulin sensitivity, they have been associated with a risk of
tors that mediate beta-cell destruction, namely, TNF-α and hypertension in pregnancy. In fact, the normal physiological
interleukin-6.93 Moreover, the onset of diabetic nephropa- response to pregnancy has several components, such as insu-
thy might contribute to insulin resistance/metabolic syn- lin resistance, hyperlipidemia, increase in coagulation fac-
drome via mechanisms of low-grade inflammation and tors, and upregulation of the inflammatory cascade, which
increased oxidative stress.94,95 These concepts are included may contribute to a transient and earlier than expected
in the “accelerator hypothesis” on the role of insulin resis- excursion into metabolic syndrome108 (Figure 53.5).
tance and overweight in type 1 DM.96
Gestational diabetes and hypertensive disorders

Insulin resistance and hypertension The study of both GDM and PIH has suffered from the lack
in pregnancy of international consensus about classification, definitions,
and nomenclature, leading to difficulties in comparing stud-
In normal pregnancy, insulin resistance results in a meta- ies that used different diagnostic criteria. Nevertheless, epi-
bolic advantage for the fetus. The mother enters a state of demiological and physiological evidence suggests that GDM
accelerated starvation in which she increases her reliance and PIH are etiologically distinct entities and that GDM is
on lipolysis and protein catabolism as a source of energy. strongly associated with insulin resistance and glucose intol-
Thus, glucose is reserved for the fetus, which uses it as its erance, whereas preeclampsia is probably not.
Insulin resistance and hypertension in pregnancy 447
Complicated similar study in China wherein higher serum insulin con-

pregnancy, e.g., PE, centrations were detected in women with preeclampsia.122
IUGR, miscarriage,
preterm delivery
Besides insulin resistance, women with GDM and women
Threshold for with preeclampsia have similar hemodynamic profiles,
vascular or namely, significant left ventricular hypertrophy and reduced
Cardiovascular risk
metabolic disease
diastolic function.123 The mechanism by which insulin resis-
tance may link these physiological findings is still unknown.
One possibility is an interference of insulin resistance with
the function of an endogenous sodium pump inhibitor or
Healthy digitalis-like factor. Graves et al.124 demonstrated that levels
of serum digoxin-like immunoreactive factor are higher in
women with preexisting diabetes and preeclampsia than in
normotensive diabetic women.
In hypertensive diabetic patients, some of the physiologi-
Pregnancies Middle age
cal changes that occur during pregnancy may persist after
Age
pregnancy. In one study, women with a preeclamptic preg-
Figure 53.6 Factors influencing the generation of insulin nancy showed greater plasma insulin responses and steady-
resistance and its clinical correlates during pregnancy. (From state plasma glucose levels 2 months after delivery than
Sattar, N. and Greer, I.A., Br. Med. J., 325, 157, 2002. With women with uncomplicated pregnancy.125 A longer-term
permission.) study reported persistent mild hyperinsulinemia in women
17 years after a preeclamptic pregnancy, despite their current
Epidemiological studies normoglycemic state.126 Other investigators, however, failed
Diabetic pregnancy is associated with a higher rate of to detect these changes at 3–6 months after delivery.127
hypertensive complications than normal pregnancy109,110
and a slightly increased risk of preeclampsia (15%–20% vs
5%–7%).111–113 The latter holds true even when the diagnosis of Pregestational diabetes and hypertensive complications
GDM is based on the 2 hour 75 g oral glucose tolerance test.114 In most cases, pregestational diabetes refers to type 1 DM.
Mean arterial pressure is further increased in the presence of The incidence of type 1 DM in pregnancy ranges from 0.2%
early diagnosis of GDM and the need for insulin therapy.115 to 0.5%.128,129 Affected women contribute heterogeneous
The increased risk of hypertensive disorders in GDM is group in terms of duration of diabetes, White’s classifica-
probably a result of the combination of insulin resistance and tion, the presence of hypertension, and end-organ damage,
a genetic predisposition (Figure 53.6). A genetic predisposi- especially to the eye (retinopathy) and kidney (nephropa-
tion to PIH was described in southwestern Navajo Indians, thy). Pregnancy in women with type 1 DM is associated
who like other Native Americans are also at increased risk with increased risks of preeclampsia, intrauterine growth
of hypertension, obesity, and DM.116 Preeclampsia was also restriction (IUGR), neonatal morbidity, and perinatal mor-
reported to be associated with increased fasting plasma insu- tality.129–135 The diagnosis of preeclampsia is difficult in
lin levels in African-American women.117 However, these women with preexisting hypertension and proteinuria,129,136
findings have not been confirmed in more heterogeneous and women with chronic hypertension are at increased risk
populations.118 of superimposed preeclampsia independent of the pres-
ence of diabetes.137 The rate of hypertensive disorders (PIH
Physiological studies and preeclampsia) in the various studies ranged from 9%
Patients with the severest form of glucose intolerance are to 66%. The lowest rate occurred in women with milder
more likely to exhibit preeclampsia119 than patients with forms of DM (class B) and the highest in women with dia-
milder forms.118 Controlled studies of the association betic nephropathy. Table 53.1 summarizes the reported rates
between insulin resistance and preeclampsia have been per- of preeclampsia in women with type 1 DM.128,130,138–141 Four
formed in several populations. Martinez et al.120 found that of the six studies noted that rates of preeclampsia increased
among women with normal glucose tolerance in the third with an increasing severity of diabetes, with a mean of 16%
trimester, those who subsequently developed severe pre- (range 9%–24%). Rates were higher in patients with diabetic
eclampsia had similar fasting and postprandial glucose lev- nephropathy (mean 52%, range 35%–66%) (Table 53.2).
els to normotensive controls, but their fasting plasma insulin The risk factors identified for preeclampsia in women with
levels were twofold higher and their postload insulin con- type 1 DM were as follows: duration of diabetes, preexist-
centrations fourfold higher. Moreover, Joffe et al.121 reported ing hypertension, microalbuminuria prior to pregnancy,
that the level of plasma glucose at 1 hour after a 50 g oral glycemic control prior to 20 weeks, nulliparity, minimal
glucose challenge was an important predictor of preeclamp- proteinuria (190–499 mg/dL) before 20 weeks, and nephrop-
sia, even if it was within the normal range. This suggests that athy.118,131,132,142–144 Siddiqi et al.131 listed nulliparity, duration
there is a continuum of insulin resistance also in women of diabetes, and poor glycemic control, and Caritis et al.142
with normal glucose tolerance that may predispose them to nulliparity and mean arterial pressure. Combs et al.132
hypertensive disease. This assumption was supported by a added glycohemoglobin (HbAlc) level >9% at 12–16 weeks of
Table 53.1 Rate of preeclampsia in women with type 1 diabetes

(excluding those with nephropathy)
Preeclampsiaa
Authors No. of women White’s class No. %
Hanson and Persson128 463 B–R 53 11.5
Garner129 107 B,C 13 12.2
Greene et al.130 361 B–R 86 23.8a
Miodovnik et al.138 136 B–R 12 9
Kovilam et al.139 238 B–D 36 15
Sibai et al.140 462 BD 92 20
Total 1767 292 16.4
a Includes women with pregnancy-induced hypertension.
Table 53.2 Rate of preeclampsia in women with diabetic nephropathy
Preeclampsiaa
Authors No. of women No. %
Hanson and Persson 128 31 18 58
Greene et al.130 59 39 66
Gordon et al.133 45 24 53
Reece et al.135 31 11 35
Miodovnik et al.138 46 30 65
Kovilam et al.139 73a 32 44
Sibai et al.140 58 21 36
Bar et al.144 24 11 46
Total 367 186 52
a Includes women with retinopathy and nephropathy.
gestation and proteinuria >190 mg/dL. Accordingly, Hanson pregnancy outcome and long-term effect.144 However, con-
and Persson128 noted a preeclampsia rate of 31% when HbA1c trary to findings in earlier studies of a nonteratogenic effect
levels were >10.1% and a rate of 10.2% when HbA1c levels of ACE-I in early pregnancy,148,149 in 2006, Cooper et al.150
were <10.1%. These findings indicate that PIH/preeclamp- reported an increased risk of major congenital malforma-
sia might be preventable by aggressive control of maternal tions in infants who were exposed to ACE-I in the first tri-
BP before and during pregnancy and prevention of micro- mester compared to infants who were not (OR = 2.71, 95%
albuminuria or reduction of proteinuria level in women CI 1.72–4.27). Other investigators found no increased risk
with diabetic nephropathy. Recently, researchers found that of major congenital malformations.151 Nevertheless, ACE-Is
suboptimal control of hypertension in early pregnancy was are still useful for patients with diabetic nephropathy for a
associated with a significant risk of preterm delivery.145 planned pregnancy. Since conception can take months, pre-
ACE-Is seem to be the ideal agent for the treatment of venting the administration of ACE-I during this period may
hypertension.144,146 According to one long-term follow-up enhance long-term renal function and pregnancy outcome.
study, the survival of patients with diabetic nephropathy has Patients should be advised to use sensitive pregnancy kits
increased from 5 to 7 years to a median of 21.7 years, most for the earliest possible detection of pregnancy, when the
likely because of improvements in aggressive antihyperten- ACE-I must be immediately stopped. The preconception
sive treatment and glycemic control.145 The same group of medication should be accompanied by strict glycemic con-
investigators showed that pregnancy itself has no adverse trol better achieved by the use of insulin pumps, rather than
impact on kidney function and survival in patients with pre- multi-injections.144 Some authors suggested the use of low-
served renal function and diabetic nephropathy.147 Therefore, dose aspirin to prevent preeclampsia, since alterations in the
in women with type 1 DM and preexisting hypertension and metabolism of prostacyclin and thromboxane A2 have been
proteinuria, stringent glycemic control should be started at reported in DM.129,152 However, Caritis et al.,152 in a study of
least 6 months before conception in order to achieve a good 462 women with type 1 DM, found no significant differences
References 449
in the rate of preeclampsia between the aspirin and placebo an excess of known and potential cardiovascular risk fac-
groups, although there was a nonsignificant trend toward tors, and it is a marker of established cardiovascular disease
a lower rate of preeclampsia in the aspirin group (19% vs in both hypertensive158 and nonhypertensive159 individu-
32%, respectively). Larger studies in patients with diabetic als. Its role in diabetic and hypertensive pregnancy is less
nephropathy are needed to clarify this issue. The steady clear,160 but becoming increasingly recognized. One study
increase in overweight in developed countries, concurrent found that the presence of microalbuminuria in the early
with the rising incidence of type 1 DM discovered a new third trimester was predictive of preeclampsia in pregnant
entity of women with pregestational DM with overweight women at risk.161 Furthermore, a significantly higher rate of
and obesity. In the subgroup with diabetic nephropathy, women whose pregnancy was complicated by preeclamp-
maternal overweight was recently found to be associated sia had microalbuminuria 5 years later (40%) than women
with increased risk of pregnancy complications.153 Therefore, with uncomplicated pregnancy.162 Microalbuminuria was
these women should be advised to reduce weight, in addition also noted in 30% of 72 women who had not conceived since
to the prepregnancy counseling program. GDM pregnancy 5–8 years previously—a significantly higher
rate than in women who did not have a history of GDM.163
Ekbom,143 in a study of 68 women with type 1 DM, found
Microalbuminuria, diabetes, and hypertension in that preeclampsia developed in 60% of those who had micro-
pregnancy albuminuria before pregnancy. When the data were fitted
The role of microalbuminuria in DM has been estab- to a logistic regression model, a significant association was
lished over the last 20 years. At the early stage of DM, revealed between microalbuminuria and duration of diabe-
when glucose metabolism is not controlled, the increase tes and preeclampsia. Similar in patients with hypertension,
in glomerular plasma flow and intraglomerular pressure is the progression to diabetic nephropathy in patients with
probably responsible for the increased protein excretion.154 type 1 DM and microalbuminuria can be slowed by block-
Some authors believe these hemodynamic alterations are age of the renin–angiotensin system. Angiotensin II receptor
major determinants of both the initiation and progression antagonists were also recently shown to be renoprotective in
of diabetic nephropathy.155 Several studies have reported patients with type 2 DM and microalbuminuria.158 This find-
that patients with type 1156 or type 2 DM157 who have above- ing and other preliminary results144,146 further suggest that
normal urinary albumin excretion rates are more likely to pregnancy outcome may be improved with prepregnancy
acquire diabetic nephropathy, eventually progressing to ACE-I treatment (discontinued at conception) in diabetic
renal failure.156 Microalbuminuria is also associated with patients with overt proteinuria or microalbuminuria.
REFERENCES
1. Zimmet P, Alberti KG, Shaw J. Global and societal implications of 12. Frankel S, Elwood P, Sweetnam P, Yarnell J, Smith GD.
the diabetes epidemic. Nature 2001; 414: 782–787. Birthweight, adult risk factors and incident coronary heart dis-
2. Kylin E. Studien. Hypertonie-hyperglykamie-hyperurikamie syn- ease: The Caerphilly Study. Public Health 1996; 110: 139–143.
drome. Zentralblatt fur innere Medizine 1923; 44: 105–127. 13. Eriksson JG, Forsen T, Tuomilehto J, Osmond C, Barker DJ. Early
3. Reaven GM. Banting lecture 1988. Role of insulin resistance in growth and coronary heart disease in later life: Longitudinal
human disease. Diabetes 1988; 37: 1595–1607. study. Br Med J 2001; 322: 949–953.
4. DeFronzo RA, Ferrannini E. Insulin resistance. A multifaceted 14. Martyn CN, Barker DJ, Osmond C. Mothers’ pelvic size, fetal
syndrome responsible for NIDDM, obesity, hypertension, dys- growth, and death from stroke and coronary heart disease in men
lipidemia, and atherosclerotic cardiovascular disease. Diabetes in the UK. Lancet 1996; 348: 1264–1268.
Care 1991; 14: 173–194. 15. Barker DJ, Hales CN, Fall CH, Osmond C, Phipps K, Clark PM.
5. Kaplan NM. The deadly quartet. Upper-body obesity, glucose Type 2 (non-insulin-dependent) diabetes mellitus, hypertension
intolerance, hypertriglyceridemia, and hypertension. Arch Intern and hyperlipidaemia (syndrome X): Relation to reduced fetal
Med 1989; 149: 1514–1520. growth. Diabetologia 1993; 36: 62–67.
6. Barker DJP. Mothers, Babies and Health in Later Life, Churchill 16. Bavdekar A, Yajnik CS, Fall CH et al. Insulin resistance syndrome
Livingstone: London, U.K., 1998. in 8-year-old Indian children: Small at birth, big at 8 years, or
7. Leon DA, Lithell HO, Vagero D et al. Reduced fetal growth rate both? Diabetes 1999; 48: 2422–2429.
and increased risk of death from ischaemic heart disease: Cohort 17. Law CM, de Swiet M, Osmond C et al. Initiation of hypertension
study of 15 000 Swedish men and women born 1915–29. Br Med in utero and its amplification throughout life. Br Med J 1993; 306:
J 1998; 317: 241–245. 24–27.
8. Rich-Edwards JW, Stampfer MJ, Manson JE et al. Birth weight and 18. Moore VM, Miller AG, Boulton TJ et al. Placental weight, birth
risk of cardiovascular disease in a cohort of women followed up measurements, and blood pressure at age 8 years. Arch Dis Child
since 1976. Br Med J 1997; 315: 396–400. 1996; 74: 538–541.
9. Stein CE, Fall CH, Kumaran K, Osmond C, Cox V, Barker DJ. Fetal 19. Barker DJ, Osmond C, Golding J, Kuh D, Wadsworth ME. Growth
growth and coronary heart disease in south India. Lancet 1996; in utero, blood pressure in childhood and adult life, and mortality
348: 1269–1273. from cardiovascular disease. Br Med J 1989; 298: 564–567.
10. Eriksson JG, Forsen T, Tuomilehto J, Winter PD, Osmond C, 20. Hales CN. Non-insulin-dependent diabetes mellitus. Br Med Bull
Barker DJ. Catch-up growth in childhood and death from cor- 1997; 53: 109–122.
onary heart disease: Longitudinal study. Br Med J 1999; 318: 21. Whincup P, Cook D, Papacosta O, Walker M. Birth weight and
427–431. blood pressure: Cross sectional and longitudinal relations in
11. Forsen T, Eriksson JG, Tuomilehto J, Osmond C, Barker DJ. childhood. Br Med J 1995; 311: 773–776.
Growth in utero and during childhood among women who 22. Barker DJ, Martyn CN, Osmond C, Wield GA. Abnormal liver
develop coronary heart disease: Longitudinal study. Br Med J growth in utero and death from coronary heart disease. Br Med J
1999; 319: 1403–1407. 1995; 310: 703–704.
23. Taylor SJ, Whincup PH, Cook DG, Papacosta O, Walker M. Size 49. Brosens I, Pijnenborg R, Vercruysse L, Romero R. The “Great
at birth and blood pressure: Cross sectional study in 8–11 year Obstetrical Syndromes” are associated with disorders of deep
old children. Br Med J 1997; 314: 475–480. placentation. Am J Obstet Gynecol 2011; 204: 193–201.
24. Gupta R, Gupta VP, Ahluwalia NS. Educational status, coronary 50. Baschat AA. Fetal responses to placental insufficiency: An
heart disease, and coronary risk factor prevalence in a rural pop- update. BJOG 2004; 111: 1031–1041.
ulation of India. Br Med J 1994; 309: 1332–1336. 51. Pardi G, Marconi AM, Cetin I. Placental-fetal interrelationship in
25. McKeigue PM, Shah B, Marmot MG. Relation of central obesity IUGR fetuses—A review. Placenta 2002; 23 (Suppl A): S136–S141.
and insulin resistance with high diabetes prevalence and cardio- 52. Salafia CM, Vintzileos AM, Silberman L, Bantham KF, Vogel CA.
vascular risk in South Asians. Lancet 1991; 337: 382–386. Placental pathology of idiopathic intrauterine growth retardation
26. Pais P, Pogue J, Gerstein H et al. Risk factors for acute myocar- at term. Am J Perinatol 1992; 9: 179–184.
dial infarction in Indians: A case-control study. Lancet 1996; 348: 53. Vedmedovska N, Rezeberga D, Teibe U, Melderis I, Donders
358–363. GG. Placental pathology in fetal growth restriction. Eur J Obstet
27. Hales CN, Barker DJ, Clark PM et al. Fetal and infant growth Gynecol Reprod Biol 2010; 155: 36–40.
and impaired glucose tolerance at age 64. Br Med J 1991; 303: 54. Fok RY, Pavlova Z, Benirschke K, Paul RH, Platt LD. The correla-
1019–1022. tion of arterial lesions with umbilical artery Doppler velocimetry
28. Yajnik CS, Fall CH, Vaidya U et al. Fetal growth and glucose and in the placentas of small-for-dates pregnancies. Obstet Gynecol
insulin metabolism in four-year-old Indian children. Diabet Med 1990; 75: 578–583.
1995; 12: 330–336. 55. Dicke JM, Huettner P, Yan S, Odibo A, Kraus FT. Umbilical artery
29. McLeod KI, Goldrick RB, Whyte HM. The effect of maternal Doppler indices in small for gestational age fetuses: Correlation
malnutrition on the progeny in the rat. Studies on growth, body with adverse outcomes and placental abnormalities. J Ultrasound
composition and organ cellularity in first and second generation Med 2009; 28: 1603–1610.
progeny. Aust J Exp Biol Med Sci 1972; 50: 435–446. 56. Mayhew TM, Ohadike C, Baker PN, Crocker IP, Mitchell C, Ong
30. Lemonnier D, Suquet JP, Aubert R, Rosselin G. Long term effect SS. Stereological investigation of placental morphology in preg-
of mouse neonate food intake on adult body composition, insulin nancies complicated by pre-eclampsia with and without intra-
and glucose serum levels. Horm Metab Res 1973; 5: 223–224. uterine growth restriction. Placenta 2003; 24: 219–226.
31. Wallace JM, Aitken RP, Milne JS, Hay WW Jr. Nutritionally medi- 57. Kovo M, Schreiber L, Ben-Haroush A et al. The placental fac-
ated placental growth restriction in the growing adolescent: tor in early- and late-onset normotensive fetal growth restriction.
Consequences for the fetus. Biol Reprod 2004; 71: 1055–1062. Placenta 2013; 34: 320–324.
32. Moore SE, Halsall I, Howarth D, Poskitt EM, Prentice AM. 58. Bellamy L, Casas JP, Hingorani AD, Williams DJ. Pre-eclampsia
Glucose, insulin and lipid metabolism in rural Gambians exposed and risk of cardiovascular disease and cancer in later life:
to early malnutrition. Diabet Med 2001; 18: 646–653. Systematic review and meta-analysis. Br Med J 2007; 335: 974.
33. Stene LC, Magnus P, Lie RT, Sovik O, Joner G. Birth weight and 59. Xu Y, Shen S, Sun L, Yang H, Jin B, Cao X. Metabolic syndrome
childhood onset type 1 diabetes: Population based cohort study. risk after gestational diabetes: A systematic review and meta-
Br Med J 2001; 322: 889–892. analysis. PLOS ONE 2014; 9: e87863.
34. Leon DA. Twins and fetal programming of blood pressure. 60. Ben-Ami S, Oron G, Ben-Haroush A, Blickstein D, Hod M, Bar
Questioning the role of genes and maternal nutrition. Br Med J J. Primary atherothrombotic occlusive vascular events in pre-
1999; 319: 1313–1314. menopausal women with history of adverse pregnancy outcome.
35. Langley-Evans SC, Gardner DS, Welham SJ. Intrauterine pro- Thromb Res 2010; 125: 124–127.
gramming of cardiovascular disease by maternal nutritional sta- 61. Shalom G, Shoham-Vardi I, Sergienko R, Wiznitzer A, Sherf M,
tus. Nutrition 1998; 14: 39–47. Sheiner E. Is preeclampsia a significant risk factor for long-term
36. Kramer MS. Determinants of low birth weight: Methodological hospitalizations and morbidity? J Matern Fetal Neonatal Med
assessment and meta-analysis. Bull World Health Organ 1987; 2013; 26: 13–15.
65: 663–737. 62. Srinivas SK, Edlow AG, Neff PM, Sammel MD, Andrela CM,
37. Gliksman MD, Kawachi I, Hunter D et al. Childhood socioeco- Elovitz MA. Rethinking IUGR in preeclampsia: Dependent or
nomic status and risk of cardiovascular disease in middle aged independent of maternal hypertension? J Perinatol 2009; 29:
US women: A prospective study. J Epidemiol Community Health 680–684.
1995; 49: 10–15. 63. Bhargava SK, Sachdev HS, Fall CH et al. Relation of serial changes
38. Susser M, Levin B. Ordeals for the fetal programming hypothesis. in childhood body-mass index to impaired glucose tolerance in
The hypothesis largely survives one ordeal but not another. Br young adulthood. N Engl J Med 2004; 350: 865–875.
Med J 1999; 318: 885–886. 64. Barker DJ, Osmond C, Forsen TJ, Kajantie E, Eriksson JG.
39. Lucas A, Fewtrell MS, Cole TJ. Fetal origins of adult disease-the Trajectories of growth among children who have coronary events
hypothesis revisited. Br Med J 1999; 319: 245–249. as adults. N Engl J Med 2005; 353: 1802–1809.
40. Dwyer T, Blizzard L, Morley R, Ponsonby AL. Within pair asso- 65. Bateson P, Barker D, Clutton-Brock T et al. Developmental plas-
ciation between birth weight and blood pressure at age 8 in twins ticity and human health. Nature 2004; 430: 419–421.
from a cohort study. Br Med J 1999; 319: 1325–1329. 66. Gluckman PD, Hanson MA, Cooper C, Thornburg KL. Effect of
41. Poulter NR, Chang CL, MacGregor AJ, Snieder H, Spector TD. in utero and early-life conditions on adult health and disease.
Association between birth weight and adult blood pressure in N Engl J Med 2008; 359: 61–73.
twins: Historical cohort study. Br Med J 1999; 319: 1330–1333. 67. Novakovic B, Saffery R. The ever growing complexity of placen-
42. Martin H, Hu J, Gennser G, Norman M. Impaired endothelial tal epigenetics—Role in adverse pregnancy outcomes and fetal
function and increased carotid stiffness in 9-year-old children programming. Placenta 2012; 33: 959–970.
with low birthweight. Circulation 2000; 102: 2739–2744. 68. Eckel RH, Grundy SM, Zimmet PZ. The metabolic syndrome.
43. Singhal A, Kattenhorn M, Cole TJ, Deanfield J, Lucas A. Preterm Lancet 2005; 365: 1415–1428.
birth, vascular function, and risk factors for atherosclerosis. 69. Haffner SM. Do interventions to reduce coronary heart disease
Lancet 2001; 358: 1159–1160. reduce the incidence of type 2 diabetes? A possible role for
44. Nelson RG. Intrauterine determinants of diabetic kidney disease in inflammatory factors. Circulation 2001; 103: 346–347.
disadvantaged populations. Kidney Int Suppl 2003; (83): S13–S16. 70. Alberti KG, Zimmet PZ. Definition, diagnosis and classification
45. Koleganova N, Benz K, Piecha G, Ritz E, Amann K. Renal, car- of diabetes mellitus and its complications. Part 1: Diagnosis and
diovascular and metabolic effects of fetal programming. Nephrol classification of diabetes mellitus provisional report of a WHO
Dial Transplant 2012; 27: 3003–3007. consultation. Diabet Med 1998; 15: 539–553.
46. Huxley R, Owen CG, Whincup PH et al. Is birth weight a risk fac- 71. Executive summary of the third report of the National Cholesterol
tor for ischemic heart disease in later life? Am J Clin Nutr 2007; Education Program (NCEP) expert panel on detection, evaluation,
85: 1244–1250. and treatment of high blood cholesterol in adults (adult treatment
47. Skogen JC, Overland S. The fetal origins of adult disease: A nar- panel III). J Am Med Assoc 2001; 285: 2486–2497.
rative review of the epidemiological literature. JRSM Short Rep 72. Tanaka S, Horimai C, Katsukawa F. Ethnic differences in abdomi-
2012; 3: 59. nal visceral fat accumulation between Japanese, African-
48. Khong Y, Brosens I. Defective deep placentation. Best Pract Res Americans, and Caucasians: A meta-analysis. Acta Diabetol
Clin Obstet Gynaecol 2010; 25: 301–311. 2003; 40(Suppl 1): S302–S304.
References 451
73. Aubert H, Frere C, Aillaud MF, Morange PE, Juhan-Vague I, Alessi 98. Bergman RN, Phillips LS, Cobelli C. Physiologic evaluation of fac-
MC. Weak and non-independent association between plasma tors controlling glucose tolerance in man: Measurement of insulin
TAFI antigen levels and the insulin resistance syndrome. J Thromb sensitivity and beta-cell glucose sensitivity from the response to
Haemost 2003; 1: 791–797. intravenous glucose. J Clin Invest 1981; 68: 1456–1467.
74. Modan M, Halkin H, Almog S et al. Hyperinsulinemia. A link 99. Buchanan TA, Metzger BE, Freinkel N, Bergman RN. Insulin
between hypertension obesity and glucose intolerance. J Clin sensitivity and B-cell responsiveness to glucose during late preg-
Invest 1985; 75: 809–817. nancy in lean and moderately obese women with normal glu-
75. O’Hare JA. The enigma of insulin resistance and hypertension. cose tolerance or mild gestational diabetes. Am J Obstet Gynecol
Insulin resistance, blood pressure, and the circulation. Am J Med 1990; 162: 1008–1014.
1988; 84: 505–510. 100. Yen SS, Tsai CC, Vela P. Gestational diabetogenesis: Quantitative
76. Schmidt MI, Watson RL, Duncan BB et al. Clustering of dyslip- analyses of glucose-insulin interrelationship between normal
idemia, hyperuricemia, diabetes, and hypertension and its asso- pregnancy and pregnancy with gestational diabetes. Am J Obstet
ciation with fasting insulin and central and overall obesity in a Gynecol 1971; 111: 792–800.
general population. Atherosclerosis Risk in Communities Study 101. Freinkel N, Metzger BE, Phelps RL et al. Gestational diabetes
Investigators. Metabolism 1996; 45: 699–706. mellitus. Heterogeneity of maternal age, weight, insulin secre-
77. Harano Y, Suzuki M, Shinozaki K et al. Clinical impact of insulin tion, HLA antigens, and islet cell antibodies and the impact of
resistance syndrome in cardiovascular diseases and its therapeu- maternal metabolism on pancreatic B-cell and somatic develop-
tic approach. Hypertens Res 1996; 19(Suppl 1): S81–S85. ment in the offspring. Diabetes 1985; 34(Suppl 2): 1–7.
78. Crave JC, Fimbel S, Lejeune H, Cugnardey N, Dechaud H, Pugeat 102. Metzger BE, Cho NH, Roston SM, Radvany R. Prepregnancy
M. Effects of diet and metformin administration on sex hormone- weight and antepartum insulin secretion predict glucose toler-
binding globulin, androgens, and insulin in hirsute and obese ance five years after gestational diabetes mellitus. Diabetes Care
women. J Clin Endocrinol Metab 1995; 80: 2057–2062. 1993; 16: 1598–1605.
79. Lanzone A, Caruso A, Di Simone N, De Carolis S, Fulghesu AM, 103. O’Sullivan JB. The Boston gestational diabetes studies: Review
Mancuso S. Polycystic ovary disease. A risk factor for gestational and perspectives. In: Sutherland HW, Stowers JM, Pearson
diabetes? J Reprod Med 1995; 40: 312–316. DWM, eds., Carbohydrate Metabolism in Pregnancy and the
80. Urman B, Sarac E, Dogan L, Gurgan T. Pregnancy in infertile Newborn, Springer-Verlag: London, U.K., 1989, pp. 287–294.
PCOD patients. Complications and outcome. J Reprod Med 104. Peters RK, Kjos SL, Xiang A, Buchanan TA. Long-term diabeto-
1997; 42: 501–505. genic effect of single pregnancy in women with previous gesta-
81. Talbott E, Guzick D, Clerici A et al. Coronary heart disease risk tional diabetes mellitus. Lancet 1996; 347: 227–230.
factors in women with polycystic ovary syndrome. Arterioscler 105. Abbasi F, McLaughlin T, Lamendola C, Lipinska I, Tofler G,
Thromb Vasc Biol 1995; 15: 821–826. Reaven GM. Comparison of plasminogen activator inhibitor-1
82. Sutherland JP, McKinley B, Eckel RH. The metabolic syndrome concentration in insulin-resistant versus insulin-sensitive healthy
and inflammation. Metab Syndr Relat Disord 2004; 2: 82–104. women. Arterioscler Thromb Vasc Biol 1999; 19: 2818–2821.
83. Fernandez-Real JM, Ricart W. Insulin resistance and chronic 106. Segal KR, Landt M, Klein S. Relationship between insulin sensi-
cardiovascular inflammatory syndrome. Endocr Rev 2003; 24: tivity and plasma leptin concentration in lean and obese men.
278–301. Diabetes 1996; 45: 988–991.
84. Trayhurn P, Wood IS. Adipokines: Inflammation and the pleiotro- 107. Fernandez-Real JM, Broch M, Ricart W et al. Plasma levels of the
pic role of white adipose tissue. Br J Nutr 2004; 92: 347–355. soluble fraction of tumor necrosis factor receptor 2 and insulin
85. Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, resistance. Diabetes 1998; 47: 1757–1762.
Ferrante AW Jr. Obesity is associated with macrophage accumu- 108. Sattar N, Greer IA. Pregnancy complications and maternal car-
lation in adipose tissue. J Clin Invest 2003; 112: 1796–1808. diovascular risk: Opportunities for intervention and screening? Br
86. Xu H, Barnes GT, Yang Q et al. Chronic inflammation in fat plays Med J 2002; 325: 157–160.
a crucial role in the development of obesity-related insulin resis- 109. Greco P, Loverro G, Selvaggi L. Does gestational diabetes rep-
tance. J Clin Invest 2003; 112: 1821–1830. resent an obstetrical risk factor? Gynecol Obstet Invest 1994; 37:
87. Capaldo B, Lembo G, Napoli R et al. Skeletal muscle is a primary 242–245.
site of insulin resistance in essential hypertension. Metabolism 110. Rudge MV, Calderon IM, Ramos MD, Peracoli JC, Pim A.
1991; 40: 1320–1322. Hypertensive disorders in pregnant women with diabetes mel-
88. Levy J, Zemel MB, Sowers JR. Role of cellular calcium metabo- litus. Gynecol Obstet Invest 1997; 44: 11–15.
lism in abnormal glucose metabolism and diabetic hypertension. 111. Sacks DA, Greenspoon JS, Abu-Fadil S, Henry HM, Wolde-
Am J Med 1989; 87: 7S–16S. Tsadik G, Yao JF. Toward universal criteria for gestational dia-
89. Resnick LM. Cellular ions in hypertension, insulin resistance, betes: The 75-gram glucose tolerance test in pregnancy. Am J
obesity, and diabetes: A unifying theme. J Am Soc Nephrol 1992; Obstet Gynecol 1995; 172: 607–614.
3: S78–S85. 112. Sermer M, Naylor CD, Gare DJ et al. Impact of increasing carbo-
90. Gilbert RE, Jerums G, Cooper ME. Diabetes and hypertension: hydrate intolerance on maternal-fetal outcomes in 3637 women
Prognostic and therapeutic considerations. Blood Press 1995; 4: without gestational diabetes. The Toronto Tri-Hospital Gestational
329–338. Diabetes Project. Am J Obstet Gynecol 1995; 173: 146–156.
91. DeFronzo RA, Goodman AM. Efficacy of metformin in patients 113. Pennison EH, Egerman RS. Perinatal outcomes in gestational
with non-insulin-dependent diabetes mellitus. The Multicenter diabetes: A comparison of criteria for diagnosis. Am J Obstet
Metformin Study Group. N Engl J Med 1995; 333: 541–549. Gynecol 2001; 184: 1118–1121.
92. Nolan JJ, Ludvik B, Beerdsen P, Joyce M, Olefsky J. Improvement 114. Schmidt MI, Duncan BB, Reichelt AJ et al. Gestational diabetes
in glucose tolerance and insulin resistance in obese subjects mellitus diagnosed with a 2-h 75-g oral glucose tolerance test and
treated with troglitazone. N Engl J Med 1994; 331: 1188–1193. adverse pregnancy outcomes. Diabetes Care 2001; 24: 1151–1155.
93. Banerji MA. Impaired beta-cell and alpha-cell function in African- 115. Schaffir JA, Lockwood CJ, Lapinski R, Yoon L, Alvarez M.
American children with type 2 diabetes mellitus—“Flatbush dia- Incidence of pregnancy-induced hypertension among gestational
betes.” J Pediatr Endocrinol Metab 2002; 15(Suppl 1): 493–501. diabetics. Am J Perinatol 1995; 12: 252–254.
94. Thorn LM, Forsblom C, Fagerudd J et al. Metabolic syndrome in type 116. Levy MT, Jacober SJ, Sowers JR. Hypertensive disorders of preg-
1 diabetes: Association with diabetic nephropathy and glycemic nancy in southwestern Navajo Indians. Arch Intern Med 1994;
control (the FinnDiane study). Diabetes Care 2005; 28: 2019–2024. 154: 2181–2183.
95. Svensson M, Eriksson JW. Insulin resistance in diabetic 117. Sowers JR, Saleh AA, Sokol RJ. Hyperinsulinemia and insu-
nephropathy—Cause or consequence? Diabetes Metab Res Rev lin resistance are associated with preeclampsia in African-
2006; 22: 401–410. Americans. Am J Hypertens 1995; 8: 1–4.
96. Wilkin TJ. The accelerator hypothesis: Weight gain as the missing 118. Cioffi FJ, Amorosa LF, Vintzileos AM et al. Relationship of insulin
link between type I and type II diabetes. Diabetologia 2001; 44: resistance and hyperinsulinemia to blood pressure during preg-
914–922. nancy. J Matern Fetal Med 1997; 6: 174–179.
97. Kalkhoff RK, Kissebah AH, Kim HJ. Carbohydrate and lipid 119. Solomon CG, Graves SW, Greene MF, Seely EW. Glucose intoler-
metabolism during normal pregnancy: Relationship to gestational ance as a predictor of hypertension in pregnancy. Hypertension
hormone action. Semin Perinatol 1978; 2: 291–307. 1994; 23: 717–721.
120. Martinez Abundis E, Gonzalez Ortiz M, Quinones Galvan A, 141. Garner PR, D’Alton ME, Dudley DK, Huard P, Hardie M.
Ferrannini E. Hyperinsulinemia in glucose-tolerant women with Preeclampsia in diabetic pregnancies. Am J Obstet Gynecol
preeclampsia. A controlled study. Am J Hypertens 1996; 9: 1990; 163: 505–508.
610–614. 142. Caritis S, Sibai B, Hauth J et al. Predictors of pre-eclampsia in
121. Joffe GM, Esterlitz JR, Levine RJ et al. The relationship between women at high risk. National Institute of Child Health and Human
abnormal glucose tolerance and hypertensive disorders of preg- Development Network of Maternal-Fetal Medicine Units. Am J
nancy in healthy nulliparous women. Calcium for Preeclampsia Obstet Gynecol 1998; 179: 946–951.
Prevention (CPEP) Study Group. Am J Obstet Gynecol 1998; 179: 143. Ekbom P. Pre-pregnancy microalbuminuria predicts pre-
1032–1037. eclampsia in insulin-dependent diabetes mellitus. Copenhagen
122. Gu H, Rong L, Sai JY. [Insulin resistance and pregnancy induced Pre-eclampsia in Diabetic Pregnancy Study Group. Lancet 1999;
hypertension]. Zhonghua Fu Chan Ke Za Zhi 1994; 29: 711–713, 353: 377.
760. 144. Bar J, Chen R, Schoenfeld A et al. Pregnancy outcome in patients
123. Oren S, Golzman B, Reitblatt T, Turkot S, Kogan J, Segal S. with insulin dependent diabetes mellitus and diabetic nephrop-
Gestational diabetes mellitus and hypertension in pregnancy: athy treated with ACE inhibitors before pregnancy. J Pediatr
Hemodynamics and diurnal arterial pressure profile. J Hum Endocrinol Metab 1999; 12: 659–665.
Hypertens 1996; 10: 505–509. 145. Astrup AS, Tarnow L, Rossing P, Pietraszek L, Riis Hansen P,
124. Graves SW, Lincoln K, Cook SL, Seely EW. Digitalis-like factor Parving HH. Improved prognosis in type 1 diabetic patients with
and digoxin-like immunoreactive factor in diabetic women with nephropathy: A prospective follow-up study. Kidney Int 2005;
preeclampsia, transient hypertension of pregnancy, and normo- 68: 1250–1257.
tensive pregnancy. Am J Hypertens 1995; 8: 5–11. 146. Hod M, van Dijk DJ, Karp M et al. Diabetic nephropathy and
125. Fuh MM, Yin CS, Pei D et al. Resistance to insulin-mediated pregnancy: The effect of ACE inhibitors prior to pregnancy
glucose uptake and hyperinsulinemia in women who had pre- on fetomaternal outcome. Nephrol Dial Transplant 1995; 10:
eclampsia during pregnancy. Am J Hypertens 1995; 8: 768–771. 2328–2333.
126. Laivuori H, Tikkanen MJ, Ylikorkala O. Hyperinsulinemia 17 years 147. Rossing K, Jacobsen P, Hommel E et al. Pregnancy and progres-
after preeclamptic first pregnancy. J Clin Endocrinol Metab 1996; sion of diabetic nephropathy. Diabetologia 2002; 45: 36–41.
81: 2908–2911. 148. Bar J, Hod M, Merlob P. Angiotensin converting enzyme inhibi-
127. Jacober SJ, Morris DA, Sowers JR. Postpartum blood pressure and tors use in the first trimester of pregnancy. Int J Risk Saf Med 1997;
insulin sensitivity in African-American women with recent pre- 10: 23–26.
eclampsia. Am J Hypertens 1994; 7: 933–936. 149. Lip GY, Churchill D, Beevers M, Auckett A, Beevers DG.
128. Hanson U, Persson B. Epidemiology of pregnancy-induced Angiotensin-converting-enzyme inhibitors in early pregnancy.
hypertension and preeclampsia in type 1 (insulin-dependent) dia- Lancet 1997; 350: 1446–1447.
betic pregnancies in Sweden. Acta Obstet Gynecol Scand 1998; 150. Cooper WO, Hernandez-Diaz S, Arbogast PG et al. Major con-
77: 620–624. genital malformations after first-trimester exposure to ACE inhibi-
129. Garner P. Type I diabetes mellitus and pregnancy. Lancet 1995; tors. N Engl J Med 2006; 354: 2443–2451.
346: 157–161. 151. Bullo M, Tschumi S, Bucher BS, Bianchetti MG, Simonetti GD.
130. Greene MF, Hare JW, Krache M et al. Prematurity among insulin- Pregnancy outcome following exposure to angiotensin-converting
requiring diabetic gravid women. Am J Obstet Gynecol 1989; enzyme inhibitors or angiotensin receptor antagonists: A system-
161: 106–111. atic review. Hypertension 2012; 60: 444–450.
131. Siddiqi T, Rosenn B, Mimouni F, Khoury J, Miodovnik M. 152. Caritis S, Sibai B, Hauth J et al. Low-dose aspirin to prevent
Hypertension during pregnancy in insulin-dependent diabetic preeclampsia in women at high risk. National Institute of Child
women. Obstet Gynecol 1991; 77: 514–519. Health and Human Development Network of Maternal-Fetal
132. Combs CA, Rosenn B, Kitzmiller JL, Khoury JC, Wheeler BC, Medicine Units. N Engl J Med 1998; 338: 701–705.
Miodovnik M. Early-pregnancy proteinuria in diabetes related to 153. Yogev Y, Chen R, Ben-Haroush A, Hod M, Bar J. Maternal over-
preeclampsia. Obstet Gynecol 1993; 82: 802–807. weight and pregnancy outcome in women with type-1 diabe-
133. Gordon M, Landon MB, Samuels P, Hissrich S, Gabbe SG. tes mellitus and different degrees of nephropathy. J Matern Fetal
Perinatal outcome and long-term follow-up associated with mod- Neonatal Med 2010; 23: 999–1003.
ern management of diabetic nephropathy. Obstet Gynecol 1996; 154. Viberti G, Mackintosh D, Keen H. Determinants of the pen-
87: 401–409. etration of proteins through the glomerular barrier in insulin-
134. Hanson U, Persson B. Outcome of pregnancies complicated by dependent diabetes mellitus. Diabetes 1983; 32(Suppl 2): 92–95.
type 1 insulin-dependent diabetes in Sweden: Acute pregnancy 155. Hostetter TH, Rennke HG, Brenner BM. The case for intrarenal
complications, neonatal mortality and morbidity. Am J Perinatol hypertension in the initiation and progression of diabetic and
1993; 10: 330–333. other glomerulopathies. Am J Med 1982; 72: 375–380.
135. Reece EA, Leguizamon G, Homko C. Pregnancy performance 156. Viberti GC, Hill RD, Jarrett RJ, Argyropoulos A, Mahmud U, Keen
and outcomes associated with diabetic nephropathy. Am J H. Microalbuminuria as a predictor of clinical nephropathy in
Perinatol 1998; 15: 413–421. insulin-dependent diabetes mellitus. Lancet 1982; 1: 1430–1432.
136. Diamond MP, Shah DM, Hester RA, Vaughn WK, Cotton RB, 157. Schmitz A, Vaeth M. Microalbuminuria: A major risk factor in
Boehm FH. Complication of insulin-dependent diabetic pregnan- non-insulin-dependent diabetes. A 10-year follow-up study of
cies by preeclampsia and/or chronic hypertension: Analysis of 503 patients. Diabet Med 1988; 5: 126–134.
outcome. Am J Perinatol 1985; 2: 263–267. 158. Parving HH, Lehnert H, Brochner-Mortensen J, Gomis R,
137. Sibai BM, Lindheimer M, Hauth J et al. Risk factors for pre- Andersen S, Arner P. The effect of irbesartan on the development
eclampsia, abruptio placentae, and adverse neonatal out- of diabetic nephropathy in patients with type 2 diabetes. N Engl J
comes among women with chronic hypertension. National Med 2001; 345: 870–878.
Institute of Child Health and Human Development Network 159. Yudkin JS, Forrest RD, Jackson CA. Microalbuminuria as pre-
of Maternal-Fetal Medicine Units. N Engl J Med 1998; 339: dictor of vascular disease in non-diabetic subjects. Islington
667–671. Diabetes Survey. Lancet 1988; 2: 530–533.
138. Miodovnik M, Rosenn BM, Khoury JC, Grigsby JL, Siddiqi TA. 160. Bar J, Hod M, Erman A, Friedman S, Ovadia Y. Microalbuminuria:
Does pregnancy increase the risk for development and progres- Prognostic and therapeutic implications in diabetic and hyper-
sion of diabetic nephropathy? Am J Obstet Gynecol 1996; 174: tensive pregnancy. Diabet Med 1995; 12: 649–656.
1180–1189; discussion 1189–1191. 161. Bar J, Hod M, Erman A et al. Microalbuminuria as an early pre-
139. Kovilam O, Rosenn B, Miodovnik M et al. Is proliferative retinop- dictor of hypertensive complications in pregnant women at high
athy a risk factor for adverse pregnancy outcome in women with risk. Am J Kidney Dis 1996; 28: 220–225.
Type 1 diabetes? J Soc Gynecol Invest 1997; 4(Suppl 1): 152A. 162. Bar J, Kaplan B, Wittenberg C et al. Microalbuminuria after preg-
140. Sibai BM, Caritis S, Hauth J et al. Risks of preeclampsia and nancy complicated by pre-eclampsia. Nephrol Dial Transplant
adverse neonatal outcomes among women with pregestational 1999; 14: 1129–1132.
diabetes mellitus. National Institute of Child Health and Human 163. Friedman S, Rabinerson D, Bar J et al. Microalbuminuria follow-
Development Network of Maternal-Fetal Medicine Units. Am J ing gestational diabetes. Acta Obstet Gynecol Scand 1995; 74:
Obstet Gynecol 2000; 182: 364–369. 356–360.
54 Diabetic retinopathy
Nir Melamed and Moshe Hod
Diabetic retinopathy is the most common chronic complica- Pathogenesis and natural history
tion of diabetes mellitus1 and is the most common cause of Diabetic retinopathy is the result of several pathological
blindness in middle-aged subjects in the United States and the processes that include loss of capillary pericytes, damage
United Kingdom.2–5 The mutual effects of pregnancy and reti- to capillary wall resulting in increased permeability and
nopathy, though long a subject of research and debate, remain weakness of capillary wall, microvascular occlusion lead-
unclear, and data on methods of diagnosis and management ing to retinal ischemia, and proliferation of new blood ves-
in pregnancy remain scarce.6–10 The purpose of this chapter is sels.25–28 It is currently believed that chronic hyperglycemia
to review the current literature on the effects of pregnancy on is the primary cause of diabetic retinopathy.29–31 Although
diabetic retinopathy and to provide guidelines for the man- the exact mechanisms by which hyperglycemia initiates
agement of pregnancies complicated by diabetic retinopathy. these pathological processes remain unclear, putative mech-
anisms include impaired autoregulation of retinal blood
flow (RBF),32 increased production of sorbitol by the enzyme
Diabetic retinopathy in nonpregnant aldose reductase,33 accumulation of advanced glycation
end products,34 oxidative stress,35–37 increased platelet aggre-
patients gation and hypercoagulability,38 and alteration in cell signal-
Epidemiology ing pathways such as diacylglycerol-induced protein kinase
C activity.39–41
The prevalence of diabetic retinopathy is almost 100% in
patients with type 1 diabetes and over 60% in patients with
type 2 diabetes in whom the disease has been present for Nonproliferative diabetic retinopathy
more than 20 years.11–13 In the population-based Wisconsin Microaneurysms, hypercellular outpouchings from the cap-
Epidemiologic Study of Diabetic Retinopathy, the prevalence illary wall, are the earliest signs of diabetic retinopathy. They
of proliferative diabetic retinopathy was 1.2% and 67% in appear as small red dots (Figure 54.1) and can be identified
persons with type 1 diabetes for less than 10 years and more in 2% and 98% of the patients with type 1 diabetes in whom
than 35 years, respectively. the disease has been present for 2 and 15 years, respectively.42
It is estimated that 20%–30% of diabetic women in the Possible mechanisms for the formation of microaneurysms
reproductive age group have some evidence of retinopathy.1 include weakness of capillary wall and increased intralu-
Risk factors for diabetic retinopathy include longer duration minal pressure.25,26 Rupture of microaneurysms results in
of the disease, onset of disease before 30 years of age, poor retinal hemorrhages that may appear as either small dots or
glycemic control manifested by higher levels of glycosylated larger flame-shaped hemorrhages. As a result of the increased
hemoglobin, hypertension, hypertriglyceridemia, anemia, capillary permeability, lipids and proteins leak into the outer
and evidence of diabetic nephropathy.4,11,14–21 retina, appearing as well-defined yellow-white deposits,
known as hard exudates (Figure 54.2). Extravasation of flu-
ids may lead to retinal thickening and edema. Although, in
Classification of diabetic retinopathy general, nonproliferative retinopathy does not impair vision,
Diabetic retinopathy is a progressive disorder and is gener- hard exudates or retinal edema involving the macular area
ally categorized as either nonproliferative (NPDR, or back- may lead to severe loss of central vision.43 Thus, diabetic
ground) or proliferative, although more detailed grading macular edema (DME) is the most common cause of vision
systems exist,22,23 as presented in Table 54.1. Grading is per- loss in NPDR.44
formed according to the semiquantitative assessment of the
morphological lesions on fundus photographs. The grading Preproliferative diabetic retinopathy
system considers mainly the type and number of retinopathy As the disease progresses, retinal capillaries become
lesions, while the diagnostic value of the regional distribu- occluded, leading to retinal ischemia. This stage, known
tion of the lesions is largely unknown.24 as preproliferative retinopathy, is characterized by the
453
Table 54.1 Diabetic retinopathy severity scale
Classification Description
Normal No evidence of retinopathy
Nonproliferative
Minimal Microaneurysms only
Mild Microaneurysms plus hard exudates, soft exudates (cotton wool spots), mild retinal
hemorrhages, venous loops
Moderate Microaneurysms plus mild intraretinal microvascular abnormalities (IRMA), moderate
retinal hemorrhages, venous beading
Severe Microaneurysms plus moderate to severe IRMA, severe retinal hemorrhages, venous beading
Proliferative
Proliferative New vessels in the retina or on the optic disc
High-risk proliferative Proliferative plus preretinal hemorrhage, fibrous tissue, or other high-risk characteristics
Advanced diabetic eye dis. Proliferative plus vitreous hemorrhage, retinal detachment, or rubeosis iridis
Source: Adapted from Early Treatment Diabetic Retinopathy Study Research Group, Ophthalmology, 98, 786, 1991.
Figure 54.2 Nonproliferative diabetic retinopathy.
membrane of the retina and become adherent, with their

Figure 54.1 Normal fundus. accompanying fibrovascular tissue, to the posterior vitreous.
Visual loss in patients with proliferative retinopathy may
appearance of white-grayish cotton wool spots (soft exu- be caused by either vitreous hemorrhage (VH), tractional
dates), reflecting the accumulation of axoplasmic debris due retinal detachment following vitreous contraction, or neo-
to ischemia-induced interruption of axoplasmic transport, vascular glaucoma.44,46,47 Several growth factors have been
and by the presence of microvascular abnormalities such as shown to be involved in the process of neovascularization,
venous beading and loops.22 including vascular endothelial growth factor (VEGF),48–57
insulin-like growth factor 1 (IGF-1),58–65 growth hormone
Proliferative diabetic retinopathy (GH),66–72 pigment-epithelium-derived factor (PEDF),73–79
The hallmark of proliferative retinopathy is retinal and placental growth factor,80 basic fibroblast growth fac-
preretinal neovascularization that occurs in response to tor,81 hepatocyte growth factor,82–84 and, recently, also
the increased retinal ischemia (Figure 54.3).45 Presumably, erythropoietin.85–87
the nutrient-starved retina sends out a chemical message The Diabetic Retinopathy Study Research Group88 has
to stimulate the growth of new blood vessels (neovascu- identified four risk factors for severe visual loss in diabetic
larization). These vessels often grow on the surface of the retinopathy, as shown in Table 54.2. Untreated, affected
retina, at the optic nerve, or on the iris, and are character- patients have a 30% and 50% risk of visual acuity deteriora-
ized by increased fragility. They may penetrate the outlining tion to less than 5/200 within 3 and 5 years, respectively.
Diabetic retinopathy in nonpregnant patients 455
type 1 diabetes. Subsequent studies supported this finding90

and demonstrated similar beneficial effects in patients with
type 2 diabetes.91–93
Unexpectedly, it was found that some of the patients with
preexisting retinopathy had a transient worsening of their
retinopathy after the institution of intensive glycemic con-
trol.30,94–97 This deterioration manifested as an increase in soft
exudates and blot hemorrhages, but progression to prolifera-
tive retinopathy was uncommon. Although the reason for this
transient worsening in not clear, possible mechanism include
reduced plasma volume or increased levels of growth factors
such IGF-1 and VEGF.98–100 Among patients with mild to
moderate nonproliferative retinopathy, the long-term benefits
of intensive glycemic control exceed the risk of this transient
worsening. However, in patients with severe nonproliferative
or proliferative retinopathy, photocoagulation may be indi-
cated before intensive glycemic control is initiated.94
Multifactorial risk reduction

Figure 54.3 Proliferative retinopathy.
Control of hypertension, a risk factor for diabetic retinop-
athy, appears to delay the progression of retinopathy in
patients with type 2 diabetes.16,21,101 High serum lipid levels
Table 54.2 High-risk characteristics for severe
are associated with increased risk of hard exudates, macular
visual loss from proliferative retinopathy
edema, and proliferative retinopathy.102,103 Normalization of
lipid levels may reduce this risk.
1. New vessels on or within one disc diameter of the optic
disc and at least one-fourth of the disc area in extent,
with or without preretinal or vitreous hemorrhage Antiplatelet agents
2. Any new vessels on or within one disc diameter of the The increased platelet aggregation observed in diabetic
optic disk, with preretinal or vitreous hemorrhage patients and the evidence of microthromboses in retinal
3. New vessels anywhere in the retina more than one disc capillaries38 suggested that treatment with antiplatelet agents
diameter from the optic disk and at least one-half of the may be beneficial in diabetic retinopathy. However, in the
disc area in extent (including the total area of all new
vessels in the ocular fundus), with preretinal or vitreous Early Treatment of Diabetic Retinopathy Study (ETDRS),
hemorrhage treatment with daily aspirin had no effect on the develop-
4. Extensive preretinal or vitreous hemorrhage that hides ment or progression of diabetic retinopathy.104,105
probable new vessels meeting the aforementioned
criteria Laser photocoagulation
Source: The Diabetic Retinopathy Study Research Group. Arch Argon-laser photocoagulation, which was introduced in the
Ophthalmol 1979. 1950s, is the primary treatment for advanced retinopathy.
In the Diabetic Retinopathy Study, panretinal photocoagu-
lation in patients with proliferative retinopathy was associ-
Treatment of diabetic retinopathy ated with 50% reduction in severe vision loss.88,106,107 In the
ETDRS, the use of panretinal photocoagulation at earlier
Several preventative and therapeutic interventions have been
stages of retinopathy (mild to moderate nonproliferative
shown to be effective in reducing the morbidity associated
retinopathy) was not found to be beneficial and was asso-
with diabetic retinopathy. Nevertheless, because of the limita-
ciated with significant loss of visual acuity and peripheral
tions of these interventions, efforts are being made to improve
vision.108 However, focal photocoagulation in areas of vas-
the understanding of the mechanisms responsible for diabetic
cular abnormality was found to be effective in reducing the
retinopathy. This may lead to the development of new thera-
risk of moderate visual loss in cases of macular edema.108
pies, several of which are currently under clinical trials.
Although the mechanism responsible for the beneficial effect
of panretinal photocoagulation is unclear, it was suggested
Glycemic control that the destruction of hypoxic regions in the peripheral ret-
Preventing diabetic retinopathy from developing is the ina reduces the stimulus for neovascularization.29 Because
most effective approach to preserve vision. The results of the of the possible side effects associated with panretinal pho-
Diabetes Control and Complications Trial (DCCT)30,89 have tocoagulation (loss of peripheral vision and impaired night
provided clear evidence that intensive glycemic control is vision), this treatment is usually recommended only in cases
associated with a remarkable reduction in the development of high-risk proliferative retinopathy.108 However, patients
and progression of diabetic retinopathy in patients with with type 2 diabetes can benefit from laser photocoagulation
at earlier stages of the disease (severe nonproliferative or whether pregnancy influences the development or progres-
early proliferative retinopathy).108,109 sion of diabetic retinopathy is still unanswered.145 Although
most studies have reported that pregnancy is associated with
Vitrectomy progression of retinopathy in 17%–70% of the cases,146–158
Vitrectomy, which includes removal of the vitreous gel and several studies have failed to support these findings.159–163
may be accompanied by removal of fibrous tissue and repair
of retinal detachment, is associated with dramatic improve- Studies supporting a deteriorating effect of pregnancy on
ment of vision in cases of severe VH,110–112 as well as in cases of diabetic retinopathy
very severe neovascularization with no evidence of VH.113–117 Moloney et al.148 prospectively followed 53 pregnant women
Vitrectomy may also be achieved by means of pharmacologi- with type 1 diabetes every 6 weeks during pregnancy and
cal interventions using enzymes such as hyaluronidase, plas- until 6 months postpartum. Of the 20 women who had no
min, and chondroitinase that have been shown to be effective evidence of retinopathy, 8 (40%) developed retinopathy dur-
in clearing VH, thereby avoiding the need for surgery. Plasmin ing pregnancy. Progression of preexisting retinopathy dur-
has been used to cleave the vitreoretinal junction and allow ing pregnancy was common, with a moderate increase in the
less traumatic removal of the posterior hyaloids.118,119 These number of microaneurysms and the appearance of retinal
interventions are still under investigation. hemorrhages and soft exudates in 56% and 28% of the cases,
respectively. The risk for the development and progression of
Experimental therapies retinopathy was related to the duration of diabetes.
Better understanding of the mechanisms and factors involved Dibble et al.147 followed 55 pregnant diabetics who were
in the pathogenesis of diabetic retinopathy (as described managed by strict glycemic control. Progression of retinopa-
earlier) has led to exploration of new drugs directed against thy was related to the duration of diabetes and the severity of
these factors, some of which are currently under clinical tri- retinopathy before pregnancy. Thus, progression of retinopa-
als. Examples include aldose reductase inhibitors,33,120–122 thy occurred in 16% (3 of 19) of the women with minimal or
protein kinase C inhibitors,123 antioxidants, inhibition of mild nonproliferative retinopathy compared with 86% (6 of 7)
nonenzymatic glycation of proteins (aminoguanidine),124 of the women with untreated preexisting proliferative reti-
VEGF inhibitors,125,126 intraocular gene therapy with the nopathy. Similarly, in a prospective cohort study of 145 preg-
PEDF gene,127,128 and somatostatin analogues (reduce GH nant women with type 1 diabetes,152 Serup et al. have found
and IGF-1 levels).129 that none of the women with White classes B or C (no evi-
There is some evidence that intravitreal injection of dence of retinopathy) developed persistent retinopathy as a
glucocorticoids such as triamcinolone acetonide may be consequence of pregnancy, while deterioration was noticed in
effective in decreasing neovascularization. However, gluco- 50% of the women with preexisting retinopathy. Additional
corticoids may be associated with significant risks includ- support to the role of the severity of retinopathy as a risk fac-
ing increased intraocular pressure, cataract formation, and tor for progression during pregnancy comes from a retrospec-
endophthalmitis.130,131 tive study of 23 pregnant diabetic women.149 While none of
Inhibition of VEGF has been shown to prevent neovas- the women with either no evidence of retinopathy (n = 10) or
cularization in animal models.125,132,133 Several human- nonproliferative retinopathy (n = 8) experienced progression
ized antibodies directed against VEGF have demonstrated of retinopathy during pregnancy, four of the five women with
promising results in the treatment of proliferative retinopa- proliferative retinopathy required photocoagulation during
thy. Examples include bevacizumab and ranibizumab that pregnancy due to the progression of retinopathy.
are nonselective anti-VEGF antibodies that are currently In a prospective study of 234 intensively treated preg-
being used in the treatment of age-related macular degen- nant diabetics,146 Jervell et al. reported that the development
eration. Pegaptanib is an oligonucleotide ligand that spe- or progression retinopathy occurred in 68 women (29%).
cifically inhibits VEGF-165, the VEGF isoform that is most Nevertheless, progression was rarely associated with loss of
important in promoting neovascularization and vascular visual acuity.
permeability.134,135 There are now several clinical studies In order to quantify the effect of pregnancy on diabetic
demonstrating regression of neovascularization after intra- retinopathy, Soubrane et al. followed 22 pregnant diabet-
vitreal injection of VEGF inhibitors.136–144 However, there are ics using serial fluorescein angiography examinations.150
only limited data regarding the long-term efficacy and safety The mean number of microaneurysms increased from 42.7
of these agents, and more randomized trials are required. before pregnancy to 56.7 at the 28th week and to 79.7 at the
35th week. This number decreased to 62.7 and 60.3 at 6 and
15 months postpartum, respectively. Similarly, Hellstedt
Diabetic retinopathy in pregnancy et al.156 followed the number of microaneurysms in 21 dia-
betic women with mild retinopathy. Microaneurysms were
Impact of pregnancy on the development and assessed during pregnancy and at 3 and 6 months postpar-
progression of diabetic retinopathy tum using red-free photographs. The number of microaneu-
Understanding the effect of pregnancy on diabetic retinop- rysms increased progressively during pregnancy and peaked
athy is essential for determining the optimal management at 3 months postpartum. This increase was related to the
of pregnancies complicated by diabetic retinopathy. However, degree of improvement in glycemic control.
Diabetic retinopathy in pregnancy 457
Phelps et al.151 prospectively studied 35 pregnant women While most of the data come from patients with type 1
with type 1 diabetes under tight glycemic control. The pro- diabetes, there is more recent evidence that pregnancy
gression of nonproliferative retinopathy, noticed in 55% of the can affect the progression of diabetic retinopathy also in
women, correlated with the levels of plasma glucose at entry women with type 2 diabetes. In a recent study, Rasmussen
and with the magnitude of improvement in glycemic control. et al.157 studied the progression of diabetic retinopathy
Klein et al.153 assessed the effect of pregnancy on diabetic during pregnancy in 80 women with type 2 diabetes, 11
retinopathy using a control group of nonpregnant women of whom had evidence of retinopathy in early pregnancy.
with type 1 diabetes. In addition to poor glycemic control, The progression of retinopathy was observed in 11 (14%)
pregnancy per se was a significant and independent risk fac- women, and although it was mild in most cases, one case
tor for progression of retinopathy. was complicated by progression from mild retinopathy to
Rosenn et al.154 reported a progression rate of 33% in a sight-threatening retinopathy and impaired vision in both
prospective study of 154 pregnant women with type 1 dia- eyes. Risk factors for the progression of retinopathy were
betes. They have found that hypertension, either chronic a longer duration of diabetes and insulin treatment before
or pregnancy induced (PIH), was a significant and inde- pregnancy.
pendent risk factor for the progression of retinopathy. The
level of glycemic control and improved glycemic control in Studies not supporting a significant effect of pregnancy
early pregnancy was also associated with the progression of on progression of retinopathy
retinopathy. There are several studies that failed to demonstrate a signifi-
In the Diabetes in Early Pregnancy Study,155 the risk of cant effect of pregnancy on the natural history of diabetic
progression was strongly correlated with the severity of reti- retinopathy. Other studies have found that although reti-
nopathy before pregnancy. Thus, progression of more than nopathy progressed during pregnancy, complete or partial
two steps was noted in 10.3%, 21.1%, 18.8%, and 54.8% of regression was observed after delivery.
the women with no, minimal (microaneurysms only), mild, Stephens et al.159 found the rate of development of reti-
and moderate-to-severe retinopathy at baseline, respectively. nopathy during pregnancy to be as low as 2%. Horvat et al.,160
Progression to proliferative retinopathy was observed in 6.3% in a prospective study extending over 12 years, followed 172
and 29% of the women with mild and moderate-to-severe diabetic women during pregnancy. Of the 40 women with
retinopathy, respectively. Other factors associated with the nonproliferative retinopathy, only 4 (10%) progressed to pro-
progression of retinopathy included elevated baseline glyco- liferative retinopathy. They have found retinopathy to fluc-
sylated hemoglobin and the magnitude of improvement of tuate during pregnancies, with simultaneous progression
glycemic control in early pregnancy. and regression of retinopathy in different parts of the same
In another prospective study, 8 65 diabetic women were eye. Their conclusion was that pregnancy is not associated
followed during pregnancy and 12 months postpartum. with an increased risk for progression of retinopathy and
While the overall progression rate was 77.5%, only 26% of visual loss.
the women with no retinopathy at baseline had evidence of In another retrospective study,161 the progression of
retinopathy during pregnancy. Duration of diabetes, poor retinopathy in 65 pregnant women with type 1 diabetes
glycemic control, anemia, and elevated systolic blood pres- was compared to a matched control group of 56 nonpreg-
sure were also risk factors for progression of retinopathy. nant women with type 1 diabetes. The rate of sight-threat-
Lovestam et al.161 retrospectively compared a group of ening deterioration of retinopathy was similar for the two
65 pregnant women with type 1 diabetes with a matched groups (13% and 11%, respectively). Similarly, Temple
group of nonpregnant women. et al. prospectively studied 179 pregnancies of women
They have found preeclampsia to be a potent independent with type 1 diabetes.163 The overall rate of progression of
risk factor for progression of retinopathy. Moreover, Gordin retinopathy during pregnancy was only 5%, and the rate
et al.164 have recently investigated the long-term impact of of progression to proliferative retinopathy necessitating
preeclampsia and PIH complicating pregnancies of 158 laser therapy was 2.2%.
women with type 1 diabetes. Based on a follow-up of 16 years In the DCCT,162 although pregnancy was associated with
on average, preeclampsia and PIH were found to be indepen- a greater risk for worsening of retinopathy, this worsening
dently associated with an increased risk of developing severe was transient, and the long-term risk for progression of reti-
diabetic retinopathy later in life, with a hazard ratio of 3.5 nopathy did not appear to be increased by pregnancy.
and 3.2, respectively. Some of the studies cited earlier have reported partial
More recently, Vestgaard et al.158 reported on the rate of or complete regression of the pregnancy-related wors-
progression of diabetic retinopathy in a prospective study of ening of retinopathy. Moloney et al.148 reported that by
102 pregnant women with type 1 diabetes who were man- 6 months postpartum, the nonproliferative changes had
aged by tight glycemic. Overall, progression of retinopathy regressed to control levels, and neovascularization showed
occurred in 28 (27%) women. Sight-threatening progression some regression. Similarly, Serup et al.152 reported that
occurred in six women and was associated with the presence postpartum regression was common and that prolifera-
of macular edema, impaired visual acuity, and higher blood tive changes that developed during pregnancy disappeared
pressure in early pregnancy, but not with HbA1c, decline in spontaneously after delivery in most of the cases. For that
HbA1c, or prevalence of severe hypoglycemia. reason, the authors recommended that treatment with
photocoagulation during pregnancy and in the early post-

Table 54.3 Risk factors for progression of
partum period should be restricted. In the study of Rosenn
retinopathy during pregnancy
et al.,154 postpartum regression was observed in 13 of the 51
women (25%) in whom progression of retinopathy occurred
Risk factor References
during pregnancy. In a more recent study, Arun et al.165 fol-
lowed 59 women with type 1 diabetes for 5–10 years post- Nonmodifiable
pregnancy. Although four women required laser therapy Duration of diabetes [8,147,148]
during pregnancy, none of the women required laser ther- Severity of retinopathy before [147,152,155]
apy over the 5 years postpregnancy, although retinopathy conception
was worsened in 14 women. On 10-year follow-up, only Poor glycemic control before conception [8,151,154,155]
1 of 22 women required laser therapy. The authors’ conclu- Modifiable
sion was that pregnancy is not associated with significant
Chronic or pregnancy-induced [8,154,161]
postpartum worsening of retinopathy. hypertension
Rapid normalization of blood glucose [151,154–156]
Suggested mechanism values
Several mechanisms have been suggested by which preg- Anemia [8]
nancy may lead to progression of retinopathy.
Pregnancy is associated with a dramatic change in the
hormonal milieu. Human placental lactogen (hPL) is pro- risk of progression of retinopathy (odds ratio 2.0, 95% con-
duced in enormous amounts by the placenta, reaching a fidence interval 1.1–3.6).
production rate of about 1 g/day at term. Due to its GH-like Finally, it has been recently hypothesized that pregnancy-
activity, hPL may play an important role in the effect of preg- induced changes in the immune system may contribute to
nancy on diabetic retinopathy, as mentioned earlier.66,67,69 the progression of retinopathy during pregnancy.169 These
The vascular changes induced by the elevated levels of estro- changes, which have been previously implicated in the
gen and progesterone during pregnancy may also contribute pathogenesis of diabetic retinopathy, include generalized
to the progression of retinopathy. leukocyte activation and upregulation of adhesion molecules
Pregnancy is also associated with marked physiologi- that may lead to leukocyte–endothelial interaction and the
cal changes. The pregnancy-induced increase in cardiac consequent leukostasis with capillary occlusion, ischemia,
output may lead to increased RBF, which could damage and vascular leakage. In addition, pregnancy is associated
diabetic vessels. Chen et al. studied the changes in RBF with an increase in the levels of certain cytokines that are
during pregnancy in diabetic and nondiabetic women known to cause blood–retina barrier breakdown and to pro-
using laser Doppler velocimetry.166 Among the diabetic mote angiogenic and proliferation and thereby contribute to
women, only those whose retinopathy progressed had a progression of retinopathy.
significant increase in RBF during pregnancy. However, as
these patients also had worse glycemic control (higher gly- Summary
cosylated hemoglobin levels prior to and throughout preg- In summary, it appears that pregnancy is associated with the
nancy), it is not clear if the deterioration can be attributed progression of diabetic retinopathy and that the risk for pro-
to the increase in RBF alone. gression is related to the factors presented in Table 54.3. One
As mentioned earlier, hypertension is a risk factor for the major drawback of many of these studies presented earlier is
progression of diabetic retinopathy.11,14,16,20,21,154 Hypertensive the lack of a matched control group of nonpregnant diabetic
disorders are common in diabetic pregnancies, complicating women. The progression of retinopathy during pregnancy
up to 40% of pregnancies in women with type 1 diabetes.167 may be the result of the normal hormonal and physiological
Thus, hypertension may contribute to the progression of ret- changes during pregnancy as well as the improved glycemic
inopathy during pregnancy. control and high rate of hypertensive disorders observed
Another possible mechanism is the improved glycemic in diabetic pregnancies. Nevertheless, these changes may
control achieved by most of the diabetic women before or at regress postpartum, especially in cases of nonproliferative
the onset of pregnancy. As discussed earlier, rapid normal- retinopathy, and may not affect the long-term progression of
ization of blood glucose is associated with the progression of retinopathy.162
retinopathy in both nonpregnant94–97 and pregnant151,154–156
patients.
In a recent study, Ringholm et al.168 measured blood Diabetic retinopathy and perinatal outcome
IGF-I levels in 88 women with type 1 diabetes at differ- Several investigators have addressed the issue of a pos-
ent time points along pregnancy. Overall, 22 (25%) women sible relationship between diabetic retinopathy and perina-
had evidence of new onset or progression of retinopathy. tal outcome. Moloney et al.148 noted that maternal retinal
Although IGF-I levels increased throughout pregnancy hemorrhages or neovascularization was associated with
in all women, the increase was significantly more rapid in increased infant morbidity. McElvy et al.170 performed a
women in whom retinopathy progressed. Elevated IGF-I prospective study of 205 pregnant women with type 1 dia-
level was independently associated with an increase in the betes. The development or progression of retinopathy during
Diabetic retinopathy in pregnancy 459
pregnancy was significantly associated with reduced fetal photocoagulation is beneficial mainly in cases of high-risk
growth, manifested by lower mean birth weight and a higher proliferative retinopathy and clinically significant macular
rate of small-for-gestational-age (SGA) and low birth weight edema. However, it is recommended that in cases of impend-
infants. There was no correlation between progression of ing pregnancy, during which rapid progression of retinopa-
retinopathy and gestational age at delivery, preterm delivery, thy may occur, photocoagulation should be instituted at
respiratory distress syndrome, neonatal hypoglycemia, or earlier stages, as in cases of severe nonproliferative or early
neonatal death. proliferative retinopathy.108,109 In one study of women with
Klein et al.171 reported adverse perinatal outcome in 33 proliferative retinopathy, treatment with laser photocoagu-
of 179 diabetic pregnancies. Severity of retinopathy was the lation before conception was associated with significant pro-
only variable that significantly predicted adverse perinatal gression of retinopathy in 26% of the women compared with
outcome. Lauszus et al.172 reviewed the records of 26 dia- 58% when treatment was initiated in early pregnancy.173
betic pregnancies with preexisting proliferative retinopathy.
Proliferative retinopathy was associated with increased pre- Antenatal follow-up
term delivery rate (27%) and serious neonatal morbidity. The concerns regarding rapid optimization of glycemic con-
trol with respect to progression of retinopathy in diabetic
women with poor glycemic control who did not have precon-
Management of diabetic retinopathy during pregnancy ception counseling have been described earlier. Nevertheless,
As discussed earlier, diabetic retinopathy can worsen during according to the 2008 National Institute for Health and Care
pregnancy. Thus, patients with diabetic retinopathy should Excellence (NICE) guidelines, diabetic retinopathy should
undergo preconception counseling and follow-up during not be considered a contraindication to rapid optimiza-
pregnancy by a multidisciplinary team that includes a peri- tion of glycemic control in women who present with a high
natologist, endocrinologist, and ophthalmologist that are HbA1c in early pregnancy.174
experienced in the management of diabetic retinopathy. The frequency of eye examinations during pregnancy
depends on the severity of retinopathy prior to conception,
Preconception counseling evidence of progression of retinopathy, glycemic control,
Patients with diabetes who are planning to become preg- duration of diabetes, and the presence of other risk fac-
nant should be given a thorough explanation on the risk tors such as hypertension. Thus, patients with no or mini-
of the development or progression of diabetic retinopathy mal retinopathy should be evaluated in the first and third
during pregnancy and the importance of glycemic control trimesters.174 Patients with mild to moderate retinopathy
throughout pregnancy. Patients with high-risk characteris- should undergo evaluation each trimester. In cases of severe
tics (long-standing diabetes, severe retinopathy, coexisting nonproliferative or proliferative retinopathy, when there is
hypertension, poor glycemic control) should be identified evidence of progression of retinopathy, and in patients with
and followed appropriately. poor glycemic control or hypertension, follow-up may be
A comprehensive eye examination prior to conception needed every 2–4 weeks.153,155,162,175,176
should be performed by an ophthalmologist experienced in Fluorescein angiography is a sensitive tool to assess the
the care of diabetic retinopathy. This examination carries great extent of capillary nonperfusion and early neovasculariza-
importance since it determines the risk for progression of reti- tion and may aid in guiding treatment of macular edema,
nopathy, the frequency of follow-up visits during pregnancy, although ophthalmoscopic examination is satisfying in most
and the need for laser photocoagulation prior to conception. of the cases for the diagnosis of proliferative retinopathy. In
Glycemic control should be achieved prior to conception in addition, although detrimental effects of fluorescein dye
order to reduce the risk of progression of retinopathy as well on the fetus have not been documented,177 fluorescein does
as to avoid the adverse maternal and fetal outcome associ- cross the placenta into the fetal circulation. Thus, fluorescein
ated with poorly controlled diabetes during pregnancy. One angiography during pregnancy is generally not indicated.
goal is to achieve a glycosylated hemoglobin level of less than
six standard deviations above normal prior to conception.155 Laser photocoagulation during pregnancy
In the presence of proliferative or severe nonproliferative The use of laser photocoagulation during pregnancy is
retinopathy, normalization of blood glucose levels should be controversial. In general, the indications for treatment
achieved gradually over a period of weeks to months in order and the response to laser photocoagulation are the same
to avoid progression of retinopathy that may result from rapid as for nonpregnant women.178,179 As described earlier, sev-
normalization. However, in patients presenting after concep- eral studies reported that postpartum regression of the
tion, blood glucose should be normalized as soon as possible pregnancy-induced progression of retinopathy is com-
since the benefits of good glycemic control in early pregnancy mon,148,152 and the authors recommended that treatment
far outweigh the risk of transient progression of retinopathy. with photocoagulation during pregnancy and in the early
Other risk factors, such as uncontrolled chronic hypertension, postpartum period should be restricted. In contrast, others
should be followed and treated adequately prior to conception. believe that because the progression of retinopathy during
One question that needs to be answered is the role of laser pregnancy can be at times rapid and aggressive, treat-
photocoagulation prior to conception. As discussed earlier, ment with laser photocoagulation should be applied dur-
the ETDRS have found that in type 1 diabetes, panretinal ing pregnancy and after delivery when indicated.147,180,181
Thus, the decision whether to treat should be made on an Insulin lispro has a 1.5-fold higher affinity toward the IGF-1
individual basis, taking into account the severity of reti- receptor compared with human insulin, and the initial sugges-
nopathy, evidence of progression, lack of glycemic con- tion that insulin lispro may worsen retinopathy was made by
trol, the presence of additional risk factors, and, on the Kitzmiller et al.184 In this study, of the ten patients who were
other hand, the risks and side effects associated with laser treated with insulin lispro and had no evidence of retinopa-
photocoagulation. thy prior to conception, three developed bilateral proliferative
According to the ETDRS,182 in patients with DME, focal changes, and two developed VH during pregnancy. However,
photocoagulation should be performed during pregnancy, these patients had poor glycemic control at baseline and expe-
because spontaneous regression rarely occurs after delivery. rienced significant improvement in glycemic control during
If the edema does not respond well to photocoagulation, hos- pregnancy, two major risk factors that may be responsible
pitalization is required, with diuretic treatment and occa- for the progression of retinopathy in these cases. Subsequent
sionally steroids.183 studies did not support an effect of insulin lispro on retinop-
athy.185–188 Insulin aspart is similar to insulin lispro in many
Insulin analogues during pregnancy and diabetic aspects, although its affinity to the IGF-1 receptor is the same
retinopathy as human insulin.189 There are only little data on insulin aspart
In addition to its metabolic effects that are mediated by the in pregnancy, and currently it remains a category C medica-
insulin receptor, insulin also has a mitogenic effect, currently tion for pregnancy.185 The results of multicentric study now in
thought to be mediated by the IGF-1 receptor. The possible role progress on the efficacy and safety of insulin aspart in pregnant
of IGF-1 in the pathogeneses of diabetic retinopathy (discussed patients with type 1 diabetes will provide important informa-
earlier) and the change in the relative affinity of the insulin tion on the safety of insulin aspart during pregnancy.190 Insulin
analogues toward the IGF-1 receptor have raised concerns glargine has a higher affinity to IGF-1 receptor and a greater
regarding the effect of these analogues on diabetic retinopathy. mitogenic activity compared with human insulin (6.5-fold and
Preconception counseling
• Patient education—Explanation on the risk of development or progression of diabetic retinopathy during

pregnancy, and the importance of glycemic control during pregnancy.
• Risk stratification—Identification of patients with high-risk characteristics
• Comprehensive eye examination
• Improve of glycemic control—Should be achieved gradually, especially in patients with advanced
retinopathy.
• Control of co-existing hypertension
• Laser photocoagulation should be considered in patients with severe non-proliferative or proliferative
retinopathy
Antenatal care
• Multidisciplinary team—Perinatologist, endocrinologist and ophthalmologist that are experienced in the

management of diabetic retinopathy.
• Frequency of eye examinations:
No—Minimal retinopathy 1st, 3rd trimesters
Mild—Moderate retinopathy 1st, 2nd, 3rd trimesters
Severe NPDR, PDR
Evidence of progression
Every 2–4 weeks
Poor glycemic control
Other risk factors
• Fluorescein angiography is usually avoided during pregnancy
• Laser photocoagulation—Because postpartum regression is common, laser photocoagulation should be
considered in cases severe non-proliferative or proliferative retinopathy on an individual basis considering
the severity of retinopathy, evidence of progression, lack of glycemic control, and the presence of additional
risk factors.
• Macular edema should be treated with focal laser photocoagulation, diuretics, or steroids.
Delivery
• There is no need for labor induction in women with good glycemic control.
• In cases of high-risk proliferative retinopathy, elective cesarean section or shortening of the second stage
should be considered due to the risk of vitreous hemorrhage.
Figure 54.4 Recommendations for the management of pregnancies complicated by diabetic retinopathy.
References 461
8-fold, respectively),189 which raised concerns on its effects in Summary

diabetic retinopathy. Two randomized trials demonstrated an
increased risk of progression of retinopathy191 and a higher inci- Diabetic retinopathy is the most common chronic complica-
dence of new onset macular edema.192 However, there was lack tion of diabetes mellitus, and it is estimated that 20%–30%
of consistency between the methods of assessment, and subse- of diabetic women in the reproductive age group has some
quent prospective randomized studies did not support these evidence of retinopathy.
observations. A large 5-year randomized multicenter study The purpose of this chapter was to review the current lit-
is in progress.185 For the moment, the use of insulin glargine erature on the effects of pregnancy on diabetic retinopathy
during pregnancy is not recommended. The affinity of insulin and to provide guidelines for the management of pregnan-
detemir to IGF-I receptor is fivefold lower than human insulin, cies complicated by diabetic retinopathy.
and its mitogenic activity is reduced by tenfold.189 There are no It appears that pregnancy is associated with the progres-
data on the use of insulin detemir in pregnancy. sion of diabetic retinopathy, especially in women with the
risk factors presented in Table 54.3, although these changes
Delivery may regress postpartum, especially in cases of nonpro-
Earlier studies recommended early delivery, as soon as lung liferative retinopathy, and may not affect the long-term
maturation was achieved, in pregnancies complicated by ret- progression of retinopathy. In addition, the presence of
inopathy and other vascular diseases. However, more recent diabetic retinopathy is associated with adverse pregnancy
research has shown that a well-controlled diabetic preg- outcome, including fetal growth restriction, and preterm
nancy can be allowed to continue to term in order to avoid delivery.
iatrogenic prematurity, the risk of amniocentesis, failed Patients with diabetic retinopathy should undergo pre-
induction of labor due to unfavorable cervix, and unneces- conception counseling and follow-up during pregnancy by
sary cesarean section.193 a multidisciplinary team as summarized in Figure 54.4. It is
The Valsalva maneuver during labor might induce VH now clear that White’s advice at 1971 to terminate pregnan-
from active neovascularization.194,195 The role of elective cies in diabetic patients with progressive proliferative reti-
cesarean section in these cases is controversial. The mode of nopathy is no longer valid196 and that with careful profession
delivery should be discussed in advance between the patient care, a favorable pregnancy outcome can usually be expected
and obstetrician. According to the 2008 NICE guidelines, with minimal or no deterioration in ophthalmologic status.
diabetic retinopathy should not be considered a contraindi- Ongoing trials may provide better answers regarding the use
cation to vaginal birth.174 A summary of recommendations of laser photocoagulation and the safety of insulin analogues
for management is presented in Figure 54.4. during pregnancy.
REFERENCES
1. Reece EA, Homko CJ, Hagay Z. Diabetic retinopathy in preg- 12. Aiello LP, Gardner TW, King GL et al. Diabetic retinopathy.
nancy. Obstet Gynecol Clin North Am 1996; 23: 161–171. Diabetes Care 1998; 21: 143–156.
2. Kahn HA, Hiller R. Blindness caused by diabetic retinopathy. Am 13. Fong DS, Aiello L, Gardner TW et al. Diabetic retinopathy.
J Ophthalmol 1974; 78: 58–67. Diabetes Care 2003; 26 (Suppl. 1): S99–S102.
3. Foulds WS, McCuish A, Barrie T et al. Diabetic retinopathy in 14. Klein R, Klein BE, Moss SE, Cruickshanks KJ. The Wisconsin
the West of Scotland: Its detection and prevalence, and the cost- Epidemiologic Study of Diabetic Retinopathy: XVII. The 14-year
effectiveness of a proposed screening programme. Health Bull incidence and progression of diabetic retinopathy and associ-
(Edinb) 1983; 41: 318–326. ated risk factors in type 1 diabetes. Ophthalmology 1998; 105:
4. Kahn HA, Bradley RF. Prevalence of diabetic retinopathy. 1801–1815.
Age, sex, and duration of diabetes. Br J Ophthalmol 1975; 59: 15. Nguyen HT, Luzio SD, Dolben J et al. Dominant risk factors for
345–349. retinopathy at clinical diagnosis in patients with type II diabetes
5. Klein R, Klein B. Vision disorders in diabetes. In: Harris MI et al. mellitus. J Diabetes Complications 1996; 10: 211–219.
eds., Diabetes in America, DHHS publication number 85–1468, 16. Okudaira M, Yokoyama H, Otani T, Uchigata Y, Iwamoto Y.
United States Government Printing Office: Washington, DC, Slightly elevated blood pressure as well as poor metabolic con-
1985. trol are risk factors for the progression of retinopathy in early-
6. Cassar J, Kohner EM, Hamilton AM, Gordon H, Joplin GF. Diabetic onset Japanese type 2 diabetes. J Diabetes Complications 2000;
retinopathy and pregnancy. Diabetologia 1978; 15: 105–111. 14: 281–287.
7. Elman KD, Welch RA, Frank RN, Goyert GL, Sokol RJ. Diabetic 17. Klein R, Klein BE, Moss SE. The Wisconsin epidemiological study
retinopathy in pregnancy: A review. Obstet Gynecol 1990; 75: of diabetic retinopathy: A review. Diabetes Metab Rev 1989; 5:
119–127. 559–570.
8. Axer-Siegel R, Hod M, Fink-Cohen S et al. Diabetic retinopathy 18. Klein R, Moss SE, Klein BE, Davis MD, DeMets DL. The
during pregnancy. Ophthalmology 1996; 103: 1815–1819. Wisconsin epidemiologic study of diabetic retinopathy. XI.
9. Hare JW, White P. Pregnancy in diabetes complicated by vascular The incidence of macular edema. Ophthalmology 1989; 96:
disease. Diabetes 1977; 26: 953–955. 1501–1510.
10. Carstensen LL, Frost-Larsen K, Fugleberg S, Nerup J. Does 19. Henricsson M, Berntorp K, Berntorp E, Fernlund P, Sundkvist
pregnancy influence the prognosis of uncomplicated insulin- G. Progression of retinopathy after improved metabolic control
dependent diabetes mellitus? Diabetes Care 1982; 5: 1–5. in type 2 diabetic patients. Relation to IGF-1 and hemostatic
11. Klein R, Klein BE, Moss SE, Davis MD, DeMets DL. The Wisconsin variables. Diabetes Care 1999; 22: 1944–1949.
epidemiologic study of diabetic retinopathy. II. Prevalence and 20. Rand LI, Krolewski AS, Aiello LM, Warram JH, Baker RS, Maki
risk of diabetic retinopathy when age at diagnosis is less than T. Multiple factors in the prediction of risk of proliferative dia-
30 years. Arch Ophthalmol 1984; 102: 520–526. betic retinopathy. N Engl J Med 1985; 313: 1433–1438.
21. Knowler WC, Bennett PH, Ballintine EJ. Increased incidence 45. Garner A. Histopathology of diabetic retinopathy in man. Eye
of retinopathy in diabetics with elevated blood pressure. A six- 1993; 7 (Pt 2): 250–253.
year follow-up study in Pima Indians. N Engl J Med 1980; 302: 46. Davis MD. Vitreous contraction in proliferative diabetic retinopa-
645–650. thy. Arch Ophthalmol 1965; 74: 741–751.
22. Grading diabetic retinopathy from stereoscopic color fundus 47. Tolentino FI, Lee PF, Schepens CL. Biomicroscopic study of vitre-
photographs—An extension of the modified Airlie House clas- ous cavity in diabetic retinopathy. Arch Ophthalmol 1966; 75:
sification. ETDRS report number 10. Early Treatment Diabetic 238–246.
Retinopathy Study Research Group. Ophthalmology 1991; 98: 48. Sydorova M, Lee MS. Vascular endothelial growth factor levels in
786–806. vitreous and serum of patients with either proliferative diabetic
23. Wilkinson CP, Ferris FL 3rd, Klein RE et al. Proposed international retinopathy or proliferative vitreoretinopathy. Ophthalmic Res
clinical diabetic retinopathy and diabetic macular edema disease 2005; 37: 188–190.
severity scales. Ophthalmology 2003; 110: 1677–1682. 49. Caldwell RB, Bartoli M, Behzadian MA et al. Vascular endothelial
24. Bek T, Helgesen A. The regional distribution of diabetic retinopa- growth factor and diabetic retinopathy: Role of oxidative stress.
thy lesions may reflect risk factors for progression of the disease. Curr Drug Targets 2005; 6: 511–524.
Acta Ophthalmol Scand 2001; 79: 501–505. 50. Rakoczy PE, Brankov M, Fonceca A, Zaknich T, Rae BC, Lai CM.
25. Engerman RL. Pathogenesis of diabetic retinopathy. Diabetes Enhanced recombinant adeno-associated virus-mediated vascu-
1989; 38: 1203–1206. lar endothelial growth factor expression in the adult mouse ret-
26. Frank RN. On the pathogenesis of diabetic retinopathy. A 1990 ina: A potential model for diabetic retinopathy. Diabetes 2003;
update. Ophthalmology 1991; 98: 586–593. 52: 857–863.
27. Palmberg P, Smith M, Waltman S et al. The natural history of 51. Hogeboom van Buggenum IM, Polak BC, Reichert-Thoen JW,
retinopathy in insulin-dependent juvenile-onset diabetes. de Vries-Knoppert WA, van Hinsbergh VW, Tangelder GJ.
Ophthalmology 1981; 88: 613–618. Angiotensin converting enzyme inhibiting therapy is associated
28. Mizutani M, Kern TS, Lorenzi M. Accelerated death of retinal with lower vitreous vascular endothelial growth factor con-
microvascular cells in human and experimental diabetic retinop- centrations in patients with proliferative diabetic retinopathy.
athy. J Clin Invest 1996; 97: 2883–2890. Diabetologia 2002; 45: 203–209.
29. Frank RN. Diabetic retinopathy. N Engl J Med 2004; 350: 48–58. 52. Simo R, Lecube A, Segura RM, Garcia Arumi J, Hernandez C.
30. The effect of intensive treatment of diabetes on the development Free insulin growth factor-I and vascular endothelial growth fac-
and progression of long-term complications in insulin-dependent tor in the vitreous fluid of patients with proliferative diabetic reti-
diabetes mellitus. The Diabetes Control and Complications Trial nopathy. Am J Ophthalmol 2002; 134: 376–382.
Research Group. N Engl J Med 1993; 329: 977–986. 53. Hernandez C, Burgos R, Canton A, Garcia-Arumi J, Segura RM,
31. Sheetz MJ, King GL. Molecular understanding of hyperglyce- Simo R. Vitreous levels of vascular cell adhesion molecule and
mia’s adverse effects for diabetic complications. J Am Med Assoc vascular endothelial growth factor in patients with proliferative
2002; 288: 2579–2588. diabetic retinopathy: A case-control study. Diabetes Care 2001;
32. Kohner EM, Patel V, Rassam SM. Role of blood flow and impaired 24: 516–521.
autoregulation in the pathogenesis of diabetic retinopathy. 54. Funatsu H, Yamashita H, Shimizu E, Kojima R, Hori S. Relationship
Diabetes 1995; 44: 603–607. between vascular endothelial growth factor and interleukin-6 in
33. A randomized trial of sorbinil, an aldose reductase inhibitor, in diabetic retinopathy. Retina 2001; 21: 469–477.
diabetic retinopathy. Sorbinil Retinopathy Trial Research Group. 55. Hammes HP, Lin J, Bretzel RG, Brownlee M, Breier G.
Arch Ophthalmol 1990; 108: 1234–1244. Upregulation of the vascular endothelial growth factor/vascu-
34. Brownlee M, Cerami A. The biochemistry of the complications of lar endothelial growth factor receptor system in experimental
diabetes mellitus. Annu Rev Biochem 1981; 50: 385–432. background diabetic retinopathy of the rat. Diabetes 1998; 47:
35. Kunisaki M, Bursell SE, Clermont AC et al. Vitamin E prevents 401–406.
diabetes-induced abnormal retinal blood flow via the diac- 56. Burgos R, Simo R, Audi L et al. Vitreous levels of vascular
ylglycerol-protein kinase C pathway. Am J Physiol 1995; 269: endothelial growth factor are not influenced by its serum con-
E239–E246. centrations in diabetic retinopathy. Diabetologia 1997; 40:
36. Bursell SE, Clermont AC, Aiello LP et al. High-dose vitamin E 1107–1109.
supplementation normalizes retinal blood flow and creatinine 57. Clermont AC, Aiello LP, Mori F, Aiello LM, Bursell SE. Vascular
clearance in patients with type 1 diabetes. Diabetes Care 1999; endothelial growth factor and severity of nonproliferative diabetic
22: 1245–1251. retinopathy mediate retinal hemodynamics in vivo: A potential
37. Kowluru RA, Tang J, Kern TS. Abnormalities of retinal metabolism role for vascular endothelial growth factor in the progression of
in diabetes and experimental galactosemia. VII. Effect of long- nonproliferative diabetic retinopathy. Am J Ophthalmol 1997;
term administration of antioxidants on the development of reti- 124: 433–446.
nopathy. Diabetes 2001; 50: 1938–1942. 58. Loukovaara S, Immonen IJ, Koistinen R et al. The insulin-like
38. Boeri D, Maiello M, Lorenzi M. Increased prevalence of micro- growth factor system and type 1 diabetic retinopathy during
thromboses in retinal capillaries of diabetic individuals. Diabetes pregnancy. J Diabetes Complications 2005; 19: 297–304.
2001; 50: 1432–1439. 59. Poulaki V, Joussen AM, Mitsiades N, Mitsiades CS, Iliaki EF,
39. Xia P, Inoguchi T, Kern TS, Engerman RL, Oates PJ, King GL. Adamis AP. Insulin-like growth factor-I plays a pathogenetic role
Characterization of the mechanism for the chronic activation of in diabetic retinopathy. Am J Pathol 2004; 165: 457–469.
diacylglycerol-protein kinase C pathway in diabetes and hyper- 60. Simo R, Hernandez C, Segura RM et al. Free insulin-like growth
galactosemia. Diabetes 1994; 43: 1122–1129. factor 1 in the vitreous fluid of diabetic patients with proliferative
40. Derubertis FR, Craven PA. Activation of protein kinase C in glo- diabetic retinopathy: A case-control study. Clin Sci (Lond) 2003;
merular cells in diabetes. Mechanisms and potential links to the 104: 223–230.
pathogenesis of diabetic glomerulopathy. Diabetes 1994; 43: 61. Inokuchi N, Ikeda T, Imamura Y et al. Vitreous levels of insulin-
1–8. like growth factor-I in patients with proliferative diabetic retinop-
41. Ishii H, Jirousek MR, Koya D et al. Amelioration of vascular dys- athy. Curr Eye Res 2001; 23: 368–371.
functions in diabetic rats by an oral PKC beta inhibitor. Science 62. Janssen JA, Jacobs ML, Derkx FH, Weber RF, van der Lely AJ,
1996; 272: 728–731. Lamberts SW. Free and total insulin-like growth factor I (IGF-I),
42. Klein R. The epidemiology of diabetic retinopathy: Findings from IGF-binding protein-1 (IGFBP-1), and IGFBP-3 and their rela-
the Wisconsin Epidemiologic Study of Diabetic Retinopathy. Int tionships to the presence of diabetic retinopathy and glomeru-
Ophthalmol Clin 1987; 27: 230–238. lar hyperfiltration in insulin-dependent diabetes mellitus. J Clin
43. Klein R, Klein BE, Moss SE, Cruickshanks KJ. The Wisconsin Endocrinol Metab 1997; 82: 2809–2815.
Epidemiologic Study of Diabetic Retinopathy. XV. The long-term 63. Wang Q, Dills DG, Klein R, Klein BE, Moss SE. Does insulin-like
incidence of macular edema. Ophthalmology 1995; 102: 7–16. growth factor I predict incidence and progression of diabetic reti-
44. Fong DS, Ferris FL 3rd, Davis MD, Chew EY. Causes of severe nopathy? Diabetes 1995; 44: 161–164.
visual loss in the early treatment diabetic retinopathy study: 64. Arner P, Sjoberg S, Gjotterberg M, Skottner A. Circulating insulin-
ETDRS report no. 24. Early Treatment Diabetic Retinopathy Study like growth factor I in type 1 (insulin-dependent) diabetic patients
Research Group. Am J Ophthalmol 1999; 127: 137–141. with retinopathy. Diabetologia 1989; 32: 753–758.
References 463
65. Chantelau E. Evidence that upregulation of serum IGF-1 con- 86. Katsura Y, Okano T, Matsuno K et al. Erythropoietin is highly
centration can trigger acceleration of diabetic retinopathy. Br J elevated in vitreous fluid of patients with proliferative diabetic
Ophthalmol 1998; 82: 725–730. retinopathy. Diabetes Care 2005; 28: 2252–2254.
66. Powell ED, Frantz AG, Rabkin MT, Field RA. Growth hormone in 87. Watanabe D, Suzuma K, Matsui S et al. Erythropoietin as a retinal
relation to diabetic retinopathy. N Engl J Med 1966; 275: 922–925. angiogenic factor in proliferative diabetic retinopathy. N Engl J
67. Goebel FD, Ehrenreich H. Growth hormone and diabetic reti- Med 2005; 353: 782–792.
nopathy. Lancet 1985; 2: 666. 88. Four risk factors for severe visual loss in diabetic retinopathy. The
68. Hattori N, Moridera K, Ishihara T, Hino M, Ikekubo K, Kurahachi third report from the Diabetic Retinopathy Study. The Diabetic
H. Is growth hormone associated with diabetic retinopathy? Retinopathy Study Research Group. Arch Ophthalmol 1979; 97:
J Diabetes Complications 1993; 7: 12–14. 654–655.
69. Alzaid AA, Dinneen SF, Melton LJ 3rd, Rizza RA. The role of 89. Retinopathy and nephropathy in patients with type 1 diabetes
growth hormone in the development of diabetic retinopathy. four years after a trial of intensive therapy. The Diabetes Control
Diabetes Care 1994; 17: 531–534. and Complications Trial/Epidemiology of Diabetes Interventions
70. McGowan LM, Kiser WR. Regarding the role of growth hormone and Complications Research Group. N Engl J Med 2000; 342:
in the development of diabetic retinopathy. Diabetes Care 1995; 381–389.
18: 422–423. 90. Reichard P, Nilsson BY, Rosenqvist U. The effect of long-term
71. The effect of a growth hormone receptor antagonist drug on intensified insulin treatment on the development of microvascu-
proliferative diabetic retinopathy. Ophthalmology 2001; 108: lar complications of diabetes mellitus. N Engl J Med 1993; 329:
2266–2272. 304–309.
72. Chantelau E. Effect of a growth hormone receptor antagonist on 91. Klein R, Klein BE, Moss SE. Relation of glycemic control to dia-
proliferative diabetic retinopathy. Ophthalmology 2002; 109: betic microvascular complications in diabetes mellitus. Ann
2187; author reply 2187–2188. Intern Med 1996; 124: 90–96.
73. Ogata N, Tombran-Tink J, Nishikawa M et al. Pigment epithelium- 92. Ohkubo Y, Kishikawa H, Araki E et al. Intensive insulin therapy
derived factor in the vitreous is low in diabetic retinopathy and prevents the progression of diabetic microvascular complications
high in rhegmatogenous retinal detachment. Am J Ophthalmol in Japanese patients with non-insulin-dependent diabetes mel-
2001; 132: 378–382. litus: A randomized prospective 6-year study. Diabetes Res Clin
74. Ogata N, Nishikawa M, Nishimura T, Mitsuma Y, Matsumura M. Pract 1995; 28: 103–117.
Unbalanced vitreous levels of pigment epithelium-derived factor 93. Intensive blood-glucose control with sulphonylureas or insulin
and vascular endothelial growth factor in diabetic retinopathy. compared with conventional treatment and risk of complications
Am J Ophthalmol 2002; 134: 348–353. in patients with type 2 diabetes (UKPDS 33). UK Prospective
75. Boehm BO, Lang G, Feldmann B et al. Proliferative diabetic reti- Diabetes Study (UKPDS) Group. Lancet 1998; 352: 837–853.
nopathy is associated with a low level of the natural ocular anti- 94. Early worsening of diabetic retinopathy in the Diabetes Control
angiogenic agent pigment epithelium-derived factor (PEDF) in and Complications Trial. Arch Ophthalmol 1998; 116: 874–886.
aqueous humor. A pilot study. Horm Metab Res 2003; 35: 382–386. 95. Dahl-Jorgensen K, Brinchmann-Hansen O, Hanssen KF, Sandvik
76. Boehm BO, Lang G, Volpert O et al. Low content of the natural L, Aagenaes O. Rapid tightening of blood glucose control leads
ocular anti-angiogenic agent pigment epithelium-derived factor to transient deterioration of retinopathy in insulin dependent dia-
(PEDF) in aqueous humor predicts progression of diabetic reti- betes mellitus: The Oslo study. Br Med J (Clin Res Ed) 1985; 290:
nopathy. Diabetologia 2003; 46: 394–400. 811–815.
77. Zhang SX, Wang JJ, Gao G, Parke K, Ma JX. Pigment epithelium- 96. Lauritzen T, Frost-Larsen K, Larsen HW, Deckert T. Effect of 1
derived factor downregulates vascular endothelial growth factor year of near-normal blood glucose levels on retinopathy in insu-
(VEGF) expression and inhibits VEGF-VEGF receptor 2 binding in lin-dependent diabetics. Lancet 1983; 1: 200–204.
diabetic retinopathy. J Mol Endocrinol 2006; 37: 1–12. 97. Diabetic retinopathy after two years of intensified insulin treat-
78. Yamagishi S, Matsui T, Nakamura K, Inoue H. Pigment epithe- ment. Follow-up of the Kroc Collaborative Study. The Kroc
lium-derived factor is a pericyte mitogen secreted by micro- Collaborative Study Group. J Am Med Assoc 1988; 260: 37–41.
vascular endothelial cells: Possible participation of angiotensin 98. Chantelau E, Kohner EM. Why some cases of retinopathy worsen
II-elicited PEDF downregulation in diabetic retinopathy. Int J when diabetic control improves. Br Med J 1997; 315: 1105–1106.
Tissue React 2005; 27: 197–202. 99. Lu M, Amano S, Miyamoto K et al. Insulin-induced vascular
79. Matsuoka M, Ogata N, Minamino K, Matsumura M. Expression endothelial growth factor expression in retina. Invest Ophthalmol
of pigment epithelium-derived factor and vascular endothelial Vis Sci 1999; 40: 3281–3286.
growth factor in fibrovascular membranes from patients with 100. Carnesecchi S, Carpentier JL, Foti M, Szanto I. Insulin-induced
proliferative diabetic retinopathy. Jpn J Ophthalmol 2006; 50: vascular endothelial growth factor expression is mediated by the
116–120. NADPH oxidase NOX3. Exp Cell Res 2006; 312: 3413–3424.
80. Khaliq A, Foreman D, Ahmed A et al. Increased expression of 101. Tight blood pressure control and risk of macrovascular and
placenta growth factor in proliferative diabetic retinopathy. Lab microvascular complications in type 2 diabetes: UKPDS 38.
Invest 1998; 78: 109–116. UK Prospective Diabetes Study Group. Br Med J 1998; 317:
81. Frank RN, Amin RH, Eliott D, Puklin JE, Abrams GW. Basic fibro- 703–713.
blast growth factor and vascular endothelial growth factor are 102. Klein BE, Moss SE, Klein R, Surawicz TS. The Wisconsin
present in epiretinal and choroidal neovascular membranes. Am Epidemiologic Study of Diabetic Retinopathy. XIII. Relationship
J Ophthalmol 1996; 122: 393–403. of serum cholesterol to retinopathy and hard exudate.
82. Katsura Y, Okano T, Noritake M et al. Hepatocyte growth fac- Ophthalmology 1991; 98: 1261–1265.
tor in vitreous fluid of patients with proliferative diabetic reti- 103. Ferris FL 3rd, Chew EY, Hoogwerf BJ. Serum lipids and dia-
nopathy and other retinal disorders. Diabetes Care 1998; 21: betic retinopathy. Early Treatment Diabetic Retinopathy Study
1759–1763. Research Group. Diabetes Care 1996; 19: 1291–1293.
83. Hernandez C, Carrasco E, Garcia-Arumi J, Maria Segura R, 104. Effects of aspirin treatment on diabetic retinopathy. ETDRS report
Simo R. Intravitreous levels of hepatocyte growth factor/scat- number 8. Early Treatment Diabetic Retinopathy Study Research
ter factor and vascular cell adhesion molecule-1 in the vitreous Group. Ophthalmology 1991; 98: 757–765.
fluid of diabetic patients with proliferative retinopathy. Diabetes 105. Chew EY, Klein ML, Murphy RP, Remaley NA, Ferris FL 3rd.
Metab 2004; 30: 341–346. Effects of aspirin on vitreous/preretinal hemorrhage in patients
84. Simo R, Lecube A, Garcia-Arumi J, Carrasco E, Hernandez C. with diabetes mellitus. Early Treatment Diabetic Retinopathy
Hepatocyte growth factor in the vitreous fluid of patients with Study report no. 20. Arch Ophthalmol 1995; 113: 52–55.
proliferative diabetic retinopathy: Its relationship with vascular 106. Preliminary report on effects of photocoagulation therapy. The
endothelial growth factor and retinopathy activity. Diabetes Care Diabetic Retinopathy Study Research Group. Am J Ophthalmol
2004; 27: 287–288. 1976; 81: 383–396.
85. Demiroglu H. The balance between anemia, erythropoietin treat- 107. Photocoagulation treatment of proliferative diabetic retinopathy.
ment, and elevated erythrocyte aggregation in patients with dia- Clinical application of Diabetic Retinopathy Study (DRS) findings,
betic retinopathy and nephropathy: A hematologic point of view. DRS report number 8. The Diabetic Retinopathy Study Research
Am J Ophthalmol 2003; 136: 776; author reply 776–777. Group. Ophthalmology 1981; 88: 583–600.
108. Early photocoagulation for diabetic retinopathy. ETDRS report 130. Bandello F, Polito A, Pognuz DR, Monaco P, Dimastrogiovanni A,
number 9. Early Treatment Diabetic Retinopathy Study Research Paissios J. Triamcinolone as adjunctive treatment to laser panreti-
Group. Ophthalmology 1991; 98: 766–785. nal photocoagulation for proliferative diabetic retinopathy. Arch
109. Ferris F. Early photocoagulation in patients with either type I or Ophthalmol 2006; 124: 643–650.
type II diabetes. Trans Am Ophthalmol Soc 1996; 94: 505–537. 131. Jonas JB. Intravitreal triamcinolone acetonide for diabetic reti-
110. Machemer R, Norton EW. A new concept for vitreous surgery. 3. nopathy. Dev Ophthalmol 2007; 39: 96–110.
Indications and results. Am J Ophthalmol 1972; 74: 1034–1056. 132. Adamis AP, Shima DT, Tolentino MJ et al. Inhibition of vascular
111. Machemer R. A new concept for vitreous surgery. 2. Surgical endothelial growth factor prevents retinal ischemia-associated
technique and complications. Am J Ophthalmol 1972; 74: iris neovascularization in a nonhuman primate. Arch Ophthalmol
1022–1033. 1996; 114: 66–71.
112. Machemer R, Parel JM, Buettner H. A new concept for vitreous 133. Robinson GS, Pierce EA, Rook SL, Foley E, Webb R, Smith LE.
surgery. I. Instrumentation. Am J Ophthalmol 1972; 73: 1–7. Oligodeoxynucleotides inhibit retinal neovascularization in a
113. Two-year course of visual acuity in severe proliferative dia- murine model of proliferative retinopathy. Proc Natl Acad Sci U
betic retinopathy with conventional management. Diabetic S A 1996; 93: 4851–4856.
Retinopathy Vitrectomy Study (DRVS) report #1. Ophthalmology 134. Ng EW, Adamis AP. Anti-VEGF aptamer (pegaptanib) therapy for
1985; 92: 492–502. ocular vascular diseases. Ann N Y Acad Sci 2006; 1082: 151–171.
114. Early vitrectomy for severe vitreous hemorrhage in diabetic 135. Ng EW, Shima DT, Calias P, Cunningham ET Jr., Guyer DR,
retinopathy. Two-year results of a randomized trial. Diabetic Adamis AP. Pegaptanib, a targeted anti-VEGF aptamer for ocular
Retinopathy Vitrectomy Study report 2. The Diabetic Retinopathy vascular disease. Nat Rev Drug Discov 2006; 5: 123–132.
Vitrectomy Study Research Group. Arch Ophthalmol 1985; 103: 136. Avery RL. Regression of retinal and iris neovascularization after
1644–1652. intravitreal bevacizumab (Avastin) treatment. Retina 2006; 26:
115. Early vitrectomy for severe proliferative diabetic retinopathy in 352–354.
eyes with useful vision. Results of a randomized trial—Diabetic 137. Avery RL, Pearlman J, Pieramici DJ et al. Intravitreal bevacizumab
Retinopathy Vitrectomy Study report 3. The Diabetic Retinopathy (Avastin) in the treatment of proliferative diabetic retinopathy.
Vitrectomy Study Research Group. Ophthalmology 1988; 95: Ophthalmology 2006; 113: 1695 e1–1695 e15.
1307–1320. 138. Minnella AM, Savastano CM, Ziccardi L, Scupola A, Falsini B,
116. Early vitrectomy for severe proliferative diabetic retinopathy in Balestrazzi E. Intravitreal bevacizumab (Avastin) in proliferative
eyes with useful vision. Clinical application of results of a ran- diabetic retinopathy. Acta Ophthalmol 2008; 86: 683–687.
domized trial—Diabetic Retinopathy Vitrectomy Study report 4. 139. Mirshahi A, Roohipoor R, Lashay A, Mohammadi SF,
The Diabetic Retinopathy Vitrectomy Study Research Group. Abdoallahi A, Faghihi H. Bevacizumab-augmented retinal
Ophthalmology 1988; 95: 1321–1334. laser photocoagulation in proliferative diabetic retinopathy:
117. Early vitrectomy for severe vitreous hemorrhage in diabetic A randomized double-masked clinical trial. Eur J Ophthalmol
retinopathy. Four-year results of a randomized trial: Diabetic 2008; 18: 263–269.
Retinopathy Vitrectomy Study report 5. Arch Ophthalmol 1990; 140. Arevalo JF, Wu L, Sanchez JG et al. Intravitreal bevacizumab
108: 958–964. (Avastin) for proliferative diabetic retinopathy: 6-months follow-
118. Bhavsar AR. Diabetic retinopathy: The latest in current manage- up. Eye (Lond) 2009; 23: 117–123.
ment. Retina 2006; 26: S71–S79. 141. Gonzalez VH, Giuliari GP, Banda RM, Guel DA. Intravitreal
119. Hirata A, Takano A, Inomata Y, Yonemura N, Sagara N, Tanihara injection of pegaptanib sodium for proliferative diabetic retinop-
H. Plasmin-assisted vitrectomy for management of proliferative athy. Br J Ophthalmol 2009; 93: 1474–1478.
membrane in proliferative diabetic retinopathy: A pilot study. 142. Ishikawa K, Honda S, Tsukahara Y, Negi A. Preferable use of
Retina 2007; 27: 1074–1078. intravitreal bevacizumab as a pretreatment of vitrectomy for
120. Kador PF, Akagi Y, Takahashi Y, Ikebe H, Wyman M, Kinoshita severe proliferative diabetic retinopathy. Eye (Lond) 2009; 23:
JH. Prevention of retinal vessel changes associated with diabetic 108–111.
retinopathy in galactose-fed dogs by aldose reductase inhibitors. 143. Smith JM, Steel DH. Anti-vascular endothelial growth factor
Arch Ophthalmol 1990; 108: 1301–1309. for prevention of postoperative vitreous cavity haemorrhage
121. Robinson WG Jr., Laver NM, Jacot JL et al. Diabetic-like reti- after vitrectomy for proliferative diabetic retinopathy. Cochrane
nopathy ameliorated with the aldose reductase inhibitor WAY- Database Syst Rev 2011: CD008214.
121,509. Invest Ophthalmol Vis Sci 1996; 37: 1149–1156. 144. Zhao LQ, Zhu H, Zhao PQ, Hu YQ. A systematic review and
122. The sorbinil retinopathy trial: Neuropathy results. Sorbinil meta-analysis of clinical outcomes of vitrectomy with or without
Retinopathy Trial Research Group. Neurology 1993; 43: 1141–1149. intravitreal bevacizumab pretreatment for severe diabetic reti-
123. Harhaj NS, Felinski EA, Wolpert EB, Sundstrom JM, Gardner TW, nopathy. Br J Ophthalmol 2011; 95: 1216–1222.
Antonetti DA. VEGF activation of protein kinase C stimulates 145. Sheth BP. Does pregnancy accelerate the rate of progression
occludin phosphorylation and contributes to endothelial perme- of diabetic retinopathy?: An update. Curr Diab Rep 2008; 8:
ability. Invest Ophthalmol Vis Sci 2006; 47: 5106–5115. 270–273.
124. Brownlee M, Vlassara H, Kooney A, Ulrich P, Cerami A. 146. Jervell J, Moe N, Skjaeraasen J, Blystad W, Egge K. Diabetes mel-
Aminoguanidine prevents diabetes-induced arterial wall protein litus and pregnancy—Management and results at Rikshospitalet,
cross-linking. Science 1986; 232: 1629–1632. Oslo, 1970–1977. Diabetologia 1979; 16: 151–155.
125. Aiello LP, Pierce EA, Foley ED et al. Suppression of retinal neo- 147. Dibble CM, Kochenour NK, Worley RJ, Tyler FH, Swartz M. Effect
vascularization in vivo by inhibition of vascular endothelial of pregnancy on diabetic retinopathy. Obstet Gynecol 1982; 59:
growth factor (VEGF) using soluble VEGF-receptor chimeric pro- 699–704.
teins. Proc Natl Acad Sci U S A 1995; 92: 10457–10461. 148. Moloney JB, Drury MI. The effect of pregnancy on the natu-
126. Cunningham ET Jr., Adamis AP, Altaweel M et al. A phase II ran- ral course of diabetic retinopathy. Am J Ophthalmol 1982; 93:
domized double-masked trial of pegaptanib, an anti-vascular 745–756.
endothelial growth factor aptamer, for diabetic macular edema. 149. Price JH, Hadden DR, Archer DB, Harley JM. Diabetic retinopa-
Ophthalmology 2005; 112: 1747–1757. thy in pregnancy. Br J Obstet Gynaecol 1984; 91: 11–17.
127. Mori K, Duh E, Gehlbach P et al. Pigment epithelium-derived 150. Soubrane G, Canivet J, Coscas G. Influence of pregnancy on the
factor inhibits retinal and choroidal neovascularization. J Cell evolution of background retinopathy. Preliminary results of a pro-
Physiol 2001; 188: 253–263. spective fluorescein angiography study. Int Ophthalmol 1985; 8:
128. Rasmussen H, Chu KW, Campochiaro P et al. Clinical protocol. 249–255.
An open-label, phase I, single administration, dose-escalation 151. Phelps RL, Sakol P, Metzger BE, Jampol LM, Freinkel N. Changes
study of ADGVPEDF.11D (ADPEDF) in neovascular age-related in diabetic retinopathy during pregnancy. Correlations with
macular degeneration (AMD). Hum Gene Ther 2001; 12: regulation of hyperglycemia. Arch Ophthalmol 1986; 104:
2029–2032. 1806–1810.
129. Grant MB, Mames RN, Fitzgerald C et al. The efficacy of octreo- 152. Serup L. Influence of pregnancy on diabetic retinopathy. Acta
tide in the therapy of severe nonproliferative and early prolif- Endocrinol Suppl (Copenh) 1986; 277: 122–124.
erative diabetic retinopathy: A randomized controlled study. 153. Klein BE, Moss SE, Klein R. Effect of pregnancy on progression of
Diabetes Care 2000; 23: 504–509. diabetic retinopathy. Diabetes Care 1990; 13: 34–40.
References 465
154. Rosenn B, Miodovnik M, Kranias G et al. Progression of diabetic 175. Diabetic Retinopathy: Preferred Practice Pattern. American
retinopathy in pregnancy: Association with hypertension in preg- Academy of Ophthalmology, 2015. http://www.aao.org/
nancy. Am J Obstet Gynecol 1992; 166: 1214–1218. summary-benchmark-detail/diabetic-retinopathy-summary-
155. Chew EY, Mills JL, Metzger BE et al. Metabolic control and progres- benchmark–october-20.
sion of retinopathy. The Diabetes in Early Pregnancy Study. National 176. Puza SW, Malee MP. Utilization of routine ophthalmologic exam-
Institute of Child Health and Human Development Diabetes in inations in pregnant diabetic patients. J Matern Fetal Med 1996;
Early Pregnancy Study. Diabetes Care 1995; 18: 631–637. 5: 7–10.
156. Hellstedt T, Kaaja R, Teramo K, Immonen I. The effect of preg- 177. Halperin LS, Olk RJ, Soubrane G, Coscas G. Safety of fluores-
nancy on mild diabetic retinopathy. Graefes Arch Clin Exp cein angiography during pregnancy. Am J Ophthalmol 1990; 109:
Ophthalmol 1997; 235: 437–441. 563–566.
157. Rasmussen KL, Laugesen CS, Ringholm L, Vestgaard M, Damm 178. Hercules BL, Wozencroft M, Gayed II, Jeacock J. Peripheral reti-
P, Mathiesen ER. Progression of diabetic retinopathy during preg- nal ablation in the treatment of proliferative diabetic retinopathy
nancy in women with type 2 diabetes. Diabetologia 2010; 53: during pregnancy. Br J Ophthalmol 1980; 64: 87–93.
1076–1083. 179. Frank RN. Diabetic retinopathy: Current concepts of evaluation
158. Vestgaard M, Ringholm L, Laugesen CS, Rasmussen KL, Damm P, and treatment. Clin Endocrinol Metab 1986; 15: 933–969.
Mathiesen ER. Pregnancy-induced sight-threatening diabetic reti- 180. Conway M, Baldwin J, Kohner EM, Schulenburg WE, Cassar J.
nopathy in women with type 1 diabetes. Diabet Med 2010; 27: Postpartum progression of diabetic retinopathy. Diabetes Care
431–435. 1991; 14: 1110–1111.
159. Stephens JW, Page OC, Hare RL. Diabetes and pregnancy. A 181. Chan WC, Lim LT, Quinn MJ, Knox FA, McCance D, Best RM.
report of experiences in 119 pregnancies over a period of ten Management and outcome of sight-threatening diabetic retinopa-
years. Diabetes 1963; 12: 213–219. thy in pregnancy. Eye 2004; 18: 826–832.
160. Horvat M, Maclean H, Goldberg L, Crock GW. Diabetic retinop- 182. Treatment techniques and clinical guidelines for photocoagulation
athy in pregnancy: A 12-year prospective survey. Br J Ophthalmol of diabetic macular edema. Early Treatment Diabetic Retinopathy
1980; 64: 398–403. Study report number 2. Early Treatment Diabetic Retinopathy
161. Lovestam-Adrian M, Agardh CD, Aberg A, Agardh E. Pre- Study Research Group. Ophthalmology 1987; 94: 761–774.
eclampsia is a potent risk factor for deterioration of retinopathy 183. Cassar J, Hamilton AM, Kohner EM. Diabetic retinopathy in preg-
during pregnancy in type 1 diabetic patients. Diabet Med 1997; nancy. Int Ophthalmol Clin 1978; 18: 179–188.
14: 1059–1065. 184. Kitzmiller JL, Main E, Ward B, Theiss T, Peterson DL. Insulin lispro
162. Effect of pregnancy on microvascular complications in the dia- and the development of proliferative diabetic retinopathy during
betes control and complications trial. The Diabetes Control and pregnancy. Diabetes Care 1999; 22: 874–876.
Complications Trial Research Group. Diabetes Care 2000; 23: 185. Zib I, Raskin P. Novel insulin analogues and its mitogenic poten-
1084–1091. tial. Diabetes Obes Metab 2006; 8: 611–620.
163. Temple RC, Aldridge VA, Sampson MJ, Greenwood RH, Heyburn 186. Persson B, Swahn ML, Hjertberg R et al. Insulin lispro therapy in
PJ, Glenn A. Impact of pregnancy on the progression of diabetic pregnancies complicated by type 1 diabetes mellitus. Diabetes
retinopathy in type 1 diabetes. Diabet Med 2001; 18: 573–577. Res Clin Pract 2002; 58: 115–121.
164. Gordin D, Kaaja R, Forsblom C, Hiilesmaa V, Teramo K, Groop 187. Buchbinder A, Miodovnik M, McElvy S et al. Is insulin lispro
PH. Pre-eclampsia and pregnancy-induced hypertension are associated with the development or progression of diabetic
associated with severe diabetic retinopathy in type 1 diabetes retinopathy during pregnancy? Am J Obstet Gynecol 2000; 183:
later in life. Acta Diabetol 2013; 50: 781–787. 1162–1165.
165. Arun CS, Taylor R. Influence of pregnancy on long-term progres- 188. Bhattacharyya A, Vice PA. Insulin lispro, pregnancy, and retinop-
sion of retinopathy in patients with type 1 diabetes. Diabetologia athy. Diabetes Care 1999; 22: 2101–2104.
2008; 51: 1041–1045. 189. Kurtzhals P, Schaffer L, Sorensen A et al. Correlations of
166. Chen HC, Newsom RS, Patel V, Cassar J, Mather H, Kohner EM. receptor binding and metabolic and mitogenic potencies of
Retinal blood flow changes during pregnancy in women with dia- insulin analogs designed for clinical use. Diabetes 2000; 49:
betes. Invest Ophthalmol Vis Sci 1994; 35: 3199–3208. 999–1005.
167. Cundy T, Slee F, Gamble G, Neale L. Hypertensive disorders of 190. Lapolla A, Dalfra MG, Fedele D. Insulin therapy in pregnancy
pregnancy in women with type 1 and type 2 diabetes. Diabet complicated by diabetes: Are insulin analogs a new tool?
Med 2002; 19: 482–489. Diabetes Metab Res Rev 2005; 21: 241–252.
168. Ringholm L, Vestgaard M, Laugesen CS, Juul A, Damm P, 191. Rosenstock J, Schwartz SL, Clark CM Jr., Park GD, Donley DW,
Mathiesen ER. Pregnancy-induced increase in circulating IGF-I Edwards MB. Basal insulin therapy in type 2 diabetes: 28-week
is associated with progression of diabetic retinopathy in women comparison of insulin glargine (HOE 901) and NPH insulin.
with type 1 diabetes. Growth Horm IGF Res 2011; 21: 25–30. Diabetes Care 2001; 24: 631–636.
169. Kastelan S, Tomic M, Pavan J, Oreskovic S. Maternal immune sys- 192. Yki-Jarvinen H, Dressler A, Ziemen M. Less nocturnal hypoglyce-
tem adaptation to pregnancy—A potential influence on the course mia and better post-dinner glucose control with bedtime insulin
of diabetic retinopathy. Reprod Biol Endocrinol 2010; 8: 124. glargine compared with bedtime NPH insulin during insulin com-
170. McElvy SS, Demarini S, Miodovnik M, Khoury JC, Rosenn B, bination therapy in type 2 diabetes. HOE 901/3002 Study Group.
Tsang RC. Fetal weight and progression of diabetic retinopathy. Diabetes Care 2000; 23: 1130–1136.
Obstet Gynecol 2001; 97: 587–592. 193. Jovanovic R, Jovanovic L. Obstetric management when normo-
171. Klein BE, Klein R, Meuer SM, Moss SE, Dalton DD. Does the glycemia is maintained in diabetic pregnant women with vascu-
severity of diabetic retinopathy predict pregnancy outcome? lar compromise. Am J Obstet Gynecol 1984; 149: 617–623.
J Diabet Complications 1988; 2: 179–184. 194. Kassoff A, Catalano RA, Mehu M. Vitreous hemorrhage and the
172. Lauszus FF, Gron PL, Klebe JG. Pregnancies complicated by dia- Valsalva maneuver in proliferative diabetic retinopathy. Retina
betic proliferative retinopathy. Acta Obstet Gynecol Scand 1998; 1988; 8: 174–176.
77: 814–818. 195. Jones WL. Valsalva maneuver induced vitreous hemorrhage.
173. Sunness JS. The pregnant woman’s eye. Surv Ophthalmol 1988; J Am Optom Assoc 1995; 66: 301–304.
32: 219–238. 196. White P. Pregnancy and diabetes. In: Marble A, White P, Bradley
174. NICE clinical guideline 63: Diabetes in pregnancy, 2008. https:// RF eds., Joslins Diabetes Mellitus, Lea & Febiger: Philadelphia,
www.nice.org.uk/guidance/cg63. PA, 1971, pp. 581–598.
55 Diabetic nephropathy
Elisabeth R. Mathiesen, Lene Ringholm, and Peter Damm
defined as a urinary albumin excretion of 30–300 mg/24 h,

Introduction corresponding to a spot urine albumin to creatinine ratio of
By the end of the last century, diabetic nephropathy was pres- 30–300 μg/mg. Left untreated, microalbuminuria tends to
ent in up to 7% among unselected pregnant women with progress to overt diabetic nephropathy characterized by persis-
type 1 diabetes, but in a recent study, it was reported that the tent proteinuria, hypertension, and a relentless decline in glo-
prevalence of diabetic nephropathy is now 2.5% and 2.3% merular filtration rate.14 Histological changes in the glomeruli
in pregnancies complicated by type 1 and type 2 diabetes, with increased basal membrane thickness and glomeruloscle-
respectively. In line with this, the current prevalence of micro- rosis are characteristic. Elevated blood pressure is a marker of
albuminuria is now 3.4% and 4.5%, respectively.1,2 poor prognosis and is associated with progression of the kidney
For many years, pregnancy in women with diabetic disease. If left untreated, progression to end-stage renal dis-
nephropathy was associated with a very high risk of severe ease occurs within a median of 7 years after onset of diabetic
pregnancy complications including perinatal mortality as nephropathy. Meanwhile, progression of diabetic nephropathy
well as deterioration of maternal kidney function leading to can be slowed by the use of strict antihypertensive treatment
end-stage renal disease.3,4 However, maternal and perinatal including renin–angiotensin system inhibition in terms of
mortality and morbidity rates in pregnancies complicated by angiotensin-converting enzyme (ACE) inhibitors or angioten-
diabetic nephropathy have declined substantially during the sin II receptor antagonists as first-line drugs. Accordingly, the
last decade. Successful pregnancy outcome with fetal sur- prognosis has improved considerably.4,15–17 It is often necessary
vival rates of up to 95%–100% is now the norm in developed to combine this treatment with diuretics, beta-blockers, and/
countries.5–12 Nonetheless, maternal and perinatal complica- or calcium antagonists in order to control the blood pressure
tions in pregnancies of women with diabetic nephropathy and the urinary albumin excretion sufficiently. Strict antihy-
are often related to elevated blood pressure and are still more pertensive treatment in patients with diabetic nephropathy
common than in diabetic women with normal urinary albu- results in the preservation of kidney function documented by
min excretion at conception.2,4 a reduction in the decline of glomerular filtration rate to less
The present chapter highlights factors of importance than one-third of the decline in untreated patients.18 Inhibition
for the clinical care of pregnant women with diabetic of the renin–angiotensin system in normotensive patients with
nephropathy or microalbuminuria with particular focus on microalbuminuria may reduce and possibly eliminate albu-
the possible role of strict antihypertensive treatment dur- minuria in patients with type 1 diabetes treated with ACE
ing pregnancy. The majority of data on this topic are from inhibitors18,19 as well as in patients with type 2 diabetes treated
women with type 1 diabetes. Most likely the findings are with angiotensin II receptor antagonists.20 The treatment goal
similar in women with type 2 diabetes, and the clinical rec- includes blood pressure below 130/80 mmHg21 and to lower or
ommendations given in this chapter are probably useful in normalize urinary albumin excretion.
both types of diabetes. Renin–angiotensin system inhibition is, however, con-
traindicated in pregnancy, and alternative antihypertensive
drugs tolerated in pregnancy must be used as discussed in
the following text.
Pathophysiology and treatment of
diabetic nephropathy outside of
pregnancy Pathophysiology of preeclampsia and
hypertension in diabetic pregnancy
Diabetic nephropathy is a progressive disease that affects
approximately 30% of patients with diabetes and is the most Preeclampsia is characterized by hypertension, proteinuria,
common cause of end-stage renal disease in the United States. and peripheral edema developing after 20 gestational weeks.
In Denmark, 22% of patients with end-stage kidney disease Patients with diabetic nephropathy or microalbumin-
have diabetes.13 The first clinical sign is microalbuminuria, uria prior to pregnancy are at increased risk of developing
466
Effect of diabetic nephropathy on pregnancy outcome 467
preeclampsia and may already present with elevated blood nine gestational weeks as compared with women not devel-
pressure in early pregnancy.22 Women with diabetes developing preeclampsia. Throughout pregnancy, ANP levels were
oping preeclampsia are generally characterized by higher 34% higher in these women.37 However, the function of pla-
urinary albumin excretion, blood pressure, and HbA1c in centa in the early stage of pregnancy judged by the level of
early pregnancy compared with women who do not develop activin A and inhibin A is often well preserved in diabetic
preeclampsia.23 Women with type 1 diabetes developing women developing preeclampsia.38 Similarly, growth restric-
preeclampsia have impaired vasodilatatory capacity present tion of the fetus is rare in diabetic women with preeclampsia.38
already in early pregnancy, several weeks prior to the clini-
cal presentation of preeclampsia. In addition, the markers
of endothelial dysfunction vascular cell adhesion molecule Pathogenesis of preeclampsia
(VCAM) and intercellular adhesion molecule (I-CAM) are
elevated in early pregnancy.23 Signs of vascular dysfunction in women with diabetes and
thus precede development of clinical preeclampsia in women nephropathy
with type 1 diabetes who are prone to the condition. In addi-
tion, several other pathophysiological aspects are involved We suggest that the increased prevalence of preeclampsia
such as increased oxidative stress and reduced antioxidative in women with type 1 diabetes complicated with diabetic
defenses that may be related to levels of vitamin C and E.24,25 nephropathy or microalbuminuria is mainly related to
However, supplementation with vitamin C and E in a ran- common maternal constitutional factors with an increased
domized study including 762 women with type 1 diabetes susceptibility to endothelial dysfunction. The pathogenesis
did not reduce the risk of preeclampsia.25 As in preeclamp- of development of preeclampsia in women with diabetic
sia in women without diabetes, preeclampsia in women with nephropathy or microalbuminuria and type 1 diabetes does
type 1 diabetes is also associated with elevated levels of anti- thus include presence of endothelial dysfunction,23 impaired
angiogenic factors in the third trimester.26 maximal vasodilatation,23 increased levels of components
of the renin–angiotensin system,32 and markers of cardiac
overload.37 All these can be modulated by antihypertensive
treatment. Thus, tight antihypertensive treatment during
Preeclampsia and the pregnancy to prevent increase in blood pressure and/or uri-
renin–angiotensin system nary albumin excretion theoretically should be beneficial in
these women.
During the early stages of normal pregnancy, activation of
the local27,28 and systemic29,30 renin–angiotensin systems
is present. At 3–6 gestational weeks, the plasma prorenin Effect of diabetic nephropathy on
level increases tenfold, with a much lower prorenin level
from 9 weeks onward.27 This is consistent with the role of pregnancy outcome
the renin–angiotensin system, particularly prorenin, in Diabetic nephropathy may adversely affect the outcome of
embryonic and fetal development and in placentation.27 In pregnancy mainly by three mechanisms: (1) preterm deliv-
preeclampsia, disturbance in the renin–angiotensin sys- ery due to maternal hypertensive complications, (2) fetal
tem is seen with increased vascular responsiveness to intrauterine growth restriction and fetal distress caused by
angiotensin II. 31 The prorenin levels in 108 pregnant women placental dysfunction, or (3) deterioration of maternal kid-
with type 1 diabetes have been prospectively investigated, ney function. In addition, congenital malformations have
and prorenin concentrations at 8 gestational weeks were posi- been described with a slightly higher prevalence in women
tively associated with later development of preeclampsia.32 with diabetic nephropathy compared to diabetic women
Likewise, throughout pregnancy, prorenin concentrations with normal kidney function.39
were 30% higher in the nine women with type 1 diabetes who The prevalence of preeclampsia in women with diabetic
developed preeclampsia compared to those who did not.32 nephropathy is up to 64%4,5,7,12 especially in the presence
of reduced kidney function,40 hypertension at the start of
pregnancy, or nephrotic proteinuria.5,7 Also, women with
Preeclampsia and vasoactive markers type 1 diabetes and microalbuminuria are at increased risk
of developing preeclampsia compared to women with type 1
The vasoactive marker of cardiac overload, atrial natriuretic diabetes and normal urinary albumin excretion,26,41 and
peptide (ANP), is synthesized in cardiac tissue in response these women are also at particular risk of preterm delivery41
to volume expansion and ventricular pressure overload.33,34 where preterm delivery before 34 gestational weeks has been
In nondiabetic women, increased levels of ANP and brain reported in up to 45%.9,41 Severe disabilities of the children
natrioretic protein (BNP) are seen in late pregnancy when born to mothers with diabetic nephropathy have also been
the diagnosis of preeclampsia has been established.35,36 In described. In a follow-up study of 35 children born between
a small, prospective series of women with type 1 diabetes 1982 and 1992 by women with diabetic nephropathy, the
followed throughout pregnancy, preeclampsia developed majority were developmentally normal, but neurodevelop-
in six women (7%) with significantly higher ANP levels at mental problems were seen in children born preterm with a
birth weight less than 2000 g.9 The risk of perinatal mortality creatinine, blood pressure, and proteinuria is necessary to
in pregnancies complicated by diabetic nephropathy is now estimate the risk for complications during pregnancy in a
close to that of women with type 1 diabetes without diabetic woman with diabetic nephropathy. Smoking and alcohol
nephropathy.2,4,5,7,12 Recently, similar pregnancy outcomes consumption are discouraged strongly during pregnancy.
were reported in women with kidney involvement and either Both the level of blood pressure and the actual number of
type 1 or type 2 diabetes.1 antihypertensive agents used to control the blood pressure
sufficiently prior to pregnancy are of importance to predict
and plan the need for further intensification of antihyper-
Effects of pregnancy on diabetic tensive treatment during pregnancy.
nephropathy
Glycemic control
Only a few studies have addressed the long-term impact
Poor glycemic control before pregnancy is associated with
of pregnancy on renal function in women with diabetic
pregnancy complications such as congenital malforma-
nephropathy. The most recent study is a prospective
tions,46,47 preeclampsia,48–50 and preterm delivery.46 Strict
cohort study including women with diabetic nephropa-
glycemic control is therefore the goal, and HbA1c as close to
thy, and normal serum creatinine offered strict blood
normal as possible at least below 7% is recommended. The
pressure control during the whole follow-up period of
risk of severe hypoglycemia has to be taken into account.51
16 years. In total, 26 women had at least one pregnancy
during the period and 67 had no pregnancies in the obser-
vation period. In these women with normal serum creati- Low-dose aspirin
nine, pregnancy was not associated with a greater decline
Low-dose aspirin treatment is indicated in women with dia-
in kidney function or impaired long-term maternal sur-
betic nephropathy or microalbuminuria to prevent cardio-
vival.42 However, other studies report that in women with
vascular events, and it might also prevent preeclampsia.52
a reduced creatinine clearance before pregnancy, there is
If a woman is already on low-dose aspirin treatment before
increased risk of deterioration of kidney function during
pregnancy, this treatment may be continued, while initiation
pregnancy.6,22,43
of low-dose aspirin treatment in pregnancy normally is post-
In general, pregnancy outcome is favorable in women
poned until after organogenesis (10–12 weeks).
with small elevations in serum creatinine, i.e., less than
124 μmol/L (1.4 mg/dL), with proteinuria less than 1 g/24 h,
and with normal blood pressure.44 In contrast, serum creati- Folic acid
nine above 176 μmol/L and severe hypertension or protein-
Daily folic acid supplementation should be advised as it
uria in the nephrotic range (>3 g/24 h) and/or preexisting
might reduce the risk of fetal malformations.53 There is no
cardiovascular disease are associated with a high risk for
consensus on the dose of folic acid; at least 400 μg/day is rec-
poor maternal and fetal outcome.44 The long-term survival of
ommended in Denmark, while in Canada and the United
a mother with diabetic nephropathy has improved consider-
States, the recommendation is up to 5 mg/day.54
ably in recent years, but the long-term likelihood of compli-
cations including visual impairment and renal dysfunction
is still increased.4,9,11,42 Blood pressure control
Prepregnancy treatment with ACE inhibitors combined
with strict metabolic control for at least 6 months resulting
Prepregnancy counseling of women in low levels of urinary albumin excretion has been found
with diabetic nephropathy to be associated with a high rate of successful pregnancy
outcome.5
Careful counseling of the woman with diabetes and kidney Whether treatment with ACE inhibitors or angioten-
involvement concerning the risk for herself and the newborn sin II receptor antagonists in early pregnancy is associated
is important for a well-considered decision regarding preg- with increased risk of congenital malformations is cur-
nancy. The use of safe contraception in the planning phase is rently being debated.39,55–57 However, treatment with ACE
recommended.45 inhibitors during the last part of pregnancy is associated
To minimize the risk of pregnancy-induced maternal with abnormal fetal renal development and neonatal renal
progression to end-stage renal failure, focusing on the pre- failure.55,58 Treatment with ACE inhibitors or angiotensin
pregnancy level of serum creatinine is important, since II receptor antagonists should therefore ideally be stopped
serum creatinine above 176 µmol/L is the best predictor of prior to conception,55,58 but if these drugs are given during
the risk of pregnancy-induced decline in maternal kidney organogenesis, interruption of pregnancy is not indicated
function leading to end-stage renal disease during preg- on this background.56 If the severity of diabetic nephropathy
nancy or shortly after.44 In addition, an updated diabetes sta- indicates continuous treatment with inhibitors of the renin–
tus including self-monitored glucose values, HbA1c, risk of angiotensin system, in the case of unplanned pregnancy, a
severe hypoglycemia, degree of diabetic retinopathy, serum shift to other antihypertensive agents can successfully take
Treatment of women with diabetic nephropathy during pregnancy 469
place in early pregnancy when the pregnancy test is posi- Treatment of women with diabetic
tive.5 Types of antihypertensive agents that are considered
safe in pregnancy are methyldopa, beta-blockers (such as nephropathy during pregnancy
labetalol), and the calcium antagonists nifedipine and diltia-
Glycemic control
zem.4 Although the use of diuretics throughout pregnancy is
controversial,59 we have good clinical experience with con- Strict glycemic control during pregnancy is of utmost
tinuation of diuretic treatment initiated before pregnancy importance but may be difficult because pregnant women
in stable doses during pregnancy in women with diabetic with type 1 diabetes have an increased risk of severe hypo-
nephropathy.4,60 glycemia, particularly in early pregnancy.51 Development of
preeclampsia is more frequent in women with higher levels
of HbA1c in early pregnancy,48–50 but it is reassuring that
Cholesterol-lowering drugs improvement of glycemic control during pregnancy is asso-
Treatment with statins or other cholesterol-lowering agents ciated with less preeclampsia.64 The British NICE guidelines
during pregnancy may be associated with malformations or recommend HbA1c below 6.0% during pregnancy.65
changes in the development of the central nervous system
and should ideally be discontinued prior to pregnancy61 or
as soon as pregnancy is diagnosed. Low-dose aspirin
In women with high risk of developing preeclampsia, treat-
ment with low-dose aspirin has some preventive effect52
Diabetic retinopathy if initiated before 16 gestational weeks. It is not known what
Sight-threatening diabetic retinopathy is prevalent in is the optimal gestational age to start low-dose aspirin, but
women with diabetic nephropathy, and it is well known generally it is initiated after the organogenesis. Low-dose
that a pregnancy-induced deterioration of retinopathy may aspirin treatment in women with diabetic nephropathy is
occur. Treatment for diabetic retinopathy should therefore therefore recommended in American66 and British65 guide-
be performed if indicated to stabilize the retinopathy prior lines. The British NICE guidelines now recommend 75 mg
to pregnancy.62,63 aspirin daily from 12 gestational weeks to all pregnant
women with diabetes and/or kidney disease.65 Preferably
this treatment should be stopped 1 week before delivery,
Summary of prepregnancy care i.e., at 36–37 weeks.
Intensive glycemic control, folic acid supplementation, and
intensive antihypertensive treatment are of importance
prior to pregnancy in women with diabetic nephropathy or Folic acid
microalbuminuria. Renin–angiotensin system inhibitors Daily folic acid supplementation during the first 12 gesta-
and cholesterol-lowering agents should ideally be stopped tional weeks might reduce the incidence of congenital mal-
before pregnancy or at the latest when pregnancy is diag- formations in pregnant women with diabetes.53
nosed. Screening for diabetic retinopathy and treatment, if
indicated, is important (Table 55.1).
Blood pressure control
A gradual increase in both blood pressure and urinary albu-
Table 55.1 Prepregnancy counseling of women with min excretion has been demonstrated prior to onset of pre-
diabetic nephropathy eclampsia in women with type 1 diabetes.48 With the aim
to reduce the high prevalence of preeclampsia in pregnant
• Use of safe contraception in the planning phase women with diabetic nephropathy or microalbuminuria,
• Evaluation of the risk of pregnancy-induced maternal
our group in the year 2000 decided to initiate antihyperten-
kidney failure sive treatment if blood pressure exceeded 140/90 mmHg or
• Evaluation of the risk of preeclampsia and preterm urinary albumin excretion exceeded 2000 mg/24 h. If the
delivery women already were on antihypertensive treatment, the
• Intensive glycemic control with HbA1c as near to normal drugs were changed to antihypertensive agents well tolerated
levels as possible (<7%, ~50 mmol/mol) in pregnancy such as methyldopa, labetalol, or nifedipine. In
• Supplementation with folic acid an unselected cohort of 20 normotensive pregnant women
• Tight target for antihypertensive treatment, i.e., with type 1 diabetes and microalbuminuria treated accord-
BP <130/80 mmHg and urinary albumin excretion ing to this strategy, a significant reduction in preterm deliv-
<300 mg/24 h ery before 34 gestational weeks was seen when compared
• Consider change to pregnancy-friendly antihypertensive to a previous cohort where antihypertensive treatment
agents before conception for elevated blood pressure was less rigorous.60 However,
• Review the medication list for drugs contraindicated in the prevalence of preeclampsia and preterm delivery was
pregnancy, i.e., cholesterol-lowering agents
still high, and therefore, we decided in 2004 to intensify the
• Screen for sight-threatening diabetic retinopathy
strategy even more by initiating antihypertensive treatment
when blood pressure exceeds 135/85 mmHg or urinary albu- Diabetic retinopathy
min excretion exceeds 300 mg/24 h. This strategy seems to Besides protecting the kidney function, focusing on diabetic
be associated with further improvement as fewer women retinopathy is important in these women since progression
with type 1 diabetes and microalbuminuria developed pre- to sight-threatening diabetic retinopathy is prevalent during
eclampsia or delivered preterm in the following audit.2 pregnancy.62,63 Laser treatment should be performed during
Furthermore, in women with diabetic nephropathy, pregnancy, if indicated.63
early-onset and intensive antihypertensive treatment most
likely also reduces the severity of preeclampsia and preterm
delivery.1,2,4 It is often necessary to use a combination of dif- Summary of treatment recommendations
ferent antihypertensive agents to control blood pressure and Strict glycemic control, low-dose aspirin, folic acid supple-
urinary albumin excretion.4 Methyldopa, labetalol, nifedip- mentation, and intensive antihypertensive treatment with
ine, and diltiazem are often used and are apparently safe in pregnancy-friendly drugs are of importance during preg-
pregnancy.4,60,67 In addition, diuretics, both thiazides and nancy in women with diabetic nephropathy or microalbu-
loop diuretics, may be used with caution during pregnancy, minuria. The goal for antihypertensive treatment includes
and we often find it necessary to continue with unchanged both lowering of blood pressure and urinary albumin excre-
dose of diuretics if the women are already treated with this tion and is stricter compared to pregnant women without
class of drug prior to pregnancy due to diabetic nephropa- diabetes. Close obstetric surveillance and screening for dia-
thy,4,60 while we normally do not use diuretics in other preg- betic retinopathy is important to improve pregnancy out-
nant women. Many women with diabetic nephropathy can comes in these high-risk pregnancies (Table 55.2).
be controlled with one or two antihypertensive agents, but
as many as four different antihypertensive classes of drugs,
including diuretics, are used for selected pregnant women Concluding remarks
at our center in order to stabilize the blood pressure.2,4,60 Pregnancy outcome in women with diabetic nephropathy or
Although antihypertensive agents have been reported to be microalbuminuria has improved considerably over the last
associated with intrauterine growth restriction,68 this seems decade with a take-home-baby rate of approximately 95%.
not to be the case in women with diabetes.1,2 Most information in the literature comes from women with
Carr et al.6 described a cohort of 43 pregnant women with type 1 diabetes and diabetic nephropathy or microalbu-
diabetic nephropathy where suboptimal control of hyperten- minuria, but the same numbers are probably also valid for
sion in early pregnancy was associated with increased risk of women with type 2 diabetes. Careful counseling of women
preterm delivery compared to women with well-controlled with diabetic nephropathy or microalbuminuria prior to
blood pressure on medical treatment (38% vs. 5%). pregnancy with estimation of the risk for the mother and
Although prospective, randomized trials are not avail- fetus is important. Pregnancy does not result in worsening
able, studies from our center1,2,60 in combination with studies of kidney function in women with diabetic nephropathy
by Kimmerle9 and Carr6 strongly suggest that women with and normal serum creatinine, but pregnancy complications
diabetes and diabetic nephropathy or microalbuminuria are common.
receiving early and intensive antihypertensive treatment Strict glycemic control before and during pregnancy,
have a better pregnancy outcome compared to women initi- low-dose aspirin from 10 to 12 gestational weeks onward,
ating antihypertensive treatment in late pregnancy. folic acid supplementation in early pregnancy, and inten-
The mechanism of the effect of early and intensive sive antihypertensive treatment are important factors for
antihypertensive treatment in pregnant women with dia- optimizing pregnancy outcome. Methyldopa, labetalol,
betic nephropathy or microalbuminuria is not known. nifedipine, and diltiazem are regarded safe in pregnancy,
Antihypertensive treatment may stabilize the urinary albu- whereas ACE inhibitors, angiotensin II receptor antagonists,
min excretion and the universal leakage of albumin from or cholesterol-lowering agents should not be used during
the microcirculation and thus improve the endothelial func-
tion. Thereby, the antihypertensive treatment reduces not
Table 55.2 Treatment of women with diabetic
only blood pressure and urinary albumin excretion but also
nephropathy during pregnancy
the other clinical manifestations of preeclampsia associated
with maternal endothelial dysfunction. The beneficial effect
• Intensive glycemic control
of antihypertensive treatment of diabetic nephropathy or
microalbuminuria in normotensive patients with type 1 dia- • Low-dose aspirin from 10–12 weeks until 36–37 weeks
betes outside pregnancy is well documented.14 • Supplementation with folic acid during the first 12 weeks
• Tight target for antihypertensive treatment, i.e., BP
<135/85 and urinary albumin excretion <300 mg/24 h
Obstetric surveillance • Use of pregnancy-friendly antihypertensive agents
• Review the medication list for drugs contraindicated in
In late pregnancy, close obstetrical surveillance including pregnancy, i.e., cholesterol-lowering agents
frequent ultrasound examinations of fetal growth and non-
• Tight obstetric surveillance
stress testing are important to diagnose complications and
• Screen for sight-threatening diabetic retinopathy
plan the time and mode of delivery.45
References 471
pregnancy. Case series and pathophysiological studies sup- before and during pregnancy is mandatory, and laser treat-
port a stringent goal for blood pressure and urinary albumin ment should be performed if indicated. Further research on
excretion in pregnant women with diabetic nephropathy the benefit and risk of intensive antihypertensive treatment
or microalbuminuria. Screening for diabetic retinopathy in this population is needed.
REFERENCES
1. Damm JA, Asbjornsdottir B, Callesen NF, Mathiesen JM, 18. Mathiesen ER, Hommel E, Giese J, Parving HH. Efficacy of cap-
Ringholm L, Pedersen BW, Mathiesen ER. Diabetic nephropathy topril in postponing nephropathy in normotensive insulin depen-
and microalbuminuria in pregnant women with type 1 and type dent diabetic patients with microalbuminuria. Br Med J 1991;
2 diabetes: Prevalence, antihypertensive strategy, and pregnancy 303: 81–87.
outcome. Diabetes Care 2013; 36: 3489–3494. 19. Mathiesen ER, Ronn B, Storm B, Foght H, Deckert T. The natu-
2. Nielsen LR, Damm P, Mathiesen ER. Improved pregnancy out- ral course of microalbuminuria in insulin-dependent diabetes:
come in type 1 diabetic women with microalbuminuria or dia- A 10-year prospective study. Diabet Med 1995; 12: 482–487.
betic nephropathy: Effect of intensified antihypertensive therapy? 20. Estacio RO, Coll JR, Tran ZV, Schrier RW. Effect of intensive blood
Diabetes Care 2009; 32: 38–44. pressure control with valsartan on urinary albumin excretion in
3. Kitzmiller M and Montoro MN. Diabetic nephropathy and preg- normotensive patients with type 2 diabetes. Am J Hypertens
nancy. In: Kitzmiller JL, Jovanovic L, Brown F, Coustan D, Reader 2006; 19: 1241–1248.
DM ed., Managing Preexisting Diabetes and Pregnancy, American 21. Standards of medical care in diabetes—2014. Diabetes Care
Diabetes Association: Alexandria, VA, 2008, pp. 374–386. 2014; 37(Suppl 1): S14–S80.
4. Mathiesen ER, Ringholm L, Feldt-Rasmussen B, Clausen P, Damm 22. Biesenbach G, Stoger H, Zazgornik J. Influence of pregnancy on
P. Obstetric nephrology: Pregnancy in women with diabetic progression of diabetic nephropathy and subsequent requirement
nephropathy—The role of antihypertensive treatment. Clin J Am of renal replacement therapy in female type I diabetic patients
Soc Nephrol 2012; 7: 2081–2088. with impaired renal function. Nephrol Dial Transplant 1992; 7:
5. Bar J, Chen R, Schoenfeld A, Orvieto R, Yahav J, Ben-Rafael Z, 105–109.
Hod M. Pregnancy outcome in patients with insulin dependent 23. Clausen P, Ekbom P, Damm P, Feldt-Rasmussen U, Nielsen B,
diabetes mellitus and diabetic nephropathy treated with ACE Mathiesen ER, Feldt-Rasmussen B. Signs of maternal vascular
inhibitors before pregnancy. J Pediatr Endocrinol Metab 1999; dysfunction precede preeclampsia in women with type 1 diabe-
12: 659–665. tes. J Diabetes Complications 2007; 21: 288–293.
6. Carr DB, Koontz GL, Gardella C, Holing EV, Brateng DA, 24. Holmes VA, McCance DR. Could antioxidant supplementation
Brown ZA, Easterling TR. Diabetic nephropathy in pregnancy: prevent pre-eclampsia? Proc Nutr Soc 2005; 64: 491–501.
Suboptimal hypertensive control associated with preterm deliv- 25. McCance DR, Holmes VA, Maresh MJ, Patterson CC, Walker JD,
ery. Am J Hypertens 2006; 19: 513–519. Pearson DW, Young IS. Vitamins C and E for prevention of pre-
7. Dunne FP, Chowdhury TA, Hartland A, Smith T, Brydon PA, eclampsia in women with type 1 diabetes (DAPIT): A randomised
McConkey C, Nicholson HO. Pregnancy outcome in women placebo-controlled trial. Lancet 2010; 376: 259–266.
with insulin-dependent diabetes mellitus complicated by 26. Yu Y, Jenkins AJ, Nankervis AJ et al. Anti-angiogenic factors and
nephropathy. QJM 1999; 92: 451–454. pre-eclampsia in type 1 diabetic women. Diabetologia 2009; 52:
8. Gordon M, Landon MB, Samuels P, Hissrich S, Gabbe SG. 160–168.
Perinatal outcome and long-term follow-up associated with mod- 27. Itskovitz J, Rubattu S, Levron J, Sealey JE. Highest concentrations
ern management of diabetic nephropathy. Obstet Gynecol 1996; of prorenin and human chorionic gonadotropin in gestational
87: 401–409. sacs during early human pregnancy. J Clin Endocrinol Metab
9. Kimmerle R, Zass RP, Cupisti S, Somville T, Bender R, Pawlowski 1992; 75: 906–910.
B, Berger M. Pregnancies in women with diabetic nephropathy: 28. Nielsen AH, Schauser KH, Poulsen K. Current topic: The utero-
Long-term outcome for mother and child. Diabetologia 1995; 38: placental renin-angiotensin system. Placenta 2000; 21: 468–477.
227–235. 29. Al Kadi H, Nasrat H, Broughton PF. A prospective, longitu-
10. Landon MB. Diabetic nephropathy and pregnancy. Clin Obstet dinal study of the renin-angiotensin system, prostacyclin and
Gynecol 2007; 50: 998–1006. thromboxane in the first trimester of normal human pregnancy:
11. Reece EA, Coustan DR, Hayslett JP, Holford T, Coulehan J, Association with birthweight. Hum Reprod 2005; 20: 3157–3162.
O’Connor TZ, Hobbins JC. Diabetic nephropathy: Pregnancy 30. Sealey JE, Itskovitz-Eldor J, Rubattu S, James GD, August P,
performance and fetomaternal outcome. Am J Obstet Gynecol Thaler I, Levron J, Laragh JH. Estradiol- and progesterone-related
1988; 159: 56–66. increases in the renin-aldosterone system: Studies during ovarian
12. Reece EA, Leguizamon G, Homko C. Stringent controls in dia- stimulation and early pregnancy. J Clin Endocrinol Metab 1994;
betic nephropathy associated with optimization of pregnancy 79: 258–264.
outcomes. J Matern Fetal Med 1998; 7: 213–216. 31. Wang A, Rana S, Karumanchi SA. Preeclampsia: The role of
13. Sorensen VR, Mathiesen ER, Heaf J, Feldt-Rasmussen B. Improved angiogenic factors in its pathogenesis. Physiology (Bethesda)
survival rate in patients with diabetes and end-stage renal disease 2009; 24: 147–158.
in Denmark. Diabetologia 2007; 50: 922–929. 32. Ringholm L, Pedersen-Bjergaard U, Thorsteinsson B, Boomsma
14. Parving HH, Smidt UM, Hommel E, Mathiesen ER, Rossing P, F, Damm P, Mathiesen ER. A high concentration of prorenin
Nielsen F, Gall MA. Effective antihypertensive treatment post- in early pregnancy is associated with development of pre-
pones renal insufficiency in diabetic nephropathy. Am J Kidney eclampsia in women with type 1 diabetes. Diabetologia 2011;
Dis 1993; 22: 188–195. 54: 1615–1619.
15. Brenner BM, Cooper ME, de ZD, Keane WF, Mitch WE, Parving HH, 33. McDonagh TA, Robb SD, Murdoch DR, Morton JJ, Ford I,
Remuzzi G, Snapinn SM, Zhang Z, Shahinfar S. Effects of losartan Morrison CE, Tunstall-Pedoe H, McMurray JJ, Dargie HJ.
on renal and cardiovascular outcomes in patients with type 2 diabe- Biochemical detection of left-ventricular systolic dysfunction.
tes and nephropathy. N Engl J Med 2001; 345: 861–869. Lancet 1998; 351: 9–13.
16. Lewis EJ, Hunsicker LG, Bain RP, Rohde RD. The effect of angio- 34. Tulevski II, Groenink M, van der Wall EE, van Veldhuisen DJ,
tensin-converting-enzyme inhibition on diabetic nephropa- Boomsma F, Stoker J, Hirsch A, Lemkes JS, Mulder BJ. Increased
thy. The Collaborative Study Group. N Engl J Med 1993; 329: brain and atrial natriuretic peptides in patients with chronic right
1456–1462. ventricular pressure overload: Correlation between plasma neuro-
17. Lewis EJ, Hunsicker LG, Clarke WR, Berl T, Pohl MA, Lewis hormones and right ventricular dysfunction. Heart 2001; 86: 27–30.
JB, Ritz E, Atkins RC, Rohde R, Raz I. Renoprotective effect of 35. Borghi C, Cicero AF, Degli ED, Immordino V, Bacchelli S, Rizzo
the angiotensin-receptor antagonist irbesartan in patients with N, Santi F, Ambrosioni E. Hemodynamic and neurohumoral pro-
nephropathy due to type 2 diabetes. N Engl J Med 2001; 345: file in patients with different types of hypertension in pregnancy.
851–860. Intern Emerg Med 2011; 6: 227–234.
36. Resnik JL, Hong C, Resnik R, Kazanegra R, Beede J, Bhalla V, 52. Duley L, Henderson-Smart D, Knight M, King J. Antiplatelet drugs
Maisel A. Evaluation of B-type natriuretic peptide (BNP) levels for prevention of pre-eclampsia and its consequences: Systematic
in normal and preeclamptic women. Am J Obstet Gynecol 2005; review. Br Med J 2001; 322: 329–333.
193: 450–454. 53. Wilson RD, Davies G, Desilets V, Reid GJ, Summers A, Wyatt P,
37. Ringholm L, Pedersen-Bjergaard U, Thorsteinsson B, Boomsma F, Young D. The use of folic acid for the prevention of neural tube
Damm P, Mathiesen ER. Atrial Natriuretic Peptide (ANP) in early defects and other congenital anomalies. J Obstet Gynaecol Can
pregnancy is associated with development of preeclampsia in 2003; 25: 959–973.
type 1 diabetes. Diabetes Res Clin Pract 2011; 93: e106–e109. 54. Capel I, Corcoy R. What dose of folic acid should be used for
38. Ekbom P, Damm P, Andersson AM, Skakkebaek NE, Feldt- pregnant diabetic women? Diabetes Care 2007; 30: e63.
Rasmussen U, Mathiesen ER. Serum levels of activin A and 55. Cooper WO, Hernandez-Diaz S, Arbogast PG, Dudley JA, Dyer
inhibin A are not related to the increased susceptibility to pre- S, Gideon PS, Hall K, Ray WA. Major congenital malformations
eclampsia in type I diabetic pregnancies. Acta Obstet Gynecol after first-trimester exposure to ACE inhibitors. N Engl J Med
Scand 2006; 85: 143–147. 2006; 354: 2443–2451.
39. Bell R, Glinianaia SV, Tennant PW, Bilous RW, Rankin J. Peri- 56. Porta M, Hainer JW, Jansson SO et al. Exposure to candesar-
conception hyperglycaemia and nephropathy are associated tan during the first trimester of pregnancy in type 1 diabetes:
with risk of congenital anomaly in women with pre-existing dia- Experience from the placebo-controlled DIabetic REtinopathy
betes: A population-based cohort study. Diabetologia 2014 Feb; Candesartan Trials. Diabetologia 2011; 54: 1298–1303.
57(2): 285–294. 57. Tennant PW, Glinianaia SV, Bilous RW, Rankin J, Bell R. Should
40. Khoury JC, Miodovnik M, LeMasters G, Sibai B. Pregnancy out- women with diabetic nephropathy considering pregnancy con-
come and progression of diabetic nephropathy. What’s next? tinue ACE inhibitor or angiotensin II receptor blocker therapy
J Matern Fetal Neonatal Med 2002; 11: 238–244. until pregnancy is confirmed? Reply to Lewis G and Maxwell AP
41. Ekbom P, Damm P, Feldt-Rasmussen B, Feldt-Rasmussen U, [letter]. Diabetologia 2014; 57: 1084–1085.
Molvig J, Mathiesen ER. Pregnancy outcome in type 1 dia- 58. Tabacova S, Little R, Tsong Y, Vega A, Kimmel CA. Adverse
betic women with microalbuminuria. Diabetes Care 2001; 24: pregnancy outcomes associated with maternal enalapril anti-
1739–1744. hypertensive treatment. Pharmacoepidemiol Drug Saf 2003; 12:
42. Rossing K, Jacobsen P, Hommel E, Mathiesen E, Svenningsen A, 633–646.
Rossing P, Parving HH. Pregnancy and progression of diabetic 59. Sibai BM. Chronic hypertension in pregnancy. Obstet Gynecol
nephropathy. Diabetologia 2002; 45: 36–41. 2002; 100: 369–377.
43. Purdy LP, Hantsch CE, Molitch ME, Metzger BE, Phelps RL, 60. Nielsen LR, Muller C, Damm P, Mathiesen ER. Reduced preva-
Dooley SL, Hou SH. Effect of pregnancy on renal function in lence of early preterm delivery in women with Type 1 diabetes
patients with moderate-to-severe diabetic renal insufficiency. and microalbuminuria—Possible effect of early antihyper-
Diabetes Care 1996; 19: 1067–1074. tensive treatment during pregnancy. Diabet Med 2006; 23:
44. Sibai BM. Diabetic nephropathy in pregnancy. In: McCance 426–431.
D, Maresh M, Sacks D, eds., A Practical Manual of Diabetes 61. Edison RJ, Muenke M. Mechanistic and epidemiologic consid-
in Pregnancy, Blackwell Publishing: Oxford, U.K., 2008, erations in the evaluation of adverse birth outcomes following
pp. 153–156. gestational exposure to statins. Am J Med Genet A 2004; 131:
45. Ringholm L, Mathiesen ER, Kelstrup L, Damm P. Managing type 1 287–298.
diabetes mellitus in pregnancy—From planning to breastfeeding. 62. Lauszus FF, Gron PL, Klebe JG. Pregnancies complicated by dia-
Nat Rev Endocrinol 2012; 8: 659–667. betic proliferative retinopathy. Acta Obstet Gynecol Scand 1998;
46. Boulot P, Chabbert-Buffet N, d’Ercole C et al. French multicentric 77: 814–818.
survey of outcome of pregnancy in women with pregestational 63. Vestgaard M, Ringholm L, Laugesen CS, Rasmussen KL, Damm P,
diabetes. Diabetes Care 2003; 26: 2990–2993. Mathiesen ER. Pregnancy-induced sight-threatening diabetic reti-
47. Evers IM, de Valk HW, Visser GH. Risk of complications of preg- nopathy in women with Type 1 diabetes. Diabet Med 2010; 27:
nancy in women with type 1 diabetes: Nationwide prospective 431–435.
study in the Netherlands. Br Med J 2004; 328: 915. 64. Hiilesmaa V, Suhonen L, Teramo K. Glycaemic control is associ-
48. Ekbom P, Damm P, Nogaard K, Clausen P, Feldt-Rasmussen U, ated with pre-eclampsia but not with pregnancy-induced hyper-
Feldt-Rasmussen B, Nielsen LH, Molsted-Pedersen L, Mathiesen tension in women with type I diabetes mellitus. Diabetologia
ER. Urinary albumin excretion and 24-hour blood pressure as 2000; 43: 1534–1539.
predictors of pre-eclampsia in type I diabetes. Diabetologia 65. Management of diabetes from preconception to the postna-
2000; 43: 927–931. tal period: Summary of NICE guidance. Br Med J 2008; 336:
49. Hanson U, Persson B. Epidemiology of pregnancy-induced 714–717.
hypertension and preeclampsia in type 1 (insulin-dependent) dia- 66. Catalano PM, Conway, DL. Management of pregnancy complica-
betic pregnancies in Sweden. Acta Obstet Gynecol Scand 1998; tions. In: Kitzmiller JL, Jovanovic L, Brown F, et al. eds., Managing
77: 620–624. Preexisting Diabetes and Pregnancy, American Diabetes
50. Holmes VA, Young IS, Patterson CC, Pearson DW, Walker Association: Alexandria, VA, 2008, pp. 578–584.
JD, Maresh MJ, McCance DR. Optimal glycemic control, pre- 67. Khandelwal M, Kumanova M, Gaughan JP, Reece EA. Role of dil-
eclampsia, and gestational hypertension in women with type tiazem in pregnant women with chronic renal disease. J Matern
1 diabetes in the diabetes and pre-eclampsia intervention trial. Fetal Neonatal Med 2002; 12: 408–412.
Diabetes Care 2011; 34: 1683–1688. 68. Magee LA, Ornstein MP, von DP. Fortnightly review: Management
51. Nielsen LR, Pedersen-Bjergaard U, Thorsteinsson B, Johansen of hypertension in pregnancy. Br Med J 1999; 318: 1332–1336.
M, Damm P, Mathiesen ER. Hypoglycemia in pregnant women
with type 1 diabetes: Predictors and role of metabolic control.
Diabetes Care 2008; 31: 9–14.
56 Diabetic ketoacidosis
Annunziata Lapolla and Maria Grazia Dalfrà
Diabetic ketoacidosis (DKA) is a serious complication of dia- metabolized as an alternative energy source in the process of
betes resulting often in a medical emergency. Nowadays, the ketogenesis.10,12 In particular, the excessive beta-oxidation of
occurrence of DKA is rare in pregnant women with diabe- fatty acids prompts the formation of large quantities of acetyl
tes, occurring mainly in patients with type 1 diabetes, but CoA, which is then converted by the liver into ketone bod-
also in type 2 or, more rarely, in women with gestational ies (3-beta-hydroxy-butyrate and acetoacetate). Acetoacetate
diabetes mellitus (GDM). Maternal and fetal complications undergoes decarboxylation and conversion into acetone. The
determined by DKA can be life threatening, so prompt rec- abundant ketone bodies, and particularly 3-beta-hydroxy-
ognition and treatment of this condition is mandatory to butyrate and lactic acid (which is also used in gluconeogen-
avoid these complications. esis), are the main contributors to the metabolic acidosis.12
It is important to bear in mind that the metabolic changes
characteristics of DKA trigger a chain of events that become
Incidence self-perpetuating in a vicious cycle. Hyperglycemia gives rise
to an osmotic gradient that results in an excessive diuresis,
The reported incidence of diabetes in pregnancy ranges leading to severe dehydration and hypovolemia.10–12 This
from 6% to 7% with 90% of cases constituted by women further exacerbates the hyperglycemia and acidosis due to
affected by GDM.1,2 In this context, DKA has a reported the activation of other counterregulatory hormones (corti-
incidence between 0.5% and 3% of all diabetic pregnan- sol growth hormone), and the osmotic diuresis causes low
cies.3–9 Previously considered typical of type 1 diabetes is sodium levels. The shortage of insulin also leads to protein
now reported also to occur in patients with type 2 diabe- breakdown and interferes with cellular potassium uptake
tes and GDM. Furthermore, the increase of the utilization so that normal or high serum potassium levels can coincide
of assisted reproductive technology resulting often in mul- with low total body potassium.10–12
tifetal gestation can be a risk factor for the development of
DKA; in fact, this gestation can be associated with preterm
labor determining the necessity to utilize corticosteroid for Ketoacidosis and pregnancy
fetal lung maturation, a therapy that in diabetes disease can
determine hyperglycemia and subsequently DKA. Pregnancy is characterized by insulin resistance, accelerated
starvation, and respiratory alkalosis, especially in the second
and third trimesters.13 It has been demonstrated that sensi-
Pathophysiology tivity to insulin decreases in pregnancy, reaching its nadir in
the third trimester and rapidly returning to prepregnancy
DKA is a complex metabolic disorder characterized by hyper- levels after delivery.13–15 Although the specific mechanisms
glycemia, acidosis, and ketonemia.10 It frequently occurs as a behind the gradual onset of insulin resistance, often with a
consequence of absolute or relative insulin deficiency accom- concomitant increase in insulin secretion, during pregnancy
panied by an increased secretion of the counterregulatory have not been fully clarified, an important contribution
hormones (cortisol, glucagon, epinephrine).11 These hormone comes from the endocrine changes characteristic of preg-
changes lead to a more marked hepatic gluconeogenesis and nancy, including increased levels of estrogens, progesterone,
glycogenolysis, prompting severe hyperglycemia. Glycogen human placental lactogen (HPL), cortisol, and TNFα.16,17
stores are depleted, and gluconeogenesis is enhanced, partly The gradual decline in insulin sensitivity occurring in
because of high levels of glucose precursors (and glycerol normal pregnancy was demonstrated in a clamp study by
in particular), as a result of the increased lipolysis and of Catalano et al.,15 who found insulin sensitivity 56% lower in
amino acids from muscle breakdown. Insulin resistance is normal pregnant women during their third trimester. This is
responsible for the increased lipolysis, which in turn reduces considered a physiological mechanism to facilitate the supply
the adipocytes’ capacity to store free fatty acids, which are of glucose to the fetus, and it coincides with a parallel gradual
473
increase in insulin secretion to maintain a normal glucose

Table 56.1 Factors triggering diabetic ketoacidosis
tolerance.13,14 The gradient of the glucose concentration from
mother to fetus is raised by means of higher postprandial glu-
Starvation
cose levels in pregnant women. During fasting, their insulin
Protracted vomiting
resistance promotes the release of fatty acids that the mother
can use as an alternative fuel, thereby keeping the glucose for Infections
the fetus, and this explains the states of accelerated starvation Insulin pump failure
and facilitated anabolism. High levels of human chorionic Undiagnosed diabetes
gonadotropin are also often associated with nausea and vom- Poor control of diabetes and/or poor compliance
iting, which may give rise to a state of starvation, dehydration, Use of beta-sympathomimetic drugs for tocolysis
and acidosis, with the subsequent activation of stress-related Use of steroid drugs for fetal lung maturation
hormones.18 Then progesterone reduces gastrointestinal Diabetic gastroparesis
motility and increases carbohydrate absorption, thus raising
plasma glucose levels. In pregnancy complicated by type 1
diabetes, the total absence of exogenous insulin hinders the Factors triggering DKA
achievement of an appropriate balance between the acceler-
ated starvation and facilitated anabolism characteristic of The most common factors involved in the onset of DKA in
pregnancy. This explains why it is difficult to keep glucose pregnant women are listed in Table 56.1.
levels within the normal range, and, when hyperglycemia is In a review conducted by Rodgers and Rodgers,21 the use
not treated promptly, the pregnancy-induced lipolysis makes of beta-sympathomimetic agents for tocolysis accounted
patients more susceptible to ketoacidosis. In pregnant women for 57% of cases, while patient compliance issues and phy-
with type 2 diabetes, the pregnancy-related drop in insulin sicians’ patient management errors were responsible for
sensitivity overlaps the patient’s preexisting insulin resis- 24% and 16% of cases, respectively. The authors reported
tance, meaning that these patients need insulin therapy early that diabetic gastroparesis, when associated with nausea,
in their pregnancy to avoid ketoacidosis. We need to bear in vomiting, and dehydration, contributes to triggering keto-
mind that maternal ketone body levels are elevated by 33% acidosis. Infections and a history of missing doses of insulin
higher during the third trimester of pregnancy than in the therapy were the most important causes of ketoacidosis in
postpartum period.12,19 Finally, a greater alveolar ventilation a retrospective analysis conducted by Schneider, account-
in pregnant women gives rise to a state of respiratory alka- ing, respectively, for 27% and 18% of cases.22 Other causes
losis that is overcome by means of an increased renal excre- include insulin pump failure (in 13% of the cases reported by
tion of bicarbonate, which leads to a lower buffering capacity. Chen23), chronic corticosteroid therapy, undiagnosed diabe-
These changes can trigger the onset of ketoacidosis at lower tes, and smoking.24,25
glycemic levels than those seen in diabetic women who are
not pregnant20 (Figure 56.1).
Euglycemic ketoacidosis
Circulating insulin Euglycemic ketoacidosis is a severe form of ketoacidosis with
Insulin resistance serum bicarbonate levels of 10 mEq/L or less, in the absence
Counterregulatory hormones of hyperglycemia (blood glucose levels below 10 mmol/L).
First described by Munro et al. in 1973,26 it has since been
reported by several clinicians26–30 in women with both pre-
Muscle glucose Hepatic glucose Ketogenesis Lipolysis gestational (type 1 and type 2) diabetes and GDM.26–30 True
utilization production euglycemic ketoacidosis is rare, with only 0.8%–1.1% of all
Gluconeogenesis episodes meeting the aforementioned plasma bicarbonate
Glycogenolysis concentration criterion.5 The fetal and placental demand for
glucose (approximately 150 g/day), the increased renal excre-
Hyperglycemia tion of glucose, the mother’s higher glucose consumption,
and the physiological hemodilution occurring in pregnancy
all contribute to explaining why DKA can occur rapidly and
Glucosuria Acidosis at lower glucose levels in women who are pregnant than in
those who are not.
Loss water Emesis2 and decreased
and electrolytes oral intake
Clinical presentation
The symptoms of diabetic DKA in pregnancy are no different
Dehydration
from those seen in women who are not pregnant, except that
Figure 56.1 Pathophysiology of diabetic ketoacidosis in they tend to develop more rapidly in pregnancy (Table 56.2).
pregnancy. Patients usually present with a generalized malaise, nausea,
Maternal and perinatal complications 475
who are not pregnant (euglycemic ketoacidosis)18–23; fur-

Table 56.2 Clinical presentation of diabetic
thermore, ketoacidosis in diabetic pregnant women devel-
ketoacidosis
ops more rapidly than in nonpregnant women, and this can
result in a delayed recognition and treatment.27 The main
Nausea/vomiting
ketones to be increased in ketoacidosis is beta-OHB, and nei-
Polydipsia
ther acetone nor beta-OHB react as strongly as acetoacetate
Polyuria with nitroprusside (used to test for the presence of ketosis
Weakness in urine), so the severity of a patient’s hyperketonemia may
Weight loss be underestimated if only urine is tested, and that is why it
Abdominal pain is essential to test plasma ketone bodies too. After insulin
Hyperventilation/fruity breath infusion, beta-OHB is rapidly converted into acetoacetate so
Dry mucous membranes the ketone body levels in urine increase—even though the
Hypotension patient’s ketoacidosis is regressing. Patients may have normal
Mental status changes or high potassium levels, but it is important to consider that
Coma their total body potassium is generally low, and they become
dehydrated and hypokalemic. Creatinine levels and blood
urea nitrogen may be high as a result of renal dysfunction.10
vomiting, polyuria, polydipsia, weakness, tachypnea, and
signs of dehydration (flaccid skin, dry mucous mem-
branes, tachycardia, hypotension, change in mental status).
Abdominal pain due to a reduced peripheral perfusion may
Maternal and perinatal complications
be severe and accompanied by uterine contractions. When DKA is fortunately uncommon in pregnancy, but it can
infection is the precipitating factor, patients may have fever result in such serious maternal complications as acute
or, paradoxically, hypothermia in some cases (due to the renal failure, respiratory distress syndrome, myocardial
vasodilatory effects of the hydrogen ions in excess when ischemia, cerebral edema, and death.7,8 The rate of occur-
severe ketoacidosis develops).31 In the advanced stages of the rence of these complications depends on the mother’s
disease, patients may have Kussmaul respiration, a breath condition at the time of onset of DKA and on how the
with a classic fruity smell, lethargy, and signs of central ner- condition is managed. Recent data indicate a DKA-related
vous system involvement such as disorientation, obtunda- maternal mortality rate of less than 1% and a fetal mortal-
tion, and coma due to cerebral edema.32 ity rate of 9%–36%.6,22,24
The rates of fetal complications such as fetal loss, preterm
delivery, hypoxia, and acidosis are still high among the new-
Laboratory findings born of diabetic mothers with DKA.
The laboratory tests performed in cases of diabetic DKA Animal studies have shown that maternal ketoacidosis
are shown in Table 56.3. The analyses must include testing and hyperglycemia can lead to lactic acidosis and hypoxia
serum glucose levels, serum electrolytes, osmolarity, blood in the fetus.33,34 This acidosis can occur via several pathways:
urea nitrogen, creatinine, arterial blood gas and serum fetal hyperglycemia results in fetal osmotic diuresis and
bicarbonate levels, anion gap, and serum ketones, a com- hypovolemia, contributing to lactic acidosis; fetal hyper-
plete blood cell count, liver function tests, and urine analy- insulinemia increases the fetal oxygen demand; and the
sis. Hyperglycemia (plasma glucose levels higher than 300 mother’s hemoglobin has a greater affinity for oxygen due
mg/dL), an increased anion gap, an arterial pH lower than to decreased 2,3-DPG levels, and this lowers the amount of
7.3, ketosis (positivity for serum and urine ketosis), and oxygen available to the fetus.35
decreased serum bicarbonate levels are characteristic of High keto acid levels in pregnancy have been associated
DKA. It is worth emphasizing, however, that plasma glucose with a lower IQ score as well as a worse mental development
levels may be lower in pregnancy than in diabetic women score during the second year of life.36 An association has also
been reported between ketonuria measured during prenatal
visits and adverse neurobehavioral outcomes even for non-
Table 56.3 Laboratory findings in diabetic diabetic pregnant women.37
ketoacidosis The fetal brain is particularly susceptible to higher than
normal levels of β-hydroxybutyrate and lactate concentra-
Hyperglycemia >300 mg/dL tions, which limit its glucose uptake. 38 These substances
16.7 mmol/L are known to accumulate in the basal ganglia of children
Arterial pH acidosis <7.3 during episodes of DKA. 39,40 The acidotic environment that
Serum HC03 <15 mEq/L develops has been associated with a poor myelination and
Serum acetone Positive a weak cortical connectivity as well as with aberrations
Anion gap >12 mEq/L in hippocampal neurons—findings linked to expressive
Osmolality >280 mOsm/kg language impairments in children with autistic spectrum
disorders.41
DKA management can affect insulin absorption.8 Initial insulin boluses should
contain approximately 10–15 U of regular insulin (0.1 U/kg),
DKA during pregnancy is an obstetric and medical emer- and the follow-on continuous insulin infusion could be at
gency that warrants intensive treatment at a specialized care 0.1 U/kg/h, testing glucose levels hourly. If glucose levels
unit. The principles of DKA management in pregnancy are drop too quickly, then there may be excessive displacements
the same as for patients who are not pregnant. Treatment for of water inside cells due to the rapid drop in serum osmo-
DKA includes fluid replacement, insulin therapy, correcting larity. Serum glucose levels should ideally drop by approxi-
acidosis and abnormal electrolytes, and treating the under- mately 50–75 mg/dL/h. Continuous insulin infusions should
lying disease, associated with intensive monitoring of mater- be given until metabolic alterations have been corrected and
nal and fetal conditions (Table 56.4). patients are able to resume their regular diet and switch to
their regular subcutaneous insulin treatment regimen.
Fluid replacement
Patients with DKA may have a fluid deficit amounting to a Potassium
mean 100 mL/kg of body weight42 and ranging from 6 to 10 L,42
Potassium depletion may be substantial, and the correction
with a total fluid loss of 100–150 mL/kg actual body weight.43
of acidosis, the administration of insulin, and fluid replace-
Initial intravenous replacement solutions should consist
ments can shift the ion from the extracellular to the intracel-
of isotonic saline (0.9% NaCl) solution at 1000 mL/h for at
lular medium. In DKA, the total potassium deficiency is about
least 2 hours. The use of hypotonic saline solution (0.45%
5–10 mEq/L. To prevent fatal arrhythmias, it is important to
NaCl) in the early phase of fluid replacement can raise the
keep potassium levels between 4 and 5 mEq/L. Potassium
risk of cerebral edema.44
levels should be monitored every 2–4 hours, and replace-
After administering an initial 2 L of isotonic saline, in the
ment potassium should be administered slowly (≤20 mEq/h)
presence of hypernatremia, the solution may be changed to a
to avoid hyperkalemia and the risk of cardiac arrhythmia.
hypotonic saline (0.45% NaCl) delivered at 250 mL/h, a rate
Potassium replacement can be achieved either by adding KCl
matching the electrolyte loss during osmotic dieresis, until
(40 mEq/L) to the replacement fluid administered, in an infu-
serum glucose levels are between 200 and 250 mg/dL. Once
sion administered at 140–250 mL/h, or by means of intermit-
glucose levels have dropped to lower than 250 mg/dL, the
tent intravenous boluses of 10 mEq/h every 4–6 hours.43
saline solution could be replaced with a 5% dextrose solution
to prevent an excessively rapid drop of glucose levels.
Acidosis correction
Insulin therapy Correcting acidosis with intravenous bicarbonate is rarely
Insulin treatment has to be initiated to correct the metabolic recommended in the protocols for DKA management.
disorder. It is preferable to start with intravenous boluses The administration of bicarbonate may be associated with
and follow up with continuous intravenous infusions rather profound alkalosis or worsening acidosis secondary to an
than to administer insulin via an intramuscular or subcu- increase in the partial pressure of carbon dioxide, which can
taneous route because a lower than normal tissue perfusion lead in turn to an impaired fetal oxygen transfer.8
For a pH <7.0 or serum bicarbonate levels <5 mEq/L, it is
advisable to administer 50 mEq of bicarbonate.
Table 56.4 Treatment for diabetic ketoacidosis in
pregnancy
Oxygenation
1. Fluids
a. Isotonic saline solution Maximizing tissue oxygenation in the mother and oxygen
b. Total replacement in the first 12 hours with 4–6 L supply to the uterus and placenta are essential aspects of
i. 1 L in the first hour. acute DKA management. Fetal hypoxia and fetal distress
ii. 500–1000 mL/hour for 2–4 hours. may occur during maternal acidemia. Oxygen administered
iii. 250 mL/hour up to 80% fluid replacement. through a face mask, hydration, and left lateral decubitus of
iv. In cases of hypernatremia, use 0.45% saline
solution. the mother can help to maximize uteroplacental perfusion
c. Glucose, starting with a 5% dextrose infusion, when and fetal oxygenation.
glucose levels drop < 250 mg/dL (14 mmol/L)
2. Electrolytes
a. Potassium
i. If normal or low, begin with 15–20 mEq/hour. Managing of precipitating factors
ii. If high, wait until it drops to within normal
range, then 20–30 mEq/L. Among precipitating factors for the onset of DKA infec-
b. Bicarbonate: 44 mEq if pH < 7.0 tions must be bear in mind. Possible sources of infection
3. Insulin include the urinary tract, the upper airways, the gallbladder,
a. An initial bolus of 10–15 U of regular insulin the ear, a tooth abscess, cellulitis, and wound infections after
(0.2–0.4 U/kg)
b. Intravenous infusion with 2–10 U/hour surgery. It is mandatory to identify and treat these issues
with appropriate antibiotics and/or surgery.
References 477
Fetal monitoring and timing of delivery Prevention

Maternal DKA results in fetal hypoxemia and acidosis. Prepregnancy counseling for diabetic women should include
Fetal heart rate monitoring is recommended from 24 providing information on their risk of DKA. Women with
weeks of gestation onward. During episodes of DKA, fetal pregestational diabetes should be thoroughly informed of
heart rate monitoring shows minimal variability, absent the importance of compliance with dietary restrictions,
accelerations, and repetitive variable and late decelera- exercise, and therapy (especially their insulin dosage and its
tions. Furthermore, the fetal biophysical profile may be means of administration, e.g., pumps). All women have to be
abnormal and the blood flow redistributed. The frequency aware that if their glucose levels are higher than 200 mg/dL,
and severity of these abnormalities depends on the sever- it becomes necessary for them to test their blood or urine
ity and duration of the DKA.45 ketone body levels and to contact their physician if the test
As concerns the timing of delivery, this will depend on result is positive. Infections, nausea and vomiting, diarrhea,
the gestational age of the fetus, the condition of mother and and polyuria are all conditions that carry a high risk of DKA
child, and their response to treatment. It is essential to stabi- and warrant hospitalization.
lize the mother’s condition.
REFERENCES
1. American College of Obstetricians and Gynecologists. 21. Rodgers BD, Rodgers DE. Clinical variables associated with dia-
Gestational diabetes mellitus. Practice Bullettin No 137. Obstet betic ketoacidosis during pregnancy. J Reprod Med 1991; 36:
Gynecol 2013; 122: 406–416. 449–451.
2. Wier LM, Witt E, Burges J, Elixhauder A. Hospitalization related 22. Schneider M, Umpierrez G, Ramsey R et al. Pregnancy com-
to diabetes in pregnancy, 2008. Statistical Brief 102, Healthcare plicated by diabetic ketoacidosis: Maternal and fetal outcomes.
Cost and Utilization Project (HCUP). Rockville, MD: Agency for Diabetes Care 2003; 26: 958–959.
Healthcare Policy and Research (US); 2006 Feb-2010 Dec. 23. Chen R, Ben-Haroush A, Weissmann-Brenner A et al. Level of
3. Montoro MN, Myers VP, Mestman JH et al. Outcome of preg- glycemic control and pregnancy outcome in type 1 diabetes: A
nancy in diabetic ketoacidosis. Am J Perinatol 1993; 10: 17–20. comparison between multiple daily insulin injections and contin-
4. Kilvert JA, Nicholson HO, Wrigtht AD. Ketoacidosis in diabetic uous subcutaneous insulin infusions. Am J Obstet Gynecol 2007;
pregnancy. Obstet Gynecol 2003; 102: 278–281. 197: 404–405.
5. Chauhan SP, Perry KG, Mc Laughlin BN et al. Diabetic ketoacido- 24. Hawtorne G. Maternal complications in diabetic pregnancy. Best
sis complicating pregnancy. J Perinatol 1996; 16: 173–175. Pract Res Clin Obstet Gynecol 2011; 25: 77–90.
6. Cullen MT, Reece EA, Homko CJ et al. The changing presentation 25. Bernstein IM, Catalano PM. Ketoacidosis in pregnancy associated
of diabetic ketoacidosis during pregnancy. Am J Perinatol 1996; with parenteral administration of terbutaline and betamethasone.
13: 449–451. A case report. J Reprod Med 1990; 35: 818–820.
7. Kamalakannan D, Baskar V, Barton DM et al. Diabetic ketoacido- 26. Munro JF, Campbell IW, McCuish AC et al. Euglycemic diabetic
sis in pregnancy. Postgrad Med J 2003; 79: 454–457. ketoacidosis. Br Med J 1973; 2: 578–580.
8. Parker JA, Conway DL. Diabetic ketoacidosis in pregnancy. 27. Madaan M, Aggarwal K, Sharma R et al. Diabetic ketoacidosis
Obstet Gynecol Clin North Am 2007; 34: 533–543. occurring with lower blood glucose levels in pregnancy: A report
9. Ramin KD. Diabetic ketoacidosis in pregnancy. Obstet Gynecol of two cases. J Reprod Med 2012; 57: 452–455.
Clin North Am 1999; 26: 481–488. 28. Darbhamulla S, Shah N, Bosio P. Euglycaemic ketoacidosis in a
10. Kitabchi AE, Umpierrez GE, Murphy MB et al. Hyperglycemic patients with gestational diabetes. Eur J Obstet Gynecol Reprod
crises in adult patients with diabetes: A consensus statement Biol 2012; 163: 118–119.
from the American Diabetes Association. Diabetes Care 2006; 29. Guo R, Yang L, Li L et al. Diabetic ketoacidosis in pregnancy
29: 2739–2748. tends to occur at lower blood glucose levels. Case-control study
11. Foster DW, McGarry JD. The metabolic derangements and treat- and a case report of euglycemic diabetic ketoacidosis in preg-
ment of diabetic ketoacidosis. N Engl J Med 1983; 309: 159–169. nancy. J Obstet Gynecol Res 2008; 34: 324–330.
12. Laffel L. Ketone bodies: A review of physiology, pathophysiology 30. Tarif N, Al Badr W. Euglycemic diabetic ketoacidosis in preg-
and application of monitoring to diabetes. Diabetes Metab Res nancy. Saudi J Kidney Dis Transpl 2007; 18: 590–593.
Rev 1999; 15: 412–426. 31. Johnston DG, Alberti KG. Diabetic emergencies. Clin Endocrinol
13. Lain KY, Catalano PM. Metabolic changes in pregnancy. Clin Metab 1980; 9: 437–460.
Obstet Gynaecol 2007; 50: 938–948. 32. Fulop M, Tannenbaum H, Dreyer N. Ketotic hyperosmolar coma.
14. Di Cianni G, Miccoli R, Volpe L et al. Intermediate metabolism in Lancet 1973; 2: 635–639.
normal pregnancy and in gestational diabetes. Diab Metab Res 33. Miodovnik M, Skillman CA, Hertzberg V et al. Effect of mater-
Rev 2003; 19: 259–270. nal hyperketonemia in hyperglycemic pregnant ewes and their
15. Catalano PM, Tzybir ED, Wolfe RR et al. Carbohydrate metabo- fetuses. Obstet Gynecol 1986; 154: 394–401.
lism during pregnancy in control subjects and women with ges- 34. Miodovnik M, Lavin JP, Harrington DJ et al. Cardiovascular and
tational diabetes. Am J Physiol 1999; 264: E60–E67. biochemical effects of infusion of beta hydroxybutyrate into the
16. Ryan EA, Enns L. Role of gestational hormones in the induction of fetal lamb. Obstet Gynecol 1982; 144: 594–600.
insulin resistance. J Clin Endocrinol Metab 1988; 67: 341–347. 35. Ditzel J, Standl E. The oxygen transport system of red blood cell
17. Kirwan JP, Hauguel-De Mouzon S, Lepercq J et al. TNF-alfa is during diabetic ketoacidosis and recovery. Diabetologia 1975; 11:
a predictor of insulin resistance in human pregnancy. Diabetes 255–260.
2002; 51: 2207–2213. 36. Rizzo T, Metzer BE, Burns WJ et al. Correlations between ante-
18. Tamay AG, Kuscu N. Hyperemesis gravidarum: Current aspect. partum maternal metabolism and child intelligence. N Engl J Med
J Obstet Gynaecol 2011; 31: 708–712. 1991; 325: 911–916.
19. Montelongo A, Lasuncion MA, Pallardo LF, Herrera E. Longitudinal 37. Churchill IA, Berendes HW, Nemore J. Neuropsychological defi-
study of plasma lipoprotein and hormones during pregnancy in cits in children of diabetic mothers. A report from the Collaborative
normal and diabetic women. Diabetes 1992; 41: 1651–1659. Study of Cerebral Palsy. Am J Obstet Gynecolol 1969; 105: 257–268.
20. Moore TR. Diabetes in pregnancy. In: Creasy RK, Resnik R, 38. Harding JE, Charlton VE. Effect of lactate and beta-hydroxybutyr-
Iams JD, eds., Maternal-Fetal Medicine, 5th ed., Saunders: ate infusions on brain metabolism in the fetal sheep. J Dev Physiol
Philadelphia, PA, 2004, pp. 1031–1032. 1990; 14: 139–146.
39. Wooton-Gorges SL, Buonocore MH, Kuppermann N et al. 42. Carroll MA, Yeomans ER. Diabetic ketoacidosis in pregnancy.
Detection of cerebral (beta)-hydroxy butyrate, acetoacetate and Crit Care Med 2005; 33(Suppl.): S347–S353.
lactate on proton MR spectroscopy in children with diabetic 43. Winkler C, Coleman F. Endocrine emergencies. In: Belfort M, Saade
ketoacidosis. Am J Neuroradiol 2005; 26: 1286–1291. G, Foley M, Phelan J, Didly G, eds., Critical Care Obstetrics, 5th
40. Wooton-Gorges SL, Buonocore MH, Kuppermann N et al. ed., Blackwell Publishing Ltd.: Oxford, U.K., 2010, pp. 425–437.
Cerebral proton magnetic resonance spectroscopy in chil- 44. Van der Meulen JA, Klip A, Grinstein S. Possible mechanisms for
dren with diabetic ketoacidosis. Am J Neuroradiol 2007; 28: cerebral oedema in diabetic ketoacidosis. Lancet 1987; 2: 306–308.
895–899. 45. Takahashi Y, Kawabata I, Shinohara A et al. Transient fetal blood
41. Bedalow A, Balasubramnyam A. Glucocorticoid-induced keto- flow redistribution induced by maternal diabetic ketoacidosis
acidosis in gestational diabetes: Sequelae of the acute treatment diagnosed by Doppler ultrasonography. Prenat Diagn 2000; 20:
of preterm labor. Diabetes Care 1997; 20: 922–924. 524–525.
57 Thyroid disease in pregnancy
Yoel Toledano and Gabriella Solomon
Pregnancy has a profound effect on the thyroid gland Thyroxine binding globulin
function. Numerous physiological changes occur during During pregnancy, estrogen increases serum TBG concen-
pregnancy including hormonal alterations and increased tration about twofold by elevated production and decreased
metabolic demands. Thyroid disorders may affect the progres- clearance. To maintain adequate free thyroid hormone con-
sion of pregnancy, the pregnant woman, and the developing centrations during this period, thyroid hormone production
fetus. On the other hand, pregnancy may also affect the course increases until a new equilibrium is reached and plateaus at
and the treatment of thyroid disorders. Hence, diagnosis and approximately 16 weeks of gestation.1
treatment of thyroid disorders in pregnancy parallel that of
nonpregnant women, but present unique problems.
Iodine and pregnancy
Iodine requirements are higher in pregnancy due to enhanced
Maternal thyroid physiology T4 production, increased renal iodine clearance, and fetal
iodine demands. Maternal iodine deficiency can result in
Thyroid hormone production and secretion is regulated by impaired maternal and fetal thyroid hormone synthesis. Severe
pituitary thyroid-stimulating hormone (TSH), with a nega- iodine deficiency in a given population has been associated
tive feedback loop, i.e., TSH production and secretion are with decreased fertility, spontaneous abortion, increased peri-
suppressed when thyroid hormone levels are increased. For natal and infant mortality, endemic goiter, and endemic men-
the vast majority of clinical situations, TSH is considered as tal retardation.4 Moreover, it can lead to endemic cretinism
the single best screening test for primary thyroid dysfunc- characterized by irreversible mental retardation, with either a
tion. Thus, a low serum TSH level, in the presence of a high neurological syndrome such as deaf-mutism and motor rigid-
serum–free T4 (FT4) and/or free T3 (FT3) levels, is charac- ity or predominant hypothyroidism or a combination.4
teristic of primary hyperthyroidism. Alternatively, a high The prophylactic action of treatment with iodine on the
serum TSH level in the presence of a reduced serum FT4 incidence of cretinism demonstrates the fundamental etio-
level is characteristic of primary hypothyroidism. The main logical role of iodine deficiency. Iodine deficiency is the
changes in thyroid physiology during pregnancy include leading cause of preventable mental retardation worldwide.5
thyrotropin (TSH) receptor stimulation by human chorionic Prolonged enhanced thyroid stimulation by even mild to
gonadotropin (hCG) and increased serum thyroxine bind- moderate iodine deficiency can lead to gradual decline in
ing globulin (TBG) concentrations and iodine requirements. total body iodine stores and goiter formation.6 Iodine intake
and methods to treat and avoid iodine deficiency differ
between countries. The World Health Organization recom-
Human chorionic gonadotropin mends 250 μg of iodine daily during pregnancy and lactation
As TSH and hCG are both glycoprotein hormones with and 150 μg/day for women of reproductive age.7 Accordingly,
considerable homology, the rising hCG level during early the American Thyroid Association (ATA) has recommended
pregnancy, with peak at 10–12 weeks, probably has a weak a 150 μg daily iodine supplementation for pregnant and
thyroid stimulatory activity. Consequently, the elevated breastfeeding women in North America in order to achieve
serum hCG concentration is accompanied by a reciprocal the total of 250 μg iodine ingestion per day.8
fall in TSH level. This effect is more pronounced at higher
hCG levels as is seen in twin pregnancies.1 Near the end
of the first trimester (peak circulating hCG), there is a Thyroid function in the fetus
small and transient increase in FT4 levels, usually within At 10th–12th week of gestation, fetal TSH appears in the
the normal range. Regarding physiological changes in pituitary gland. At the same time, the thyroid gland is
pregnancy, transient gestational hyperthyroidism and/or capable of concentrating iodine and synthesizing thyroid
hyperemesis gravidarum may reflect the extreme end of hormones. Hormone synthesis is little until midgestation.
the spectrum.2,3 Thereafter, fetal thyroid production and secretion increase
479
Mother neurodevelopment. Concordantly, there is evidence that

maternal thyroid hormones partially cross the placenta.13
TBG
Assessment of thyroid function in pregnancy

Total T4
hCG The following thyroid function tests are affected by the
Free T4 altered thyroid physiology in pregnancy.
Thyrotropin
Thyroid-stimulating hormone
There is mounting evidence that the reference range for
serum TSH levels is lower in pregnancy compared with the
Fetus nonpregnant state. The largest TSH level decrease is observed
TBG during the first trimester (Figure 57.2). Recently, the upper
normal limit during the first trimester has been accepted as
Total T4 2.5 mIU/L, as recommended in the guidelines of several
professional societies (Table 57.1).
Thyrotropin
Thyroid hormones
Free T4 Total T3 Serum FT4 level is usually within the normal range.
However, measuring FT4 by commonly used immunoas-
Free T3
say methods may be unreliable due to alterations in serum
10 20 30 40 proteins during pregnancy. Hence, awareness is needed. It is
Week of pregnancy
recommended to use method-specific and trimester-specific
laboratory range if available.18 As expected, total T4 concen-
Figure 57.1 Relative changes in maternal and fetal thyroid tration reference range is approximately 1.5 times the non-
function during pregnancy. hCG, human chorionic gonado- pregnant level by 16 weeks of gestation.19
tropin; TBG, thyroxin binding globulin. (Adapted from Burrow,
G.N. et al., N. Engl. J. Med., 331(16), 1072, 1994. With
permission.)
Hypothyroidism
gradually 9 (Figure 57.1). Thyroid hormones have a major Although maternal hypothyroidism is associated with men-
role in the normal growth and development of the brain and strual disturbance and decreased fertility, it is not rare to
central nervous system.10–12 conceive.20 Women with limited thyroid reserve or iodine
During early pregnancy, maternal thyroid hormones deficiency can develop hypothyroidism due to the increased
can be an important source necessary for early fetal metabolic demands during pregnancy. When iodine intake
TSH changes during pregnancy

5
4.80
4.16
4
3.67
Serum TSH (mlU/L)
3 2.93
2.51
2.15
2
1.23 1.35
1.03
1
0.53
0.12 0.31
0.03 0.09 0.20
0
First trimester Second trimester Third trimester
(n = 7) (n = 5) (n = 2)
Figure 57.2 TSH changes during pregnancy-trimester-specific TSH reference intervals. Median values (rectangle) versus the range
of 2.5th (ellipse, bottom) and 97.5th (ellipse, top) percentiles for each trimester of pregnancy taken from eight studies of trimes-
ter specific TSH reference intervals, reported between 2004 and 2009, for women without thyroid peroxidase autoantibodies
from iodine-sufficient populations. The dotted horizontal lines show the typical nonpregnant reference range (0.4–4.1 mU/mL).
(Adapted from Glinoer, D. and Spencer, C.A., Nat. Rev. Endocrinol., 6(9), 527, 2010. With permission.)
Hypothyroidism 481
Table 57.1 Thyroid-stimulating hormone trimester-specific reference ranges:

Recommendations in recent clinical guidelines
Trimester-specific reference range for TSH (mIU/L)

ATA 201115 AACE and ATA 201216 Endocrine Society 201217
First trimester 0.1–2.5 Upper normal < 2.5 0.1–2.5
Second trimester 0.2–3.0 Upper normal < 3.0 0.1–2.5
Third trimester 0.3–3.0 Upper normal < 3.5
Note: These cutoff values are recommended in case the local laboratory does not provide trimester
specific reference ranges.
Abbreviations: ATA, American Thyroid Association; AACE, American Association of Clinical
Endocrinologists.
is adequate, the most common cause of hypothyroidism Overall, the prevalence of overt and SCH in pregnancy is
during pregnancy is autoimmune thyroid disease (AITD) estimated to be 0.3%–0.5% and 2%–2.5%, respectively. The
(Hashimoto’s thyroiditis).21 Also, the presence of other auto- prevalence may be higher in iodine insufficient areas.24,25
immune disorders like type 1 diabetes mellitus raises the Screening programs detect congenital hypothyroidism in
risk of AITD.22 In iodine-poor areas, iodine deficiency itself approximately 1:2000 to 1:4000 newborns.26
is the most frequent cause of hypothyroidism. In fact, iodine
deprivation is the most common cause of hypothyroidism
worldwide. Next, other etiologies can occur in pregnant Maternal and fetal complications of hypothyroidism
women including prior radioiodine ablation of the thyroid, Hypothyroidism during pregnancy may adversely affect
thyroidectomy, or secondary hypothyroidism e.g., Sheehan’s maternal and fetal outcomes. It has been associated with an
syndrome. Finally, an important cause for neonatal thyroid increased risk of spontaneous abortion. In continuing preg-
hypofunction is congenital hypothyroidism. Most cases of nancies, hypothyroidism has been associated with increased
this disorder are due to agenesis or dysgenesis of the fetal risk of several complications such as preterm delivery, gesta-
thyroid gland, congenital dyshormonogenesis, or iodine tional hypertension, placental abruption, low birth weight,
deficiency in endemic areas. cesarean section, and postpartum hemorrhage.4,20,27–29
Symptoms and signs of hypothyroidism during pregnancy Conversely, no difference in adverse pregnancy outcomes
are the same as in the nonpregnant state. They include sensi- was detected in a study of women with antenatal severe
tivity to cold, dry skin, constipation, fatigue, muscle cramps, hypothyroidism, defined as TSH level above 20.0 mIU/L,
weight gain, voice change, and drowsiness. The symptoms compared to a control euthyroid group. Hirsch et al. con-
may be nonspecific or may go unnoticed. Notably, many cluded that intense follow-up and levothyroxine replacement
patients are asymptomatic. Regarding congenital hypothy- therapy during gestation may improve pregnancy outcomes,
roidism, it is a very common cause of intellectual disability even when target serum TSH levels are not reached.30
that can be prevented. Thyroid function tests confirm the Women with SCH were also at increased risk for adverse
diagnosis of hypothyroidism. An elevated serum TSH level, in pregnancy outcomes in some (but not all) studies, though
combination with a subnormal serum FT4 level, characterizes lower than overt hypothyroidism. The complications
overt/primary hypothyroidism. Furthermore, serum TSH included spontaneous abortion, preterm delivery, placental
level elevation above 10 mIU/L, irrespective of serum FT4 lev- abruption, and pregnancy loss.20,25,31,32 The adverse outcomes
els, is also considered as overt hypothyroidism by the ATA.15 have varied between studies.
Next, subclinical hypothyroidism (SCH) is character- Importantly, maternal hypothyroidism has been associ-
ized by a serum TSH above the upper reference limit (using ated with impairment in psychomotor development in the
pregnancy-specific reference ranges) in combination with a offspring.12,32,33 The full IQ scores of children at age 7–9 years,
normal serum FT4 concentration. However, in secondary born to hypothyroid women (mean TSH 13.2 mIU/L), were
and tertiary (central) hypothyroidism, the diagnosis should four points lower in average compared with euthyroid con-
be based on a low serum FT4 and low-normal serum TSH trols. Strikingly, the test scores of children of untreated
levels. hypothyroid women were seven points lower than controls,
An additional thyroid function disorder during gesta- but the test scores of children whose mothers were treated
tion is isolated hypothyroxinemia (low T4). It is defined as a (although inadequately) were similar in most categories to
maternal FT4 concentration in the lower 5th or 10th percen- the control children.33 Correspondingly, the association of
tile of the reference range, in conjunction with a normal TSH. cognitive development and SCH in pregnancy is still under
The effect of isolated maternal hypothyroxinemia on perinatal study. Recently, a randomized study was reported on the cog-
and neonatal outcome is yet unclear. Indeed, in mothers with nitive examination of children whose mothers received levo-
isolated hypothyroxinemia treated or not with T4 before thyroxine therapy during pregnancy for impaired thyroid
20 weeks gestation, the IQ of children was not affected.23 function tests (mean maternal serum TSH was 3.8 mIU/L).
No benefit in cognitive function, assessed at age 3 years, treatment for all pregnant women with SCH as we expect
was detected versus controls.34 More randomized trials are that the benefits outweigh any potential risks. However, close
needed to determine the benefit of SCH treatment on neuro- monitoring of thyroid function is highly recommended in
cognitive outcomes in early pregnancy. case the therapeutic decision was not to treat.
Finally, assessment of treatment in secondary and ter-
tiary (central) hypothyroidism should be guided by FT4 and
Treatment of hypothyroidism not TSH.
It is well accepted that maternal hypothyroidism should be
treated considering the risks to both the mother and the fetus.
The recommended treatment for hypothyroidism is oral syn- Screening for hypothyroidism
thetic l-T4 (levothyroxine or l-thyroxine). The goal of l-T4 Universal screening for hypothyroidism in asymptomatic
replacement during gestation is to restore euthyroidism as soon women who are pregnant or are planning pregnancy is con-
as possible. Treatment with l-T4 is usually initiated with a dose troversial due to insufficient evidence that it is beneficial.39
of 100–150 µg/day or close to full replacement doses (1.6 mcg/ Indeed, screening practices and guidelines vary widely.
kg body weight/day), but individual dose requirements vary Some professional societies updated these clinical practice
widely. Milder hypothyroidism may require a lower dose. The guidelines and recommended “targeted” or “aggressive case
normal trimester reference ranges for TSH are used as target of finding” rather than universal screening.15,17
treatment during pregnancy.15–17 Concordantly, the ATA recommends thyroid function
In most of the pregnant women already treated with testing in women at high risk for overt hypothyroidism15:
l-T4, early dose adjustment is needed as soon as pregnancy
is confirmed in order to meet the increased requirement
during pregnancy. Requirements increase as early as the History of thyroid dysfunction or prior thyroid surgery
fifth week of gestation.35 Optimizing preconception TSH Age > 30 years
level (<2.5 mIU/L) is recommended for hypothyroid women Symptoms of thyroid dysfunction or the presence of goiter
anticipating pregnancy.15 There is an indication that lower TPO-Ab positivity
TSH levels before pregnancy (to the lower quartile of nor- Type 1 diabetes or other autoimmune disorders
mal) might reduce the risk of elevated levels in pregnancy.36 History of miscarriage or preterm delivery
Adjustment may require 30% increase in dosage or more.37 History of head or neck radiation
Strategy for increasing l-T4 dose should be individual- Family history of thyroid dysfunction
ized. Factors like preconception TSH level and etiology of Morbid obesity (BMI > 40 kg/m2)
maternal hypothyroidism may provide insight into the mag- Use of amiodarone or lithium or recent administration of
nitude of the needed increase. Dose increase is more likely iodinated radiological contrast
to be required in patients without a functional thyroid tissue Infertility
(e.g., due to radio-ablation or surgery) than in patients with Residing in an area of known moderate to severe iodine
Hashimoto’s thyroiditis.38 insufficiency
Next, treatment of hypothyroidism should be monitored.
l-T4 has a long half-life of almost seven days. Hence, a peak
therapeutic effect at a given dose may not be attained for The time for screening is in early pregnancy. Indeed, the
4–6 weeks. Correspondingly, monitoring TSH levels every Endocrine Society recommends TSH screening either on the
4–6 weeks is recommended during the first half of preg- ninth week of gestation or on the first prenatal visit.17
nancy. Actually, monitoring every 4 weeks can detect 92% of
abnormal values.37 In the second half of pregnancy, it is rea-
sonable to monitor TSH levels at least once between 26–32 Hyperthyroidism in pregnancy
weeks gestation and also 4–6 weeks after changing dosage
or brand, as new equilibrium is expected.37 After delivery, Etiology
most hypothyroid women need to decrease the l-T4 dosage The prevalence of hyperthyroidism during pregnancy is
administered during pregnancy to the prepregnancy dose. relatively uncommon. It ranges between 0.1% and 1%.
Regarding treatment of SCH during pregnancy, there is Approximately 0.4% and 0.6% of pregnant women present
no consensus.39,40 Reduction in adverse pregnancy outcomes with clinical and subclinical hyperthyroidism, respectively.41
was demonstrated in a study of women with SCH and positive Etiologies of thyrotoxicosis include those that present specif-
TPO-Ab treated with l-T4.39 Recently, several professional ically during pregnancy, as well as others found in the gen-
societies addressed this issue but had controversial opin- eral population.
ions. For example, the ATA recommended l-T4 treatment The most common etiology of hyperthyroidism in the
in pregnant women with SCH who are also thyroid peroxi- childbearing age is Graves’ disease (GD). This cause accounts
dase antibody positive (TPO-Ab+).15 Instead, the Endocrine for 85% of clinical hyperthyroidism in pregnancy. Also,
Society recommended replacement in all pregnant women another common etiology of thyrotoxicosis is hypereme-
with SCH.17 From our perspective (Toledano and Solomon, sis gravidarum.42 In contrast, uncommon etiologies d uring
the authors of this chapter), we also recommend offering pregnancy include silent or subacute thyroiditis, toxic
Hyperthyroidism in pregnancy 483
multinodular goiter, single toxic adenoma, molar disease, TRAb titer between gestational weeks 22 and 26. When the
and struma ovarii.43 titer is three to five times above normal, the fetus should be
evaluated for detection of potential fetal hyperthyroidism.
This condition can be assessed by several fetal ultrasono-
Clinical diagnosis graphic parameters: tachycardia (persistent fetal heart rate of
The clinical characteristics are the same in pregnant and >160 bpm), the presence of fetal goiter, increased fetal motil-
nonpregnant patients. However, the clinical diagnosis of ity, growth retardation, and accelerated bone maturation.43
hyperthyroidism during pregnancy may be challenging. A pattern of the fetal heart monitor tracing unique to fetal
Many nonspecific symptoms can be mimicked by normal thyrotoxicosis47 includes a sustained baseline of 170–180
pregnancy, e.g., anxiety, fatigue, heat intolerance, tachycar- beats/minute combined with moderate variability. The lat-
dia, tremor, warm moist skin, and systolic murmur. Instead, ter exhibits acceleration with a lack of deceleration. If the
several physical findings such as goiter, ophthalmopathy, and information to be gained will change therapy, it is possible
pretibial myxoedema are highly suggestive of GD. Another to perform serial cordocentesis for diagnosis and monitoring
suspicious symptom of hyperthyroidism during gestation is drug therapy.48,49 Suspected fetal hyperthyroidism in utero
weight loss or the absence of weight gain despite an increased can be treated with a combination of mercaptizol (MMI,
appetite. When hyperemesis gravidarum is associated with 20 mg/day) and thyroxine, which is occasionally required to
weight loss, the differential diagnosis should include hCG- maintain maternal euthyroidism. Fortunately, neonatal GD
induced hyperthyroidism. with hyperthyroidism is presented in only 1%–5% of neo-
nates born to mothers with GD. As discussed previously, the
etiology is related to the transplacental passage of stimulat-
Laboratory diagnosis ing maternal TRAb. In fact, it is recommended to perform
Measurements of serum TSH, T4 and T3 levels, and anti-TSH cord serum-free T4 and TSH determinations in all deliveries
receptor antibodies (TRAbs) are required when suspect- of mothers with a history of GD.
ing hyperthyroidism. Patients with prominent symptoms Fetal hypothyroidism: Maternal GD treated with anti-
almost always have a suppressed serum TSH level <0.1 mU/L thyroid drugs (ATDs) may induce fetal hypothyroidism.
concurrently with elevated serum free T4 and T3 levels. Conceivably, keeping maternal circulating thyroid hormone
Importantly, the evaluation of thyroid function tests must levels in the upper quartile of the normal range is the best
consider the hCG-mediated suppression of serum TSH that way to avoid fetal hypothyroidism.41,43 Inhibitory TRAb pro-
occurs during pregnancy. Furthermore, thyroid antibodies duction has been shown to cause hypothyroidism transiently
(thyroglobulin [Tg]-Ab and TPO-Ab) are usually positive in neonates born to mothers with GD.50,51
in patients with GD. Hence, the differential diagnosis must
include AITD as the etiology of hyperthyroidism. Not sur-
prisingly, TRAbs are detectable in the majority of patients Management of Graves’ disease in pregnancy
with GD. During the second half of pregnancy, TRAb pro- As a general rule, gradual improvement of GD occurs dur-
duction tends to undergo immunological remission. Thus, ing gestation. Several causes were alluded to this spontane-
measurement of these antibodies is dependent on gestational ous improvement including immunosuppressive effect of
age at determination.44,45 the pregnancy, elevation of maternal serum TBG levels that
reduce serum-free T4 and T3 fractions, and iodine losses
characteristic to the gestation period.
Pregestational counseling for women with Graves’ The key management for GD during pregnancy are the
disease ATDs.41,43 In general, the two main targets of therapy are
Conceivably, pregestational counseling is extremely impor- (1) controlling hyperthyroidism in the mother by main-
tant in the treatment plan of young women with GD. taining free thyroid hormone levels at the upper limit of
Contraception counseling is strongly recommended in all normal range or borderline high levels and (2) using the
women of childbearing age affected with GD. The main aim smallest possible dose of ATD in order to minimize the
is to avoid pregnancy while having hyperthyroidism.41,42 risk of fetal hyperthyroidism or hypothyroidism. Beneficial
The different therapeutic choices should be presented and response to therapy includes weight gain and improvement
discussed with women planning a pregnancy: medical treat- in symptoms. Several adverse effects of ATD have been doc-
ment, radioiodine therapy, or surgery. A full discussion on umented. First, propylthiouracil (PTU) is associated with
the different choices can be found in reference 46. risk of hepatic toxicity necessitating liver transplantation in
a number of cases and may be fatal.52,53 Second, methima-
zole (MMI) treatment is infrequently associated with aplasia
Fetal and neonatal adverse effects cutis in neonates. This is a localized lesion in the parietal
Fetal hyperthyroidism is a serious and rare complication. area of the scalp. It is characterized by congenital absence
However, this potentially fatal disorder is preventable. of the skin and a punch-out, “ulcer” lesion.54 Third, cho-
Several risk factors were identified including active maternal anal atresia and/or esophageal atresia, minor dysmorphic
Graves’ hyperthyroidism (GH) and a history of GD treated features, and developmental delay are characteristic to the
with ablation therapy. It is recommended to determine the “methimazole embryopathy.” Taken together, the FDA and
the ATA recommended to use PTU only in the first trimes- Diagnostic evaluation and management of a thyroid
ter of pregnancy. By the second trimester, PTU should to be nodule in pregnancy
switched to MMI.55 The first investigation of choice is fine-needle aspiration
Simultaneously, symptoms of acute hyperthyroid disease biopsy (FNAB). It can detect a malignancy/suspicious result
may be controlled transiently by propranolol. in 35%.64 Several management guidelines suggested the
The second treatment option in pregnancy is surgery. following approach for a single thyroid nodule detected on
Subtotal thyroidectomy is the preferred surgery and typi- physical examination or for a dominant nodule in a multi-
cally is performed in the second trimester. Indications to nodular gland, confirmed by ultrasonography:
the surgical alternative are few: consumption of large ATD
doses, allergy to ATD,56 and patient preference. 1. Solid lesion <1 cm, follow-up in the postpartum.
As far as breastfeeding in mothers treated with ATD 2. Nodule >1–1.5 cm, should be considered for FNAB if
for GD is concerned, the risk to the neonates is insig- there are suspicious findings on ultrasound.
nificant. However, this hazard is low as long as the doses 3. In the presence of tracheal obstruction, immediate
of ATD can be kept moderate (MMI <20 mg/dL; PTU surgery.
<250–300 mg/dL).43 4. If the FNAB is diagnostic of malignancy or it is a suspi-
cious lesion, surgery may be postponed until after deliv-
ery, unless there is one of the following signs: presence of
Gestational transient thyrotoxicosis lymph node metastases, the lesion is a large primary or
Gestational transient thyrotoxicosis (GTT) or nonauto- extensive lymph node involvement related to a medullary
immune gestational hyperthyroidism is a transient syn- cancer.
drome of hyperthyroidism occurring near the end of the 5. Surgery and FNAB could both be postponed until after
first trimester of pregnancy, usually in healthy women. It is delivery with probable safety.
frequently related to excessive vomiting.43 As its nature is 6. A woman with a malignant lesion or rapid growth
transient, GTT may be clinically inapparent. The clinical should be offered surgery in the second trimester of
and biochemical characteristics of thyrotoxicosis are usually gestation.
mild and variable. They include mild thyroid enlargement 7. Some authors recommend that women with follicular
and lack of thyroid autoantibodies. lesions or early-stage papillary carcinoma may postpone
The etiology of GTT was discussed earlier. Its prevalence the surgery until postpartum since these lesions are not
is highly variable and ranges between 0.3% in Japan and 11% expected to progress rapidly.15,65,66
in Hong Kong.58,59 It may reach 2%–3% of unselected preg-
nancies in Europe, i.e., tenfold more prevalent than GD.57
Since GTT improves in parallel to the decline in hCG levels, Pregnancy and coexisting thyroid malignancy
this syndrome is always transient. As a rule, partial or total Currently, evidence suggests that treated differentiated thyroid
suppression of serum TSH can last several weeks after serum cancer without the evidence of residual disease should not
free T4 reverts to normal.60,61 GTT usually has no adverse hamper planning pregnancy. Accordingly, Hirsch et al.67
pregnancy outcome. reported on 63 consecutive women (90 births), followed
However, the most severe cases are commonly associated between 1992 and 2009 for papillary thyroid cancer. All
with hyperemesis gravidarum, e.g., twin pregnancy. Usually, women had delivered at least once after total thyroid-
no specific treatment is necessary. In some cases, admin- ectomy, and I-131 was also administered in 58 of them.
istration of beta-adrenergic blocking agents, hydration, A comparison of neck ultrasound and serum thyroglobulin
and antiemetics are required for a short period to relieve (Tg) levels was performed before and after pregnancy. The
symptoms. Importantly, GD should be excluded in patients conclusion was that gestation does not cause thyroid can-
with a severe clinical presentation and clear symptomatic cer recurrence in PTC survivors who have no structural or
hyperthyroidism by measurement of TRAb. Transient PTU biochemical evidence of disease persistence at the time of
administration for a few weeks should be considered. conception. Hence, pregnancy has a minimal effect on thy-
roid cancer. Instead, it is possible that residual cancer at the
time of conception will progress. More evidence is needed
Nodular and malignant thyroid disease before making a firm recommendation.
As far as thyroid hormone administration is concerned,
The prevalence of nodular thyroid disease is not so rare. Up suppression and maintenance of serum TSH levels between
to 10% of pregnant women have thyroid nodules. Nodular 0.1 and 0.5 mU/L is prudent. Specifically, it is important in
thyroid disease was not associated with pregnancy in an patients with a high-risk tumor recurrence but with no clini-
iodine-deficient area.62 Furthermore, the incidence of thy- cal or biochemical evidence of current disease. In analogy
roid cancer was reported in a retrospective study from the with this, TSH suppression is recommended in pregnant
California Cancer Registry during the years 1991–1999: women with a positive or suspicious FNAB for cancer that
3.3/100,000, 0.3/100,000, and 10.8/100,000 were diagnosed prefer delaying surgery until postpartum. However, the
before pregnancy, at the time of delivery and within 1 year level of serum TSH in this case should be detectable, i.e.,
after delivery, respectively.63 suppressed only mildly. Next, adverse events with previous
Thyroid autoimmunity 485
I-131 therapy were not reported in subsequent pregnancies.68 nondiabetics. As expected, PPT is strongly correlated to thy-
Following I-131 ablation therapy, it is strongly recommended roid antibodies in women with type 1 diabetes. As discussed,
to delay pregnancy for at least 1 year. 10% of pregnant women have positive antithyroid antibod-
ies at 14–16 gestational weeks. PPTD occur in 50% of these
women during the first 6 months after delivery.71 Of note, up
Postpartum thyroid dysfunction to 50% of these patients may develop permanent hypothy-
roidism at 7–10 years postpartum.
Prominent changes in the immune system are detected The etiology of PPT is a destructive thyroiditis. It may
during pregnancy.69 Hence, the change in immune cellu- present up to 9 months after delivery with a transient hyper-
lar status consists of an alteration from the Th1 to the Th2 thyroidism and/or hypothyroidism.72 Using the correlated
state. Also, the humoral immune system alters. Generally, thyroid function tests, it is possible to diagnose the specific
10% of pregnant women have positive anti-TPO-Ab titer at thyroid dysfunction, i.e., hyperthyroidism or hypothyroidism.
14–16 weeks gestation. However, it decreases considerably The clinical course include hyperthyroidism alone, hypothy-
during the second and third trimesters. After delivery, the roid phase alone, and hyperthyroidism followed by hypo-
Th2 position reverts abruptly back to the nonpregnant Th1 thyroidism in 19%, 49%, and 32% of patients, respectively.
state. Notably, this immune rebound phenomenon consists Remarkably, prominent symptoms in the hyperthyroid phase,
of a very pronounced elevation in the anti-TPO-Ab titer. and even in antibody-positive patients without any thyroid
Indeed, it reaches a peak between 3 and 6 months postpar- dysfunction, are irritability and lack of energy. Conversely, the
tum. Similarly, TRAb titer present in early pregnancy has hypothyroid status may present with intense symptomatology.
the same pattern through gestation and postpartum. The The diagnosis of PPT may be established by radioiodine
postpartum immunological alterations discussed earlier uptake. After isotope administration, the uptake will be very
can lead to the development of postpartum thyroid dysfunc- low both at early and late times, consistent with the destruc-
tion (PPTD), and they may be either transient or perma- tive nature of this disorder. Although antithyroid antibodies
nent. Moreover, the dysfunction may be related to either a are frequently positive, TRAbs are negative.
destructive or stimulating pathological process (Figure 57.3). Regarding the management of PPT, no specific therapy
is generally required in the hyperthyroid phase as it is rela-
tively asymptomatic. Beta-adrenoreceptor blocking agents
Postpartum Graves’ disease may be administered for hyperthyroid symptoms and signs,
Significantly, women with GD in remission after ATD e.g., palpitations, anxiety, sweating, or tachycardia. However,
therapy have an increased risk for relapsing hyperthyroid- symptomatic and persistent hypothyroidism should be
ism postpartum.70 The differential diagnosis of postpartum treated with l-thyroxine. As 50% of women with transient
hyperthyroidism requires clinical examination, measure- thyroid dysfunction postpartum will develop permanent
ment of circulating TRAb, thyroid scintiscanning, and radio- hypothyroidism after 7–10 years, annual thyroid function
iodine uptake. test should be performed.
Postpartum thyroiditis Thyroid autoimmunity

It is estimated that 5%–9% of women have postpartum thy-
roiditis (PPT). Risk factors for this disease include a prior Infertility
episode of PPT and type 1 diabetes. In fact, PPT is three- Immunological factors may have adverse impact on normal
fold more prevalent in women with this diabetes versus fertility and reproduction processes including fertilization,
(III) Destructive thyrotoxicosis

(I) Persistent thyrotoxicosis
Postpartum
Graves’ disease
with elevated RAIU

Thyroid function
(II) Transient thyrotoxicosis

with low RAIU
2 4 6 (III) Destructive thyrotoxicosis

Delivery Months (IV) Transient hypothyroidism Postpartum
exacerbation
of autoimmune
thyroiditis
(V) Persistent hypothyroidism
Figure 57.3 Patterns of postpartum thyroid dysfunction. (Adapted from Lazarus, J.H., Eur. Thyroid J., 1, 24, 2012. With
permission.)
implantation, and early embryo development. Indeed, the When thyroid hormone levels are compatible with euthy-
correlation between abnormal immunological laboratory roidism, it is possible that in women with AITD these lev-
tests, including antithyroid antibodies, and infertility is well els are actually inappropriately low for the given gestational
established.73,74 period. Indeed, some reports suggested that l-thyroxine
Regarding thyroid dysfunction, AITD, called also chronic treatment may correct such minor deficiency. Moreover, this
lymphocytic thyroiditis or Hashimoto’s disease, is the most treatment may influence the systemic immune disorder and
common cause of hypothyroidism in women of reproductive the placental–decidual environment.85–88 Of note, the asso-
age. Simultaneously, female infertility is correlated to clinical ciation between thyroid autoimmunity and miscarriage still
hypothyroidism. Medically assisted conception and onset of does not imply causation.89,90
gestation is not hampered by AITD, but a successful outcome
of the ongoing pregnancies is significantly reduced in those
IVF
women with AITD due to greater early pregnancy loss.75
Currently, screening women with infertility and/or miscar-
riage for the presence of thyroid autoimmunity and dys-
Miscarriage function is routinely performed in many centers. However,
Several studies suggest that thyroid-antibody-positive several reports on women achieving a pregnancy through
women have a higher propensity to pregnancy loss, placen- an IVF procedure have shown no correlation between the
tal abruption, and even preterm delivery and become preg- presence of thyroid autoimmunity and a risk for spontane-
nant at an older age (approximately, 3–4 years older).76–84 ous miscarriage.75,91–96
REFERENCES
1. Glinoer D. The regulation of thyroid function in pregnancy: 16. Garber R, Cobin RH, Gharib H et al. Clinical practice guide-
Pathways of endocrine adaptation from physiology to pathology. lines for hypothyroidism in adults: Co-sponsored by American
Endocr Rev 1997; 18: 404–433. Association of Clinical Endocrinologists and the American
2. Hershman JM. Physiological and pathological aspects of the Thyroid Association. Thyroid 2012; 22(12): 1200–1235.
effect of human chorionic gonadotropin on the thyroid. Best 17. De Groot L, Abalovich M, Alexander EK et al. Management
Pract Res Clin Endocrinol Metab 2004; 18: 249–265. of thyroid dysfunction during pregnancy and postpartum: An
3. Goodwin TM, Montoro M, Mestman JH. The role of chorionic Endocrine Society clinical practice guideline. J Clin Endocrinol
gonadotropin in transient hyperthyroidism of hyperemesis gravi- Metab 2012; 97(8): 2543–2565.
darum. J Clin Endocrinol Metab 1992; 75: 1333–1337. 18. Berta E, Samson L, Lenkey A et al. Evaluation of the thyroid func-
4. Delange F. The disorders induced by iodine deficiency. Thyroid tion of healthy pregnant women by five different hormone assays.
1994; 4: 107–128. Pharmazie 2010; 65: 436–439.
5. Pharoah P, Buttfield IH, Hetzel BS. Neurological damage to the 19. Weeke J, Dybkjaer L, Granlie K et al. A longitudinal study of
fetus resulting from severe iodine deficiency during pregnancy. serum TSH and total and free iodothyronines during normal preg-
Int J Epidemiol 2012; 41(3): 589–592. nancy. Acta Endocrinol 1982; 101: 531–537.
6. Hetzel BS. Iodine deficiency disorders (IDD) and their eradica- 20. Abalovich M, Gutierrez S, Alcaraz G et al. Overt and subclini-
tion. Lancet 1983; 2(8359): 1126–1129. cal hypothyroidism complicating pregnancy. Thyroid 2002; 12(1):
7. WHO Secretariat, Andersson M, de Benoist B, Delange F, Zupan 63–68.
J. Prevention and control of iodine deficiency in pregnant and lac- 21. Mandel SJ. Hypothyroidism and chronic autoimmune thyroid-
tating women and in children less than 2-years-old: Conclusions itis in the pregnant state: Maternal aspects. Best Pract Res Clin
and recommendations of the Technical Consultation. Public Endocrinol Metab 2004; 18: 213–224.
Health Nutr 2007; 10(12A): 1606–1611. 22. Gallas PR, Stolk RP, Bakker K et al. Thyroid dysfunction during
8. Public Health Committee of the American Thyroid Association, pregnancy and in the first postpartum year in women with diabe-
Becker DV, Braverman LE et al. Iodine supplementation tes mellitus type 1. Eur J Endocrinol 2002; 147: 443–451.
for pregnancy and lactation—United States and Canada: 23. Pop VJ, Brouwers EP, Vader HL et al. Maternal hypothyroxinae-
Recommendations of the American Thyroid Association. Thyroid mia during early pregnancy and subsequent child development:
2006; 16: 949–951. A 3-year follow-up study. Clin Endocrinol (Oxf) 2003; 59(3):
9. Burrow GN, Fisher DA, Larsen PR. Maternal and fetal thyroid 282–288.
function. N Engl J Med 1994; 331(16): 1072. 24. Klein RZ, Haddow JE, Faix JD et al. Prevalence of thyroid defi-
10. Bernal J, Nunez J. Thyroid hormones and brain development. Eur ciency in pregnant women. Clin Endocrinol (Oxf) 1991; 35:
J Endocrinol 1995; 133(4): 390–398. 41–46.
11. Williams GR. Neurodevelopmental and neurophysiological actions 25. Cleary-Goldman J, Malone FD, Lambert-Messerlian G et al.
of thyroid hormone. J Neuroendocrinol 2008; 20(6): 784–794. Maternal thyroid hypofunction and pregnancy outcome. Obstet
12. Man EB, Brown JF, Serunian SA. Maternal hypothyroxinemia psy- Gynecol 2008; 112(1): 85–92.
choneurological deficits of progeny. Ann Clin Lab Sci 1991; 21: 26. Harris KB, Pass KA. Increase in congenital hypothyroidism in
227–239. New York State and in the United States. Mol Genet Metab 2007;
13. Vulsma F, Gons MH, Devijldern JJM. Maternal fetal transfer of 91(3): 268–277.
thyroxine in congenital hypothyroidism due to a total organifica- 27. Leung AS, Millar LK, Koonings PP et al. Perinatal outcome in
tion defect or thyroid dysgenesis. N Engl J Med 1989; 321: 13–16. hypothyroid pregnancies. Obstet Gynecol 1993; 81: 349–353.
14. Glinoer D, Spencer CA. Serum TSH determinations in pregnancy: 28. Männistö T, Mendola P, Grewal J et al. Thyroid diseases and
How, when and why? Nat Rev Endocrinol 2010; 6(9): 527. adverse pregnancy outcomes in a contemporary US cohort.
15. Stagnaro-Green A, Abalovich M, Alexander E et al. Guidelines J Clin Endocrinol Metab 2013; 98: 2725–2733.
of the American Thyroid Association for the diagnosis and man- 29. Davis LE, Leveno KJ, Cunningham FG. Hypothyroidism compli-
agement of thyroid disease during pregnancy and postpartum cating pregnancy. Obstet Gynecol 1988; 72: 108–112.
American Thyroid Association Taskforce on Thyroid Disease 30. Hirsch D, Levy S, Nadler V et al. Pregnancy outcomes in
During Pregnancy and postpartum. Thyroid 2011; 21(10): women with severe hypothyroidism. Eur J Endocrinol 2013; 169:
1081–1125. 313–320.
References 487
31. Negro R, Schwartz A, Gismondi R et al. Increased pregnancy 56. Bruner J, Landon MB, Gabbe SG. Diabetes mellitus and Graves’
loss rate in thyroid antibody negative women with TSH levels disease in pregnancy complicated by maternal allergies to anti-
between 2.5 and 5.0 in the first trimester of pregnancy. J Clin thyroid medication. Obstet Gynecol 1988; 72: 443–445.
Endocrinol Metab 2010; 95(9): E44–E48. 57. Glinoer D. The systematic screening and management of hypo-
32. Casey BM, Dashe JS, Wells CE et al. Subclinical hypothyroid- thyroidism and hyperthyroidism during pregnancy. Trends
ism and pregnancy outcomes. Obstet Gynecol 2005; 105: Endocrinol Metab 1998; 9: 403–411.
239–245. 58. Tanaka S, Yamada H, Kato EH et al. Gestational transient hyper-
33. Haddow JE, Palomaki GE, Allan WC et al. Maternal thyroid defi- thyroxinaemia (GTH): Screening for thyroid function in 23163
ciency during pregnancy and subsequent neuropsychological pregnant women using dried blood spots. Clin Endocrinol 1998;
development of the child. N Engl J Med 1999; 341: 549–555. 49: 325–329.
34. Lazarus JH, Bestwick JP, Channon S et al. Antenatal thyroid 59. Yeo CP, Khoo DH, Eng PH et al. Prevalence of gestational thyro-
screening and childhood cognitive function. N Engl J Med 2012; toxicosis in Asian women evaluated in the 8th to 14th weeks of
366(6): 493–501. pregnancy: Correlations with total and free beta human chori-
35. Alexander EK, Marqusee E, Lawrence J et al. Timing and mag- onic gonadotrophin. Clin Endocrinol 2001; 55: 391–398.
nitude of increases in levothyroxine requirements during preg- 60. Ramsey PS, Ramin KD. 34-year-old pregnant woman with nau-
nancy in women with hypothyroidism. N Engl J Med 2004; 351: sea and vomiting. Mayo Clin Proc 2000; 75(12): 1317–1320.
241–249. 61. Tan JY, Loh KC, Yeo GS et al. Transient hyperthyroidism of hyper-
36. Rotondi M, Mazziotti G, Sorvillo F et al. Effects of increased thy- emesis gravidarum. BJOG 2002; 109: 683–688.
roxine dosage pre-conception on thyroid function during early 62. Karger S, Schötz S, Stumvoll M et al. Impact of pregnancy on
pregnancy. Eur J Endocrinol 2004; 151(6): 695–700. prevalence of goitre and nodular thyroid disease in women living
37. Yassa L, Marqusee E, Fawcett R et al. Thyroid hormone early in a region of borderline sufficient iodine supply. Horm Metab
adjustment in pregnancy (the THERAPY) trial. J Clin Endocrinol Res 2010; 42: 137–142.
Metab 2010; 95: 3234–3241. 63. Smith LH, Danielsen B, Allen ME et al. Cancer associated with
38. Kaplan MM. Monitoring thyroxine treatment during pregnancy. obstetric delivery: Results of linkage with California cancer regis-
Thyroid 1992; 2: 147–152. try. Am J Obstet Gynecol 2003; 189: 1128–1135.
39. Negro R, Schwartz A, Gismondi R et al. Universal screening ver- 64. Hay I. Nodular thyroid disease diagnosed during pregnancy:
sus case finding for detection and treatment of thyroid hormonal How and when to treat. Thyroid 1999; 9: 667–670.
dysfunction during pregnancy. J Clin Endocrinol Metab 2010; 95: 65. Pacini T, Schlumberger M, Dralle H et al. European consensus
1699–1707. of the management of patients with differentiated thyroid car-
40. Committee on Patient Safety and Quality Improvement, cinoma of the follicular epithelium. Eur J Endocrinol 2006; 154:
Committee on Professional Liability. ACOG Committee Opinion 787–803.
No. 381: Subclinical hypothyroidism in pregnancy. Obstet 66. Cooper DS, Doherty GM, Haugen BR et al. Revised American
Gynecol 2007; 110(4): 959–960. Thyroid Association management guidelines for patients with
41. Marx H, Amin P, Lazarus JH. Hyperthyroidism and pregnancy. Br thyroid nodules and differentiated thyroid cancer. Thyroid 2009;
Med J 2008; 336: 663–667. 19: 1167–1214.
42. Krassas GE, Poppe K, Glinoer D. Thyroid function and human 67. Hirsch D, Levy S, Tsvetov G et al. Impact of pregnancy on out-
reproductive health. Endocr Rev 2010; 31: 702–755. come and prognosis of survivors of papillary thyroid cancer.
43. Mestman JH. Hyperthyroidism in pregnancy. Best Pract Res Clin Thyroid 2010; 20: 1179–1185.
Endocrinol Metab 2004; 18(2): 267–288. 68. Lin JD, Wang HS, Weng HF et al. Outcome of pregnancy after
44. Thung SF, Funai EF, Grobman WA. The cost-effectiveness of uni- radioactive iodine treatment for well differentiated thyroid carci-
versal screening in pregnancy for subclinical hypothyroidism. nomas. J Endoc Invest 1998; 21: 662–667.
Am J Obstet Gynecol 2009; 200: 267.e1–267.7. 69. Weetman AP. Immunity, thyroid function and pregnancy:
45. Dosiou C, Sanders GD, Araki SS et al. Screening pregnant women Molecular mechanisms. Nat Rev Endocrinol 2010; 6(6): 311–318.
for autoimmune thyroid disease: A cost-effectiveness analysis. 70. Rotondi M, Cappelli C, Pirali B et al. The effect of pregnancy on
Eur J Endocrinol 2008; 158: 841–851. subsequent relapse from Graves’ disease after a successful course
46. Lazarus JH. Pre-conception counselling in Graves’ disease. Eur of antithyroid drug therapy. J Clin Endocrinol Metab 2008; 93(10):
Thyroid J 2012; 1: 24–29. 3985–3988.
47. Towers CV, Thomas S, Steiger RM. The fetal heart monitor tracing 71. Lazarus JH, Hall R, Othman S et al. The clinical spectrum of post-
in pregnancies complicated by fetal thyrotoxicosis. Am J Perinatal partum thyroid disease. QJM 1996; 89: 429–435.
2009; 26: 373–377. 72. Muller AF, Drexhage HA, Berghout A. Postpartum thyroiditis and
48. Kilpatrick S. Umbilical cord sampling in women with thyroid dis- autoimmune thyroiditis in women of childbearing age: Recent
ease in pregnancy: Is it necessary? Am J Obstet Gynecol 2003; insights and consequences for antenatal and postnatal care.
189: 1–2. Endocr Rev 2001; 22(5): 605–630.
49. Polak M, Leger J, Luton D et al. Fetal cord blood sampling in the 73. Van Steirteghem A. Current status of assisted reproductive tech-
diagnosis and treatment of fetal hyperthyroidism in the offsprings nology (ART) 30 years after the first IVF birth. In: Lazarus J, Pirags
of a euthyroid mother producing thyroid stimulating antibodies. V, Butz S, eds., The Thyroid and Reproduction. Thieme: Stuttgart,
Ann Endocrinol (Paris) 1997; 58: 338. Germany, 2009, pp. 7–14.
50. Matsuura N, Harada S, Ohyama Y et al. The mechanisms of tran- 74. Dosiou C, Giudice LC. Natural killer cells in pregnancy and
sient hypothyroxinemia in infants born to mothers with Graves’ recurrent pregnancy loss: Endocrine and immunologic perspec-
disease. Pediatr Res 1997; 42: 214–218. tives. Endocr Rev 2005; 26: 44–62.
51. Evans C, Jordan NJ, Owens G et al. Potent thyrotropin receptor- 75. Poppe K, Glinoer D, Tournaye H et al. Assisted reproduction and
blocking antibodies: A cause of transient congenital hypothyroid- thyroid immunity: An unfortunate combination? J Clin Endocrinol
ism and delayed thyroid development. Eur J Endocrinol 2004; Metab 2003; 88: 4149–4152.
150(3): 265–268. 76. Stagnaro-Green A, Roman SH, Cobin RH et al. Detection of at-
52. Rivkees SA, Szarfman A. Dissimilar hepatotoxicity profiles of risk pregnancy by means of highly sensitive assays for thyroid
propylthiouracil and methimazole in children. J Clin Endocrinol autoantibodies. J Am Med Assoc 1990; 19: 1422–1425.
Metab 2010; 95: 3260–3267. 77. Prummel MF, Wiersinga WM. Thyroid autoimmunity and miscar-
53. Cooper DS, Rivkees SA. Putting propylthiouracil in perspective. riage. Eur J Endocrinol 2004; 150: 751–755.
J Clin Endocrinol Metab 2009; 94: 1881–1882. 78. Krassas GE, Perros P, Kaprara A. Thyroid autoimmunity, infertility
54. Mandel S, Brent GA, Larsen PR. Review of antithyroid drugs use and miscarriage. Expert Rev Endocrinol Metab 2008; 3: 127–136.
during pregnancy and report of a case of aplasia cutis. Thyroid 79. Abbassi-Ghanavati M, Casey BM, Spong CY et al. Pregnancy
1994; 4: 129–133. outcomes in women with thyroid peroxidase antibodies. Obstet
55. Bahn RS, Burch HS, Cooper DS et al. The role of propylthiouracil Gynecol 2010; 116: 381–386.
in the management of Graves’ disease in adults: Report of a meet- 80. Stagnaro-Green A, Chen X, Bogden JD et al. The thyroid and
ing jointly sponsored by the American Thyroid Association and pregnancy: A novel risk factor for very preterm delivery. Thyroid
the Food and Drug Administration. Thyroid 2009; 19: 673–674. 2005; 15: 351–357.
81. Stagnaro-Green A. Maternal thyroid disease and preterm deliv- 89. Poppe K, Velkeniers B, Glinoer D. The role of thyroid autoimmu-
ery. J Clin Endocrinol Metab 2009; 94: 21–25. nity in fertility and pregnancy. Nat Clin Pract Endocrinol Metab
82. Haddow JE, Cleary-Goldman J, McClain MR et al. First- and Second- 2008; 4: 394–405.
Trimester Risk of Aneuploidy (FaSTER) Research Consortium. 90. Kaprara A, Krassas GE. Thyroid autoimmunity and miscarriage.
Thyroperoxidase and thyroglobulin antibodies in early pregnancy Hormones 2008; 7: 294–302.
and preterm delivery. Obstet Gynecol 2010; 116: 58–62. 91. Toulis KA, Goulis DG, Venetis CA et al. Risk of spontaneous
83. Negro R, Schwartz A, Gismondi R et al. Thyroid antibody posi- miscarriage in euthyroid women with thyroid autoimmunity
tivity in the first trimester of pregnancy is associated with nega- undergoing IVF: A meta-analysis. Eur J Endocrinol 2010; 162:
tive pregnancy outcomes. J Clin Endocrinol Metab 2011; 96: 643–652.
E920–E924. 92. Muller AF, Verhoeff A, Mantel MJ et al. Thyroid autoimmunity
84. He X, Wang P, Wang Z et al. Thyroid antibodies and risk of pre- and abortion: A prospective study in women undergoing in vitro
term delivery: A meta-analysis of prospective cohort studies. Eur fertilization. Fertil Steril 1999; 71: 30–34.
J Endocrinol 2012; 167: 455–464. 93. Negro R, Mangieri T, Coppola L et al. Levothyroxine treatment in
85. Thangaratinam S, Tan A, Knox E et al. Association between thy- thyroid peroxidase antibody-positive women undergoing assisted
roid autoantibodies and miscarriage and preterm birth: Meta- reproduction technologies: A prospective study. Hum Reprod
analysis of evidence. Br Med J 2011; 9: 342. 2005; 20: 1529–1533.
86. Korevaar TI, Schalekamp-Timmermans S, de Rijke YB et al. 94. Kilic S, Tasdemir N, Yilmaz N et al. The effect of anti-thyroid anti-
Hypothyroxinemia and TPO-antibody positivity are risk factors bodies on endometrial volume, embryo grade and IVF outcome.
for premature delivery: The generation R study. J Clin Endocrinol Gynecol Endocrinol 2008; 24: 649–655.
Metab 2013; 98: 4382–4390. 95. Shoenfeld Y, Carp HJ, Molina V et al. Autoantibodies and pre-
87. Negro R, Formoso G, Coppola L et al. Euthyroid women with diction of reproductive failure. Am J Reprod Immunol 2006; 56:
autoimmune disease undergoing assisted reproduction technolo- 337–344.
gies: The role of autoimmunity and thyroid function. J Endocrinol 96. Zhong YP, Ying Y, Wu HT et al. Relationship between antithyroid
Invest 2007; 30: 3–8. antibody and pregnancy outcome following in vitro fertilization
88. Negro R, Formoso G, Mangieri T et al. Levothyroxine treatment and embryo transfer. Int J Med Sci 2012; 9: 121–125.
in euthyroid pregnant women with autoimmune thyroid disease:
Effects on obstetrical complications. J Clin Endocrinol Metab
2006; 91(7): 2587–2591.
58 Quality of care for the woman
with diabetes at pregnancy
Alberto de Leiva, Rosa Corcoy, Alejandra de Leiva-Pérez, and Eulàlia Brugués
At a meeting held in St. Vincent, Italy, in October 1989,

Quality assessment and improvement representatives of government health departments and
in diabetes care patients’ organizations from all European countries met
diabetes experts to discuss a set of recommendations—
The aim of healthcare is to achieve the best health outcomes the St. Vincent Declaration,1 a joint initiative of the World
in the most efficient manner, and the challenge for today’s Health Organization–Europe and the International Diabetes
health delivery systems is to increase productivity and qual- Federation–Europe (WHO/IDF)—with the intention of
ity of care without increasing the economic costs. creating conditions allowing major reductions in deaths
Assessment of the quality of healthcare needs complex and the burden caused by diabetes mellitus. The declaration
measures of the structure (staff, equipment, organization), meant an important step forward in the general improve-
the process (technical quality), and the outcomes (effective- ment in the quality of delivery of diabetes healthcare.
ness, satisfaction, functional status, quality of life). One of the main targets of the declaration was to estab-
Healthcare delivery depends on efficient communica- lish monitoring and control systems using state-of-the-art
tion and cooperation among patients, healthcare services, information technology (IT) for quality assurance of dia-
and professionals; this matter is particularly critical with betes healthcare provision. A European group of experts
regard to chronic disorders in which effective shared care is was established to design and implement mechanisms for
pursued by multiple healthcare providers and professionals, the continuous improvement of the quality of diabetes care
enabling the patient to become actively involved in the pro- in Europe. The term continuous quality improvement was
cess of their care. The effective share of related information accepted to emphasize the progressive nature of the never-
is highly facilitated by the operation of an electronic patient ending process after reaching a determined standard. The
record and a telematics infrastructure. assessment requires the comparison of care with standards
Healthcare delivery is moving toward disease manage- that are derived from scientific evidence, consensus, good
ment, focused on a patient-oriented approach, illness pre- practice, and clinical experience.
vention promoting good health, and managing long-term The St. Vincent Declaration pointed out that self-monitoring
care, all of which require integrated activities from general- results in very effective control of treatment. Later in this
ists, specialists, and other healthcare professionals. This type chapter, the quality assessment of the procedures for self-
of care requires effective coordination and an interrelated, monitoring glycemic control will be reviewed in some detail.
multidisciplinary approach.
In addition, the implementation of effective strategies
for continuous quality improvement takes advantage of European DiabCare quality network
four main areas: (1) efficient use of healthcare resources
(e.g., eliminating practices that are clearly harmful, or with- A subgroup of the St. Vincent Declaration Steering
out known benefits); (2) linking clinical research to clini- Committee was established to develop instruments and
cal practice (evidence-based care); (3) application of new mechanisms for quality assurance in diabetes care. The first
concepts for improvement of care (the process of care must initiative of the DiabCare program was the development of
comply with the “best practice,” including solid methods to the St. Vincent Diabetes Dataset from three main sources:
monitor and assess the outcomes); and (4) changing clini- (1) EuroDiabeta2, a research project on modeling healthcare
cal practice (design of appropriate models for the man- and the implementation of IT in diabetes; (2) the specific rec-
agement of healthcare services, based on valid, scientific ommendations provided by the different working groups of
information). the St. Vincent Declaration Steering Committee; and (3) the
489
490 Quality of care for the woman with diabetes at pregnancy
advice provided by more than 130 expert diabetologists from data collection, data aggregation, and analysis (Figure 58.3)
213 European countries. of proper indicators (clinical, analytical, etc.) allows a local
DiabCare Basic Information Sheet (BIS) contains 141 evaluation (internal comparison); then, sending the aggre-
fields that include all the necessary data for the analy- gated data in anonymous fashion to a server, the compari-
sis of the quality of diabetes care (Figure 58.1). The person with all the other teams sharing the network is possible
tinent analysis provides the performance of care in both (external comparison).
aspects of process and outcomes (intermediate and final). After all of this, the members of the local quality circle are
Demographic data (age, sex, etc.) are required for a num- in the situation to propose and debate measures for quality
ber of purposes. True patient outcomes include the burden improvement. These measures are implemented in the fol-
of the medical end points of the St. Vincent Declaration lowing period and the evaluation of their effects will be then
(such as amputation, blindness). Symptoms of diabetes- analyzed, following the scheme of continuous assurance and
related problems (e.g., painful neuropathy, angina pec- improvement (Figure 58.4).
toris) are also recorded. Specific outcomes regarding The present authors’ use of the DiabCare program,
pregnancies are also included. For the measurement of adapted to a net environment in a hospital-based out-
quality of life, the DiabCare data sets only include infor- patient consultation, has provided a variety of benefits,
mation related to the duration of hospital admissions and including Diabetes Data Set exploitation as a registry, dia-
the number of days without the ability to perform normal betes-type characterization, assessment of self–blood glu-
activities. Assessment of diabetic complications (retinopa- cose monitoring (SBGM) status, St. Vincent Declaration
thy, nephropathy, neuropathy), cardiovascular risk factors, targets, treatment characterization, outcome for diabetic
pharmacological treatment, and metabolic outcomes (gly- pregnancies, completeness assessment of medical records,
cated hemoglobin [HbA1c], lipid profile) was considered cardiovascular risk factors, and identification of groups of
essential.4 The computer database (Figure 58.2) contains patients at risk.8
all the data items of the BIS and additional information On the basis of this information, a quality assurance
with easy access by a single key stroke. circle on diabetes care has been operating in the present
Once a year, at least, the data of all patients under care authors’ center since then, following the protocol proposed
must be collected in the DiabCare BIS. The performance by the EU Consortium DiabCare Quality Network, inte-
of the diabetes team is compared with the gold standards grated in a comprehensive disease management program
of the St. Vincent Declaration Program. The evaluation of (the Optidiab System). A published report about the infor-
the level of quality should cover the structure (housing, mation provided by the annual evaluation (the 141 param-
human resources, equipment, logistics) and the process eters of the DiabCare BIS) of >1000 subjects confirmed the
(the way the care is organized, from the first call to the burden of type 2 diabetes patients compared to type 1 dia-
treatment plan; the annual measurements of indicators, betes patients undergoing intensive and specialized care on
HbA1c, blood pressure, etc.; the way the treatment is initi- a regular basis.9
ated, use of antihypertensive drugs, cholesterol-lowering Interestingly, aggregated and compared data from
agents, etc.). the central server, integrating national centers from the
The DiabCare program was designed for those services European DiabCare Quality Network (22,000 patients),
not having a computer database but having access to com- lead to the conclusion that the long-term metabolic
puters. In 1991, the feasibility phase, integrating the infor- outcome of patients under intensive management in
mation from 4000 patients of 29 centers in 19 European European specialized centers is far short of achieving their
countries, was completed. After some minor modifications, desired goal (HbA1c < mean + four standard deviations
it gained widespread adoption by centers and local, regional, [4 SD] of the nondiabetic population); aggregated HbA1c
and national diabetes task forces all over Europe.5–7 levels (Diabetes Control and Complication Trial [DCCT]
The DiabCare Feasibility Study5 demonstrated the adjusted) recorded at the annual evaluation were optimal
achievements obtained by the implementation of local docu- for only 26.9% of cases, acceptable for 23.2%, and poor in
mentation compatible to the DiabCare Diabetes Data Set. the remaining 49.9% of subjects.10
It made possible the assessment of the quality of care and The DiabCare program allows a simple registration
to install regional/national quality networks, along with procedure for collecting basic data from pregnant diabetic
establishing a standard documentation to be used in various women; the system has also been demonstrated to be useful
healthcare settings in different countries. for limited evaluation of quality assurance in the broad field
A quality circle is a group of motivated and commit- of diabetes and pregnancy.11–13
ted people acting as a structured forum to solve on-the-job
problems affecting the quality of their work. Prerequisites
for the constitution of the circle are the political awareness DiabCare BIS for diabetes and
and the involvement of the decision-makers to get things pregnancy
going. The implementation of pilots or demonstration proj-
ects make clear what the benefits are and the economic cost. One of the main recommendations of the St. Vincent
The quality circle must select targets according to the Declaration was the following: “Achieve pregnancy outcome
local health requirements. The information gathered after in diabetic women that approximates to that of nondiabetic
DiabCare BIS for diabetes and pregnancy 491
Diabetes
Basic Information Sheet
Internat. Centre: Implementation of the St. Vincent Declaration
Basic Patient N+. : Initials: Date of Birth Sex: M F
Data 1st name last name Mon. Year
IDDM NIDDM Other Diabetes OAD Insulin
since: since: since:
Reason for Consultation Routine visit Y Stabilisation Y Complications Other

Consultation/ Y Y
Admission or: Admission Newly diagnosed Y Pregnancy Y Emergency Y
Pregnancies Ending within last
Y N Normal Abortions Major malfomat. Perinatal deaths
12 months
Risk Factors Smoker N if yes: cig./day Alcohol N g/day
current status Y Y
Self Monitoring Self-monitoring Y N Blood glucose (nr/week) Urin glucose (nr/week)
Education/ Healthy eating Y N Foot care Y N Complications Y N Self-monitoring Y N

Diab. Pat. Org. Hypoglycemia. Y N Self adjustment Y N Member of a diabetic patient organisation Y N
Measurements
most recent Weight kg Blood Press. / mm Hg Cholesterol
value in the Height cm HDL-Cholest.
last 12 months BG Creatinine
Triglycerides
HbA1 % Microalbum.
Fasting Y N
HbA1c % Proteinuria
Blindness Y N if yes: occurred last 12 mo. Y N End-stage renal fail. Y N if yes: occurred last 12 mo. Y N
St. Vincent MI/CABG/Angiopl. N if yes: occurred last 12 mo. Y N N if yes: occurred last 12 mo. Y N
Targets Y Leg amput. ab. ankle Y
Cerebral stroke Y N if yes: occurred last 12 mo. Y N Leg amput. bel. ankle Y N if yes: occurred last 12 mo. Y N
Symptoms Postural hypotension Y N Anginal chest pain N Peripheral neuropathy Y N Leg claudication N
last 12 months Y Y
Examinations YES Examined last 12 months N FEET Examined last 12 months N

Y Y
left right left right
Photocoagulation last 12 months Y N Y N
Cataract N N Normal vibration sensitivity Y N Y N
Y Y
N N Normal pin prick sensitivity Y N Y N
Retina seen Y Y
N N Foot pulses present Y N Y N
– if yes: Maculopathy Y Y
Retinopathy Y N Y N
– if Rp. : Non-proliferative Rp. Y N Y N Healed ulcer Y N Y N
Preproliferative Rp. Y N Y N Acute ulcer/gangrene Y N Y N
Proliferative Rp. N N Bypass/angioplasty Y N Y N
Y Y
Advanced diab. eye disease Y N Y N
Visual acuity
Quality of Life Hypoglycemia Hypoglycemia Sick leave Hospital days

Emergencies (no/yr) (no/yr) (??) (??)
Management up to now from now on up to now from now on

Diet only Y N Y N
N+. of Insulin-
Biguanides Since Y N Y N Injections/day
Sulphonylureas Since Y N Y N Insulin-pump Y N Y N
Glucosid. inhibit. Since Y N Y N Other treatment Y N Y N
Additional Hypertension Cardiac failure Isch. heart dis. Dyslipidaemia Nephropathy Neuropathy Other
Treatment Y N Y N Y N Y N Y N Y N Y N Y N Y N Y N Y N Y N Y N Y N
Figure 58.1 Basic Information Sheet, DiabCare.

Figure 58.2 DiabCare data for windows.
women.” In consequence, WHO/IDF guidelines for care and Blood glucose monitoring in diabetic
management of pregnant diabetic women have been proposed
by an invited group of international experts in the field.14 pregnancies
The document brought attention to the important dif-
Portable meters are used by healthcare workers and by
ferences in the provision of diabetes and obstetrical care in
patients. Because of their imprecision and variability, they
different European countries. Specifically, the relevance of
should not be used for diagnosing diabetes and their value in
intensive metabolic care before conception and during preg-
screening must be limited.
nancy and parturition and the needs of special training and
SBGM is recommended for all insulin-treated patients.
education of the diabetic women contemplating pregnancy
Glucose can be measured in whole blood, serum, or plasma,
were addressed. For the purpose of developing the quality
but plasma is recommended for diagnosis. Although red
assurance program, a DiabCare BIS for diabetes and preg-
blood cells are freely permeable to glucose, the concentration
nancy was proposed by members of the WHO/IDF Working
of water in plasma is around 11% higher than that of whole
Group on Pregnancy Outcomes in the Diabetic Woman
blood; as a consequence, the glucose concentration in plasma
(Figure 58.5), with data fields addressing diabetes diagnosis,
is higher than in whole blood (being the hematocrit normal).
obstetrical history, prepregnancy counseling, status at enter-
The glucose concentration decreases with time in the assay
ing the specialized interdisciplinary clinic, maternal and
tube because of in vitro glycolysis (on average by 5%–7%,
newborn outcomes, and reclassification after pregnancy.
or 0.6 mmol/L or 10 mg/dL/hour), which can be attenu-
The OBStetrical Quality Indicators and Data (OBSQID)
ated by inhibition of enolase with sodium fluoride used in
Perinatal Aggregated Data (PAD) protocol, mainly focused
combination with anticoagulants. Therefore, when plasma
on outcomes, represents a valid alternative proposed by the
glucose is going to be analyzed, plasma should be separated
Quality of Care and Technologies Program, WHO–Europe,
from cells within 1 hour. Glucose is almost exclusively mea-
being exploited for epidemiological studies.
sured by enzymatic methods (hexokinase, glucose oxidase).
Blood glucose monitoring in diabetic pregnancies 493
Q-Networking
Internet client
Local
client
Modem Modem
WWW—Server
Dial up
server
TCP/IP (first level
node)
- Guidelines Modem TCP/IP

- Benchmarking (Internet)
Aggregation server
(second level node)
WHO—Node
Qualicare-server
(top level node)
Figure 58.3 DiabCare Q-Net, system architecture.
Guidelines
Quality-
Patient-care
circles
Feedback
Documentation communication
Evaluation
Indicators
Quality-
Results
benchmarking
Figure 58.4 Operation of the DiabCare quality circle.

Around Conception:
Basic Information Sheet for Diabetes and Pregnancy
Insulin: O Yes O No OAD: O Yes O No Lowering BP Med: O Yes O No ACE Inhibitor: O Yes O No
Data Set Number: Country: Centre Name:

Mother
Treating Physician: Date of record:
Gestational Week of Termination: Third Trimester Severe Hypos:
Weight Gain in Pregnancy: HbA1c: % Ketoacidotic Episodes:
Basic Personal Data Self Measure BG/wk Last Trimester: Acute Urinary Tract Infections:
DCCT/UKPDS: O Yes O No
Family Name: Given Name: Date of Birth:

GDM: O Yes O No
Residence: O Urban O Rural Distance from Clinic (km): Single: O Yes O No Progress of: Chronic Hypertension: O Yes O No
Preconceptional:
Education: Retinopathy: O Yes O No Chronic HT and PIH: O Yes O No
First Trimester:
O None O Primary O Secondary O University Nephropathy: O Yes O No Pregnancy Induced
Hypertension (PIH): O Yes O No Second Trimester:
Diabetes Diagnosis Third Trimester:
DM Diagnosis: Diagnosis Based on: Diagnosis: Birth

O DM-Type 1 O GDM O NDDG criteria
Date: Stillbirth: O Yes O No
O LADA O Secondary-DM O IADPSG criteria
Gestational Week: Abortion: Labour: Maternal Discharge:
O DM-Type 2 O Other
O Induced Spontaneous: O Yes O No Instrument Used: O Yes O No O Alive
O Spontaneous Induced: O Yes O No Cesarean Section: O Yes O No O Dead
Past Obstetrical History
Previous Pregnancies: Induced Abortions:
Newborn
Total Live Births: Stillbirths: Early Infant Mortality:
Spontaneous Abortions: Pregnancy with Diabetes: Late Infant Mortality: Male: Female: Neonatal Weight: g Length: cm Head circle: cm
Deaths in First Year: Discharge:

Apgar (1 min): Apgar (5 min):
Born with Major Malformation: O Healthy O Ill O Dead
Pre-Pregnancy Counselling Pathologies:

Obstetric Trauma: O Yes O No Neonatal Asphyxia: O Yes O No Severe Hypoglycemia: O Yes O No
Pre-Pregnancy Counselling: O Yes O No HbA1c: %
Hyperbilirubinemia: O Yes O No Fetal Distress: O Yes O No Hyaline Membrane Disease: O Yes O No
Structured Information: O Yes O No DCCT/UKPDS criteria (standard): O Yes O No
Severe Hypocalcaemia: O Yes O No Polycythemia: O Yes O No
Microalbuminuria: O Yes O No Proteinuria: O Yes O No

Date of Referral for Index Pregnancy: Neonatal Malformations:
mg/24h mg/24h Cardiovascular: O Yes O No Genitourinary: O Yes O No Skeletal: O Yes O No
Neurotube: O Yes O No Gastrointestinal: O Yes O No Other: O Yes O No
Booking
Feeding: O Breast O Bottle
Date of First Visit to the Specialist: Pregnancy Advised Against: O Yes O No

Last Menstruation Period: Planned Pregnancy: O Yes O No Reclassification
Conception: Booking: Has the Type of Diabetes Been Reclassified? O Yes O No
HbA1c: % HbA1c: % Postpartum: After 5 years:

DCCT/UKPDS criteria (standard): O Yes O No DCCT/UKPDS criteria (standard): O Yes O No O Normal GT O DM-Type 2 O Normal GT O DM-Type 2
O DM-Type 1 O Other O DM-Type 1 O Other

Microalbuminuria before pregnancy: O Yes O No Proteinuria before pregnancy: O Yes O No O LADA O LADA
mg/24h mg/24h
Previous Retina Laser Treated: O Yes O No
Figure 58.5 DiabCare Basic Information Sheet for diabetes and pregnancy. (Adapted from Tamás, G. et al., Av. Diabetol., 5, 137,
1992; de Leiva, A. and Hernando, M.E., Diabetes Care, 32(Suppl. 2), S211, 2009.)
For plasma glucose, a coefficient of variation <2.2% is recom- reports not recorded in the memory of the meter), usually
mended as a target for imprecision. depicting a trend toward correcting results of readings.15,16
The evaluation of SBGM devices may be misleading; in Of course, adequate training, visual acuity, hypoxia,
general, several units should be tested to explore interdevice altitude, hemolysis, hematocrit, hypertriglyceridemia,
variability. The evaluation should include the analysis of the adequate sample volume,17 and other technical elements
mean difference of the device reading, with respect to a ref- can influence the results. Reinforcing patient education at
erence procedure at low, medium, and high blood glucose regular clinic visits, evaluating his or her technique, and
concentrations. Other items of the evaluation will include cus- frequent comparison of SBGM profiles with concurrent
tomer acceptability (size, weight, portability, calibration, dura- laboratory blood glucose analysis will assess the reliability
tion of a test performance, economic cost). Then, a validation of patient reports.
protocol using an adequate sample size of recruited patients Many different brands of meters are commercially avail-
will follow, covering a wide range of blood glucose levels from able. They use strips containing glucose oxidase or hexokinase;
the hypoglycemic range to extreme hyperglycemia. some meters contain a porous membrane that separates eryth-
The utilization of memory meters has shown that patients rocytes, the analysis being carried out in plasma. The meters
often make incomplete recordings of their daily blood glu- provide a digital readout using reflectance photometry or elec-
cose profiles (inaccurate readings, omission of outliers, false trochemistry for the measurements of glucose concentration.
Impact of advanced technologies in the quality of care of diabetic pregnancies 495
subcutaneously, reverse iontophoresis uses a low-level elec-

Table 58.1 Blood glucose targets
trical current, which by electro-osmosis moves glucose
across the skin, with the glucose concentration measured by
The American Diabetes Association and the American
College of Obstetricians and Gynecologists recommend a glucose-oxidase detector.25,26 Total noninvasive technology
the following blood glucose targets: <95 mg/dL at for glucose sensing, including techniques of near-infrared
fasting, <140 mg/dL at 1 hour after meals, and spectroscopy, light scattering, and photoacoustic spectros-
<120 mg/dL at 2 hours after meals. copy, is in progress. It is anticipated that important prog-
ress in these methods for noninvasive or minimally invasive
glucose monitoring will be made in the near future.
There is a wide variability in the performance of the differ-
ent meters and the level of imprecision remains high. The
American Diabetes Association (ADA) goal of analytical devi- Impact of advanced technologies
ation remains <5% from reference values.18–20
SBGM should be performed at least four times per day in
in the quality of care of diabetic
patients with type 1 DM. It has been demonstrated that moni- pregnancies
toring with a lower frequency than this is associated with dete-
rioration of glycemic control. In both GDM and pregestational Telemedicine in pregnancy complicated by diabetes:
diabetes, it is usually recommended to measure blood glucose Overview on quality care
levels at fasting and 1 or 2 hours after meals (Table 58.1).21,22 One of the first experiences of telemedicine (TM) applied to
Noninvasive or minimally invasive continuous monitor- diabetes management was the DIACRONO device, a micro-
ing of blood glucose is a high priority (both to detect unsus- computer that allowed data interchange between doctor and
pected hypoglycemia and as a further step in the development patient; it already included the function of downloading glu-
of an artificial pancreas), allowing automatic measurement cose data from the glucometer.27 The same group of research-
of blood glucose and adjustment of insulin administration ers developed DIABTel, which is made possible by remote
(Figure 58.6). The method for sampling in minimally inva- supervision and teleconsulting to improve patient–physi-
sive systems takes advantage of the correlation between cian communication28 and to provide decision support in
the concentration of glucose in the interstitial fluid and diabetes management.29 Following initiatives used Internet:
in blood.23,24 Whereas microdialysis systems are inserted TIDDM,30 M2DM,31 all of them EU Projects, which, finally,
Optimal accuracy criteria Sensor Meter

No. of paired Correlation Mean absolute No. of Average STO Sensor range No. of Average STO Meter range
sensor Mtr coefficient difference readings (mg/dL) (mg/dL) (mg/dL) Readings (mg/dL) (mg/dL) (mg/dL)
Date/Time readings (r) (%)
13-Nov-00 x 6 30 29 154 164 67 40–271 6 159 32 100–204
14-Nov-00 8 .97 11 288 107 39 45–202 8 102 46 49–183
15-Nov-00 10 .95 13 288 164 59 48–294 10 129 73 43–283
16-Nov-00 x 0 118 202 49 80–271 1 179 0 179–179
All days 25 .87 16 848 150 63 40–294 25 130 60 43–283
400
Meter value paired meter value Sensor value
350 Insulin Meal Exercise Other
300
Glucose concentration (mg/dL)
250
200
150
100
50
–50
12.00 a.m. 4.00 a.m. 8.00 a.m. 12.00 p.m. 4.00 p.m. 8.00 p.m. 12.00 a.m.
Figure 58.6 Registry of continuous glucose monitoring in interstitial subcutaneous tissue.

integrated for the first time, continuous glucose monitoring So far, clinical trials show that closed-loop insulin delivery
(CGM) into a smart TM platform and various predictive in diabetic women during pregnancy and out of pregnancy
algorithms with positive results in glycemic control and improves stability of glucose control at night and prevents
glucose variability (EU INCA Project).32,33 the “dawn phenomenon,” but delayed insulin action at meal-
A recent systematic review and meta-analysis has investi- time results in hyper-/hypoglycemia. Therefore, in spite of
gated the impact of TM on the management of diabetes (4207 the great research effort invested into the development of
patients, type 1 or 2). TM was associated with a statistically closed-loop control algorithms, their continuous ambulatory
significant decline in HbA1c (mean difference −0.44%). Low utilization is still far from being a short-term reality; patient
density cholesterol level (LDL-c) was reduced in 6.6 mg/dL, safety reasons for this initiative have also been argued.
just a tendency to reduce body mass index (BMI), and there The telemedical artificial pancreas (TAP) is built on two
were no changes in either systolic or diastolic blood pressure.34 interlinked loops. The “personal loop” allows wireless commu-
The use of TM can facilitate the management of diabetic nication between the PA—a personal device assistant-based
pregnancies in multiple aspects: smart assistant—an insulin pump, and a CGM device. The
“remote loop” connects the diabetic subject to the healthcare
1. Education of patients and relatives professional via the PA and its wireless connection to the tele-
2. Teleconsultations (e-mail, online text messages, mobile medicine central server (TMCS). The platform implies two
phones, etc.) control modes:
3. Videoconferences, excellent tools for training and com-
munication to multiple patients 1. The patient decides changes in the insulin pump program
4. Greater involvement of other healthcare professionals by using the information coming from the glucosensor
(dietitians, midwifes), saving physicians’ time, without (free mode).
affecting the quality of care35 2. The patient gets advice, through the PA, on the insulin
5. Reducing the number of visits to diabetes clinic and bolus before each main meal (advisory mode). The patient
emergency department uses the PA to register and transfer data related to tim-
ing, pre- and postprandial glucose excursions, amount of
Having real-time access to medical devices can improve ingested carbohydrates, and additional relevant informa-
decision-support tools with immediate feedback to patients. tion (e.g., previous exercise). From this information, the ad
A personal assistant (PA) integrated in a portable device hoc algorithm computes the corresponding insulin dose
such as a smartphone is able to communicate different medi- that has to be approved by the physician, through the web
cal devices and/or sensors in a personal wireless network.36 interface, before it is shown to the patient (Figure 58.7).
The system is able to support diabetes care in a distributed The patient uses the PA to get results of the proposed
and ambulatory environment.37 The mobile application can insulin bolus and decides to follow it or not.43
provide advice from automatic processing tools based on
monitoring data as well as expert feedback to patients. The
whole system has been successfully tested in two random- Telemedical control-loop strategies
ized and crossover clinical experiments with type 1 diabetic
1
patients showing that the wireless PA joined to the TM sys- 2
Patient control
tem allows better glycemic control in insulin pump–treated Doctor control
diabetic patients.38 Patient
Healthcare
professional
Continuous subcutaneous insulin infusion
Continuous subcutaneous insulin infusion (CSII) and mul-
tiple doses of insulin (MDI) are both safe and valid treat-
ments for women with type 1 diabetes during pregnancy.39,40
Nevertheless, the main predictors of glucose optimization
in CSII therapy are high HbA1c levels before CSII, pregesta- 3
tional care, younger age, and the use of more basal segments 4 Control algorithms
over the daily insulin profile.41 Control algorithms (remote loop)
(personal loop)
Closed-loop insulin delivery
Telemedical
In recent years, the acceptable accuracy of continuous glucose central server
sensing in association with the model predictive control (MPC)
algorithm has been demonstrated; overnight closed-loop insu-
lin delivery under these premises can be used safely in preg- Figure 58.7 Telemedical artificial pancreas. (Adapted from
Telemedical Artificial Pancreas; EC-INCA Project: Gómez, E.J.
nancy.42 Nevertheless, large multicenter randomized studies et al., IEEE Trans. Inf. Technol. Biomed., 12(4), 470, 2008; de
are needed to confirm the benefits of closed-loop insulin deliv- Leiva, A. and Hernando, M.E., Diabetes Care, 32(Suppl. 2),
ery in comparison with sensor-augmented pump therapy. S211, 2009.)
About using HbA1c for the diagnosis of diabetes mellitus 497
The system has been, so far, technically validated and a clini- Several hemoglobinopathies interfere with some assay
cal pilot has been also satisfactorily carried out. A multi- methods; the results can be falsely increased or decreased;
center randomized ambulatory crossover study is already in non-hemoglobin-based methods for assessing long-term
process. glycemic control may represent useful alternatives in these
Very recently, the safety and effectiveness of a bihor- circumstances.46
monal, automated pancreas (“bionic pancreas”), allowing European guidelines have recommended the classifi-
the precise delivery of insulin and glucagon, have been cation of blood glucose control by the number of SD; the
compared with insulin pump (control). The clinical experi- experimental value is from the nondiabetic mean for the
ment was validated for 5 days in 20 adults and 32 adoles- particular assay.47
cents with type 1 diabetes. The bionic pancreas achieved The International Federation of Clinical Chemistry
improved mean glycemic levels with less frequent hypogly- (IFCC) has organized a working party to develop a scientifi-
cemic episodes.44 cally based method of the production of a primary reference.
Electrospray ionization mass spectrometry (ESIMS) has been
demonstrated to be a precise measurement of HbA1c, in par-
ticular glycation of the beta chain, which has been proposed
Glycated hemoglobin (HbA1c) and as a robust procedure for calibration purposes. In a protocol
glycation of human serum albumin performed with 1022 patients, the comparison of the ESIMS
with the ion-exchange chromatographic procedure showed
More than 40 years ago, it was observed that normal adult excellent agreement, with values, on average, 0.7% lower
hemoglobin could be separated by chromatographic pro- with ESIMS. The comparison with DCCT-corrected ion-
cedures into major and various minor components. In one exchange values gave good agreement, with ESIMS showing
of the minor components—HbA1c—glucose was attached, an overall lower value of mean 0.4%.48
nonenzymatically, to the terminal N-valine of the beta chain There is a general agreement that the new mass-
of HbA0. In the following years, numerous assays were spectroscopy-based method appears to be more accurate and
developed to measure HbA1c levels. to reflect “true” HbA1c. It yields normal (nondiabetic) val-
The average lifespan of erythrocytes is 100–120 days. ues that are significantly lower than those used by the NGSP,
Measuring HbA1c offers an accurate estimation of the aver- DCCT, and UKPDS (3%–5% vs. 4%–6%).
age blood glucose concentration of the past 2–3 months. The Ames DCA 2000 Analyzer measures HbA1c by an
HbA1c is used as an integrated estimation of mean glyce- agglutination inhibition immunoassay, allowing results in
mia and as a marker of risk for the development of diabetes 6 minutes with 1 mL of blood. The Ames DCA 2000 analyzer
complications. offers reliable results, although showing a trend to lightly
At present, there are more than 30 HbA1c assay methods underestimate the results in comparison with HPLC.49 This
available. Certain methods quantify HbA1c based on charge analyzer gave valid and reliable results when operated by
differences of the glycated components (cation exchange nonmedical personnel. The Primus CLC330 provides an
chromatography, agar gel electrophoresis). Other methods HPLC near-patient HbA1c method, comparable in preci-
analyze structural differences between glycated and nongly- sion and accuracy to the Ames DCA 2000 analyzer.50 Near-
cated components (affinity chromatography, immunoassay). patient testing for HbA1c has practical clinical use in the
In addition, various methods quantify total glycated hemo- diabetes clinic, avoiding the need for a second appointment
globin, including HbA1c and other hemoglobin–glucose and allowing immediate changes in therapy.
adducts.
In 1996, the American Association of Clinical Chemists
(AACC), in collaboration with the ADA, established the
National Glycohemoglobin Standardization Program About using HbA1c for the diagnosis
(NGSP). It was decided to adopt the high-performance liq- of diabetes mellitus
uid chromatography (HPLC) reference method used in the
DCCT (Bio-Rex 70; between run coefficient of variation In 2010, the ADA introduced a value of HbA1c equal or
[CV] <3%), as the designated comparison method. superior to 6.5% as diagnostic criteria of diabetes,51 primar-
The UKPDS incorporated the same standardization ily based upon the increased accuracy and reliability of the
method; therefore, its HbA1c reports were compatible with improvement of A1c assays through the NGSP. A1c correlates
those of DCCT. More than 90% of U.S. laboratories, and well with both mean glucose concentration52 and complica-
many others worldwide, are using this procedure for stan- tions of diabetes. Nevertheless, the new recommendations
dardization purposes.45 acknowledge that a cutoff for A1c equal or above 6.5% clas-
Any condition that shortens erythrocyte survival (hemo- sifies one-third fewer individuals as having diabetes in com-
lytic anemia, acute blood loss) falsely lowers HbA1c test parison with a fasting plasma glucose (FPG) value equal or
results; in contrast, iron-deficiency anemia increases the above 126 mg/dL. Some advantages of the estimation of A1c
value. Glycation may also vary between patients with similar include the following: (1) fasting is not needed, (2) A1c is not
capillary blood glucose levels and appears to be lower in sub- subjected to acute perturbations (stress, exercise, diet, etc.),
jects with a higher BMI. and (3) A1c can be used for both diagnosis and initiation of
diabetes monitoring. On the contrary, (1) standardization Standardization Program. The mean (SD) A1c result of
of A1c assay is poor; (2) A1c depicts important differences 21,064 HAPO Study participants was 4.79 (0.40)%. 57
in various ethnic groups; (3) the cost is elevated when com- In a recently published study of 335 gravidas who received
pared to glucose assay; (4) A1c levels vary not only according a 3-hour, 100 g OGTT at 24–32 weeks, it was discovered that
to glycemia but also according to the erythrocyte turnover HbA1c concentrations gradually increased in diet-treated
rate (anemia, hemoglobinopathies, malaria, etc); (5) A1c lev- and insulin-treated GDM gravidas compared with non-
els of 6.0%–6.5% do not predict diabetes as effectively as FPG GDM gravidas. The HbA1c value corresponding to an FPG
and 2-hour OGTT; and (6) using A1c may delay the diagno- of 5.1 mmol/L (diagnostic of GDM) was 2 mmol/L (0.2%)
sis of DM in more than 50% of cases.53 higher if sampling occurred at 29–32 vs. 24–28 weeks of
pregnancy, indicating that HbA1c varies with gestational
age.58 Various other reports have recommended to incorpo-
NGSP Network rate HbA1c in the diagnosing criteria. The WHO considers
The NGSP was implemented to standardize HbA1c results that HbA1c of 48 mmol/mol (6.5%) and above is diagnostic
to those of the DCCT/UKPDS. There have been consid- of DM. Individuals with HbA1c values below such a cut-
erable variations in the procedure for reporting glycated off may still have diabetes; abnormal Hb variants, anemia,
hemoglobin; in addition, an important variability among altered lifespan of the red cell, aging, and ethnicity, among
values reported in each category was displayed, which was other variables, ought to be taken into consideration. Clinical
associated with confusion in clinical practice. The DCCT judgment is required for each individual to avoid inappro-
tried to avoid the variability among results by having all priate exclusion or inclusion in the category of diabetes.
analyses of HbA1c performed in a single laboratory. Then,
the AACC decided to standardize the HbA1c assay to
facilitate that all clinical laboratories are able to compare
their own results with those of DCCT/UKPDS. For this
Glycated serum proteins
purpose, a network of NGSP consisted in a central pri- Mostly albumin and other serum proteins undergo the pro-
mary reference laboratory (CPRL), using the DCCT/EDIC cess of glycation. Glycated albumin (GA) is a ketoamine
Bio-Rex 70 HPLC method, additional primary reference formed via nonenzymatic glycation of serum albumin. The
laboratories, using the same method, and seven second- turnover of serum albumin depicts a half-life of 25 days; GA
ary reference laboratories (SRLs) using commercial meth- provides an index of a mean glycemic level of a shorter inter-
ods (ion-exchange HPLC, immunoassay, boronate-affinity val than HbA1c. The fructosamine assay is the most widely
HPLC, capillary electrophoresis). 54 used method for estimating glycated serum proteins.
The IFCC also created a working group to develop an Nevertheless, although fructosamine levels correlate with
internationally accepted reference method and HbA1c mate- HbA1c levels within a population, transference cannot apply
rials for the assay. Each laboratory of the IFCC Network for individual values.59 Also, changes in serum proteins
runs one of the two approved IFCC methods (HPLC–mass affect the readings of fructosamine60; the technique is unre-
spectrometry or HPLC–capillary electrophoresis). Results liable in diabetic patients with renal failure,61 liver cirrhosis,
obtained by the IFCC method correlates quite well with and nephrotic syndrome.62
NGSP/DCCT results. GA can be used for patients with anemia or hemoglobin-
ADA, EASD, and IDF reached the agreement that labo- opathies. It can be also used to estimate the effectiveness of
ratories should report an estimated average glucose (eAG).55 treatment before starting or changing medications for dia-
A linear relationship was established between the HbA1c betic patients. As GA is potentially an atherogenic protein,
and mean blood glucose, but the observation of too much GA measurement may be included in the routine protocol to
variability in the HbA1c/mean BG relationship did not make screen for both diabetes and atherosclerosis.63
the report of eAG advisable (it has not been accepted outside
the United States).
A new International Consensus in 2010 indicated that
HbA1c should be reported in both DCCT units (%HbA1c) Assessment of the effectiveness of
and NGSP units (millimoles/M).56 After all of these efforts, SBGM in diabetic pregnancies
the quality of HbA1c test meets the clinical needs.
The Hyperglycemia and Adverse Pregnancy Outcome Various randomized controlled trials (RCTs)64–68 and case
(HAPO) Study compared associations of maternal glu- series studies 69–72 carried out in either diabetes or obstet-
cose and A1c with adverse outcomes. A1c was measured rics departments of university hospitals have evaluated
by an HPLC method (Biomen HA 8140 instrument). The the clinical effectiveness of SBGM in women with GDM
coefficient of variation ranged from 2.8% to 5.0% across or diabetic pregnancies. Blood glucose and HbA1c deter-
the range of A1c values. For external quality control, minations, as well as maternal and fetal outcomes, were
the Central Laboratory participated in the European recorded. Some studies examined the costs of hospital care
Reference Laboratory Program, which allows standard- and home monitoring and compared the costs for a control
ization of DCCT values, and the U.K. National Glycation group receiving standard care. The largest of all protocols
Quality of life in women with GDM: The Italian DAWN Pregnancy Study 499
included a population of 153 women with GDM in the at 24 weeks of pregnancy, performed after an overnight fast.
experimental group and 2153 nondiabetics in the control Overt diabetes will be diagnosed if FPG equals or exceeds
group. The main goal of the case series studies was to assess 7.0 mmol/L (126 mg/dL) and GDM if one or more val-
the feasibility of managing pregnant women at home using ues equal or exceed the following threshold values: FPG,
SBGM. There was general agreement that women were 5.1 mmol/L (92 mg/dL); 1-hour plasma glucose, 10.0 mmol/L
able to achieve satisfactory blood glucose profiles at home (180 mg/dL); and 2-hour plasma glucose, 8.5 mmol/L
using SBGM; hospital utilization was lower, and infant (153 mg/dL).74 Accepting IADPSG recommendations means
birth weights and indicators of macrosomia were also more accurate and timely diagnosis of pregestational diabe-
more favorable in those women. Main findings from the tes and of GDM and probably will double the frequency of
RCT demonstrated that patients with type 1 DM managed pregnant women with diabetes. The economic impact of this
by SBGM at home obtained similar results regarding gly- new approach deserves thorough investigation and is likely
cemic control to those patients under intensive control in to differ across countries.75
the hospital. Maternal and fetal outcomes were also simi- The WHO in a recent report has confirmed the increased
lar in both groups. Women preferred lower use of hospi- prevalence of GDM accepting the new IADPSG crite-
tal and home management with SBGM. In particular for ria, which has implications for health services and human
GDM, monitoring blood glucose after meals, rather than and material resources. The IDF estimates that the global
before, contributed to better metabolic control and bet- prevalence of hyperglycemia in pregnancy in women aged
ter fetal outcomes (there were fewer cesarean sections for 20–49 years is 16.9%. Some health services throughout the
cephalopelvic disproportion, fewer cases of macrosomia world may experience difficulties in meeting these demands;
and large-for-gestational-age infants, and fewer episodes therefore, it is important for each health service to assess
of neonatal hypoglycemia). In addition, women who uti- the burden of diagnosing and optimally treating pregnancy
lized SBGM were less likely to require hospital admission, complicated by diabetes before adopting the convenient deci-
leading to a substantial cost saving. sion to implement programs to test and treat the condition.76
Impact of the Hyperglycemia and Quality of life in women with GDM:

Adverse Pregnancy Outcome Study The Italian DAWN Pregnancy Study
in the quality of care of diabetic The diagnosis of diabetes is a cause of anxiety, particularly
pregnancies in the case of women in pregnancy. IDF has promoted the
Diabetes Attitudes, Wishes and Needs (DAWN) study to
The HAPO Study, a prospective, blinded, multinational, assess the quality of life of diabetic patients and the satisfac-
observational study, published in 2008, investigated almost tion of healthcare professionals involved in the management
25,000 pregnancies.73 The study demonstrated a continu- of diabetes.77,78
ous association between elevated maternal blood glucose The Italian Dawn Pregnancy Study represents the first
concentrations and the risk of pregnancy complications initiative of this kind carried out in the area of diabetes and
(large for gestational age, preeclampsia, primary cesarean pregnancy. Ten specialized Italian centers with experience
section, neonatal hypoglycemia, elevated cord blood serum in the management of diabetic pregnancy collaborated in
C-peptide, premature delivery, and hyperbilirubinemia). the investigation. Recruited women were 198 Italians and
The International Association of Diabetes and Pregnancy 88 immigrants (from 27 nationalities) who answered a sur-
Study Group (IADPSG) developed, after the publication of vey containing 51 items.79
the HAPO Study, a new strategy for detection and diagnosis Most women (94%) responded they were happy about the
of hyperglycemic disorders in pregnancy in two phases. The quality of care provided in the different diabetes centers.
first is detection of women with overt diabetes not previously The main reasons for anxiety during pregnancy included
diagnosed, measuring FPG, A1c, or random plasma glucose the fear of congenital malformations and the risk for future
on all or only on high-risk women. One or more values diabetes in the newborn, as well as the potential negative
equaling or exceeding the following thresholds—7 mmol/L consequences in babies related to the insulin administration
(126 mg/dL), 6.5% (DCCT/UKPDS standardized), and 11.1 to the mother. The majority of women (73%) expressed their
mmol/L (200 mg/dL) (it should be confirmed by FPG or perception of suboptimal cooperation between gynecolo-
A1c)—will make the diagnosis of overt diabetes in preg- gists and diabetes specialists. The study revealed the effort of
nancy, and the woman will be treated as with preexisting immigrants (particularly Muslims) to change eating habits
DM. If results are not diagnostic of overt diabetes and FPG and related religious recommendations about diet in preg-
equals or exceeds 5.1 mmol/L (92 mg/dL), the diagnosis nancy. A main recommendation derived from the investiga-
of GDM will be made. If FPG concentration is lower than tion was to optimize the communication between healthcare
5.1 mmol/L (92 mg/dL), the woman will be investigated personnel and patients, as well as between gynecologists and
again in the second phase. The second phase is a 75 g OGTT diabetes specialists.
REFERENCES
1. World Health Organization (Europe) and International Diabetes 23. Bolinder J, Ungerstedt U, Arner P. Microdialysis measurement
Federation (Europe). Diabetes care and research in Europe: The of the absolute glucose concentration in subcutaneous adi-
St Vincent Declaration. Diabet Med 1990; 7: 360. pose tissue allowing glucose monitoring in diabetic patients.
2. Eurodiabeta. Information technology for diabetes care in Europe: Diabetologia 1992; 35: 1177–1180.
The Eurodiabeta initiative. Diabet Med 1990; 7: 639–650. 24. Hashiguchi Y, Sakakida M, Nishida K et al. Development of a
3. Krans HMJ, Porta M, Kee H. Diabetes care and research in miniaturized glucose monitoring system by combining a needle-
Europe. The St Vincent Declaration action programme. G Ital type glucose sensor with microdialysis sampling method. Long-
Diabetologia 1992; 12(Suppl. 2): 1–56. term subcutaneous tissue glucose monitoring in ambulatory
4. Piwernetz K, Home PD, Snorgaard O et al., for the DiabCare diabetic patients. Diabetes Care 1994; 17: 387–396.
Monitoring Group of the St Vincent Declaration Steering 25. Tamada JA, Garg J, Jovanovic L et al. Noninvasive glucose moni-
Committee. Monitoring the targets of the St Vincent Declaration toring: Comprehensive clinical results. Cygnus Research Team.
and the implementation of Quality Management in Diabetes J Am Med Assoc 1999; 282: 1839–1844.
Care: The DiabCare Initiative. Diabet Med 1993; 10: 371–377. 26. Garg SK, Potts RO, Ackerman NR et al. Correlation of finger-
5. World Health Organization (WHO). Diabetes care and research stick blood glucose measurements with Gluco Watch Biographer
in Europe: Implementation of the St Vincent Declaration. Report glucose results in young subjects with type 1 diabetes. Diabetes
on a joint WHO/IDF Meeting, Budapest, Hungary, March 9–11, Care 1999; 22: 1708–1714.
1992. (WHO Regional Office for Europe: Copenhagen, Denmark, 27. Gómez-Aguilera E, del Pozo F, Serrano RM. DIACRONO: A
1992) pp. 1–25 (EUR/ICP/ CLR0550235g). new portable microcomputer system for diabetes management.
6. World Health Organization (WHO). Recommendations to facili- In: Proceedings of Ninth Annual International IEEE Engineering
tate the implementation of the St Vincent Declaration initiatives in Medicine and Biological Society, Chicago, IL, 1987, Vol. 3,
by national, regional, and local diabetes task forces. Report on the pp. 1231–1232.
Consensus Workshop, Oslo, Norway, June 26–27, 1992. (WHO 28. Gómez EJ, del Pozo F, Hernando ME. Telemedicine for diabe-
Regional Office for Europe: Copenhagen, Denmark, 1992) EUR/ tes care: The DIABTel approach towards diabetes telecare. Med
ICPCLR0550279g. Inform (London) 1996; 21(4): 283–295.
7. Piwernetz K, Massi Benedetti M, Vermeij D et al. DiabCare 29. Hernando ME, Gómez EJ, Corcoy R, del Pozo F. Evaluation of
Thinkshop, quality network diabetes. Diabetes Nutr Metab 1993; DIABNET, a decisión support system for therapy planning in
6: 107–122. gestational diabetes. Comput Methods Programs Biomed 2000;
8. Corcoy R, Muntaner F, Pou JM et al. DiabCare data set collection: 62(3): 235–248.
Benefits and warnings. Diabetes Nutr Metab 1993; 6: 389–392. 30. Bellazzi R, Larizza C, Montani S et al. A telemedicine support for
9. Gallo G, Cermeño J, Brugués E et al. The burden of type 2 dia- diabetes management: The T-IDDM Project. Comput Methods
betes compared to the burden of type 1 diabetes in patients Programs Biomed 2002; 69(2): 147–161.
undergoing intensive and specialized care. Diabetologia 2001; 31. Bellazzi R, Arcelloni M, Bensa G et al. Design, methods, and
44(Suppl. 1): 16-A. evaluation directions of a multi-access service for the manage-
10. Cubero JM, Hernández M, Brugués E et al. Metabolic outcome ment of diabetes mellitus patients. Diabetes Technol Ther 2003;
achieved by intensive management in European Diabetes Center 5(4): 621–629.
are far from desirable targets. Diabetologia 2001; 44(Suppl. 1): 32. Gómez EJ, Hernando-Pérez ME, Vering T et al. The INCA System:
17-A. A further step towards a telemedical artificial páncreas. IEEE
11. Kerényi ZS, Tamás GY, Piwernetz K. Pregnancy complicated by Trans Inf Technol Biomed 2008; 12(4): 470–479.
diabetes: Baseline data. Diabetes Nutr Metab 1993; 6: 365–368. 33. Pérez-Gandía C, Faccinetti A, Sparacino C, Cobelli C, Gómez
12. Jówicka E, Krzymierí J, Tracz M et al. Implementation of the EJ, Rigla M, de Leiva A, Hernando ME. Artificial neural network
DiabCare Program in registration of pregnant diabetic women. algorithm for online glucose prediction from CGM. Diabetes
Diabetes Nutr Metab 1993; 6: 369–371. Technol Ther 2010; 12(1): 81–88.
13. Thaisz E, Rappai A, Fövényi J, Závodi E. Screening and care 34. Soriano-Marcolino M, Maia JX, Moreira-Alkimim MB, Boersma
of gestational and insulin-dependent diabetic pregnancies: E, Luiz-Ribeiro A. Telemedicine application in the care of dia-
The first four years experience. Diabetes Nutr Metab 1993; 6: betes patients: Systematic review and meta-analysis. PLOS ONE
373–375. 2013; 8(11): 1–13.
14. Tamás G, Hadden DR, Molsted-Pedersen L et al. WHO/IDF 35. García-Patterson A, Martín E, Ubeda J, María MA, Adelantado
Guidelines for care and management of the pregnant diabetic JM, Ginovart G, de Leiva A, Corcoy R. Nurse-based management
women. Av Diabetol 1992; 5: 137–140. in patients with gestational diabetes. Diabetes Care 2003; 26:
15. Ziegler O, Kolopp M, Got I et al. Reliability of self-monitoring 998–1001.
of blood glucose by CSII treated patients with type 1 diabetes. 36. García-Sáez G, Hernando ME, Martínez-Sarriegui I, Rigla M,
Diabetes Care 1989; 12: 184–188. Torralba V, Brugués E, de Leiva A, Gómez EJ. Architecture of a
16. Strowig SM, Raskin P. Improved glycaemic control in intensively wireless Personal Assistant for telemedical diabetes care. Int J
treated type 1 diabetic patients using blood glucose meters with Med Inform 2009; 78(6): 391–403.
storage capability and computer assisted analysis. Diabetes Care 37. García-Sáez G, Alonso JM, Molero J, Rigla M, Martínez-Sarriegui
1998; 21: 1694–1699. I, de Leiva A, Gómez EJ, Hernando ME. Mealtime blood glucose
17. Devreese K, Leroux-Roels G. Laboratory assessment of five blood classifier based on fuzzy logic for DIABTel telemedicine system.
glucose meters designed for self-monitoring of blood glucose Artif Intell Med 2009; 5651: 295–304.
concentration. Eur J Clin Chem Clin Biochem 1993; 31: 829–837. 38. Rigla M, Hernando ME, Gómez EJ, Brugués E, García-Sáez G, Capel
18. Weitgasser R, Gappmayer B, Pichler M. Newer portable glucose I, Pons B, de Leiva A. Real-time CGM together with telemedical
meters—Analytical improvement compared with previous gen- assistance improves glycemic control and glucose stability in pump-
eration devices? Clin Chem 1999; 45: 1821–1825. treated patients. Diabetes Technol Ther 2008; 10(3): 194–199.
19. Brunner GA, Ellmere M, Sendlfofer G et al. Validation of home 39. González-Romero S, González-Molero M, Fernández-Abellán
blood glucose meters with respect to clinical and analytical M, Domínguez-López ME, Ruiz-de Adana S, Olveira G, Soriguer
approaches. Diabetes Care 1998; 122: 495–502. F. CSII versus MDI in pregnant women with type 1 diabetes.
20. American Diabetes Association. Self-monitoring of blood glu- Diabetes Technol Ther 2010; 12: 263–269.
cose. Diabetes Care 1996; 19(Suppl. 1): S62–S66. 40. Volpe I, Pancani F, Aragona M, Lencioni C, Battini L, Ghio A,
21. Schiffrin A, Belmonte M. Multiple daily self-glucose monitoring: Resi V, Bertolotto A, del Prat S, Di Cianni G. CSII and MDI injec-
It is essential role in long-term glucose control in insulin-depen- tions in type 1 diabetic pregnant women: A case-control study.
dent diabetic patients treated with pump and multiple subcutane- Gynecol Endocrinol 2010; 26: 193–196.
ous injections. Diabetes Care 1982; 5: 479–484. 41. Chico A, Tundidor D, Jordana L, Saigi I, María MA, Corcoy R, de
22. Nathan DS. The importance of intensive supervision in determin- Leiva A. Changes in insulin requirements from the onset of CSII
ing the efficacy of insulin pump therapy. Diabetes Care 1983; 6: until optimization of glycemic control. J Diabetes Sci Technol
295–297. 2014; 8(2): 371–377.
References 501
42. Murphy HR, Elleri D, Allen JM et al. Closed-Loop Insulin delivery 63. Furusyo N, Hayashi J. Glycated albumin and diabetes mellitus.
during pregnancy complicated by type 1 diabetes. Diabetes Care Biochim Biophys Acta 2013; 1830: 5509–5514.
2011; 34: 406–411. 64. Varner NW. Efficacy of home glucose monitoring in diabetic
43. de Leiva A, Hernando ME. Telemedical artificial pancreas. pregnancy. Am J Med 1983; 75: 592–596.
Diabetes Care 2009; 32 (Suppl. 2): S211–S216. 65. Hanson U, Person B, Enochsson E et al. Self-monitoring of blood
44. Russell SJ, El-Khatib FH, Sinha M, Magyar KL, McKeon K, glucose by diabetic women during the third trimester of preg-
Goergen LG, Balliro C, Hillard MA, Nathan DM, Damiano ER. nancy. Am J Obstet Gynecol 1984; 150: 817–821.
Outpatient glycemic control with a bionic pancreas in type 1 dia- 66. De Veciana M, Major CA, Morgan M et al. Postprandial versus
betes. N Engl J Med 2014; 371: 313–325. prepandial blood glucose monitoring in women with gestational
45. Little RR, Rohlfing CL, Wiedmayer H-M et al. The National diabetes mellitus requiring insulin therapy. N Engl J Med 1995;
Glycohemoglobin Standardization Program (NGSP): A five-year 19: 1237–1241.
progress report. Clin Chem 2001; 47: 1985–1992. 67. Peacock M, Chunter JC, Walford S. Self-monitoring of blood glu-
46. Guerci B, Durain D, Leblanc H. Multicentre evaluation of the cose in diabetic pregnancy. Br Med J 1979; ii: 1333–1336.
DCA 2000 system for measuring glycated haemoglobin. DCA 68. Jovanovic L, Peterson CM, Saxena BB et al. Feasibility of main-
2000 Study Group. Diabetes Metab 1997; 23: 195–201. taining normal glucose profiles in insulin-dependent pregnant
47. European IDDM Policy Group. Consensus guidelines for the diabetic women. Am J Med 1980; 68: 105–112.
management of insulin-dependent (type 1) diabetes. Diabet Med 69. Jovanovic L, Druzin ML, Peterson CM. Impact of euglycaemia on
1999; 10: 990–1005. the outcome of pregnancy in insulin-dependent diabetic women
48. Roberts NB, Amara AB, Morris M, Green BN. Long-term evalu- compared with normal control subjects. Am J Med 1981; 71:
ation of electrospray ionization mass spectrometric analysis of 921–928.
glycated hemoglobin. Clin Chem 2001; 47: 316–321. 70. Espersen T, Klebe JG. Self-monitoring of blood glucose in preg-
49. Tarim O, Kucukerdogan A, Gunay U et al. Effects of iron defi- nant diabetics. A comparative study of the blood glucose level
ciency anemia on hemoglobin A1c in type 1 diabetes mellitus. and course of pregnancy in pregnant diabetics on an out-patient
Pediatr Int 1999; 41: 357–362. regime before and after the introduction of methods for home
50. Phillipov G, Charles P, Beng C, Philips PJ. Alternate site testing for analysis of blood glucose. Acta Obstet Gynecol Scand 1985; 64:
HbA1c using Primus CLC330 GHb analyzer. Diabetes Care 1997; 11–14.
20: 607–609. 71. Goldberg JD, Franklin B, Lasser D et al. Gestational diabetes:
51. ADA. Diagnosis and classification of diabetes mellitus. Diabetes Impact of home glucose monitoring on neonatal birth weight.
Care 2010; 33 (Suppl. 1): S62–S69. Am J Obstet Gynecol 1986; 154: 546–550.
52. Nathan DM, Kuenen J, Borg R, Zheng H, Schoenfeld D, Heine RJ. 72. Wecher DJ, Kaufmann RC, Amankwah KS et al. Prevention of
Translating the AC1 assay into estimated average glucose values. neonatal macrosomia in gestational diabetes by the use of inten-
Diabetes Care 2008; 31: 1473–1478. sive dietary therapy and home glucose monitoring. Am J Perinatol
53. Bonora E, Tuomilehto J. The pros and cons of diagnosing diabetes 1991; 8: 131–134.
with A1c. Diabetes Care 2011; 34(Suppl. 2): S184–S190. 73. HAPO Study Cooperative Research Group. Hyperglycemia
54. Little RR, Rohlfing CL, Sacks DV. Status of hemoglobin A1c measure- and adverse pregnancy outcomes. N Engl J Med 2008; 358:
ment and goals for improvement. Clin Chem 2011; 57(2): 205–214. 1991–2002.
55. Sacks DV, ADA EASD IDF Working Group of the HbA1c Assay. 74. IADPSG recommendations on the diagnosis and classifica-
Global Harmonization of HbA1c. Clin Chem 2005; 51: 681–683. tion of hyperglycemia in pregnancy. Diabetes Care 2010; 33(3):
56. Hanas R, John G. Consensus Statement on the Worldwide 676–682.
Standardization on the HbA1c measurement. Clin Chem 2010; 56: 75. Leary J, Pettitt DJ, Jovanovic L. Gestational diabetes guidelines in
1362–1364. a HAPO world. Best Pract Res Clin Endocrinol Metab 2010; 24:
57. Lowe LP, Metzger BE, Deyer AR et al., for the HAPO Study 673–685.
Cooperative Research Group. Association of maternal A1c and glu- 76. Colagiuri S, Falavigna M, Agarwal MM et al. Strategies for imple-
cose with pregnancy outcomes. Diabetes Care 2012; 35: 574–580. menting the WHO diagnostic criteria and classification of hyper-
58. Verhaeghe J, van Herck E, Benhalima K, Mathieu C. Glycated glycemia first detected in pregnancy. Diabetes Res Clin Pract
hemoglobin in pregnancies at increased risk for GDM. Eur J 2014; 103: 364–372.
Obstet Gynecol Reprod Biol 2012; 161: 157–162. 77. Peyrot M, Rubin NN, Lauritzen T, Snoek FJ, Matthews DR,
59. Braadvedt GD, Drury PL, Cundy T. Assessing glycaemic control Skovlund SE. Psychosocial problems and barriers to improved
in diabetes: Relationships between fructosamine and HbA1c. N Z diabetes management: Results of the Cross-National Diabetes
Med J 1997; 110: 459–462. Attitudes, Wishes and Needs (DAWN) Study. Diabet Med 2005;
60. Kruseman AC, Mercelina L, Degenaar CP. Value of fasting blood 22(10): 1379–1385.
glucose and serum fructosamine as a measure of diabetic con- 78. Peyrot M, Rubin RR, Lauritzen T et al. Patient and provider per-
trol in non-insulin-dependent diabetes mellitus. Horm Metab Res ceptions of care for diabetes: Results of the cross-national DAWN
1992; 26(Suppl.): 59–62. Study. Diabetologia 2006; 49(2): 279–288.
61. Morgan LJ, Marenah CB, Morgan AG et al. Glycated haemoglo- 79. Lapolla A, DiCianni G, Di Benedetto A, Franzetti I, Napoli A,
bin and fructosamine in non-diabetic subjects with chronic renal Sciaca L, Torlone E, Tonutti L, Vitacolonna E, Mannino D. Quality
failure. Nephrol Dial Transplant 1990; 5: 868–873. of life, wishes, and needs in women with gestational diabetes:
62. Kilpatrick ES. Problems in the assessment of glycaemic control in Italian DAWN Pregnancy Study. Int J Endocrinol 2012; 2012:
diabetes mellitus. Diabet Med 1997; 14: 819–831. 724–726.
59 Early pregnancy loss and
perinatal mortality
Kinneret Tenenbaum-Gavish, Anat Shmuely, and Moshe Hod
Live birth was defined by the WHO as “The complete expul-

Introduction sion or extraction of a product of conception, irrespective
Diabetes is one of the most prominent medical disorders of the duration of pregnancy, which after such separation
complicating pregnancy, probably affecting 1 out of 250 breathes or shows any evidence of life, such as beating of the
pregnant women.1–3 Gestational diabetes occurs in up to 18% heart, pulsation of the umbilical cord, or definite movement
of all pregnancies,4–7 and it is estimated that approximately of voluntary muscles, whether or not the umbilical cord has
10%–15% of diabetic pregnancies are due to pregestational been cut, or the placenta is attached, each product of such a
diabetes.8,9 Furthermore, diabetes is considered a major risk birth is considered live born.”18
factor for congenital malformations, stillbirth, and neona-
tal death—all constituting increased perinatal mortality
(PNM).2,5,10,11 Common causes of perinatal mortality
Achieving the desired level of glycemic control prior to
and during pregnancy is crucial in order to reduce PNM.12–17 The cause of death can be identified in 65%–75% of cases,
depending on the expertise of the multidisciplinary team
involved.21
Definitions 1. Stillbirth or fetal death—Fetal death was defined as an
Perinatal mortality involuntary loss in which the fetus (>20 weeks of gesta-
tion) showed no evidence of life (i.e., no heartbeat or res-
PNM still remains the standard measure to evaluate adverse
piration) on delivery.22 In other studies, stillbirth is the
pregnancy outcome. There is a considerable difficulty in
death of a fetus >24 or 28 weeks of gestation.4 There are
estimating PNM due to numerous systems used and sub-
considerable variations regarding the rates of stillbirth
stantial differences in definitions of PNM. Even the World
in different countries. The WHO’s report on the topic of
Health Organization’s (WHO) reports do not all follow the
stillbirths (defined there at or after 28 weeks of gestation)
same definitions.18–20 The American National Vital Statistics
reported that two-thirds of stillbirths occur in Southeast
System and the British Confidential Enquiry into Maternal
Asia and Africa and 55% occur in rural families from
and Child Health (CEMACH) both use gestational age as
these areas. The stillbirth rate has declined worldwide by
the basis for calculation of stillbirth and neonatal death
14% from 1995 to 2009, representing an annual decline
but differ in the referred gestational age (20 vs. 24 weeks of
of 1.1% per year. The rate of decline in the African region
gestation).2 In the Netherlands, gestational age >24 weeks
is 0.7% compared to 3.8% in the Western Pacific region.23
and/or birth weight > 500 g were used and a French survey
Stillbirth may account for up to 50% of cases of PNM
used either 22 weeks of gestation or 500 g at birth for their
and the underlying cause may remain obscure depend-
study.17,19 This variance between reports makes international
ing on resources applied.24 Causes for stillbirth are
comparisons somewhat more difficult and may hinder iden-
generally grouped into three categories: fetal, placental,
tification of temporal or geographic trends. Furthermore,
and maternal (Table 59.1, reference 21).
little can be said about the standard of care that is reflected
2. Neonatal death—The death of a live-born infant before
by these adverse pregnancy outcome measures.
the age of 28 days. Neonatal death can be divided into
Perinatal mortality—this definition of PNM has been
early and late neonatal deaths: early neonatal death is
adapted internationally:
defined as the death of a live-born infant during the
first 7 days after birth. Late neonatal death is the death
Stillbirths + Deaths from 0 to 6 days of age
´1000 of a live-born infant after 7 days but before 29 days of
All live births + Stillbirths birth.2,25,26

502
PNM in the diabetic population 503
Table 59.1 Categories of stillbirth
Fetal Placental Maternal Other

Percentage of 25–40 Up to 25 5–10 Unknown
cases (%)
e.g., chromosomal e.g., placental abruption, e.g., hypertension, Combined causes, e.g.,
anomalies, cord accidents, vascular disease, maternal illness
congenital infections, feto- thrombophilia, causing placental
malformations, maternal hemorrhage, trauma, substance abruption or
infections, hydrops placenta previa, etc. abuse, etc. insufficiency
Source: Wood, S.T. et al., Diabetes Care, 23, 1752, 2000.
Neonatal death is largely attributed to four factors: pre- Later in pregnancy, maternal hyperglycemia is translated
term delivery, infections (mainly sepsis and pneumonia), into fetal hyperplasia of pancreatic islet cells and hyperin-
birth asphyxia, and fetal anomalies affecting the neonatal sulinemia. Pedersen et al.34,35 linked this to early neonatal
period.25 hypoglycemia. A series of studies by Salvesen et al.36,37 dem-
onstrated that hyperinsulinemia is related to cord blood
acidemia and hypoxemia. It is suggested that hypoxemia
Early pregnancy loss and academia are related to increased rates of stillbirth
Early pregnancy loss is an indirect measure related to quality and neonatal death observed in the diabetic population.
of care before and during early pregnancy. Early pregnancy Furthermore, insulin has an anabolic effect on muscle and
loss or spontaneous abortion refers to pregnancy loss at less adipose tissue linked to fetal macrosomia.37,38 Macrosomia
than 20 weeks of gestation in the absence of elective medical in its turn is related to increased risk of PNM and shoulder
or surgical measures to terminate the pregnancy.25 dystocia.39
Etiology of PNM PNM in the diabetic population

Diabetes affects the metabolism of all nutrients (carbohy- Although steps forward in modern obstetrics allow early
drates, fatty acids, and proteins), glucose being the most detection of congenital malformation and assessment of
prominent. Glucose and those other metabolic fuels operat- fetal well-being, PNM rates in the diabetic population still
ing at the different stages of pregnancy may account for the remain up to five times higher than those of the general
multitude of pathologies inflicted upon the offspring of the population.3,29
diabetic mother. These pathological conditions range from In order to try to achieve further reduction in PNM
congenital malformations and intrauterine fetal death to rates of diabetic women, focus should be put on the vari-
macrosomia, respiratory distress, and hyperbilirubinemia. ous components constituting overall perinatal outcome
Poor metabolic control may also induce alternations in levin the different stages of the diabetic pregnancy. A closer
els of fatty acids and amino acids. This provides an altered examination of the various components comprising PNM
environment in which the embryo and fetus of the diabetic rate may help elucidate specific points at which intervention
mother may be exposed to changes in gene expression and might assist in the overall reduction of PNM in the diabetic
increased teratogensis.27,28 population.
Hyperglycemia by itself is involved in the pathogenesis of Numerous studies have demonstrated that there is an
factors contributing to PNM throughout the entire length inverse relationship between maternal blood glucose levels
of pregnancy.29 In vitro, high glucose levels generate free and adverse pregnancy outcome. Pioneer works done dur-
oxygen radicals that are linked to mechanisms of cellular ing the 1960s and 1970s allowed to plot the inverse relation-
damage.27–30 Not surprisingly, the congenital malformation ship between mean blood glucose (MBG) levels and PNM.
rate (which largely contributes to the higher rate of PNM These studies suggest that normalization of blood glucose
in pregnancies complicated by diabetes) was found to be levels might equal the PNM in the diabetic population to
inversely related to maternal age and directly related to the that of the nondiabetic population.40 Karllson and Kjellmer
level of glycemic control during early pregnancy.31,32 Good divided their diabetic gravid patients into three groups
glycemic control brings about adequate levels of glucose according to White’s classification and MBG levels. MBG
and insulin and is therefore associated with lower levels levels <100, 100–150, and >150 were related to PNM of 3.8%,
of ketones. It was suggested that maternal ketosis plays an 16%, and 24% respectively, but no association was found
important part in the occurrence of congenital malforma- between White’s classification and PNM.40 The established
tions in the fetus.33 inverse relationship between maternal metabolic control in
504 Early pregnancy loss and perinatal mortality
patients with pregestational diabetes and the risk for PNM41 proportion of spontaneous pregnancy loss in this specific
was also demonstrated in patients with gestational diabetes subgroup of diabetic patients.
(also related to maternal age and obesity).42,43 Population- Rudge et al.61 described pregnancy results in women
based studies showed that for women with PGDM, 16%–28% with abnormal glucose tolerance including women diag-
of PNM is due to congenital malformations, and an addi- nosed with GDM. They reported an abortion rate of 0.8% in
tional 21%–41% is due to preterm delivery.39,44 Since the rates the group of patients with altered diurnal glycemic profile
of both complications improve with prepregnancy care,45 it without frank diabetes (as opposed to no abortions in the
is not surprising that Wahabi et al.46 found in their meta- control group). There is evidence to support the theory that
analysis that the PNM rate in women, who attended pre- some metabolic abnormalities such as insulin resistance and
pregnancy care programs, was reduced by 66% compared to high levels of fasting blood glucose may be apparent at early
those who did not.46 stages of pregnancy in women who are subsequently diag-
Tables 59.2 and 59.3 summarize data regarding PNM and nosed with gestational diabetes.62 An alternative explanation
its various components in women with pregestational and is that this constitutes a group of women with undiagnosed
gestational diabetes, respectively. Type 2 diabetes.
These studies and others offer circumferential support
that GDM represents a metabolic disorder that interferes
Early pregnancy loss with the outcome of pregnancy from conception to delivery
The altered intrauterine conditions in women with diabe- even though it is generally diagnosed throughout the third
tes are probably associated with increase in early pregnancy trimester.
loss rate. However, the exact quantification of the added
risk in the diabetic population is difficult to assess due to
the indefinable incidence of miscarriages in the general Stillbirth (late fetal loss)
population.53,54 Stillbirth is a major component of PNM in the diabetic pop-
ulation (see Tables 59.2 and 59.3).
Pregestational diabetes Congenital malformations are thought to be the most
The incidence of pregnancy loss in the diabetic population established causes of stillbirth in the diabetic population.
has declined substantially since the introduction of insulin The most common malformations are cardiovascular and
in the early twentieth century. During the 1940s and 1950s, neural tube defects,2 but diabetes was not found to be pre-
the reported pregnancy loss rate (“sudden fetal death”) dictive of any specific malformations.24,63,64
was approximately 20%.55 This might reflect the associa- Fetal metabolic acidosis with and without hypoxemia (as
tion between enhanced glycemic control (both precon- discussed earlier) is more prevalent in the diabetic popula-
ceptionl and during the first weeks of organogenesis) and tion.36,37,65 Their presence may offer an additional explanation
the decreased rate of malformations.56 Most of the studied for late fetal demise in the diabetic population. Most stud-
reporting outcome of miscarriage included patients with pre- ies investigating the pathophysiological basis for fetal death
gestational diabetes, without a specific distinction between unrelated to congenital anomalies were done in women with
type 1 and type 2 diabetes. The reported rate of spontaneous PGDM,34,65 but it is reasonable to assume that women with
abortions in women with Type 1 diabetes is approximately GDM may share the same pathophysiological events as they
17%.57 The pooled estimate risk for early pregnancy loss in also present the same high stillbirth rates.
patients with type 1 and type 2 diabetes was calculated to be The higher prevalence of other recognized risk factors
3.23 RR (95% CI 1.64–6.36).3 for stillbirth such as maternal obesity, hypertension, and
Good glycemic control around the period of conception advanced maternal age in the diabetic population may
reduces the risk of early pregnancy loss to that observed in all contribute to higher stillbirth rates.2,66 Late fetal death
the nondiabetic population. High levels of HbA1c above occurs more often in women with poor metabolic control,
normal range showed progressively high pregnancy loss ketoacidosis, and vascular complications probably as a
rates. Lower levels of HbA1c were associated with lower result of fetal metabolic acidemia and hypoxemia.67 Landon
early pregnancy loss rates among women with Type 1 diabe- et al.67 found that women with Type 1 diabetes with vascu-
tes mellitus (DM).22,58,59 Therefore, HbA1c can be regarded lar complications were at greater risk for abnormal fetal sur-
as an indirect measure of adverse outcome and be used for veillance tests and subjected to more obstetric interventions
quality control. such as induction of labor. In the group of patients who
A threshold level of HbA1c of 10%–12% correlating with needed intervention, fetal cord pH was significantly lower
MBG of 150–170 mg/dL is suggested as the upper limit for than the nonintervention group. This may also serve as an
reducing the risk of spontaneous abortions.60 indirect measure of fetal acidemia (in diabetic women as a
possible cause of fetal compromise that may lead to intra-
Gestational diabetes mellitus uterine fetal death).
Only few studies investigating gestational diabetes mel-
litus (GDM) and pregnancy outcome have included early Pregestational diabetes
pregnancy loss as one of the outcome measures examined. The reported rate of congenital malformation in patients
Therefore, there is a scarcity of data concerning the exact with Type 1 diabetes in some population-based studies is
Table 59.2 PNM in PGDM
Congenital
Author # Total PNM Stillbirth Neonatal death Preterm Macrosomia LGA Birth trauma anomalies
CEMACH2 3808 (2767 Type 1; NA 26.8/1000 9.3/1000 36% Macrosomia 21%a; Shoulder dystocia 41.8/1000
1041 Type 2) (Type 1 (Type 1 LGA 51.7% 7.9%; Erb’s palsy
25.8/1000; 9.3/1000; 4.5/1000
Type 2 Type 2
29.2/1000) 9.2/1000)
French 435 (289 Type 1; 44/1000 34.5/1000 Type 1 3/1000; 38.2% Macrosomia 17.3% Shoulder dystocia 41/1000
multicentric17 146 Type 2) Type 2 7.6%
21/1000
Machintosh 2359 (1707 Type 1; 31.8/1000 26.8/1000 9.3/1000 NA NA NA 46/1000
et al.16 652 Type 2)
Jensen et al.39 1215 (Type 1) 31/1000 21/1000 NA 41.7% LGA 62.5% NA 50/1000
Clausen and 301 (240 Type 1; 66/1000 NA NA 38% Type 1 LGA 51% Type 1; NA 29/1000
Matheisen47 61 Type 2) (Type 2) 56% Type 2 (Type 1)
31% Type 2 Macrosomia 5% 67/1000
Type 1; 8% Type 2 (Type 2)
Evers et al.15 323 Type 1 28/1000 18.5/1000 9.25/1000 32.2% LGA 45.1% Shoulder dystocia 88/1000
14%
ATLANTIC-DIP11 104(80 Type 1, 24 25/1000 25/1000 NA 12% 32% NA 24/1000
Type 2)
Tennant et al.48 1548 (1206 Type 1, NA 26.5/1000 4/1000 NA NA NA NA
342 Type 2)
a No significant difference in macrosomia rate between type 1 and type 2.
Table 59.3 Perinatal mortality in gestational diabetes mellitus
Neonatal
Author # Total PNM Stillbirth death Preterm Macrosomia LGA Birth trauma Metabolic
Crowther et al. 49 1000 (490 NA 0/506 vs. 0/506 NA Macrosomia 10% Shoulder dystocia 1% Jaundice 9% (both)
intervention, 510 3/524 2/524 vs. 21% vs. 3% RR 0.46
routine care) LGA 13% vs. 22% (95% CI 0.19–1.10)
O’sullivan et al.42 187 GDM 64/1000 NA NA 22.5% NA NA NA
Dunne et al.50 216a NA 0 0 NA LGA 25%–37% NA NA
Ramtoola et al.51 294 116/1000 81/1000 38/1000 22% Macrosomia 16% NA Hypoglycemia 14%
Hyperbilirubinemia 39%
Ray et al.14 428 NA NA NA 19.2% LGA 15.9% Shoulder dystocia 3% NA
Tundidor et al.52 2299 5/1000 NA NA 5.7% LGA 10.4% 2.2% Jaundice 3.5%
Hypocalcemia 1.5%
a GDM + IGT.
three- to fivefold higher than that of the general popula- and hyperinsulinemia. The threshold for this pregnancy
tion.19,57 Congenital anomalies in women with Type 2 complication seems to be between 105 and 110 MBG; patients
diabetes are also significantly higher than that of the gen- who achieve glycemic control below these values have a still-
eral population and contribute to the increased PNM rate birth rate comparable to those of the general population.60
reported in this subgroup of patients.16,17 The congenital
malformation rate in patients with Type 2 diabetes is equal Neonatal death
to or even exceeds that of Type 1 diabetes.16,17,68 A meta-
Preterm delivery
analysis that included an overall number of 3088 women
with PGDM, conducted by Wahabi et al.,46 found that pre- Pregestational diabetes Preterm delivery is strongly related
pregnancy care (glycemic control being the most impor- to adverse neonatal outcome.2,4,17 Over one-third of infants
tant aspect) reduced the rate of congenital malformations of pregestational diabetic mothers were born prematurely.2
from 7.4% to 1.9%, a rate similar to that reported for the The rate of preterm delivery (both spontaneous and
background population.46 induced) in the Type 1 diabetic population is as high as
Stillbirth rate in women with pregestational diabetes 45%.17 Poor glycemic control (elevated HbA1c levels) and
may reach 28–46/1000 live births—a rate three to five times vascular complications (preeclampsia and nephropathy) are
higher than that of the general population.2,16,43,48 Lauenborg predictive factors for preterm labor.70–72
et al.,24 in an audit on 25 cases of stillbirth in Type 1 diabetes The CEMACH enquiry reported the risk for preterm
patients, found a direct relationship between poor glycemic delivery in the diabetic population (type 1 and type 2 dia-
control both before and during pregnancy (measured by betes) to be nearly fivefold higher in comparison to the
HbA1c levels) and a higher incidence of stillbirth. The rela- general population. Up to two-thirds of preterm deliv-
tive risk for stillbirth in patients with Type 2 diabetes was eries were induced (“medically indicated”), and nearly
found to be 4.7 (95% CI 3.7–6.0).16 Cundy et al.43 reported a 37% of them were due to presumed fetal compromise. 2
sevenfold increase in the risk for late fetal death in women Some studies indicate that the proportion of induced
with Type 2 diabetes that contributed to a high PNM rate of preterm delivery due to presumed fetal compromise is
46/1000 in that study. This was attributed to maternal fac- greater in the diabetic population than that of the general
tors such as obesity and advanced maternal age more preva- population. 2
lent in his study group. Furthermore, patients with Type 2 Diabetes complicated by nephropathy or vascular disease
diabetes tend to be of lower socioeconomic class and part of causes an overall higher preterm delivery rate even when
an ethnic minority group, demographic differences that are compared to a group of women with chronic hypertension,
interrelated to lack of preconception care (PCC) and may be which might suggest there are other pathophysiological
confounding in the interpretation of results.2,16,43 factors contributing to the adverse outcome.70
Tennant et al.48 found that increasing periconception
HbA1c concentration (for values above 49 mmol/mol), his- Gestational diabetes mellitus There is an association
tory of retinopathy, and lack of prepregnancy folic acid con- between increased incidence of preterm delivery and
sumption were all associated with increased odds of fetal and gestational diabetes. Hedderson et al.73 found that the risk
infant death. There was no difference in the risk of fetal and/ for preterm delivery increased with increasing levels of
or infant death in women with Type 1 diabetes compared glycemia (ranging from abnormal screening, but normal
with those with Type 2, nor was there any evidence that diagnostic glucose tolerance test to GDM). The relative
the associations with HbA1c concentration, folic acid con- risk for preterm labor was 1.42 (95% CI [1.15–1.77]) for the
sumption, or history of retinopathy were different between GDM group compared to normal controls. This association
types. Damm et al.69 defined a composite endpoint of poor was also present for both induced and spontaneous preterm
outcome in late pregnancy that included preterm delivery, labor.73–75
preeclampsia, both preterm delivery and preeclampsia, still- The degree of glucose intolerance or “early onset” of GDM
birth, and death within 1 week postnatally in women with is linked to a higher incidence of preterm labor. This may
Type 1 PGDM. They found that elevated HbA1c, high glu- suggest an inverse relation between length of pregnancy and
cose spikes, and out-of-range plasma glucose levels in the poor glycemic control and identify poor glycemic control as
third trimester, and albuminuria in early pregnancy, are a marker of increased risk for preterm delivery.13
associated with poor late-pregnancy outcomes.
Macrosomia
Gestational diabetes mellitus
Pregestational diabetes Macrosomia is associated with
The stillbirth rate in patients with GDM is greater than that increased PNM and increased birth trauma. Birth-related
of the general population. The exact extent of increase in still- trauma in women with pregestational diabetes is much higher
birth rates is dependent upon the severity of glucose intoler- than those in the general population: shoulder dystocia
ance and degree of glycemic control. Stillbirth rates are only (more than twofold), birth-related fractures, and Erb’s palsy
slightly increased when good glycemic control is achieved (more than tenfold).2,14,62 Higher rates of shoulder dystocia
(stillbirth rate of 4.8 vs. 4.2 per 1000).15 Pettitt et al.66 reported carry an increased risk of PNM. Sheiner et al. found PNM
an increased risk of stillbirth mainly related to excessive fetal rates of 3.7% in infants suffering from shoulder dystocia vs.
growth, which is also a measure of increased hyperglycemia 0.5% in those without (OR 7.4 95% C.I [3.5–14.9]).76
Mondestin et al. reported that PNM is increased among caused by hyperinsulinemia.60 There is a direct relation-
diabetic pregnancies at all birth weight categories over ship between higher rates of respiratory complications and
1250 g,77 but especially when fetal weight exceeded 4000 g, poor glycemic control,29,38 but there is little information
thus doubling the risk of fetal death in the diabetic population concerning the estimated threshold of glycemia that carries
reaching 5.9/1000 births (adjusted R.R 2, 95% CI 1.8–2.2).77 an increased risk for respiratory complications.60 Another
There was no significant difference between the rates of problem is the lack of data relating directly to the relation-
macrosomia in infants born to women with Type 1 diabetes ship between higher rates of these “metabolic” complication
compared with those born to women with Type 2 diabetes.2 and PNM. These complications may carry an increased risk
In unselected diabetic populations of patients with type 1 for morbidity and mortality, but the exact quantification of
and type 2 diabetes, up to 50% of infants of diabetic moth- this added risk is difficult due to lack of information. Dani
ers are born at weights >90th percentile. These patients were et al.80 demonstrated that infants of GDMs had lower levels
suboptimal in their glycemic control, as reflected by HbA1c of glucose and higher levels of pyruvate, histidine, alanine,
levels.2,62,76 The risk for delivery of a macrosomic child is valine, methionine, arginine, lysine, hypoxanthine, lipopro-
directly related to glycemic control measures throughout tein, and lipid than controls, but did not find any clinical
pregnancy. differences in outcomes.80
It is believed that with modern resources, the greater
Gestational diabetes mellitus High rate of macrosomia is part of these complications have a minor influence on actual
also reported in infants of mothers diagnosed with GDM mortality rates but may cause increase in morbidity.
(increased risk of two- to fourfold).5,13 Women with GDM
who attend an intensive pregnancy surveillance have a
fourfold decrease in the risk of birth-related complications What can be done to reduce the rate
compared with GDM patients receiving “standard”
prenatal care.49 of PNM in pregnancies complicated
Langer et al.78 demonstrated the relationship between gly- by diabetes?
cemic control (namely, MBG levels <105) and reduced inci-
dence of macrosomia. The St. Vincent Declaration given at 1989 set as a goal to
The aforementioned text is not direct evidence for the equalize pregnancy outcomes of women with diabetes melli-
relationship between metabolic control, macrosomia, and tus to those of nondiabetic women within 5 years. Although
reduced PNM. However, it is logical to assume that intensive there has been considerable advancement in the ability to
prenatal care results in better glycemic control that contrib- detect fetal anomalies and establish fetal well-being and
utes to improved outcome. maturity, this goal, which was perceived as feasible at the
Yee et al.79 found that in overweight and obese women time (especially in light of the conceivable improvement in
diagnosed with GDM, third trimester weight loss was asso- pregnancy outcomes of insulin-dependent diabetic women),
ciated with a decreased risk of macrosomia, as well as a has not been met. The presence of diabetes is thought to
decreased rate of cesarean delivery and NICU admission. increase the risk for congenital malformation by as much as
10-fold, stillbirth up to 5-fold, and neonatal death 3–4-fold.3
Furthermore, some reports imply that the trend toward
Respiratory and metabolic complications decline in PNM rates in the diabetic population is less than
Admission to neonatal intensive care units (NICUs) or other the decline in PNM observed in the general population dur-
special care units is more common in infants of diabetic ing the same time period.4,22
mothers and reaches up to 50% of neonates.2,17,38 Colstrup et al.81 sought to evaluate whether the goals of the
The NICU admission rate is an indirect measure of mor- 1989 St. Vincent Declaration have been obtained concerning
bidity and of quality of care in the diabetic population (both fetal and neonatal complications. Their study included 12
PGDM and GDM). The majority of admissions are due to population-based studies published between 2003 and 2013,
a myriad of conditions that may all contribute to early or with a total of 14,099 women with T1DM. The authors found
late PNM in the diabetic population, such as hypoglycemia, that the risk of adverse pregnancy outcomes (including con-
hyperbilirubinemia, and respiratory distress due to lung genital anomalies, preterm deliveries, and macrosomia) was
immaturity. Premature infants are more prone to episodes two to five times increased in women with T1DM compared
of hypoglycemia than term infants38; however, even after with the general population. They concluded that the goals
correction for preterm birth, the proportion of infants of of the St. Vincent Declaration have not been achieved.
diabetic mothers hospitalized in NICU is still significantly Good glycemic control is a fundamental part in avoid-
higher than the general population.2,38,72,74 For instance, the ing adverse pregnancy outcome. In order to elucidate pos-
rate of hypoglycemia in infants born to Type 1 diabetes may sible causes for this apparent failure to meet goals set at
reach 50% in premature infants and 23% in term infants.71 St. Vincent, two important issues must be addressed. The
Pulmonary complications ranging from transient tachy- first issue is the extent at which the diabetic population
pnea to respiratory distress are more common in infants of meets the healthcare practitioners’ criteria for “good dia-
diabetic mothers resulting from abnormal lung maturation betic control.”
References 509
There are very few randomized control trials (RCTs) population. Other studies also hold up the hypothesis that
regarding the relationship between glycemic control intensive care consisting of maternal glucose monitoring
and pregnancy outcome.82 The Diabetes Control and and decisions concerning insulin treatment based upon its
Complications Trial (DCCT) suggested that early and tight results are related with enhanced perinatal outcome.85,86
glycemic control might reduce pregnancy loss rates and con- The second issue regarding the failure to meet
genital anomaly rates equal to those of the general popula- St. Vincent goals is related to preconception counseling in
tion.83 This finding was not repeated in Cochrane’s review the diabetic population. Preconception counseling is sig-
on three other RCTs.82 Conversely, observational studies nificant due to the fact that the crucial phase of organo-
(which included less selective populations) failed to demon- genesis takes part during early pregnancy, at which timely
strate such a substantial reduction.2,17 Some of the difference and adequate metabolic control is important for optimal
can be explained by the fact that different cutoff points for pregnancy outcome.
HbA1c levels to distinguish between poor and good glycemic A meta-analysis12 reviewing 14 cohort studies concerning
control were used. There were also dissimilarities regarding patients with PGDM, calculated a major malformation rate
the time in which the predictive HbA1c measurement was of 2.1% in women attending PCC vs. 3.5% in those without
taken (first trimester, first antenatal visit, etc.) and the num- PCC (RR 0.36, 95% CI 0.22–0.59). This connection between
ber of measurements used to categorize women in “poor” vs. PCC and reduction in adverse pregnancy outcome (mainly
“optimal” control. It is clear that a single HbA1c measure- spontaneous abortions and congenital anomalies) was also
ment cannot be used as an absolute predictor for pregnancy demonstrated in other studies.17,19,39,87 The topic of PCC may
complications. HbA1c may reflect long-term average glyce- need special attention in women with Type 2 DM as they
mic control, but there are many limitations to its use, raising tend to differ in lifestyle, age, and economical background
important questions regarding its adequacy as a measure of and often require change in treatment regime (oral antigly-
glycemic control (vs. other methods such as self-monitoring cemic drugs or insulin).2,19,47
blood glucose levels, MBG, or fasting blood glucose). The route toward improved PCC attendance remains
Furthermore, the optimal HbA1c level that should be set as somewhat obscure. Some of the factors associated with low
a treatment goal is beyond the scope of this chapter. Langer attendance rates such as education level, socioeconomic
et al.84 provided support for those calling to rely on maternal status, and ethnic background are not in the hands of
glucose levels as a measure of glycemic control rather than healthcare providers. However, diabetic patients should be
HbA1c. They showed that strict glycemic control resulting in counseled against PCC failure and unplanned pregnancies
near normoglycemia leads to perinatal morbidity and mor- and provided with accessible preconception education and
tality rates almost comparable to those of the nondiabetic care programs.
REFERENCES
1. Metzger BE, Coustan DR et al. The Organising Committee. 10. Wood ST, Sauve R et al. Prediabetes and perinatal mortality.
Summary and recommendations of the 4th international Diabetes Care 2000; 23: 1752–1754.
Workshop Conference on Gestational Diabetes Mellitus. 11. Fidelma PD, Avalos G et al. ATLANTIC-DIP: Pregnancy outcome
Diabetes Care 1998; 21 (Suppl. 2): 161–167. for women with pregestational diabetes along the Irish Atlantic
2. Confidential Enquiry into maternal and child health: Pregnancy seaboard. Diabetes Care 2009; 32(7): 1205–1206.
of women with type 1 and type 2 diabetes 2002–03 in England 12. Ray JG, O’Brien TE et al. Preconception care and the risk of con-
Wales and Northern Ireland, CEMACH: London, U.K., 2005. genital anomalies in the offspring of women with diabetes mel-
3. Inkster M, Fahey TP et al. Poor glycated heamoglubin control and litus: A meta analysis. QJM 2001; 94: 435–444.
adverse pregnancy outcomes in type1 and type 2 diabetes mel- 13. Langer O, Yogev Y et al. Gestational diabetes: The consequence
litus: Systematic review of observational studied. BMC Pregnancy of not treating. Am J Obstet Gynecol 2005; 192(4): 989–997.
and Childbirth 2006; 6: 30. 14. Ray JG, Vermeulen MJ et al. Maternal and neonatal outcomes in
4. Clinical Management Guidelines for Obstetrician-Gynecologists. pregestational and gestational diabetes mellitus, and the influ-
ACOG Practice Bulletin no. 30. Washington, DC: American ence of maternal obesity and weight pain: The DEPOSIT study.
College of Obstetricians and Gynecologists, 2001. QJM 2001; 94: 347–356.
5. Jovanovic L, Pettitt DJ. Gestational diabetes mellitus. J Am Med 15. Yogev Y, Most O et al. Gestational diabetes mellitus: Is there a risk for
Assoc 2001; 286(20): 2516–2518. perinatal mortality? Am J Obstet Gynecol 2004; 191(6 Suppl.): s51.
6. Lapolla A, Dalfra MG, Ragazzi E et al. New International 16. Machintosh MC, Fleming KM et al. Perinatal mortality and con-
Association of the Diabetes and Pregnancy Study Groups genital anomalies in babies of women with type 1 or type 2 dia-
(IADPSG) recommendations for diagnosing gestational diabetes betes in England, Wales and Northern Ireland: Population based
compared with former criteria: A retrospective study on preg- study. Br Med J 2006; 333: 177–180.
nancy outcome. Diabet Med 2011; 28(9): 1074–1077. 17. Diabetes and Pregnancy group, France. French multicentric sur-
7. Holt RI, Coleman MA et al. The implications of the new inter- vey of outcome of pregnancy in women with pregestational dia-
national association of the DiABETES AND Pregnancy Study betes. Diabetes Care 2003; 26: 2990–2993.
Groups (IADPSG) diagnostic criteria for gestational diabetes. 18. World Health Organization. Manual of the International
Diabet Med 2011; 28(4): 382–385. Statistical Classification of Diseases, Injuries and Causes of
8. American Diabetes Association, Clinical Practice Death, 9th Revision, Vol. 1, World Health Organization: Geneva,
Recommendations 2004. Diabetes Care 2004; 27(1): S88–S90. Switzerland, 1977.
9. Fong A, Serra A et al. Pre-gestational versus gestational diabetes: 19. Evers IM, de Valk HW et al. Risk of complications of pregnancy
A population based study on clinical and demographic differ- in women with type 1 diabetes: Nationwide prospective study in
ences. J Diabetes Complications 2014; 28(1): 29–34. the Netherlands. Br Med J 2004; 328(7445): 915.
20. Lawn J, Shibuya K, Stein C. No cry at birth: Global estimates of 47. Clausen TD, Matheisen E. Poor pregnancy outcome in women
intrapartum stillbirths and intrapartum-related neonatal deaths. with type 2 diabetes. Diabetes Care 2005; 28: 323–328.
Bull World Health Organ June 2005; 83(6): 409–417. 48. Tennant PEG, Glinianaia SV et al. Pre-existing diabetes, mater-
21. Cunningham FG, Leveno KJ et al. Fetal death. In: Williams nal glycated haemoglobin, and the risk of fetal and infant death:
Obstetrics, 22nd ed., Suppl. 4, Appleton & Lange: Norwalk, CT, A population-based study. Diabetologia 2014; 57(2): 285–294.
2005. 49. Crowther CA, Hiller JH et al. Effect of treatment of gestational
22. Barfield W, Martin J, Hoyert D. Racial/ethnic trends in fetal mor- diabetes mellitus on pregnancy outcomes. N Engl J Med 2005;
tality in United States, 1990–2000. Morbid Mortal Wkly Rep 352: 2477–2486.
2004; 53: 529–532. 50. Dunne F, Brydon P et al. Pregnancy in women with Type 2 dia-
23. World Health Organization web site—Data, statistics and epide- betes: 12 years outcome 1990–2002. Diabet Med 2003; 20(9):
miology on maternal, newborn, child and adolescent health. 734–738.
24. Lauenborg J, Mathiesen E et al. Audit on stillbirths in women 51. Ramtoola S, Home P et al. Gestational impaired glucose toler-
with pregestational type 1 diabetes. Diabetes Care 2003; 26: ance does not increase perinatal mortality in a developing coun-
1385–1389. try: Cohort study. Br Med J 2001; 322(7293): 1025–1026.
25. Obstetrics in broad perspective. In: Canningham FG, Gant NF, 52. Tundidor D, Garcia-Patterson A et al. Perinatal maternal and
Leveno KJ, Gilstrap LC, Hauth JC, Wenstrom KD, eds., Williams neonatal outcomes in women with gestational diabetes mellitus
Obstetrics, 21st ed., McGraw-Hill: New York, 2001. according to fetal sex. Gend Med 2012; 9(6): 411–417.
26. American College of Obstetricians and Gynecologists. Perinatal 53. Griebel CP, Halvorsen J et al. Management of spontaneous abor-
and infant mortality statistics. Committee Opinion No. 167, 1995. tion. Am Fam Physician 2005; 72: 1243–1250.
27. Lee AT, Plump A et al. A role for DNA mutations in diabetes- 54. Berek JS. Early pregnancy loss and ectopic pregnancy. In:
associated teratogenesis in transgenic embryos. Diabetes 1995; Novak’s Gynecology, 13th ed., Lippincott Williams & Wilkins:
44(1): 20–24. Philadelphia, PA, 2002, Chapter 17.
28. Eriksson UJ, Borg LA et al. Diabetic embryopathy. Studies with 55. Miller HC, Hurwitz D et al. Fetal and neonatal mortality in preg-
animal and in vitro models. Diabetes 1991; 40(Suppl. 2): 94–98. nancies complicated by diabetes mellitus. J Am Med Assoc 1944;
29. Carrapato MR, Marcelino F et al. The infant of the diabetic 124: 271.
mother: The critical developmental window. Early Pregnancy 56. Miodovnik M, Skillman CA et al. Effect of maternal hyperketo-
2001; 5(1): 57–58. nemia in hyperglycemic pregnant ewes and their fetuses. Am J
30. Fine EL, Horal M et al. Evidence that elevated glucose causes Obstet Gynecol 1986; 154(2): 394–401.
altered gene expression, apoptosis and neural tube defects 57. Casson IF, Clarke CA et al. Outcome of pregnancy in insulin
in a mouse model of diabetic pregnancy. Diabetes 1999; 48: dependent diabetic women: Results of a five year population
2454–2462. cohort study. Br Med J 1997; 315: 275–278.
31. Towner D, Kojs SL et al. Congenital malformations in pregnancies 58. Mills JL, Simpson JL et al. Incidence of spontaneous abortion
complicated by NIDDM. Diabetes Care 1995; 18(11): 1446–1451. among normal women and insulin-dependent diabetic women
32. Scaefer UM, Songster G et al. Congenital anomalies in offsprings whose pregnancies were identified within 21 days of conception.
of women with hyperglycemia first detected during pregnancy. N Engl J Med 1988; 319(25): 1617–1623.
Am J Obstet Gynecol 1997; 177(5): 1165–1171. 59. Jovanovic L, Knopp RH et al. Elevated pregnancy losses at high
33. Jovanovic L, Metzger BE et al. β-Hydroxybutirate levels in type 1 and low extremes of maternal glucose in early and normal dia-
diabetic pregnancy compared with normal pregnancy. Diabetes betic pregnancy. Diabetes Care 2005; 28: 1113–1117.
Care 1998; 21: 1978–1984. 60. Langer O. Diabetes and Pregnancy Dilemma: Leading Change
34. Pedersen J, Pedersen LM et al. Assessors of fetal perinatal mor- with Proven Solutions, University Press of America Inc.: Lanham,
tality in diabetic pregnancy. Analysis of 1,332 pregnancies in MD, 2006, Chapter 14, pp. 231–235.
Copenhagen series 1946–1972. Diabetes 1974; 23(4): 302–305. 61. Rudge MV, Calderon I et al. Perinatal outcome of pregnancies
35. Pedersen J. The Pregnant Diabetic and Her Newborn, 2nd ed., complicated by diabetes and by maternal daily hyperglycemia
Williams & Wilkins: Baltimore, MD, 1977. not related to diabetes. Gynecol Obstet Invest 2000; 50: 108–112.
36. Salvesen DR, Freeman J et al. Prediction of fetal academia in 62. Smirnakis KV, Martinez A et al. Early pregnancy insulin resistance
pregnancies complicated by maternal diabetes mellitus by bio- and subsequent gestational diabetes mellitus. Diabetes Care
physical profile scoring and fetal heart rate monitoring. Br J 2005; 28(5): 1207–1208.
Obstet Gynecol 1993; 100(3): 277–233. 63. Kuhoury MJ, Becerra JE et al. Clinical-epidemiologic assessment
37. Salvesen DR, Brudenell JM et al. Fetal beta-cell function in preg- of pattern of birth defects associated with human teratogens:
nancies complicated by maternal diabetes mellitus: Relationship Application to diabetic embryopathy. Pediatrics 1989; 84(4):
to fetal academia and macrosomia. Am J Obstet Gynecol 1993; 658–665.
168(5): 1363–1369. 64. Becerra JE, Kuhoury MJ et al. Diabetes mellitus during pregnancy
38. Williams AF. Hypoglycemia of the newborn. Review of the litera- and the risks for specific birth defects: A population-based case-
ture. CHD/97.1. World Health Organization: Geneva: Switzerland, control study. Pediatrics 1990; 85(1): 1–9.
1997. Available from: http://www.who.int/child-adolescent-health/ 65. Bradley RJ, Brundenell JM et al. Fetal acidosis and hyperlactice-
publication/NUTRITION/WHO_CHD_97.1htm. mia diagnosed by cordocentesis in pregnancies complicated by,
39. Jensen DM, Damm P et al. Outcomes in type 1 diabetic pregnan- maternal diabetes mellitus. Diabet Med 1991; 8(5): 464–468.
cies. Diabetes Care 2004; 27: 2819–2823. 66. Pettitt DJ, Knowles WC et al. Gestational diabetes: Infant and
40. Karllsson K, Kjellmer I. The outcome of diabetic pregnancies in maternal complications of pregnancy in relation to third trimes-
relation to the mother’s blood sugar level. Am J Obstet Gynecol ter glucose tolerance in Pima Indians. Diabetes Care 1980; 3:
1972; 112: 213. 458–464.
41. Kitzmiller JL, Cloherty JP et al. Diabetic pregnancy and perinatal 67. Landon MB, Langer O et al. Fetal surveillance in pregnancies
morbidity. Am J Obstet Gynecol 1978; 131: 560–579. complicated by insulin-dependent diabetes mellitus. Am J Obstet
42. O’sullivan JB, Charles D et al. Gestational diabetes and perinatal Gynecol 1992; 167: 617–621.
mortality rate. Am J Obstet Gynecol 1973; 1: 901–904. 68. Roland JM, Murphy HR et al. The pregnancies of women with
43. Cundy T, Gamble G et al. Perinatal mortality in type 2 diabetes type 2 diabetes: Poor outcomes but opportunities for improve-
mellitus. Diabet Med 2000; 17(1): 33–39. ment. Diabet Med 2005; 22(12): 1774–1777.
44. Platt MJ, Stanisstreet M et al. St Vincent’s Declaration 10 years 69. Damm P, Mersebach H et al. Poor pregnancy outcome in women
on: Outcomes of diabetic pregnancies. Diabet Med 2002; 19: with type 1 diabetes is predicted by elevated HbA1c and spikes of
216–220. high glucose values in the third trimester. J Mater Fetal Neonatal
45. Wahabi HA, Alzeidan RA et al. Pre-conception care for diabetic Med 2014; 27(2): 149–154.
women for improving maternal and fetal outcomes: A system- 70. Sibai BM, Caritis SN et al. Preterm delivery in women with pre-
atic review and meta-analysis. BMC Pregnancy Childbirth 2010; gestational diabetes mellitus or chronic hypertension relative to
10: 63. women with uncomplicated pregnancies. The National Institute
46. Wahabi HA, Alzeidan RA et al. Pre-pregnancy care for women of Child Health and Human Development Maternal-Fetal
with pre-gestational diabetes mellitus: A systematic review and Medicine Units Network. Am J Obstet Gynecol 2000; 183(6):
meta-analysis. BMC Public Health 2010; 12: 792. 1520–1524.
References 511
71. Lepereq J, Coste J et al. Factors associated with preterm delivery in 80. Dani C, Bresci C et al. Metabolomic profile of term infants of
women with type 1 diabetes. Diabetes Care 2004; 27: 2824–2828. gestational diabetic mothers. J Matern Fetal Neonatal Med 2014;
72. Watson D, Rowan J et al. Admissions to neonatal intensive care 27(6): 537–542.
following pregnancies complicated by gestational or type 2 dia- 81. Colstrup M, Mathiesen ER et al. Pregnancy in women with type 1
betes. Aust N Z Obstet Gynecol 2003; 43(6): 429–432. diabetes: Have the goals of St. Vincent declaration been met con-
73. Hedderson MM, Ferrara A et al. Gestational diabetes mellitus cerning foetal and neonatal complications? J Matern Neonatal
and lesser degrees of pregnancy hyperglycemia: Association with Med 2013; 26(17): 1682–1686.
increased risk of spontaneous preterm delivery. Obstet Gynecol 82. Walkinshaw SA. Very tight versus tight glycemic control for
2003; 102(4): 850–856. diabetes in pregnancy. Cochrane database Syst Rev 2000; (2):
74. Svare JA, Hansen BB et al. Perinatal complications in women CD00226.
with gestational diabetes mellitus. Acta Obstet Gynecol Scand 83. The Diabetes Control and Complications Trial Research Group.
2001; 80(10): 899–904. The effect of intensive treatment for diabetes on development
75. Lao TT, Ho LF. Does maternal glucose intolerance affect the and progression of long-term complications in insulin dependent
length of gestation in singleton pregnancies? J Soc Gynecol diabetes mellitus. The Diabetes Control and Complications Trial
Investig 2003; 10(6): 366–371. Research Group. N Engl J Med 1993; 329(14): 977–986.
76. Sheiner E, Levy A et al. Determining factors associated with shoul- 84. Langer O, Rodriguez DA et al. Intensified versus conventional
der dystocia: A population based study. Eur J Obstet Gynecol management of gestational diabetes. Am J Obstet Gynecol 1994;
2006; 126(1): 11–15. 170: 1036–1046.
77. Mondestin MA, Ananth CV et al. Birth weight and fetal death in 85. Jovanovic L, Peterson CM et al. Maternal postprandial glucose
the United States: The effect of maternal diabetes during preg- levels and infant birth weight: The diabetes in early pregnancy
nancy. Am J Obstet Gynecol 2002; 187: 922–926. study. Am J Obstet Gynecol 1991; 164: 103–111.
78. Langer O, Yogev Y et al. Gestational diabetes: Does an associa- 86. de Veciana M, Major CA et al. Postprandial versus pre prandial
tion exist between deviant fetal growth and glycemic control? blood glucose monitoring in women with gestational diabe-
Am J Obtet Gynecol 2004; 189(6): 151. tes mellitus requiring insulin therapy. N Eng J Med 1995; 333:
79. Yee LM, Cheng YW et al. Gestational weight loss and perina- 1237–1241.
tal outcomes in overweight and obese women subsequent to 87. Steel JM, Johnstone FD et al. Can prepregnancy care of diabetic
diagnosis of gestational diabetes mellitus. Obesity 2013; 21(12): women reduce the risk of abnormal babies? Br Med J 1990; 301:
E770–E774. 1070–1074.
60 Short-term implications of
gestational diabetes mellitus:
The neonate
Diabetes mellitus during pregnancy is responsible for signif- of prostaglandins,5 particularly prostaglandin E2 that is
icant complications in the fetus and neonate. Risk is higher involved in maintaining the patency of the ductus arteriosus
in pregestational diabetes, but also if pure gestational diabe- in utero)6; (2) high glucose levels induce an excessive pro-
tes (GDM) is not recognized and well managed. The current duction of reactive oxygen species that has been shown to
worldwide pandemic of type 2 diabetes (T2D) has devastat- cause oxidative stress and subsequently increase the risk of
ing effects with T2D representing up to 25% of pregnancies fetal malformations, notably neural tube defect7; and also
with pregestational diabetes.1 The rate of fetal and neonatal (3) high glucose induces the activation of the cascades of
complications in these pregnancies is not different to those proteins involved in apoptotic cell death, including the cas-
affected by type 1 diabetes (T1D). Diabetes in pregnancy pase families. Notably, it has been shown that hyperglycemia
is still responsible for a significant perinatal mortality and upregulates caspase-8 activity (enhanced cleavage) and may
morbidity in many countries.2 be involved in hyperglycemia-associated.8 Despite a growing
Optimizing diabetic care before and during pregnancy understanding of the molecular basis of diabetic embryopa-
reduces the risk of neonatal complications. Indeed, thanks thy, mechanisms are not yet fully understood.9
to considerable advances in the management of diabetes in Likewise, the mechanisms of the impact of maternal dia-
pregnancy, whether preexisting (type 1 or type 2) or purely betes on fetal and neonatal physiology need further research.
gestational, many infants born to diabetic mothers experi- Pedersen’s hypothesis, formulated more than 50 years ago,
ence a healthy neonatal course, especially in high-income suggested that fetal overgrowth was related to increased
countries. Treatment of GDM has been shown to reduce spe- transplacental transfer of maternal glucose, thus stimulat-
cific or composite neonatal outcomes.3 However, implemen- ing the release of insulin by the fetal beta cells and inducing
tation of strategies to limit the consequences of diabetes in subsequent macrosomia.10 More recently, studies have char-
pregnancy to the offspring is still difficult to achieve in many acterized the link between maternal glycemia and neonatal
countries, particularly in low-income and developing ones. macrosomia or fat mass.11 Fetal hyperinsulinism not only
While even in countries that enjoy high standards of peri- affects fetal growth but is also currently considered as the
natal care, GDM can be misdiagnosed and neonatal complica- central factor in the pathophysiology cascade summarized
tions can occur that need a specific pediatric management. in Figure 60.1.
Furthermore, despite tight glycemic control, macrosomia
still persists in some cases. This suggests that mechanisms
other than those evoked by Pedersen contribute to fetal over-
Pathophysiology of fetal and growth. Evidence has recently emerged on the involvement
neonatal complications of the maternal metabolic environment and of placental
modifications in fetal overgrowth. In GDM, the maternal
Poorly controlled maternal pregestational diabetes sig- metabolic environment is characterized by insulin resis-
nificantly increases the risk of offspring malformations. tance (IR) and inflammation.12 Both conditions influence
Maternal hyperglycemia results in glucose metabolism in fetal growth. IR facilitates maternal hypertriglyceridemia,
the developing embryo that alters various molecular chain which in turn enhances substrate availability to the fetus.
reactions: (1) altered cell lipid metabolism (newborn infants Furthermore, specific alterations of the placental transcrip-
born to diabetic mothers have a lower status of arachidonic tome have been evidenced in the context of the altered envi-
acid concentrations,4 therefore altering the production ronment of diabetic pregnancy. For example, genes for lipid
512
Perinatal death and asphyxia 513
Maternal excess circulating

Fetal substrate transfer
glucose, lipids, amino acids
Fetal hyperinsulinemia
Fetal substrate uptake
Tissue oxygen consumption Macrosomia Lung surfactant

synthesis
Hypoxia Altered oxygen

delivery
Erythropoïetin
Myocardiopathy
Polycythemia
Stillbirth,
Respiratory distress
perinatal
syndrome
asphyxia Hyperbilirubinemia
Figure 60.1 Intrauterine exposure to maternal diabetes and short-term outcomes.
transport have been shown to be upregulated in the placenta of A relationship has been established between the risk of
women with GDM, as well as are genes for inflammatory path- malformations in GDM and maternal blood glucose level.18
ways.13 Altogether, such alterations directly or indirectly alter Such risk also increases with maternal body mass index
the availability of the substrates to the fetus either by increasing (BMI) and when GDM is diagnosed during early pregnancy.
their source or by modifying the maternal–fetal interface. These observations suggest that the increased risk could be
In addition, altered vascular structure and function, related to the inclusion of women with undiagnosed T2D in
altered microcirculation, placental immaturity, and poor tis- the GDM groups.19
sue oxygenation are important determinants of both short-
and long-term outcomes.
Perinatal death and asphyxia
Fetal malformations Both types of pregestational diabetes are associated with an
increased risk of stillbirth. In T1D, a three- to fivefold risk
Birth defects remain one of the leading causes of infant increase has been found in various countries.20. In T2D, the
mortality. It is estimated that 8 million infants in the world risk of perinatal death is even higher than in T1D (OR = 1.5
(185,000 in the United States) are born each year with major [1.15–1.96]).21 Pregnancies in women who have newly diag-
malformations. Many of these birth defects can be attributed nosed T2D have significantly worse outcomes, probably
to pregestational diabetes.9 related to obesity, hypertension, and maternal age. In T2D,
Major congenital malformations reported in babies of the BMI of women with a stillbirth has been shown to be
pregestational diabetic mothers are heart malformations higher (+2 kg/m2) than in women with live-born infants,
(transposition of great vessels, ventricular or atrial septal suggesting that obesity can strongly contribute to perinatal
defects, coarctation of the aorta), caudal regression syn- death in women with T2D.22 Unlike in pregestational dia-
drome, central nervous system defects (NTD, including betes, the increased rate of fetal deaths in the second and
anencephaly), gut malformations (duodenal and anorectal third trimesters of pregnancy is debatable in cases of GDM.
atresia, hypoplastic left colon), and skeletal and genital– Although it is difficult to demonstrate a consistent increase
urinary tract anomalies.14 in GDM, the risk for stillbirth appears slightly increased but
Poor maternal glycemic control in the periconceptual is lower than in T1D and T2D.20 Cundy et al. showed that,
period increases the risk of malformations. The rates of fetal after exclusion of newly diagnosed T2D in the GDM group,
malformations appear to be similar in maternal T1D and perinatal mortality was similar to that in the general popula-
T2D.15 The risk for congenital malformations in preexist- tion.22 According to these data, the increased risk of perina-
ing diabetes is 1.9- to 10-fold higher than in the total pop- tal death in cases of GDM, reported in some studies, may be
ulation.16 The risk is slightly increased in the case of GDM attributable to undiagnosed T2D.
compared to the general population (ORs between 1.1 and Only 50% of stillbirths (fetal death after 20–22 gestational
1.3) but is much lower than in women with pregestational weeks [GW]) can be explained by causes such as intrauter-
diabetes.16 The malformations described are similar to those ine growth restriction, preeclampsia, acute asphyxia, intra-
reported in pregestational diabetes, especially cardiovascu- uterine infection, or congenital malformation. But in the
lar defects and anomalies involving the musculoskeletal and remaining 50%, the pathophysiology is likely to be multifac-
central nervous systems.17,18 torial and may result from chronic fetal asphyxia and fetal
514 Short-term implications of gestational diabetes mellitus
cardiac dysfunction. In T1D offspring, more than 75% of macrosomia as it is characterized by an excess body fat,
perinatal deaths are attributed to congenital anomalies or an increased in muscle mass, and organomegaly with-
complications of preterm birth, whereas in T2D offspring, out increased brain size and is thus asymmetric. There is
the deaths are mainly due to stillbirth, chorioamnionitis, or a linear and continuous relationship between percentage
birth asphyxia.22 body fat in newborns, maternal glycemia, and fetal insulin
The rate of neonatal deaths, even in high-income coun- levels. 33
tries, is still around 3% in T1D or in T2D and is much lower, Hence, maternal diabetes during pregnancy, whatever
close to the rates of nondiabetic pregnancies in the case of its type, is a risk factor for macrosomia. Treatment of GDM
pure GDM. Strict glycemic control may reduce the number significantly reduces the rate of macrosomia.3,34 In case
of stillbirths in women with TD1 and TD2.23 of T1D, optimal glycemic control during pregnancy (i.e.,
Increased risk of perinatal asphyxia has been reported in HbA1c ≤ 7.0%) does not preclude a high incidence of fetal
diabetic pregnancies in a number of studies. Different risk macrosomia. Some studies have found that third trimester
factors have been identified in T1D: nephropathy developing HbA1c is an independent risk indicator for macrosomia,
during pregnancy, hyperglycemia 6 hours before delivery, but it has a weak predictive capacity. HbA1c does not reflect
preterm birth, and lower gestational age.24 In cases of GDM, intermittent hyperglycemia, mainly postprandial hyper-
the incidence of perinatal asphyxia is not increased, nor is glycemic episodes, that may be involved in accelerated fetal
it influenced by maternal treatment.25,26 Macrosomic infants growth.35,36 The frequency of macrosomia or LGA is not dif-
(>4500 g) are at increased risk of asphyxia-linked morbidi- ferent in T2D with respect to T1D,37 even if in T2D, HbA1c
ties whatever is the cause of macrosomia.27 is lower at booking and throughout gestation.21
Macrosomia, regardless of the cause, is in itself a risk
factor for adverse perinatal outcomes, such as asphyxia and
Preterm birth perinatal death, birth injury, respiratory distress, and hypo-
glycemia.27,38 Such risks increase as the birth weight or the
A number of studies have reported an increased risk of
birth percentile rise. An important complication of macro-
preterm birth in case of maternal diabetes. However, it is
somia is shoulder dystocia and subsequent birth injuries.
difficult to determine the respective parts of spontaneous
Such risk is the highest for infants with a birth weight 4,500–
preterm delivery and of induced labor or cesarean section,
4,999 g and >5,000 g, (ORs 2.4 [2.2–2.5] and 3.5 [3.0–4.2],
because of frequent interventionistic approaches. The ben-
respectively).27 Among infants with Erb’s palsy in the general
efits of early delivery to avoid fetal death or shoulder dystocia
population in the British Isles, 53% were LGA.39 Erb’s palsy is
must be balanced against the morbidity linked to preterm
tenfold higher in case of pregestational diabetes.40
birth, particularly respiratory morbidity.
In pregestational diabetes, the rate of premature birth is
Fetal growth restriction
increased up to 25%, with most of these being late preterm
birth (34–36 GW).28 In T1D, pregestational hypertension Fetal growth restriction is less often associated with mater-
is related to preterm birth, whereas in T2D, third trimester nal diabetes but has been reported in association with severe
glycosylated hemoglobin (HbA1c) is a predictor of prema- vascular complications of advanced diabetes and poor pla-
turity.29 GDM and glucose intolerance are risk factors for cental perfusion.41 An interaction between preconceptional
spontaneous preterm birth independently on other compli- HbA1c and reduced weight at birth has been found.42 The
cations,30 mean maternal glucose level was also related to the hypothesis is that high glucose levels in early pregnancy
risk of preterm birth.31,32 harm placental development, especially when it is asso-
ciated with microvascular disease. Placental growth is
reduced and placental functions are impaired, resulting in
Fetal growth fetal growth restriction. This situation exposed to the risk
of preterm birth.
Macrosomia
The concept of excessive fetal growth has been expressed
either by the word “macrosomia” or by the expression “large Other neonatal complications
for gestational age” (LGA). Macrosomia is usually defined by
an upper birth weight threshold somewhere between 4,000 Hypertrophic cardiomyopathy
and 4,500 g, depending on the study. However, in this defini- Fetuses exposed to maternal hyperglycemia and hyperinsu-
tion, gestational age is not taken into account. The term LGA linism are prone to develop hypertrophic cardiomyopathy.
corresponds to a birth weight ≥90th percentile or > +2SD It primarily affects the interventricular septum but can
(>97th percentile) for gestational age. This more precise defi- extend to the myocardium in more severe cases. Contrary
nition takes into consideration gestational age at birth and to what is described for macrosomia, the precise role of fetal
allows preterm newborns with excessive fetal growth to be growth factors and the influence of hyperglycemia on car-
identified. diomyocyte growth and function are poorly understood.43
Macrosomia observed in the case of newborns of dia- Myocardial hypertrophy is most often asymptomatic.
betic mothers can be differentiated from other forms of It can sometimes lead to severe morbidity and mortality,
Other neonatal complications 515
according to the severity and the extension of cardiac Diabetes, but also maternal BMI, is associated with a
hypertrophy, especially in the case of obstructive septal higher risk of persistent pulmonary hypertension (PPH).
hypertrophy. However, other independent risk factors like macrosomia
Myocardial hypertrophy has been reported in both and cesarean deliveries might be in the causal pathway
pregestational diabetes and GDM with a wide range of between diabetes, overweight, and PPH. 56 Furthermore,
frequencies (between 25% and 75% of infants born to dia- increased risk for meconium aspiration has been reported
betic mothers).44,45 The incidence was lower in case of GDM after diabetic pregnancy, together with increased risk of
comparing to pregestational diabetes.46 The most recent
perinatal asphyxia. However, such complication affects
studies showed that optimal maternal glycemic control does mainly pregnancies complicated by severe preexisting
not entirely preclude interventricular septum hypertrophy diabetes. 57
and minor fetal cardiac function impairment, whatever is
the type of diabetes.47,48
Neonatal glucose disorders; hypoglycemia
Antenatal ultrasounds play an important role in moni-
toring fetal cardiac anatomy and function. Babies of women It has long been known that a correlation exists between
with diabetes should have an echocardiogram in the pres- macrosomia, increased cord C-peptide levels, and neona-
ence of clinical signs associated with congenital heart mal- tal hypoglycemia. The data collected by the HAPO study
formations or cardiomyopathy. confirmed such relationship: neonatal hypoglycemia was
It is usually considered that heart hypertrophy resolves strongly associated with elevated cord serum C-peptide lev-
anatomically within few months. However, the long-term els. Infants with excessive size at birth were more likely to
effect of diabetic cardiomyopathy on heart function remains develop hypoglycemia and hyperinsulinemia.58
to be elucidated. The exact incidence of hypoglycemia in cases of maternal
diabetes is difficult to assess due to the variable definitions
used for neonatal hypoglycemia in the different studies.
Respiratory disorders
Comparisons with the risk observed in healthy newborns
Neonatal respiratory distress syndrome are difficult also because of different monitoring of blood
Newborns to diabetic mothers have increased risk of neona- glucose at birth most of the studies. The infant of diabetic
tal respiratory distress syndrome (RDS), a major cause for mother is at risk of transient hyperinsulinism, which pre-
admission into neonatal intensive care units. vents at birth the normal activation of metabolic pathways
The principal mechanism relies in altered lung surfac- producing glucose and ketone bodies and causes increased
tant synthesis due to fetal hyperinsulinism. Insulin has been glucose consumption by tissues. However, severe hypogly-
shown to alter prenatal surfactant. Delayed appearance of cemia with clinical signs is unusual nowadays, probably
phosphatidylglycerol in amniotic fluid after 34 GW has been because of the improvement of prenatal maternal care.59
found in women with poorly managed diabetes. This risk
has been found to be higher between 36 and 37 GW49 and in Glucose monitoring
cases of pregestational diabetes.50 The pathological significance of low blood glucose levels
As a consequence, late preterm birth associated with immediately after birth, in the absence of specific symptoms,
RDS is a particular characteristic of intrauterine expo- is still questioned. An immediate fall in blood glucose con-
sure to diabetes. Late preterm infants are at greater risk of centration is observed after birth, reaching a nadir between
neonatal morbidities than term infants. This risk is higher 1 and 2 hours in healthy term infants. From 2 to 3 hours
in the presence of additional maternal morbidity, particu- of age, blood glucose then rises spontaneously, even in the
larly prenatal maternal exposure to hypertensive disorders absence of any nutritional intake, due to the activation of
of pregnancy and diabetes. 51 In neonates delivered after metabolic regulatory pathways. Therefore, measurement of
34 weeks, GDM should be considered as an independent blood glucose level at this stage is not indicated in asymp-
risk factor for severe neonatal respiratory disorder. 52 tomatic infants since normal levels cannot be distinguished
Macrosomia (birth weight ≥ 4000 g), associated or from abnormal levels during this period and the incidence of
not with GDM, is a risk factor for neonatal respiratory hypoglycemia is likely to be overestimated.60 The first mea-
distress.53,54 surement in asymptomatic neonates is recommended before
Complementary studies are necessary to determine the second feed, at around 3–4 hours of age, which gener-
whether in the case of diabetic pregnancies, changes need ally allows identifying infants who cannot manage adequate
to be introduced in recommendations on antenatal steroid early glucose homoeostasis.
therapy to accelerate fetal lung maturation. Regulatory pathways involved in glucose homeostasis
are activated within the first 24 hours of life. Thus, at-risk
Other respiratory disorders infants who have not suffered hypoglycemia during the first
Besides RDS, infants born to diabetic mothers are exposed 24 hours (glucose values above 2 mmol/L), who have been
to an increased risk of transient tachypnea of the newborn, fed normally, and who are clinically stable do not present a
particularly in the context of cesarean birth, due to delayed higher risk of subsequent hypoglycemia. Then, blood glucose
reabsorption of alveolar liquid at birth.55 monitoring may be discontinued.
There is currently no consensus on the indications for increased fetal red cell production. The incidence and the
systematic glucose blood monitoring in asymptomatic severity of polycythemia are in relation with poor maternal
infants born to diabetic mothers. It seems reasonable to glycemic control.
consider that LGA or growth-restricted infants born to dia- Normovolemic polycythemia seen in infants from dia-
betic mother may benefit from a blood glucose concentration betic mother can lead to hyperviscosity. Early symptoms
check at 3–6-hour intervals during the first day of life. On are unspecific. Feeding problems, plethoric aspect, cyano-
the other hand, normal-grown infants of mothers with diet- sis, lethargy, hypotonia, respiratory distress, jitteriness and
controlled GDM should not be monitored. irritability, seizure (due to multiple cerebral infarcts), necro-
tizing enterocolitis, hyperbilirubinemia, and hypoglycemia
Prevention and treatment have all been found associated. Hypoglycemia is aggravated
Early and frequent breastfeeding remains the key to prevent in infants of diabetic mothers due to increased glucose con-
hypoglycemia, whatever the birth weight of the infant, as far sumption by the increased red cell mass. Renal thrombosis
as he or she is able to feed autonomously. Therefore, infants and other forms of thrombosis are also more frequent in
of diabetic mothers should be kept beside their mother, in infants of diabetic mothers.
the absence of any significant complications requiring a Treatment is still controversial. In asymptomatic infants,
transfer to a special care neonatal unit, which corresponds no treatment is needed except to avoid hypoglycemia.
to the majority of the cases in high-income countries. Even Polycythemia associated with clinical signs should be treated
in mildly or moderately symptomatic infants with low blood by partial exchange transfusion as early as possible.
glucose levels, sustained breastfeeding or eventually formula Hyperbilirubinemia has been traditionally described as a
supplements should be tried first, provided a satisfactory neonatal complication of maternal diabetes. It is not a seri-
clinical response is obtained. For newborns without clinical ous complication if nontoxic levels are treated, which is usu-
signs, therapeutic intervention is required when two con- ally the case. The potential danger is kernicterus, which is
secutive blood glucose levels are below 2.0 mmol/L.61 In the not classically reported in cases of diabetes. In the HAPO
case the infant is unable to feed, an IV glucose supplementa- study, hyperbilirubinemia was weakly associated with
tion (3–6 mg/kg/hour) should be provided at constant rate of maternal blood glucose levels.11 Polycythemia could be one
infusion, in order to avoid rebound hypoglycemia. of the causes, but additional mechanisms, such as preterm
In the presence of severe symptoms, such as marked hypo- birth and poor liver conjugation, are likely to be involved.
tonia, coma, seizures, or blood glucose level <1 mmol/L, an
IV bolus dose of glucose (150–200 mg/kg) should be admin-
istered urgently, followed by constant rate infusion. It is nec- Impact of maternal obesity on
essary to check that thereafter blood glucose concentrations neonatal complications of GDM
stabilize within the normal range.43
Maternal obesity is associated with worse perinatal outcomes
Neonatal hypocalcemia even in glucose-tolerant women. Macrosomia is the main
Neonatal hypocalcemia has been reported following dia- complication reported in overweight or obese women, inde-
betes during pregnancy, with an incidence up to about pendently of diabetes.64–67 It is well recognized that neonates
30% in poorly controlled diabetes.62 The mechanism is still born to obese women, even if normally glucose tolerant, have
unclear but seems to involve an abnormal calcium phos- increased fat mass.68 As in GDM, increased adiposity at birth
phorus metabolism during pregnancy with decreased cal- is related to maternal IR with excess of glucose and lipid
cemia and vitamin D concentrations especially during the availability and to increase systemic inflammation.12
third trimester. The severity of hypocalcemia is related to The risk of fetal and infant deaths are two- to threefold
the degree of maternal diabetes control. Neonatal hypo- higher for women who are obese at the start of pregnancy,
parathyroidism has been reported and may be in part sec- after excluding pregnancies affected by congenital anomalies
ondary to maternal magnesium loss.62 Furthermore, some or pregestational diabetes.69 Maternal obesity is associated
studies have reported an association between GDM and low with an increased risk of a range of structural anomalies,
maternal vitamin D status, particularly with poor blood with the higher risk for neural tube defects.70,71
glucose control. Conversely, there is growing evidence that The risk of GDM increases with maternal BMI.72 The
women who develop GDM are more likely to be vitamin overall population-attributable fraction of GDM related to
D deficient.63 Other factors like prematurity and perinatal overweight was estimated at 46.2%.73
asphyxia can contribute to low calcium levels. There is no The benefit effect of treatment of diabetes on neonatal
indication to screen healthy babies for hypocalcemia and outcomes is lower in obese women even if targeted levels
hypomagnesemia. of glycemic control are achieved. Furthermore, when GDM
is untreated or poorly controlled, overweight and obese
women have a higher risk of poor neonatal composite out-
Neonatal polycythemia and hyperbilirubinemia come, compared to normal weight GDM women.26,74
Relative fetal hypoxia, secondary to insulin-induced high The combination of GDM and obesity has a greater impact
glucose uptake and metabolic rate (Figure 60.1), causes on pregnancy outcomes than either GDM or obesity alone.
increased erythropoietin secretion and, as a consequence, This cumulative risk was shown for macrosomia, newborn
References 517
percent body fat, and birth trauma.75 It was also reported for Macrosomia and subsequent complications are the main
a composite neonatal outcome (birth weight > 4000 g, birth short-term consequences of intrauterine exposure to GDM.
trauma, shoulder dystocia, hypoglycemia, or jaundice).76 Nevertheless, maternal obesity per se is a risk factor for
adverse pregnancy outcomes and has a major impact when
associated with GDM.
Conclusions There have been considerable advances in the manage-
ment of diabetes during pregnancy in high-income coun-
Severe neonatal complications (birth defects, perinatal tries, but many remain to be done in developing countries.
deaths, or asphyxia) are mainly due to pregestational dia- Prevention of adverse outcomes in the fetus and the neo-
betes. Such serious complications reported in studies on nate would be improved with a better understanding of the
GDM are likely due to undiagnosed prepregnancy T2D. mechanisms involved in the impact of maternal diabetes
Maternal diabetes during pregnancy, regardless of the type, on embryo, fetal, and neonatal physiology. Indeed, many of
is a risk factor for macrosomia or excessive fetal growth. these mechanisms remain to be determined.
REFERENCES
1. Bell R, Bailey K, Cresswell T et al. Trends in prevalence and out- 19. Allen VM, Armson BA, Wilson RD et al. Teratogenicity associated
comes of pregnancy in women with pre-existing type I and type with pre-existing and gestational diabetes. J Obstet Gynaecol
II diabetes. BJOG 2008; 115: 445–452. Can 2007; 29: 927–934.
2. Hod M, Simeoni U. Maternal, fetal and neonatal complications 20. Dudley DJ. Diabetic-associated stillbirth: Incidence, pathophysi-
of diabetic pregnancy-delivering optimal care while awaiting for ology, and prevention. Obstet Gynecol Clin North Am 2007; 34:
cure. Semin Fetal Neonatal Med 2009; 14: 63–65. 293–307, ix.
3. Horvath K, Koch K, Jeitler K et al. Effects of treatment in women 21. Balsells M, Garcia-Patterson A, Gich I et al. Maternal and fetal
with gestational diabetes mellitus: Systematic review and meta- outcome in women with type 2 versus type 1 diabetes mellitus:
analysis. Br Med J 2010; 340: c1395. A systematic review and metaanalysis. J Clin Endocrinol Metab
4. Min Y, Lowy C, Ghebremeskel K et al. Unfavorable effect of type 2009; 94: 4284–4291.
1 and type 2 diabetes on maternal and fetal essential fatty acid 22. Cundy T, Gamble G, Townend K et al. Perinatal mortality in type 2
status: A potential marker of fetal insulin resistance. Am J Clin diabetes mellitus. Diabet Med 2000; 17: 33–39.
Nutr 2005; 82: 1162–1168. 23. Mathiesen ER, Ringholm L, Damm P. Stillbirth in diabetic preg-
5. Jawerbaum A, Gonzalez E. The role of alterations in arachidonic nancies. Best Pract Res Clin Obstet Gynaecol 2011; 25: 105–111.
acid metabolism and nitric oxide homeostasis in rat models of dia- 24. Mimouni F, Miodovnik M, Siddiqi TA et al. Perinatal asphyxia
betes during early pregnancy. Curr Pharm Des 2005; 11: 1327–1342. in infants of insulin-dependent diabetic mothers. J Pediatr 1988;
6. Schneider DJ, Moore JW. Patent ductus arteriosus. Circulation 113: 345–353.
2006; 114: 1873–1882. 25. Crowther CA, Hiller JE, Moss JR et al. Effect of treatment of ges-
7. Chang TI, Horal M, Jain SK et al. Oxidant regulation of gene tational diabetes mellitus on pregnancy outcomes. N Engl J Med
expression and neural tube development: Insights gained from 2005; 352: 2477–2486.
diabetic pregnancy on molecular causes of neural tube defects. 26. Langer O, Yogev Y, Most O et al. Gestational diabetes: The conse-
Diabetologia 2003; 46: 538–545. quences of not treating. Am J Obstet Gynecol 2005; 192: 989–997.
8. Zhao Z, Yang P, Eckert RL et al. Caspase-8: A key role in the 27. Zhang X, Decker A, Platt RW et al. How big is too big? The peri-
pathogenesis of diabetic embryopathy. Birth Defects Res B Dev natal consequences of fetal macrosomia. Am J Obstet Gynecol
Reprod Toxicol 2009; 86: 72–77. 2008; 198: 517.e1–517.e6.
9. Reece EA. Diabetes-induced birth defects: What do we know? 28. Yang J, Cummings EA, O’Connell C et al. Fetal and neonatal
What can we do? Curr Diab Rep 2012; 12: 24–32. outcomes of diabetic pregnancies. Obstet Gynecol 2006; 108:
10. Pedersen J. Weight and length at birth of infants of diabetic moth- 644–650.
ers. Acta Endocrinol (Copenh) 1954; 16: 330–342. 29. Gonzalez-Gonzalez NL, Ramirez O, Mozas J et al. Factors
11. Metzger BE, Lowe LP, Dyer AR et al. Hyperglycemia and adverse influencing pregnancy outcome in women with type 2 versus
pregnancy outcomes. N Engl J Med 2008; 358: 1991–2002. type 1 diabetes mellitus. Acta Obstet Gynecol Scand 2008; 87:
12. Catalano PM, Hauguel-De Mouzon S. Is it time to revisit the 43–49.
Pedersen hypothesis in the face of the obesity epidemic? Am J 30. Hedderson MM, Ferrara A, Sacks DA. Gestational diabetes mel-
Obstet Gynecol 2011; 204: 479–487. litus and lesser degrees of pregnancy hyperglycemia: Association
13. Radaelli T, Lepercq J, Varastehpour A et al. Differential regulation with increased risk of spontaneous preterm birth. Obstet Gynecol
of genes for fetoplacental lipid pathways in pregnancy with ges- 2003; 102: 850–856.
tational and type 1 diabetes mellitus. Am J Obstet Gynecol 2009; 31. Ostlund I, Hanson U, Bjorklund A et al. Maternal and fetal out-
201: 209.e1–209.e10. comes if gestational impaired glucose tolerance is not treated.
14. McLeod L, Ray JG. Prevention and detection of diabetic embry- Diabetes Care 2003; 26: 2107–2111.
opathy. Community Genet 2002; 5: 33–39. 32. Yogev Y, Langer O. Spontaneous preterm delivery and gesta-
15. Macintosh MC, Fleming KM, Bailey JA et al. Perinatal mortality tional diabetes: The impact of glycemic control. Arch Gynecol
and congenital anomalies in babies of women with type 1 or type Obstet 2007; 276: 361–365.
2 diabetes in England, Wales, and Northern Ireland: Population 33. Hyperglycemia and Adverse Pregnancy Outcome (HAPO) Study:
based study. Br Med J 2006; 333: 177. Associations with neonatal anthropometrics. Diabetes 2009; 58:
16. Balsells M, Garcia-Patterson A, Gich I et al. Major congenital 453–459.
malformations in women with gestational diabetes mellitus: A 34. Falavigna M, Schmidt MI, Trujillo J et al. Effectiveness of gesta-
systematic review and meta-analysis. Diabetes Metab Res Rev tional diabetes treatment: A systematic review with quality of evi-
2012; 28: 252–257. dence assessment. Diabetes Res Clin Pract 2012; 98: 396–405.
17. Aberg A, Westbom L, Kallen B. Congenital malformations among 35. Persson M, Norman M, Hanson U. Obstetric and perinatal out-
infants whose mothers had gestational diabetes or preexisting comes in type 1 diabetic pregnancies: A large, population-based
diabetes. Early Hum Dev 2001; 61: 85–95. study. Diabetes Care 2009; 32: 2005–2009.
18. Schaefer-Graf UM, Buchanan TA, Xiang A et al. Patterns of con- 36. Evers IM, de Valk HW, Mol BW et al. Macrosomia despite
genital anomalies and relationship to initial maternal fasting glu- good glycaemic control in type I diabetic pregnancy; results of
cose levels in pregnancies complicated by type 2 and gestational a nationwide study in The Netherlands. Diabetologia 2002; 45:
diabetes. Am J Obstet Gynecol 2000; 182: 313–320. 1484–1489.
37. Lapolla A, Dalfra MG, Fedele D. Pregnancy complicated by type 56. Hernandez-Diaz S, Van Marter LJ, Werler MM et al. Risk factors
2 diabetes: An emerging problem. Diabetes Res Clin Pract 2008; for persistent pulmonary hypertension of the newborn. Pediatrics
80: 2–7. 2007; 120: e272–e282.
38. Weissmann-Brenner A, Simchen MJ, Zilberberg E et al. Maternal 57. Salihu HM, Weldeselasse HE, Rao K et al. The impact of obesity
and neonatal outcomes of large for gestational age pregnancies. on maternal morbidity and feto-infant outcomes among macro-
Acta Obstet Gynecol Scand 2012; 91: 844–849. somic infants. J Matern Fetal Neonatal Med 2011; 24: 1088–1094.
39. Evans-Jones G, Kay SP, Weindling AM et al. Congenital brachial 58. Metzger BE, Persson B, Lowe LP et al. Hyperglycemia and
palsy: Incidence, causes, and outcome in the United Kingdom adverse pregnancy outcome study: Neonatal glycemia. Pediatrics
and Republic of Ireland. Arch Dis Child Fetal Neonatal Ed 2003; 2010; 126: e1545–e1552.
88: F185–F189. 59. Hawdon JM. Babies born after diabetes in pregnancy: What are
40. Weindling A. Offspring of diabetic pregnancy: Short term out- the short- and long-term risks and how can we minimise them?
comes. Semin Fetal Neonatal Med 2009; 14: 111–118. Best Pract Res Clin Obstet Gynaecol 2011; 25: 91–104.
41. Vambergue A, Fajardy I. Consequences of gestational and pre- 60. Deshpande S, Ward Platt M. The investigation and management
gestational diabetes on placental function and birth weight. of neonatal hypoglycaemia. Semin Fetal Neonatal Med 2005; 10:
World J Diabetes 2011; 2: 196–203. 351–361.
42. Glinianaia SV, Tennant PW, Bilous RW et al. HbA(1c) and birth- 61. National Collaborating Centre for Women’s and Children’s
weight in women with pre-conception type 1 and type 2 dia- Health. Diabetes in Pregnancy: Management of Diabetes and
betes: A population-based cohort study. Diabetologia 2012; 55: Its Complications from Preconception to the Postnatal Period,
3193–3203. RCOG Press: London, U.K., 2008.
43. Hay WW Jr. Care of the infant of the diabetic mother. Curr Diab 62. Demarini S, Mimouni F, Tsang RC et al. Impact of metabolic con-
Rep 2012; 12: 4–15. trol of diabetes during pregnancy on neonatal hypocalcemia: A
44. Veille JC, Sivakoff M, Hanson R et al. Interventricular septal thick- randomized study. Obstet Gynecol 1994; 83: 918–922.
ness in fetuses of diabetic mothers. Obstet Gynecol 1992; 79: 63. Alzaim M, Wood RJ. Vitamin D and gestational diabetes mellitus.
51–54. Nutr Rev 2013; 71: 158–167.
45. Oberhoffer R, Hogel J, Stoz F et al. Cardiac and extracardiac 64. Ehrenberg HM, Mercer BM, Catalano PM. The influence of obe-
complications in infants of diabetic mothers and their relation sity and diabetes on the prevalence of macrosomia. Am J Obstet
to parameters of carbohydrate metabolism. Eur J Pediatr 1997; Gynecol 2004; 191: 964–968.
156: 262–265. 65. Jolly M, Sebire N, Harris J et al. Risk factors for macrosomia
46. Ullmo S, Vial Y, Di Bernardo S et al. Pathologic ventricular and its clinical consequences: A study of 350,311 pregnancies.
hypertrophy in the offspring of diabetic mothers: A retrospec- Obstet Gynecol 2003; 111: 9–14.
tive study. Eur Heart J 2007; 28: 1319–1325. 66. Hyperglycaemia and Adverse Pregnancy Outcome (HAPO)
47. Garcia-Flores J, Janez M, Gonzalez MC et al. Fetal myocardial Study: Associations with maternal body mass index. BJOG 2010;
morphological and functional changes associated with well- 117: 575–584.
controlled gestational diabetes. Eur J Obstet Gynecol Reprod 67. Owens LA, O’Sullivan EP, Kirwan B et al. ATLANTIC DIP: The
Biol 2011; 154: 24–26. impact of obesity on pregnancy outcome in glucose-tolerant
48. Chu C, Gui YH, Ren YY et al. The impacts of maternal gesta- women. Diabetes Care 2010; 33: 577–579.
tional diabetes mellitus (GDM) on fetal hearts. Biomed Environ 68. Sewell MF, Huston-Presley L, Super DM et al. Increased neonatal
Sci 2012; 25: 15–22. fat mass, not lean body mass, is associated with maternal obesity.
49. Piper JM, Xenakis EM, Langer O. Delayed appearance of pul- Am J Obstet Gynecol 2006; 195: 1100–1103.
monary maturation markers is associated with poor glucose 69. Tennant PW, Rankin J, Bell R. Maternal body mass index and the
control in diabetic pregnancies. J Matern Fetal Med 1998; 7: risk of fetal and infant death: A cohort study from the North of
148–153. England. Hum Reprod 2011; 26: 1501–1511.
50. Moore TR. A comparison of amniotic fluid fetal pulmonary 70. Stothard KJ, Tennant PW, Bell R et al. Maternal overweight and
phospholipids in normal and diabetic pregnancy. Am J Obstet obesity and the risk of congenital anomalies: A systematic review
Gynecol 2002; 186: 641–650. and meta-analysis. J Am Med Assoc 2009; 301: 636–650.
51. Shapiro-Mendoza CK, Tomashek KM, Kotelchuck M et al. 71. Rasmussen SA, Chu SY, Kim SY et al. Maternal obesity and risk of
Effect of late-preterm birth and maternal medical conditions on neural tube defects: A metaanalysis. Am J Obstet Gynecol 2008;
newborn morbidity risk. Pediatrics 2008; 121: e223–e232. 198: 611–619.
52. Vignoles P, Gire C, Mancini J et al. Gestational diabetes: A 72. Torloni MR, Betran AP, Horta BL et al. Prepregnancy BMI and the
strong independent risk factor for severe neonatal respira- risk of gestational diabetes: A systematic review of the literature
tory failure after 34 weeks. Arch Gynecol Obstet 2011; 284(5): with meta-analysis. Obes Rev 2009; 10: 194–203.
1099–1104. 73. Kim SY, England L, Wilson HG et al. Percentage of gestational
53. Esakoff TF, Cheng YW, Sparks TN, Caughey AB. The asso- diabetes mellitus attributable to overweight and obesity. Am J
ciation between birthweight 4000 g or greater and perinatal Public Health 2010; 100: 1047–1052.
outcomes in patients with and without gestational diabetes 74. Langer O, Yogev Y, Xenakis EM et al. Overweight and obese in
mellitus. Am J Obstet Gynecol 2009; 200(6): 672.e1–672.e4. gestational diabetes: The impact on pregnancy outcome. Am J
54. Das S, Irigoyen M, Patterson MB, Salvador A, Schutzman DL. Obstet Gynecol 2005; 192: 1768–1776.
Neonatal outcomes of macrosomic births in diabetic and non- 75. Catalano PM, McIntyre HD, Cruickshank JK et al. The hyper-
diabetic women. Arch Dis Child Fetal Neonatal Ed 2009; 94: glycemia and adverse pregnancy outcome study: Associations
F419–F422. of GDM and obesity with pregnancy outcomes. Diabetes Care
55. Al-Agha R, Kinsley BT, Finucane FM et al. Caesarean section and 2012; 35: 780–786.
macrosomia increase transient tachypnoea of the newborn in 76. Roman AS, Rebarber A, Fox NS et al. The effect of maternal obe-
type 1 diabetes pregnancies. Diabetes Res Clin Pract 2010; 89: sity on pregnancy outcomes in women with gestational diabetes.
e46–e48. J Matern Fetal Neonatal Med 2011; 24: 723–727.
61 Long-term outcomes after
exposure in the offspring
An extensive literature is available, tracking health out- The risk of diabetes was significantly higher in siblings born
comes at adulthood for the offspring of diabetic mothers. after the mother developed diabetes than in those born before
Many studies have found that offspring of diabetic moth- the mother’s diagnosis of diabetes (odds ratio 3.7, = 0.02) and
ers face an increased risk of developing impaired glucose the mean of body mass index (BMI) was 2.6 kg/m2 higher in
tolerance, hypertension (HT), overweight and obesity, and offspring of diabetic than in offspring of nondiabetic preg-
dyslipidemia, as well as other noncommunicable diseases. nancies (= 0.003).3 Furthermore, there were no significant dif-
Such outcomes may be partly linked to genetic and lifestyle ferences in the risk of diabetes or in BMI between offspring
exposure. But there is increasing evidence that fetal expo- according to the paternal diabetic status and onset of diabetes.
sure to maternal diabetes is one of the multiple risk factors of These results were confirmed in other ethnic groups. In a
chronic diseases and, in particular, the metabolic syndrome large prospective Swedish cohort study, BMI at 18 years old
(MS) in adults. of men whose mothers had diabetes mellitus during their
pregnancy was on average 0.94 kg/m2 greater (95% [CI],
0.35–1.52) than in their siblings born before their mother
Effect of maternal diabetes on the was diagnosed with diabetes, after adjustment for maternal
age, parity, and education.4 These findings suggest that intra-
offspring long-term health and risk uterine exposure to maternal diabetes has a negative impact
for disease on health outcomes, beyond genetic and lifestyle factors. In
a Nordic Caucasian population, the risk for T2D and predia-
Glucose intolerance and type 2 diabetes, overweight betes was found to be increased in offspring born to mothers
and obesity with diet-treated GDM (7.76 [95% CI 2.6–23.4]) and mothers
Extensive data on the consequences of exposure to diabetes in with type 1 diabetes (T1D) (4.0 [95% CI 1.3–1.9]).5 These
utero on childhood overweight and obesity and risk of type 2 results were confirmed by others.2 Interestingly, adult off-
diabetes (T2D) have been obtained from studies on Pima spring of women with T1D are more likely to have impaired
Indians, a population with an exceptionally high prevalence glucose tolerance and a deficient insulin secretory response
of obesity and T2D due to genetic reasons. The prevalence of to glucose than the offspring of fathers with T1D.6 These
T2D in offspring of Pima women has been shown to increase data favor the effect of exposure to maternal diabetes dur-
up to sixfold in those with diabetic or prediabetic mothers, ing pregnancy on the risk of T2D, regardless of the type of
while diabetes during childhood and adolescence occurred maternal diabetes and beyond transmitted diabetic genes.
almost exclusively among the offspring of diabetic and pre- The longitudinal cohort Exploring Perinatal Outcome
diabetic mothers.1 Accordingly, the offspring born to moth- among Children (EPOCH) study found that youth exposed
ers with pregestational T2D or gestational diabetes mellitus to maternal GDM had a higher average BMI growth trajec-
(GDM) are heavier at birth and at every age than those born tory from 27 months through 13 years of age and higher
to nondiabetic mothers. There is evidence that the higher BMI growth velocity starting at age 10–13 years. But no dif-
frequency of diabetes and obesity in the offspring of diabetic ferences were noted in growth velocity until 26 months of
Pima women is not only due to their genetic susceptibility to age.7 The effects of in utero exposure to diabetes thus extend
obesity and diabetes. Studies including sibling pairs in which beyond neonatal and early childhood periods and emerge
one sibling was born before and the other was born after the during puberty, a sensitive period for the development of
onset of maternal diabetes have brought interesting data.2 glucose intolerance and obesity.
519
520 Long-term outcomes after gestational diabetes mellitus exposure in the offspring
Cardiovascular and renal diseases In GDM, apart from hyperglycemia, the maternal metabolic
Offspring of Pima mothers who had diabetes during preg- environment is characterized by insulin resistance (IR) and
nancy had higher systolic arterial blood pressure than off- inflammation.17 Both conditions influence fetal growth.
spring of mothers who did not develop T2D until after IR facilitates maternal hypertriglyceridemia that enhances
pregnancy: this was independent of the levels of adiposity.8 substrate availability to the fetus. Other mechanisms influ-
Adverse cardiovascular effects of maternal diabetes on off- ence nutrient supply to the fetus. The placental transcriptome
spring were observed early in life and in other ethnic groups. has been shown to be a target of the altered environment of
When compared with offspring not exposed to maternal dia- diabetic pregnancy. For example, genes for lipids transport
betes, exposed offspring have a worse cardiovascular risk pro- have been shown to be upregulated in the placenta of women
file with the increased levels of circulating cellular adhesion with GDM, as well as are genes for inflammatory path-
molecules, which are biomarkers of adverse endothelium per- ways.18,19 Altogether, such alterations directly or indirectly
turbation. These markers are related to the earliest preclinical change the availability of substrates, other than glucose, to
stages of atherosclerosis and diabetes.9 Epidemiological stud- the fetus either by increasing their source or by modifying
ies that focused on blood pressure values in the adult offspring the materno-fetal interface. Additionally, placental epigen-
of mothers with GDM found a low increase in systolic blood etic changes were recently reported at gene loci involved in
pressure (SBP). A recent systematic review confirmed the energy metabolism regulation like those of adipokines.20
association between exposure to maternal diabetes and SBP This epigenetic adaptation to detrimental in utero environ-
in childhood. But this association was only significant in male ment may have impacts on the short- and long-term meta-
offspring. Furthermore, there is some evidence that this asso- bolic regulations of the newborn. Maternal pregestational
ciation may be influenced by maternal prepregnancy BMI.10 overweight or obesity that frequently associated with GDM
Endothelial dysfunction (ED) has been thought to be criti- or T2D may also increase lipid availability, modulate deliv-
cal in the development of vascular disease, notably in HT.11 ery of lipid substrates to the fetus, and have programming
ED can be characterized by a loss of regulatory functions consequences.
related to vascular tone, inflammation, and oxidative stress In a multiethnic group, youth aged 10–22 years with T2D
but also by an impaired vasculogenesis and capacity of repair were more likely to have been exposed to maternal diabe-
mediated by circulating endothelial progenitor cells (EPCs). tes or obesity. Exposure to maternal diabetes (OR 5.7 [95%
These cells are now considered as a strong biomarker to CI 2.4–13.4]) and exposure to maternal obesity (2.8 [95% CI
assess ED and have been considered as endothelial colony- 1.5–5.2]) were independently associated with T2D in the off-
forming cells (ECFCs).12 It was shown in vitro and in vivo spring. Exposure to maternal diabetes in utero resulted in
that hyperglycemia or exposure of ECFCs to a diabetic intra- an attributable risk of only 4.7%, although 47.2% of T2D in
uterine environment reduced ECFC colony formation and youth could be attributed to intrauterine exposure to mater-
capillary-like tube formation13 and increased senescence and nal diabetes and obesity together.21
reduced proliferation.14 These data provide potential mecha- Concerning the link between in utero diabetes and over-
nistic insights into the development of ED and the long-term weight and obesity in offspring, the literature provides con-
cardiovascular complications observed in newborns of dia- tradictory results. Although in utero exposure to maternal
betic pregnancies. diabetes plays a role in the occurrence of overweight in the
Exposure to a diabetic intrauterine environment is also offspring, the importance and the respective significance of
considered as a strong risk factor for renal disease. Diabetic the exposure to GDM relative to other risk factors associ-
nephropathy is the major cause of end-stage renal disease.15 ated with maternal diabetes (genetic, maternal overweight,
In Pima Indians, an increase in urinary albumin excretion fetal macrosomia, lifestyle, socioeconomic) have been poorly
(UAE) in offspring of diabetic mothers has been observed: quantified. After adjusting on maternal prepregnancy
UAE was 58% in offspring of mothers with GDM, 43% in off- BMI, the association between maternal diabetes and off-
spring of mothers who develop diabetes after pregnancy, and spring BMI was no longer significant in some studies.22,23
40% in offspring of nondiabetic mothers.16 These functional Although convincing data exist in the literature on the effect
changes might result in damage to the developing glomeruli, of maternal diabetes on offspring health, many unanswered
possibly related to a similar process of decreased number of questions remain regarding the size effect of maternal intra-
nephrons. uterine exposure compared with shared genetic traits. These,
in turn, are difficult to distinguish from influences of the
child’s postnatal environment and lifestyle.
Limitations and confounding factors
Some limitations in the available studies must, however, be
taken into consideration. Most studies encounter major dif- Effect of high birth weight
ficulties to separate the responsibility of fetal exposure to on long-term outcomes
maternal hyperglycemia alone in the long-term outcome,
from fetal exposure to coexisting maternal manifestations Maternal diabetes is one of the main risk factors of having
like overweight/obesity. Different interrelated mechanisms macrosomic infants at birth. A number of publications have
play a role in the modification of fetal nutrition and metab- reported the link between high birth weight (BW) and obe-
olism and may have an impact on the long-term outcome. sity in childhood to early adulthood. A meta-analysis showed
Effect of breastfeeding 521
that BW ≥4000 g increases twofold the risk for obesity, and term because it may not be sufficient alone to prevent the
this risk is increased about 2.5-fold when BW exceeds the adverse long-term outcome on the offspring.17 As discussed
90th percentile.24 However, the association between higher earlier, other substrates influence fetal growth and probably
BW and higher adult BMI may be due to increased lean mass have long-term metabolic consequences. Excess gestational
rather than an increase in fat tissue. Evidence suggests that weight gain (GWG) is a factor associated with high BW and
although BW is positively associated with BMI, it is not nec- subsequent long-term outcomes. The comparison of differ-
essarily associated with increased adiposity, and at higher ences in BW between sibling pairs showed that for every
BW, subsequent adiposity may actually be reduced.25 additional kilogram an individual woman gained during
Being large for gestational age (LGA, BW > 90th percen- pregnancy, the BW of her offspring increased by about 25 g.31
tile), in association with GDM or maternal obesity, increases In a large prospective multicentric study, excessive GWG
the risk of MS in childhood. A longitudinal cohort study was an independent valuable predictor of macrosomia. In
analyzed the prevalence of MS in children aged 6–11 years, the subgroup of diabetic mothers, excessive GWG (accord-
accordingly they were LGA or adapted for gestational age ing to the Institute of Medicine [IOM] recommendations)
(AGA, BW 10th–90th percentile), and their mothers had was related to a 2.6-fold increased risk of developing macro-
or had no GDM. The prevalence at any time of at least two somia (OR 2.6; 95% CI 1.2–5.5). However, in this population,
components of MS was higher for the LGA/GDM group GWG lower than recommended was not associated with a
(50%), compared to the LGA/control group (29%), AGA/ reduction of macrosomia (OR 0.8; 95% CI 0.3–1.8).32 In a
GDM group (21%), and AGA/control group (18%). The risk of large Swedish prospective cohort, offspring BMI was related
developing MS with time was significantly different between to maternal GWG. However, in normal-weight women, this
LGA and AGA offspring in the GDM group, with a 3.6-fold positive association was only driven by shared familial risk
greater risk among LGA children by 11 years. In this study, factors for BMI (genetic/environment). In overweight and
children exposed to maternal obesity were also at increased obese women, a greater maternal GWG appears to be associ-
risk of developing MS.26 ated with greater offspring BMI via intrauterine mechanisms
A recent systematic review on the association between and shared familial characteristics.4
BW and risk of T2D showed that in most populations stud- Maternal weight gain during pregnancy is probably an
ied, BW was inversely related to T2D risk. There was a posi- important and challenging modifiable risk factor because
tive association between high BW (>4 kg) and risk of T2D excessive GWG is associated with higher BW. The absence
only in two native North American populations. These pop- of effect of reduced GWG on fetal growth in diabetic preg-
ulations have an exceptionally high prevalence of T2D and nancy, reported in some studies, may be due to the fact that
obesity from early ages and a very high prevalence of GDM. IOM guidelines were not conceived for diabetic mother.
Therefore, the influence of maternal diabetes/obesity on the Then, the challenge is to determine nutritional guidelines
BW–T2D association could not be excluded in these popula- specifically devised for women with diabetes, considering
tions and may overweight the effect of being LGA alone.27 the amounts of calories but also the quality of nutrients.
Other factors of prevention (prepregnancy control of
overweight, regular activity, etc.) must be recommended to
the mother as part of a comprehensive management plan.
Effect of maternal glycemic and
gestational weight gain control on
long-term prognosis in the offspring Effect of breastfeeding
Treatment of GDM may be an important determinant of off- Infancy has been suggested as another critical period for future
spring outcomes, but there is currently insufficient informa- risks in the offspring whatever was the intrauterine environ-
tion to firmly assess the effects of maternal interventions on ment. Breastfeeding compared with formula feeding showed to
infant and child health. Two trials are available in the litera- have beneficial effects on glucose tolerance, HT, dyslipidemia,
ture, with a follow-up of offspring of mothers randomized to and obesity.33 The strongest evidence for a protective effect
either minimal intervention or tight control of diabetes.28,29 of breastfeeding for later health is for a lower risk of obesity.
They did not demonstrate any differences in the frequency of Although unknown, some potential mechanisms by which
MS or in the weight of the children according to the inter- breastfeeding protects against later risk have been discussed.
vention provided to their mothers. Breastfed babies may control the amount of milk they consume
In a large cohort of 9439 mother–child pairs in a diverse and so learn to self-regulate their energy intakes better than
US population universally screened for GDM, it was shown those given formula, although whether this difference persists
that increasing hyperglycemia levels in pregnancy were into adult life is unknown. Nutritional benefits of breastfeed-
associated with increased risk of obesity in children at age ing may include differences in nutrients between human milk
5–7 years. But this risk was modifiable by treating GDM, as and formulas (lower glucose and protein, high concentrations
obesity risk was attenuated and no longer significant after in long-chain polyunsaturated fatty acid). Differences in early
multivariate adjustment in the treated GDM group.30 protein intakes that are greater in formula than in human milk
There are significant methodological difficulties in prov- could also affect later adiposity. It has been suggested that the
ing an effect of maternal glycemic control over the long benefits of breastfeeding for long-term obesity is of particular
522 Long-term outcomes after gestational diabetes mellitus exposure in the offspring
importance in the early postnatal weeks of life. This is a critical BMI growth velocity both during infancy and in the child-
period for determining levels and distribution of adiposity. The hood period, in offspring of nondiabetic mothers, as well
benefits of breastfeeding at this time may be due to a slower pat- as in offspring of diabetic mothers. These effects were inde-
tern of growth in breastfed compared with formula-fed infants pendent of sex, race/ethnicity, current childhood diet, and
(the growth acceleration hypothesis).34 Weight gain during the physical activity levels. This study indicates that the favor-
first week of life in healthy formula-fed infants was associated able effects of breastfeeding on BMI growth patterns extend
with overweight status two to three decades later. After adjust- throughout the entire childhood period and are also pres-
ment for important confounding factors, each 100 g increase in ent in youth at increased risk for obesity due to intrauterine
absolute weight gain during this period was associated with a exposure to maternal diabetes.41 Others have reported data
28% increase in the risk of becoming an overweight adult (95% that favor the benefit of breastfeeding in offspring of dia-
CI 8%–52%).35 A long-term advantage of breastfeeding was betic mothers, either on the risk of obesity42 or on the risk
further supported by a “dose–response” effect. A longer dura- of diabetes.43
tion of breastfeeding was associated with a lower tendency to Overall, there is no doubt that breastfeeding should be
later obesity; each month of breastfeeding was associated with recommended in diabetic mothers, all the more since GDM
a 4% (95% CI −6% to −2%) reduction in obesity risk.36 and T2D are increasingly spreading in emerging countries.
In diabetic mothers, there are many controversies con- The controversies concerning the benefit of breastfeed-
cerning breastfeeding benefits in their offspring, particularly ing in diabetic mothers should be reviewed in the light of
during the first week of life. Some observed a positive dose- potential confounders, especially the time point of expo-
dependent relation between the volume of diabetic mother sure, exposure quality (exclusive or not exclusive breast-
milk ingested during the first neonatal week and later risk feeding), and duration, but also ethnicity, socioeconomic
of overweight.37 But those infants who did not receive breast status, and maternal BMI. Breastfeeding may represent an
milk from their diabetic mothers were instead nourished important challenge in preventing long-term disorders in
with banked breast milk from nondiabetic donor mothers. the offspring of diabetic mothers.
This created a different “reference exposure” than in all other
studies where breastfeeding was tested against formula. The
mechanisms evoked for such a negative effect of breast milk Conclusion
from diabetic mothers was that it contains increased levels
of glucose and insulin as compared with breast milk from More and more indisputable data are evidencing long-term
healthy mothers but even hormones like insulin and leptin consequences on offspring health in case of gestational dia-
that may be absorbed from milk in the immature gut of an betes or excess growth at birth. Converging results from
infant.38 However, the composition of milk from diabetic clinical and epidemiological data, but also from animal
mothers may also be influenced by metabolic control during models (see Chapter 10), suggest that GDM contributes to
pregnancy and postpartum, and this may change the impact the current worldwide T2D pandemic. However, the specific
on the outcome of offspring. The same authors found that mechanisms remain unclear and many aspects, in particu-
neither late neonatal intakes of breast milk from diabetic lar epigenetic, need to be clarified. The specific intrauterine
mothers nor the duration of breastfeeding has an indepen- metabolic alterations that favor such long-term evolution
dent influence on childhood risk of overweight or glucose need to be investigated, and the consequences of overweight
intolerance.39 However, others found that adequate breast- and obesity among those of diabetes should be separately
feeding (≥6 months) reduces, in childhood, the increase evaluated. Such comprehensive data will offer major proved
of adiposity levels associated with exposure to diabetes in opportunities for prevention. Meanwhile, obvious preventive
utero.40 Furthermore, these results were strengthened by the measures are indicated like early diagnosis and tight control
follow-up of a longitudinal cohort. It was shown that ade- of maternal diabetes during pregnancy, limited GWG, and
quate breastfeeding reduced the overall body size and slowed promotion of breastfeeding.
REFERENCES
1. Dabelea D, Pettitt DJ. Intrauterine diabetic environment confers 5. Clausen TD, Mathiesen ER, Hansen T et al. High prevalence of
risks for type 2 diabetes mellitus and obesity in the offspring, type 2 diabetes and pre-diabetes in adult offspring of women
in addition to genetic susceptibility. J Pediatr Endocrinol Metab with gestational diabetes mellitus or type 1 diabetes: The role of
2001; 14: 1085–1091. intrauterine hyperglycemia. Diabetes Care 2008; 31: 340–346.
2. Dabelea D. The predisposition to obesity and diabetes in off- 6. Sobngwi E, Boudou P, Mauvais-Jarvis F et al. Effect of a diabetic
spring of diabetic mothers. Diabetes Care 2007; 30(Suppl. 2): environment in utero on predisposition to type 2 diabetes. Lancet
S169–S174. 2003; 361: 1861–1865.
3. Dabelea D, Hanson RL, Lindsay RS et al. Intrauterine exposure to 7. Crume TL, Ogden L, Daniels S et al. The impact of in utero expo-
diabetes conveys risks for type 2 diabetes and obesity: A study of sure to diabetes on childhood body mass index growth trajecto-
discordant sibships. Diabetes 2000; 49: 2208–2211. ries: The EPOCH study. J Pediatr 2011; 158: 941–946.
4. Lawlor DA, Lichtenstein P, Langstrom N. Association of maternal 8. Bunt JC, Tataranni PA, Salbe AD. Intrauterine exposure to dia-
diabetes mellitus in pregnancy with offspring adiposity into early betes is a determinant of hemoglobin A(1)c and systolic blood
adulthood: Sibling study in a prospective cohort of 280,866 men pressure in pima Indian children. J Clin Endocrinol Metab 2005;
from 248,293 families. Circulation 2011; 123: 258–265. 90: 3225–3229.
References 523
9. West NA, Crume TL, Maligie MA et al. Cardiovascular risk factors 27. Whincup PH, Kaye SJ, Owen CG et al. Birth weight and risk of
in children exposed to maternal diabetes in utero. Diabetologia type 2 diabetes: A systematic review. J Am Med Assoc 2008; 300:
2011; 54: 504–507. 2886–2897.
10. Aceti A, Santhakumaran S, Logan KM et al. The diabetic preg- 28. Keely EJ, Malcolm JC, Hadjiyannakis S et al. Prevalence of meta-
nancy and offspring blood pressure in childhood: A systematic bolic markers of insulin resistance in offspring of gestational dia-
review and meta-analysis. Diabetologia 2012; 55: 3114–3127. betes pregnancies. Pediatr Diabetes 2008; 9: 53–59.
11. Taddei S, Virdis A, Mattei P et al. Endothelium-dependent 29. Gillman MW, Oakey H, Baghurst PA et al. Effect of treatment of
forearm vasodilation is reduced in normotensive subjects with gestational diabetes mellitus on obesity in the next generation.
familial history of hypertension. J Cardiovasc Pharmacol 1992; Diabetes Care 2010; 33: 964–968.
20(Suppl. 12): S193–S195. 30. Hillier TA, Pedula KL, Schmidt MM et al. Childhood obesity and
12. Ingram DA, Mead LE, Tanaka H et al. Identification of a novel metabolic imprinting: The ongoing effects of maternal hypergly-
hierarchy of endothelial progenitor cells using human peripheral cemia. Diabetes Care 2007; 30: 2287–2292.
and umbilical cord blood. Blood 2004; 104: 2752–2760. 31. Hutcheon JA, Platt RW, Meltzer SJ et al. Is birth weight modi-
13. Ingram DA, Lien IZ, Mead LE et al. In vitro hyperglycemia or a fied during pregnancy? Using sibling differences to understand
diabetic intrauterine environment reduces neonatal endothelial the impact of blood glucose, obesity, and maternal weight
colony-forming cell numbers and function. Diabetes 2008; 57: gain in gestational diabetes. Am J Obstet Gynecol 2006; 195:
724–731. 488–494.
14. Blue EK, DiGiuseppe R, Derr-Yellin E et al. Gestational diabetes 32. Alberico S, Montico M, Barresi V et al. The role of gestational
induces alterations in the function of neonatal endothelial col- diabetes, pre-pregnancy body mass index and gestational
ony-forming cells. Pediatr Res 2014; 75(2): 266–272. weight gain on the risk of newborn macrosomia: Results from a
15. Atkins RC, Zimmet P. Diabetic kidney disease: Act now or pay prospective multicentre study. BMC Pregnancy Childbirth 2014;
later. Acta Diabetol 2010; 47(1): 1–4. 14: 23.
16. Nelson RG, Morgenstern H, Bennett PH. Intrauterine diabetes 33. Lanigan J, Singhal A. Early nutrition and long-term health: A prac-
exposure and the risk of renal disease in diabetic Pima Indians. tical approach. Proc Nutr Soc 2009; 68: 422–429.
Diabetes 1998; 47: 1489–1493. 34. Singhal A, Lanigan J. Breastfeeding, early growth and later obe-
17. Catalano PM, Hauguel-De Mouzon S. Is it time to revisit the sity. Obes Rev 2007; 8(Suppl. 1): 51–54.
Pedersen hypothesis in the face of the obesity epidemic? Am J 35. Stettler N, Stallings VA, Troxel AB et al. Weight gain in the first
Obstet Gynecol 2011; 204: 479–487. week of life and overweight in adulthood: A cohort study of
18. Radaelli T, Lepercq J, Varastehpour A et al. Differential regulation European American subjects fed infant formula. Circulation
of genes for fetoplacental lipid pathways in pregnancy with ges- 2005; 111: 1897–1903.
tational and type 1 diabetes mellitus. Am J Obstet Gynecol 2009; 36. Harder T, Bergmann R, Kallischnigg G et al. Duration of breast-
201: 209.e1–209.e10. feeding and risk of overweight: A meta-analysis. Am J Epidemiol
19. Radaelli T, Varastehpour A, Catalano P et al. Gestational diabe- 2005; 162: 397–403.
tes induces placental genes for chronic stress and inflammatory 37. Plagemann A, Harder T, Franke K et al. Long-term impact of neo-
pathways. Diabetes 2003; 52: 2951–2958. natal breast-feeding on body weight and glucose tolerance in
20. Houde AA, Hivert MF, Bouchard L. Fetal epigenetic program- children of diabetic mothers. Diabetes Care 2002; 25: 16–22.
ming of adipokines. Adipocyte 2013; 2: 41–46. 38. Plagemann A, Harder T. Fuel-mediated teratogenesis and breast-
21. Dabelea D, Mayer-Davis EJ, Lamichhane AP et al. Association of feeding. Diabetes Care 2011; 34(3): 779–781.
intrauterine exposure to maternal diabetes and obesity with type 39. Rodekamp E, Harder T, Kohlhoff R et al. Long-term impact of
2 diabetes in youth: The SEARCH Case-Control Study. Diabetes breast-feeding on body weight and glucose tolerance in children
Care 2008; 31: 1422–1426. of diabetic mothers: Role of the late neonatal period and early
22. Kim SY, England JL, Sharma JA et al. Gestational diabetes mel- infancy. Diabetes Care 2005; 28: 1457–1462.
litus and risk of childhood overweight and obesity in offspring: A 40. Crume TL, Ogden L, Maligie M et al. Long-term impact of neo-
systematic review. Exp Diabetes Res 2011; 2011: 541308. natal breastfeeding on childhood adiposity and fat distribution
23. Philipps LH, Santhakumaran S, Gale C et al. The diabetic preg- among children exposed to diabetes in utero. Diabetes Care
nancy and offspring BMI in childhood: A systematic review and 2011; 34: 641–645.
meta-analysis. Diabetologia 2011; 54: 1957–1966. 41. Crume TL, Ogden LG, Mayer-Davis EJ et al. The impact of neo-
24. Monasta L, Batty GD, Cattaneo A et al. Early-life determinants natal breast-feeding on growth trajectories of youth exposed and
of overweight and obesity: A review of systematic reviews. Obes unexposed to diabetes in utero: The EPOCH Study. Int J Obes
Rev 2010; 11: 695–708. (Lond) 2012; 36: 529–534.
25. Rogers I. The influence of birthweight and intrauterine environ- 42. Mayer-Davis EJ, Rifas-Shiman SL, Zhou L et al. Breast-feeding and
ment on adiposity and fat distribution in later life. Int J Obes Relat risk for childhood obesity: Does maternal diabetes or obesity sta-
Metab Disord 2003; 27: 755–777. tus matter? Diabetes Care 2006; 29: 2231–2237.
26. Boney CM, Verma A, Tucker R et al. Metabolic syndrome in 43. Pettitt DJ, Knowler WC. Long-term effects of the intrauterine
childhood: Association with birth weight, maternal obesity, and environment, birth weight, and breast-feeding in Pima Indians.
gestational diabetes mellitus. Pediatrics 2005; 115: e290–e296. Diabetes Care 1998; 21(Suppl. 2): B138–B141.
62 Metabolomics and diabetic
pregnancy
Angelica Dessì, Roberta Carboni, and Vassilios Fanos
Thus, the possibility that glucose intolerance may be prior

Introduction to or begin during pregnancy cannot be excluded. 5 This
Diabetes is one of the diseases that most frequently compli- pathology has a prevalence of 3%–10% and represents
cate the course of pregnancy, especially in developed coun- approximately 90% of all pregnancies complicated by dia-
tries where the increased incidence of obesity is one of the betes.2,6 Moreover, the increase in maternal age, the high
risk factors.1 incidence of obesity and diabetes, and the reduction of
Gestational diabetes mellitus (GDM) is defined as any physical activity with the adoption of modern lifestyles in
degree of glucose intolerance with onset, or which is recog- developing countries have contributed to an increase in the
nized for the first time, during pregnancy. In some cases, it prevalence of GDM.7
is a pregestational diabetes revealed by investigations during
pregnancy, but in other cases it is diabetes with onset during
pregnancy; the latter represents 90% of pregnancies compli- Pathophysiology and etiopathogenesis
cated by diabetes.2 Pregnancy is a condition during which a certain degree of
GDM is a disease that has important repercussions on insulin resistance develops physiologically, especially in
both the mother and the child, and, therefore, early diagno- the muscle and adipose tissues, to ensure fetal growth.8 In
sis and timely treatment may avoid acute and chronic com- the maternal organism, there is a decreased use of insulin-
plications for both.3 mediated glucose that favors an increase in carbohydrates
At present, the only laboratory method for the diagnosis for the fetus.9 In a physiological pregnancy, the basal glyce-
of gestational diabetes is glycemia measurement. However, mia remains constant up to the last 3 months and carbohy-
this is a method that does not take into account the overall drate intolerance develops only when β-cell secretion is no
metabolic condition of the organism and that is often per- longer capable of compensating for peripheral insulin resis-
formed at an advanced stage of pregnancy. Moreover, there tance.10 Most of the evidence points to a defect in the β-cell if
are still no methods that are useful in singling out women at the GDM is more than acute. The hypothesis that the β-cell
greater risk of developing GDM and assessing the effective- defect is chronic suggests that when GDM is diagnosed, it
ness of therapy in single patients. includes some women with a preexistent glucose intolerance
Metabolomics, thanks to its capacity to investigate the revealed by tests performed during pregnancy.
overall metabolic state of an entire organism, appears to be From the etiopathogenetic viewpoint, GDM is consid-
a promising technique also in the study of GDM and the ered a type 2 diabetes. GDM is frequently characterized
recognition of possible new biomarkers to reveal at an early by reduced insulin secretion accompanied by an increase
stage the women at higher risk of developing this pathologi- in peripheral insulin resistance, two conditions typical of
cal condition. type 2 diabetes. During pregnancy, the variations in insulin
secretion depend on endocrine alterations that accompany
pregnancy. Changes in β-cell function take place at the time
of development of the fetus and placenta and the production
Diabetic pregnancy of hormones such as the human chorionic growth hormone,
progesterone, cortisol, and prolactin.
Definition These hormones in a woman with GDM are capable of
By GDM is meant a condition characterized by a reduced causing insulin resistance, high triglyceride levels, and lower
tolerance to glucides, with hyperglycemia of varying seri- HDL cholesterol levels compared to what is observed in
ousness that begins during pregnancy.4 The definition is women with physiological pregnancies.11
applied regardless of the use of insulin in the treatment In GDM patients, insulin secretion is incapable of
or persistence of the condition following the pregnancy. compensating for the insulin resistance characteristic of
524
Metabolomics 525
pregnancy. Therefore, the loss of the first phase of insulin relation to the fact that GDM may be the first manifestation
secretion will cause postprandial hyperglycemia, while the of a type 2 diabetes or a metabolic syndrome; thus after preg-
reduced suppression of hepatic production of glucose will be nancy, it is important for these women to be followed up.14
responsible for fasting hyperglycemia. The early complications for the neonate are respiratory
distress, shoulder dystocia, hyperbilirubinemia, polycythe-
mia, hypocalcemia, macrosomy, and hypoglycemia; the late
Diagnosis ones are instead connected with a higher risk of overweight
As concerns diagnosis, guidelines recommend the measure- and obesity in the adolescent age as well as the higher risk
ment of blood sugar at the first appointment in pregnancy of developing type 2 diabetes, this too in the adolescent
for all women with no previous tests to single out those with period.
diabetes prior to becoming pregnant; also recommended The reduced action of insulin in GDM causes an exces-
is the screening for diabetes in cases of physiological preg- sive increase of nutrients in the blood (glucose, lipids, amino
nancy using defined risk factors. acids), which, on passing through the placenta, determine
At the 16th–18th week of gestation, an OGTT with 75 g hyperinsulinism capable of favoring a consequent increase
of glucose and a second OGTT with 75 g at 28 weeks of in adipose tissue with fetal organomegaly and macrosomia.10
gestational age if the first test gave a normal result should be On passing through the placental barrier, the excess mater-
performed on all women presenting at least one of the fol- nal glucose causes a condition of constant hyperglycemia in
lowing conditions: the fetus that will lead to chronic hyperinsulinism that, on
continuing after delivery, will result in neonatal hypogly-
●● Gestational diabetes in a previous pregnancy cemia.15 This reduced glucidic tolerance will tend to persist
●● Prepregnancy body mass index (BMI) ≥ 30 during growth and into adulthood and consequently will
●● The finding of blood sugar values between 100 and expose the children of diabetic mothers to a higher risk of
125 mg/dL (5.6–6.9 mmol/L) previous to or at the begin- developing pathologies such as obesity and type 2 diabetes.
ning of pregnancy
At the 24th–28th week of gestational age, women presenting

at least one of the following conditions must undergo a 75 g Metabolomics
OGTT: Metabolomics, one of the most recent of the omics sciences,
is defined as the study of the complex system of metabolites
●● Age ≥ 35 years
owing to its capacity of examining an individual’s entire
●● Prepregnancy BMI ≥ 25 kg/m2
metabolic profile.
●● Fetal macrosomia in a previous pregnancy (≥4.5 kg)
A metabolomic analysis is capable of detecting analytes
●● GDM in a previous pregnancy (even with a normal result
of relatively low molecular weight (up to 1000 Da), among
at the 16th–18th week)
which amino acids, oligopeptides, sugars, steroids, biliary
●● Family anamnesis of diabetes (first-degree relative with
acids, simple and compound fatty acids, and intermediate
type 2 diabetes)
compounds of many biochemical pathways.16 The metabolo-
●● Family from an area with a high rate of diabetes: Southern
mic approach consists of the sequential application of a sur-
Asia, the Caribbean, and the Middle East
veying technique with the use of nuclear magnetic resonance
Based on the results of this screening, women with fasting (NMR) and gas or liquid chromatography coupled with mass
glycemia ≥92 mg/dL, after 1 hour ≥180 mg/dL, and after spectrometry (LC-MS and GC-MS), by means of which it is
2 hours ≥153 mg/dL are defined as affected by GDM. The possible to measure simultaneously the concentration of a
guidelines also specify that in screening for GDM, the fol- large number of metabolites in the sample. The most com-
lowing are not to be used: fasting blood sugar, random glyce- mon biological samples used in metabolomics are urine,
mias, glucose challenge test (GCT) or minicurve, glycosuria, plasma, and serum, but other biofluids can be used as well
and OGTT 100 g.12 (liquor, saliva, gastric/pancreatic juices, aqueous humor) or
tissues. Each biological fluid has a unique and characteristic
biochemical composition that changes in response to physio-
Acute and chronic consequences for mother and child logical or pathophysiological stimuli to generate a metabolic
GDM causes an increased risk of morbidity in the fetus “fingerprint.” This discipline, based on the use of mathemati-
and neonate and may be the cause of a later development of cal and statistical methods to solve multivariate problems,
type 2 diabetes, in both the mother and the child.13 We can is capable of describing the chemical profile of a biological
distinguish early and late complications in connection with system in terms of low-molecular-weight metabolites present
the maternal pathology that will involve mother and child. in cells, tissues, organs, and biological fluids.
As concerns the mother, early complications are represented By means of a metabolic approach, it is thus possible to
by the risk of spontaneous abortion, polyhydramnios, hyper- photograph “in real time” the influences on the organism
tension in pregnancy, pyelonephritis and other infections, such as biochemical perturbations caused by diseases, drugs,
preterm delivery, hypoglycemia, ketoacidosis, and cesar- and toxins, thus describing the biochemical phenotype of a
ean delivery. The late complications for the mother are in biological system.17
526 Metabolomics and diabetic pregnancy
In recent years, many clinical studies have been performed As concerns GDM, metabolomics proposes to identify
on animals, adults, children, and neonates using a metabo- biomarkers useful in predicting, already at an early stage of
lomic approach that makes it possible to study important gestation, if a woman is prone to GDM and thus beginning
perinatal and postnatal pathologies such as asphyxia and timely treatment so that the disease does not develop and
hypoxia, congenital diseases, the state of nutrition, kidney that the fetus does not undergo any of the mother’s metabolic
diseases, nephrotoxicity, cardiovascular disease, and other influences that may impact on its metabolism and growth.
pathologies.18–20 In the medical field, there is growing inter- Moreover, thanks to metabolomics, it is possible to study
est in characterizing at the biochemical and molecular level more in depth the metabolic profile of pregnant women.
the different pathophysiological states that mark the lives of Up to the present, few works have been published on the
human beings in the period from embryonic development to application of metabolomics to the study of women pre-
ageing and death. “Clinical metabolomics” is proposed as a senting GDM and the assessment of the metabolic state of
new multidisciplinary scientific field developed through the the children of GDM mothers (Table 62.1), but the results
interaction of disciplines such as chemistry, physics, biol- appear promising both as regards the overall metabolic state
ogy, and medicine, which to the same extent contribute to of women and their risk of developing GDM and as regards
the gathering of up-to-now latent information, leading to the the metabolic profile of the neonate.23
identification of new phenotypical characteristics of diseases In the literature, we find four studies that focused on the
studied but not yet adequately defined. The ultimate objec- metabolomic analysis of maternal biofluids in women pre-
tive of such characterization is that of developing “tailored” senting GDM and two studies that assessed the metabolome
therapeutic approaches for the single individual and/or of the children of diabetic mothers.
group of individuals, where for group is meant persons hav- The study by Sachse et al.14 examined the urinary meta-
ing a common phenotypical profile deriving from a similar bolic profile by means of NMR of pregnant women for the
gene–environment interaction.21,22 purpose of identifying biomarkers useful in identifying the
women most at risk of developing GDM. The urinary meta-
bolic profiles of 823 pregnant women in good health were
Gestational diabetes and analyzed. Urine samples were taken in the morning while
metabolomics fasting and during three different visits: from the 8th to the
20th week of pregnancy, from the 26th to the 30th week, and
As can be seen in the guidelines, at present the diagnosis of from 10 to 16 weeks after delivery. The authors highlighted
GDM is based solely on the finding of high values of glucose the substantial changes in the composition of the urinary
in the blood. This kind of investigation does not take into metabolites during and after the pregnancy; in particular,
consideration the complete metabolic picture of a woman the most important results were the constant increase in lac-
as well as being, in cases not considered at risk, performed tose and the increasing–decreasing pattern of a plurality of
toward the end of the 6 month of pregnancy when, if it is NMR signals between 0.55 and 1.10 ppm, signals that may
found altered, the fetus is already affected by a series of meta- belong to pregnanediol and estrogens, or more probably, to
bolic influences capable of impacting on its growth. their sulfates and water-soluble glucuronoids.
For this reason, the application of an “alternative” As concerns the possibility of finding biomarkers capable
method is necessary to help in characterizing the overall of identifying GDM, the authors concluded that on the basis
metabolic state of a pregnant woman. Metabolomics in the of their results, it is not possible to find reliable biomark-
field of this pathology would appear to be the most suitable ers for GDM in a large multiethnic population; however, an
method since it is capable of providing a quantitative mea- increase in the citrate excreted in the urine in relation to the
surement as a function of time of the metabolic response of seriousness of the cases of GDM was found.
a living system to pathophysiological stimuli and/or genetic The work by Diaz et al.24 was performed on urine and
modifications.16 plasma of women in the final 3 months of pregnancy in an
Table 62.1 Studies in the literature concerning metabolomics and GDM
Metabolomic
Study Sample Population analysis Metabolites resulting
Dani et al. 28 Cord serum Infants 1H-NMR Glucose, pyruvate, histidine, alanine, valine,
methionine, arginine, lysine, hypoxanthine
Logan et al.27 Urine Infants 1H-NMR Glucose, formate, fumarate, succinate, citrate
Graça et al.26 Urine, amniotic fluid Mothers UPLC-MS None
1H-NMR
Sachse et al.14 Urine Mothers 1H-NMR Lactose, citrate
Diaz et al.24 Urine, plasma Mothers 1H-NMR 3-Hydroxyisovalerate, 2-hydroxyisobutyrate
Scioscia et al.25 Urine Mothers MS Inositolphosphoglycan P-type
References 527
attempt to characterize metabolic changes possibly caused MATLAB® software*. The statistical analysis was performed
by prenatal disorders and identify possible early biomarkers with the principal component analysis (PCA) and partial
of diseases. The metabolic profile was analyzed by means of least square-discriminant analysis (PLS-DA). The results
NMR and the metabolomic analysis revealed higher levels obtained in this study made it possible to discriminate
of 3-hydroxyisovalerate and 2-hydroxyisobutirrate in the between the two groups of neonates: the most important
urine of women who later developed GDM. differences in the urinary metabolites were those found
The work by Scioscia et al. 25 assessed by means of a for glucose, formate, fumarate, citrate, and succinate, all
metabolomic approach performed with MS on the urine of metabolites involved in the tricarboxylic acid cycle.
pregnant women if inositolphosphoglycan P-type (P-IPG), Another similar work was carried out by Dani et al.,28
a second insulin messenger that correlates with the degree who studied the metabolome of newborn children of GDM
of resistance in diabetic subjects, may have changed dur- mothers and compared it with that of the neonates of healthy
ing pregnancy. The study examined 48 women with GDM mothers. The purpose of the study was to see if there were
and 23 healthy women during pregnancy and revealed an differences between the two groups and if such differences
increase in urinary excretion of P-IPG in the GDM women remained evident despite close control over the GDM.
compared to controls. The authors also found a correla- The metabolomic analyses with 1HNMR were performed
tion between urinary values of P-IPG and blood glucose. on the serum of 30 IGDM term neonates and 40 term neo-
According to the authors, the metabolomic identification nates of healthy mothers used as controls. The results of
of P-IPG is a potential marker of insulin resistance and these investigations revealed a lower level of glucose and a
may predict fetal growth alteration in GDM patients. higher level of pyruvate, histidine, alanine, valine, methio-
Finally, the study by Graça et al. 26 was performed nine, arginine, lysine, hypoxantin, lipoproteins, and lipids
on urine and amniotic liquid of women in the second in the neonates of GDM mothers compared to the controls.
3 months of pregnancy with the use of both the UPLC-MS Despite these variations, the authors found no differences
and 1H-NMR methods. The authors investigated the pos- from the clinical standpoint.
sible metabolic effects of fetal malformations, GDM, and
preterm delivery. In particular, contrary to the other Conclusions
works published, in the metabolome in the case of GDM,
no significant changes were found, either in the urine or in A new challenge in medicine is certainly the application of
the amniotic liquid. metabolomics in the study of pathologies originating in the
The other two studies in the literature examined the fetal womb that may later lead to both short- and long-term com-
metabolism of neonates of GDM mothers to assess the con- plications, as in the case of GDM. The studies published show
sequences of the maternal pathology on the development of that metabolomics can be considered a potential and formi-
their children. dable instrument for prevention by identifying women most
In their study, Logan et al.27 hypothesized that the meta- at risk of developing GDM. An evaluation of this kind would
bolic profile of the neonates of diabetic mothers (IDM) was make it possible to indicate a therapy, a lifestyle, and diet to
different from that of children of healthy mothers. They follow in pregnancy so as to avoid the frequent complications
then studied the metabolome of 18 IDM neonates and 12 caused by GDM.
term and healthy neonates used as controls. Urine was col-
lected within 72 hours from birth using balls of absorbent * MATLAB® is a registered trademark of The MathWorks, Inc. For product
cotton and then conserved at −80°C; the metabolomic information, please contact The MathWorks, Inc. at 3 Apple Hill Drive
analysis was then performed by means of 1HNMR spec- Natick, MA 01760-2098, USA; 508-647-7000 (telephone); 508-647-7001
troscopy and the spectra obtained were analyzed using (fax); [email protected] (e-mail); www.mathworks.com (website).
REFERENCES
1. Abourawi FI. Diabetes mellitus and pregnancy. Libyan J Med 7. Ferrara A. Increasing prevalence of gestational diabetes mellitus.
2006; 1: 28–41. Diabetes Care 2007; 30: S141–S146.
2. Lapolla A, Dalfrà MG, Lencioni C et al. Epidemiology of diabe- 8. Kühl C. Etiology and pathogenesis of gestational diabetes.
tes in pregnancy: A review of Italian data. Diabetes Nutr Metab Diabetes Care 1998; 21: B19–B26.
2004; 17: 358–367. 9. Yamashita H, Shao J, Friedman JE. Physiologic and molecular alter-
3. Bloomgarden ZT. American Diabetes Association 60th Scientific ations in carbohydrate metabolism during pregnancy and gesta-
Sessions, 2000: Diabetes and pregnancy. Diabetes Care 2000; tional diabetes mellitus. Clin Obstet Gynecol 2000; 43: 87–98.
23: 1699–1702. 10. Di Cianni G, Miccoli R, Volpe L et al. Intermediate metabolism in
4. Reece EA, Leguizamon G, Wiznitzen A. Gestational diabetes: normal pregnancy and in gestational diabetes. Diab Met Res Rev
The need for a common ground. Lancet 2009; 373: 1789–1797. 2003; 19: 259–270.
5. Lawrence JM, Contreras R, Chen W et al. Trends in the preva- 11. Ryan EA, Enns L. Role of gestational hormones in the induction of
lence of preexisting diabetes and gestational diabetes melli- insulin resistance. J Clin Endocrinol Metab 1998; 67: 341–347.
tus among a racially/ethnically diverse population of pregnant 12. Ministero della Salute, Gravidanza Fisiologica, Linea Guida 20
women, 1999–2005. Diabetes Care 2008; 31: 899–904. aggiornamento 2011.
6. King H. Epidemiology of glucose intolerance and gestational 13. Hod M, Simeoni U. Maternal, fetal and neonatal complications
diabetes in women of childbearing age. Diabetes Care 1998; 21: of diabetic pregnancy-delivering optimal care while awaiting for
B9–B13. cure. Semin Fetal Neonatal Med 2009; 14: 63–65.
528 Metabolomics and diabetic pregnancy
14. Sachse D, Sletner L, Morkrid K et al. Metabolomic changes in 22. Fanos V, Van den Anker J, Noto A et al. Metabolomics in neo-
urine during and after pregnancy in a large multiethnics popula- natology: Fact or fiction? Semin Fetal Neonatal Med 2013; 18:
tion based cohort study of gestational diabetes. PLOS ONE 2012; 3–12.
7: e52399. 23. Dessì A, Puddu M, Ottonello G et al. Metabolomics and fetal-
15. Lam YY, Hatzinikolas G, Weir JM et al. Insulin-stimulated glucose neonatal nutrition: Between “not enough” and “too much.”
uptake and pathways regulating energy metabolism in skeletal Molecules 2013; 18: 11724–11732.
muscle cells: The effects of subcutaneous and visceral fat, and 24. Diaz SO, Pinto J, Graça G et al. Metabolomic biomarkers of pre-
long-chain saturated, n-3 and n-6 polyunsaturated fatty acids. natal disorders: An exploratory NMR metabonomics study of
Biochim Biophys Acta 2011; 1811: 468–475. second trimester maternal urine and blood plasma. J Proteome
16. Nicholson JK, Connelly J, Lindon JC et al. Metabonomics: A plat- Res 2011; 8: 3732–3742.
form for studying drug toxicity and gene function. Nat Rev Drug 25. Scioscia M, Kunjara S, Gumaa K et al. Urinary excretion of
Discov 2002; 1: 153–161. inositol phosphoglycan P-type in gestational diabetes mellitus.
17. Ellis DI, Dunn WB, Griffin JL et al. Metabolic fingerprinting as a Diabetic Med 2007; 24: 1300–1304.
diagnostic tool. Pharmacogenomics 2007; 8: 1243–1266. 26. Graça G, Goodfellow J, Barros AS et al. UPLC-MS metabolomic
18. Fanos V, Antonucci R, Barberini L et al. Urinary metabolomics in profiling of second trimester amniotic fluid and maternal urine
the newborn and infants. Adv Clin Chem 2012; 58: 193–223. and comparison with NMR spectral profiling for the identifica-
19. Fanos V, Antonucci R, Barberini L et al. Clinical application of tion of pregnancy disorder biomarkers. Mol Biosyst 2012; 8:
metabolism in neonatology. J Matern Fetal Neonatal Med 2012; 1243–1254.
1: 104–109. 27. Logan KM, Wijeyesekera AD, Perez IG et al. Infants of moth-
20. Dessì A, Atzori L, Noto A et al. Metabolomics in newborns with ers with diabetes have altered urinary metabolic profile at
intrauterine growth retardation (IUGR): Urine reveals markers of birth. Available from: www.neonatalsociety.ac.uk/abstracts/
metabolomic syndrome. J Matern Fetal Neonatal Med 2011; 24: logank_2012_diabeticurinaryprofile.shtml, last accessed March
35–39. 27, 2014.
21. Atzori L, Antonucci R, Barberini L et al. Metabolomics: A new 28. Dani C, Bresci C, Berti E et al. Metabolomic profile of term
tool for the neonatologist. J Matern Fetal Neonatal Med 2009; 22: infants of gestational diabetics mothers. J Matern Fetal Neonatal
50–53. Med 2014; 27: 537–542.
63 Fetal growth restriction:
Evidence-based clinical
management
Eduard Gratacós and Francesc Figueras
Introduction “Fetal growth restriction” versus

Fetal growth restriction (FGR) is defined as a failure to “small for gestational age”
achieve the endorsed growth potential. The diagnosis of fetal
Overall, a “small fetus” is associated with poorer outcome.
“smallness” is currently performed on the basis of an esti-
However, evidence shows that there are, at least, two groups
mated fetal weight below a given threshold, most commonly
of small fetuses, normally referred to as FGR versus consti-
the 10th centile. It is likely that this definition lacks sensi-
tutional SGA, also defined as just SGA. FGR is associated
tivity, so that it misses cases of growth restriction that do not
with poorer perinatal outcome, abnormal Doppler sug-
fall below the 10th centile, but it identifies a subset of preg-
gesting fetal adaptation to undernutrition/hypoxia, signs
nancies at high risk of poorer perinatal outcome. Thus, detec-
of placental disease, and higher risk of preeclampsia (PE).
tion of small fetuses is clinically relevant because as a whole
SGA fetuses do not present the aforementioned changes
this group of fetuses is associated with poorer perinatal out-
and display perinatal outcomes similar to those of normally
come, and this represents opportunities for preventing cases
grown fetuses. It is unknown whether SGA fetuses are truly
of intrauterine fetal death, perinatal brain injury, and severe
“constitutionally small” or they represent another form of
intrapartum fetal distress. In addition, evidence accumulat-
fetal abnormal smallness. The distinction between FGR ver-
ing over the last 20 years has consistently demonstrated how
sus SGA is clinically relevant because of the wide consensus
being born small has important implications for the quality
that it is reasonable to deliver electively FGR earlier, whereas
of health during adulthood. Population-based studies show
elective delivery before term offers no benefit in SGA.
that prenatal identification of small for gestational age (SGA)
The diagnosis of FGR is currently performed by means
results in a reduction of adverse perinatal outcomes and
of a combination of Doppler and estimated fetal weight. For
stillbirth.1,2 However, most SGA babies remain unnoticed
20 years, the Doppler index widely used to distinguish FGR
until birth, even when routine third trimester ultrasound
from SGA was abnormal umbilical artery (UA) Doppler.
is performed.3,4 Moreover, according to pregnancy audits,
However, UA Doppler identifies severe placental disease but
most instances of avoidable stillbirth are related with a fail-
fails to pick up instances of mild placental disease, which in
ure to antenatally detect SGA.5 However, FGR is probably
reality constitute the majority of cases of FGR.4,6 Thus, UA
among the obstetrical entities with the greatest variation in
Doppler should always be used in combination integrated in
clinical management, resulting from lack of strong support-
the cerebroplacental ratio (CPR). CPR is calculated by divid-
ive evidence, combined with the complexity of the variables
ing the middle cerebral artery (MCA) Doppler pulsatility
and indices for assessing fetal deterioration.
index (PI) by the UA Doppler PI. This index correlates bet-
In this chapter, we will summarize the main aims in the
ter with adverse outcome.7 Aside from CPR, both the uter-
clinical management of FGR, which should be (1) to distin-
ine artery Doppler PI (UtA PI)8 and a very low EFW (<p3)9
guish “true” FGR from constitutional SGA and (2) to estab-
have been shown to independently predict poorer outcome
lish whether there is an indication for elective delivery and if
in small fetuses. Thus, the definition of FGR should include
not how often monitoring should be performed.
these three parameters, and we have provided good evidence
529
530 Fetal growth restriction
that, when used in combination, the presence of at least one at stake, the distinction into “early” or “late” is too simple
of these signs can classify as the subgroup of small fetuses and difficult to integrate since the determinants of elective
with the highest risk of poor perinatal outcome.10 It is likely delivery are the severity of fetal deterioration and the risks
that in future years, composite definitions include angio- of prematurity. For this reason, in our view FGR is better
genic factors as a valid criterion,11 but the lack of normative managed with a single protocol throughout pregnancy, as
gestational-age-adjusted values is for the moment a limita- discussed later in this chapter. The cutoff to define early-
tion for its wide use in practice. versus late-onset FGR has commonly been set in an arbi-
Finally, we should stress that SGA is associated with trary fashion at about 32–34 weeks at diagnosis or 37 weeks
apparently near-normal perinatal outcomes. However, in the at delivery. A prospective study using decision tree anal-
light of current evidences, SGA is not a “constitutionally” ysis in >700 cases determined that 32 weeks at diagnosis
small fetus. Fetuses under this diagnostic category have been and 37 weeks at delivery best classified two groups where
consistently demonstrated to present signs of brain reorga- the differences in terms of adverse perinatal outcome are
nization in utero and neonatally and poor long-term neu- maximized.16
rological, cardiovascular, and endocrinological outcome.12–14 Briefly, early-onset FGR represents about 20%–30% of
Therefore, the so-defined SGA fetuses are, in the majority all FGR and it is associated with early PE in up to 50%.4 In
of instances, suffering a form of pathological deviation from early-onset FGR, there is commonly severe placental insuf-
normal development. It remains to be established whether ficiency and chronic fetal hypoxia; hence UA Doppler is usu-
SGA is a category composed by a single problem or a mix- ally abnormal.17 Fetal condition deteriorates with evolution
ture of different causes, which might include among others a to decompensated hypoxia and acidosis, which is reflected
subset of fetuses with placental insufficiency of even milder by progression of abnormalities in the UA into absent and
nature, abnormalities in hormonal pathways regulating fetal reverse diastolic flow and increased PIs in the ductus veno-
growth, genetic causes, and true constitutional smallness. sus (DV), which is used as a reflection of late-stage disease
and very high risk of fetal death. Fetal deterioration nor-
mally lasts weeks18 following a cascade of Doppler changes.
“Early-onset” versus “late-onset” FGR Management is challenging and aims at achieving the best
balance between the risks of leaving the fetus in utero versus
Aside from the clinically relevant distinction between FGR the complications of prematurity.
and SGA, FGR presents under two different phenotypes Late-onset FGR represents 70%–80% of FGR and has a
when the onset is early or late in gestation. Table 63.1 dis- low association (about 10%) with late PE.19 The degree of
plays the main differences between both clinical forms. placental disease is mild;20 thus UA Doppler is below the
Differentiating between early- and late-onset FGR has a 95th centile in virtually all cases.6 However, the CPR, which
clear value and should always be required for any research combines changes in the UA and MCA, becomes abnormal
study for the sake of comparability. Being early or late onset at some point before birth in a substantial proportion of
determines differences in the severity of placental disease,15 cases.6 Changes in the DV are virtually never observed,6,21
as well as in the fetal adaptive response and deterioration. and the cascade of sequential fetal deterioration described
These notions have been extremely important for our for early onset does not occur in late FGR. However, progres-
pathophysiological understanding of the various clinical sion to fetal deterioration may occur suddenly and there is
presentations of FGR. However, when clinical decisions are high association with intrapartum fetal distress and neona-
tal acidosis.22,23 We have proposed24 that this is possibly the
consequence of a combination of rapid placental function
Table 63.1 Summary of the main differences deterioration, lower tolerance to hypoxia in term fetuses,
between early- and late-onset forms of fetal growth and the more frequent presence of intense uterine contrac-
restriction tions at this pregnancy stage.
Early-onset FGR Late-onset FGR

Challenge: Management Challenge: Diagnosis
Fetal assessment and correlation with
Prevalence: 1%–2% Prevalence: 3%–5% perinatal outcomes
Severe placental disease: Mild placental disease: UA
UA Doppler abnormal, Doppler normal, low Some of the measures and indices discussed later are essen-
high association with PE association with PE tially used for the diagnosis/identification of FGR from
Severe hypoxia ++: systemic Mild hypoxia: central CV SGA, and consequently, they are relevant for the decision
CV adaptation adaptation as to whether delivery is indicated when pregnancy term
High mortality and Lower mortality (but common is reached. Another set of indices have a prognostic value,
morbidity cause of late stillbirth) since they are useful to determine that there is a high risk
Abbreviations: PE, preeclampsia, CV, cardiovascular. of deterioration, and consequently, they are used to indicate
delivery before term is reached.
Fetal assessment and correlation with perinatal outcomes 531
Umbilical artery Doppler of cases, abnormal DV precedes the loss of short-term vari-
UA Doppler is the only measure that provides both diag- ability in cCTG, and in about 90% of cases, it is abnormal
nostic and prognostic information for the management of 48–72 hours before the biophysical profile (BPP).18,40 Hence,
FGR. There is compelling evidence that using UA Doppler it is considered to provide a better window of opportunity
in high-risk pregnancies improves perinatal outcomes, with for delivering fetuses in critical conditions at very early ges-
a 29% reduction in perinatal deaths.25 Absent or reversed tational ages.
end-diastolic velocities, the end of the spectrum of the
abnormalities of the UA Doppler, have been reported to be
present, on average, 1 week before the acute deterioration.26 Aortic isthmus Doppler
There is an association between reversed end-diastolic flow This vessel reflects the balance between the impedance of
in the UA and adverse perinatal outcome (with a sensitivity the brain and systemic vascular systems.43,44 Reverse aortic
and specificity of about 60%), which seems to be indepen- isthmus (AoI) flow is a sign of advanced deterioration and a
dent of prematurity.27 After 30 weeks, the risk of stillbirth further step in the sequence starting with the UA and MCA
of a fetus with isolated reversed end-diastolic velocities in Doppler. AoI has a strong association with both adverse
the UA Doppler overcomes the risks of prematurity,28,29 and perinatal45 and neurological outcome.46 However, longitu-
therefore delivery seems justified. dinal studies show that the AoI precedes DV abnormalities
by 1 week,21,47 and consequently it is not as good to predict
the short-term risk of stillbirth.35 In contrast, AoI seems to
Middle cerebral artery Doppler/cerebralplacental ratio improve the prediction of neurological morbidity.30
MCA informs about the existence of brain vasodilation,
a surrogate marker of hypoxia. There is an association
between abnormal MCA-PI and adverse perinatal and neu- Fetal heart rate analysis and biophysical profile
rological outcome, but it is unclear whether delivering before Early studies on high-risk pregnancies showed that, though
term could add any benefit. MCA is particularly valuable highly sensitive, cardiotocography has a 50% rate of false
for the identification of6 and prediction of adverse outcome positives for the prediction of adverse outcome.48 In addi-
among30,31 late-onset FGR, independently of the UA Doppler, tion, a meta-analysis49 on high-risk pregnancies failed to
which is often normal in these fetuses. Fetuses with abnor- demonstrate any beneficial effect in reducing perinatal mor-
mal MCA-PI had a sixfold risk of emergency cesarean sec- tality. Hence, there is no evidence to support the use of tradi-
tion for fetal distress when compared with SGA fetuses with tional fetal heart rate (FHR) monitoring or “nonstress tests”
normal MCA-PI,32 which is particularly relevant because in FGR fetuses. A main limitation of CTG is the subjective
labor induction at term is the current standard of care of interpretation of the FHR, which is extremely challenging in
late-onset FGR.33–35 Late-FGR with abnormal MCA-PI has very preterm fetuses with a physiologically reduced variabil-
poorer neurobehavioral competence at birth and at 2 years of ity. cCGT has represented a step forward and has provided
age.23,36 MCA is considered a rather late manifestation, with new insights into the pathophysiology and management of
acceptable specificity but low sensitivity, which is improved FGR. cCGT evaluates short-term variability of the FHR, an
by the use of CPR. aspect that subjective evaluation cannot assess. Current evi-
The CPR improves remarkably the sensitivity of UA dence suggests that cCGT is sensitive to detect advanced fetal
and MCA alone. Thus, the CPR is already decreased when deterioration, and it provides a value similar to DV reverse
its individual components suffer mild changes but are still atrial flow for the short-term prediction of fetal death. Short-
within normal ranges.37,38 In late-onset SGA fetuses, abnor- tem variability becomes abnormal, coinciding with the DV,
mal CPR is present before delivery in 20%–25% of the cases,39 whereas in about half of cases, abnormal DV precedes the
and it is associated with a higher risk of adverse outcome at loss of short-term FHR variability; the latter is the first to
induction, although to a lesser degree than MCA.32 become abnormal in the other cases.18
Concerning BPP, it is calculated by combining ultra-
sound assessment of fetal tone and respiratory and body
Ductus venosus Doppler movements, with amniotic fluid index (AFI), and a conven-
DV is the strongest single Doppler parameter to pre- tional CTG. BPP was designed to improve the performance
dict the short-term risk of fetal death in early-onset FGR. of FHR. Early observational studies reported a very low
Longitudinal studies have demonstrated that DV flow wave- risk of false positives for acidosis and perinatal death, but
forms become abnormal only in advanced stages of fetal more recent studies on early-onset very preterm FGR fetuses
compromise.18,26,27,40 Absent-reversed velocities during atrial raise concerns over the false-positive rate, with up to 23% of
contraction are associated with perinatal mortality indepen- instances of IUFD in fetuses with BPP > 6 and 11% in those
dently of the gestational age at delivery,41 with a risk ranging with BPP > 8.50 A meta-analysis51 showed no significant ben-
from 40% to 100% in early-onset FGR.35,42 Thus, this sign is efit of BPP in high-risk pregnancies. Consequently, whenever
normally considered sufficient to recommend delivery at any Doppler expertise and/or cCTG is available, the incorpora-
gestational age, after completion of steroids. In about 50% tion of BPP in management protocols of FGR is questionable.
Amniotic fluid index point of view, the concerns of the clinician are to determine
AFI is used essentially as part of the BPP. Amniotic fluid vol- (1) whether there is an indication for elective delivery, i.e.,
ume is believed to be a chronic parameter. In fact, among the whether the risks of leaving the fetus in utero exceed those
components of BPP, it is the only one that is not considered of delivering electively, and (2) if there is not an indication
acute. A meta-analysis52 of 18 randomized studies demon- for elective delivery, when should be the next control sched-
strated that a reduced AFI is associated with an abnormal uled. Although when considered as groups there are clear
5-minute Apgar score, but there was no association with aci- differences between early-onset and late-onset forms, on
dosis or perinatal death in SGA (RR 1.6 [95% CI 0.9–2.6]). an individual basis there is important overlapping of clini-
Longitudinal studies in early-onset FGR fetuses have shown cal features at borderline gestational ages. In addition, cases
that the AFI fluid index progressively decreases.18 One week with the same gestational age at onset are often detected at
before acute deterioration, 20%–30% of cases have oligohy- different time points during gestation. Consequently, we
dramnios.18,40 There is limited evidence on the role of oli- believe it is most practical to define an integrated protocol
gohydramnios to predict perinatal complications in FGR for the management of FGR.
fetuses managed with Doppler so that its inclusion in man- As mentioned at the beginning of this chapter, FGR is
agement protocols is questionable. probably among the obstetrical entities with the greatest
variation in clinical practice. We have long stated that FGR
must be followed according to a stage-based protocol, which
Clinical management of FGR: integrates the best available evidence to indicate timing of
follow-up and delivery. In our experience, the application
Follow-up scheme and timing of of such protocol facilitates clinical management and com-
delivery munication among doctors about what “severity” means
and could largely reduce clinical practice variation.
No treatment has been demonstrated to be of benefit in Our suggested approach24 is to group in stages those
growth restriction.53–57 Thus, assessment of fetal well-being indices or signs that are associated with similar fetal risks,
and timely delivery remain as the main management strat- since they should indicate similar follow-up intervals and
egy. The aim behind a clinical protocol for managing FGR timing of delivery. Thus, we suggest to profile several stages,
is to combine existing evidence on various methods for or prognostic groups, which define different management
monitoring fetal well-being in order to establish the risks of strategies (Table 63.2):
fetal injury or death and to balance them against the risks of
prematurity if the fetus is delivered. SGA: Excluding infectious and genetic causes, the perinatal
As we have discussed earlier, the first aim after identify- results are good. Fortnightly Doppler and growth assess-
ing a “small fetus” is to distinguish between FGR and SGA. ment are standard practice. Labor induction should be
Once this distinction is established, from a conceptual recommended at 40 weeks. Fortnightly monitoring is safe.
Table 63.2 Stage-based classification and management of fetal growth restriction
Stage Pathophysiological correlate Criteria (any of) Monitoringa GA/mode of delivery

I Severe smallness or mild EFW <third centile Weekly 37 w LI
placental insufficiency CPR < p5
UA PI > p95
MCA PI < p5
Uterine artery PI > p95
II Severe placental insufficiency UA AEDV Biweekly 34 w CS
reverse AoI
III Low-suspicion fetal acidosis UA REDV 1–2 days 30w CS
DV-PI > p95
IV High-suspicion fetal acidosis DV reverse atrial flow 12 hours 26wb CS
cCTG <3 ms
FHR decelerations
Note: All Doppler signs described here should be confirmed at least twice, ideally at least 12 hours apart.
Abbreviations: GA, gestational age, LI, labor induction, CS, cesarean section.
a Recommended intervals in the absence of severe preeclampsia. If FGR is accompanied by this complication, strict
fetal monitoring is warranted regardless of the stage.

b Lower GA threshold recommended according to current literature figures reporting at least 50% intact survival.
Threshold could be tailored on parents’ wishes or adjusted according to local statistics of intact survival.
References 533
Stage I FGR (severe smallness or mild placental insufficiency): cesarean section after 30 weeks. Monitoring every 24–48
Either uterine artery, UA or MCA Doppler, or the CPR is is recommended.
abnormal. In the absence of other abnormalities, evidence Stage IV FGR (high suspicion of fetal acidosis and high risk
suggests a low risk of fetal deterioration before term. of fetal death): There are spontaneous FHR decelerations,
Labor induction beyond 37 weeks is acceptable, but the reduced short-term variability (<3 ms) in the cCTG, or
risk of intrapartum fetal distress is increased.51 Cervical reverse atrial flow in the DV Doppler. Spontaneous FHR
induction with Foley catheter may be recommended to deceleration is an ominous sign, normally preceded by
reduce the risk of hyperstimulation. Weekly monitoring the other two signs, and thus it is rarely observed, but
seems reasonable. if persistent it may justify emergency cesarean section.
Stage II FGR (severe placental insufficiency): This stage cCTG and DV are associated with very high risks of still-
is defined by UA AEDV or reverse AoI. Despite the birth within the next 3–7 days and disability. Deliver
evidence for UA AEDV being stronger than that for after 26 weeks by cesarean section at a tertiary care cen-
AoI, observational evidence suggests an association ter under steroid treatment for lung maturation. Intact
between the latter and abnormal neurodevelopment, survival exceeds 50% only after 26–28 weeks, and before
so that both criteria are put into a single category. this threshold, parents should be counseled by multidis-
Delivery should be recommended after 34 weeks. The ciplinary teams. Monitoring every 12–24 hours until is
risk of emergent cesarean section at labor induction recommended.
exceeds 50%, and therefore, elective cesarean section
is a reasonable option. Monitoring twice a week is Particularly at early gestational ages, and at whatever stage,
recommended. coexistence of severe PE may distort the natural history, and
Stage III FGR (advanced fetal deterioration, low-suspicion strict fetal monitoring is warranted since fetal deterioration
signs of fetal acidosis): The stage is defined by reverse may occur unexpectedly at any time.
diastolic flow in the UA (REDV) or DV-PI >95th centile. This management protocol is adjusted to available evi-
There is an association between a higher risk of still- dence at the time of writing this chapter. Obviously, as new
birth and poorer neurological outcome. However, since evidence becomes available, any protocol should be adapted
signs suggesting a very high risk of stillbirth within days accordingly. However, this should not affect the main phi-
are not present yet, it seems reasonable to delay elective losophy of structuring FGR in severity stages that determine
delivery to reduce as possible the effects of severe prema- the frequency of follow-up and the decision to wait versus
turity. We suggest delivery should be recommended by deliver electively.
REFERENCES
1. Lindqvist PG, Molin J. Does antenatal identification of small- 9. Savchev S, Figueras F, Cruz-Martinez R, Illa M, Botet F, Gratacos
for-gestational age fetuses significantly improve their outcome? E. Estimated weight centile as a predictor of perinatal outcome
Ultrasound Obstet Gynecol 2005; 25: 258–264. in small-for-gestational-age pregnancies with normal fetal and
2. Gardosi J, Madurasinghe V, Williams M, Malik A, Francis A. maternal Doppler indices. Ultrasound Obstet Gynecol 2012;
Maternal and fetal risk factors for stillbirth: Population based 39(3): 299–303.
study. Br Med J 2013; 346: f108. 10. Figueras F, Savchev S, Triunfo S, Crovetto F, Gratacos E. An
3. Skovron ML, Berkowitz GS, Lapinski RH, Kim JM, Chitkara integrated model with classification criteria to predict small-
U. Evaluation of early third-trimester ultrasound screening for for-gestational-age fetuses at risk of adverse perinatal outcome.
intrauterine growth retardation. J Ultrasound Med 1991; 10: Ultrasound Obstet Gynecol 2015; 45: 279–285.
153–159. 11. Lobmaier SM, Figueras F, Mercade I, Perello M, Peguero A,
4. Figueras F, Eixarch E, Gratacos E, Gardosi J. Predictiveness of Crovetto F, Ortiz JU, Crispi F, Gratacós E. Angiogenic factors
antenatal umbilical artery Doppler for adverse pregnancy out- versus Doppler follow up in the prediction of adverse out-
come in small-for-gestational-age babies according to custom- come among late pregnancy small-for-gestational-age fetuses.
ised birthweight centiles: Population-based study. BJOG 2008; Ultrasound Obstet Gynecol 2014; 43: 533–540.
115: 590–594. 12. Sanz-Cortes M, Figueras F, Bonet-Carne E et al. Fetal brain
5. Richardus JH, Graafmans WC, Verloove-Vanhorick SP, MRI texture analysis identifies different microstructural pat-
Mackenbach JP, EuroNatal International Audit Panel, EuroNatal terns in adequate and small for gestational age fetuses at term.
Working Group. Differences in perinatal mortality and subop- Fetal Diagn Ther 2013; 33: 122–129.
timal care between 10 European regions: Results of an interna- 13. Larroque B, Bertrais S, Czernichow P, Leger J. School difficulties
tional audit. BJOG 2003; 110: 97–105. in 20-year-olds who were born small for gestational age at term
6. Oros D, Figueras F, Cruz-Martinez R, Meler E, Munmany M, in a regional cohort study. Pediatrics 2001; 108: 111–115.
Gratacos E. Longitudinal changes in uterine, umbilical and fetal 14. Crispi F, Bijnens B, Figueras F et al. Fetal growth restriction
cerebral Doppler indices in late-onset small-for-gestational age results in remodeled and less efficient hearts in children.
fetuses. Ultrasound Obstet Gynecol 2011; 37(2): 191–195. Circulation June 8, 2010; 121(22): 2427–2436.
7. Baschat AA, Gembruch U. The cerebroplacental Doppler ratio 15. Boog G. Chronic villitis of unknown etiology. Eur J Obstet
revisited. Ultrasound Obstet Gynecol 2003; 21(2): 124–127. Gynecol Reprod Biol 2008; 136: 9–15.
8. Vergani P, Roncaglia N, Andreotti C, Arreghini A, Teruzzi M, 16. Savchev S, Figueras F, Sanz-Cortes M, Cruz-Lemini M, Triunfo
Pezzullo JC, Ghidini A. Prognostic value of uterine artery Doppler S, Botet F, Gratacos E. Evaluation of an optimal gestational age
velocimetry in growth-restricted fetuses delivered near term. Am cut-off for the definition of early- and late-onset fetal growth
J Obstet Gynecol 2002; 187(4): 932–936. restriction. Fetal Diagn Ther 2014; 36: 99–105.
17. Turam OM, Turan S, Gungor S et al. Progression of Doppler 37. Gramellini D, Folli MC, Raboni S, Vadora E, Merialdi A. Cerebral-
abnormalities in intrauterine growth restriction. Ultrasound umbilical Doppler ratio as a predictor of adverse perinatal out-
Obstet Gynecol 2008; 32: 160–167. come. Obstet Gynecol 1992; 79(3): 416–420.
18. Hecher K, Bilardo CM, Stigter RH et al. Monitoring of fetuses 38. Arbeille P, Maulik D, Fignon A, Stale H, Berson M, Bodard S,
with intrauterine growth restriction: A longitudinal study. Locatelli A. Assessment of the fetal PO2 changes by cerebral
Ultrasound Obstet Gynecol 2001; 18: 564–570. and umbilical Doppler on lamb fetuses during acute hypoxia.
19. Crovetto F, Crispi F, Scazzocchio E et al. First-trimester screen- Ultrasound Med Biol 1995; 21(7): 861–870.
ing for early and late small-for-gestational-age neonates using 39. Cruz-Martinez R, Figueras F, Hernandez-Andrade E, Puerto B,
maternal serum biochemistry, blood pressure and uterine Gratacós E. Longitudinal brain perfusion changes in near-term
artery Doppler. Ultrasound Obstet Gynecol 2014; 43: 34–40. small-for-gestational-age fetuses as measured by spectral Doppler
20. Parra-Saavedra M, Crovetto F, Triunfo S, Savchev S, Peguero indices or by fractional moving blood volume. Am J Obstet Gynecol
A, Nadal A, Parra G, Gratacos E, Figueras F. Placental findings 2010; 203(1): 42.e1–42.e6.
in late-onset SGA births without Doppler signs of placental 40. Baschat AA, Gembruch U, Harman CR. The sequence of changes in
insufficiency. Placenta 2013; 34: 1136–1141. Doppler and biophysical parameters as severe fetal growth restric-
21. Cruz-Martinez R, Figueras F, Hernandez-Andrade E, Oros D, tion worsens. Ultrasound Obstet Gynecol 2001; 18(6): 571–577.
Gratacos E. Changes in myocardial performance index and 41. Schwarze A, Gembruch U, Krapp M, Katalinic A, Germer U,
aortic isthmus and ductus venosus Doppler in term, small-for- Axt-Fliedner R. Qualitative venous Doppler flow waveform analy-
gestational age fetuses with normal umbilical artery pulsatility sis in preterm intrauterine growth-restricted fetuses with ARED
index. Ultrasound Obstet Gynecol 2011; 38: 400–405. flow in the umbilical artery—Correlation with short-term outcome.
22. Kady S, Gardosi J. Perinatal mortality and fetal growth restriction. Ultrasound Obstet Gynecol 2005; 25(6): 573–579.
Best Pract Res Clin Obstet Gynaecol 2004; 18: 397–410. 42. Baschat AA, Gembruch U, Weiner CP, Harman CR. Qualitative
23. Figueras F, Eixarch E, Meler E et al. Small-for-gestational-age venous Doppler waveform analysis improves prediction of criti-
fetuses with normal umbilical artery Doppler have subopti- cal perinatal outcomes in premature growth-restricted fetuses.
mal perinatal and neurodevelopmental outcome. Eur J Obstet Ultrasound Obstet Gynecol 2003; 22(3): 240–245.
Gynecol Reprod Biol 2008; 136: 34–38. 43. Fouron JC, Skoll A, Sonesson SE, Pfizenmaier M, Jaeggi E, Lessard
24. Figueras F, Gratacos E. Update on the diagnosis and classification M. Relationship between flow through the fetal aortic isthmus and
of fetal growth restriction and proposal of a stage-based manage- cerebral oxygenation during acute placental circulatory insuffi-
ment protocol. Fetal Diagn Ther 2014; 36: 86–98. ciency in ovine fetuses. Am J Obstet Gynecol 1999; 181(5 Pt. 1):
25. Alfirevic Z, Stampalija T, Gyte GM. Fetal and umbilical Doppler 1102–1107.
ultrasound in high-risk pregnancies. Cochrane Database Syst Rev 44. Makikallio K, Jouppila P, Rasanen J. Retrograde aortic isthmus net
2010; (1): CD007529. blood flow and human fetal cardiac function in placental insuffi-
26. Ferrazzi E, Bozzo M, Rigano S, Bellotti M, Morabito A, Pardi ciency. Ultrasound Obstet Gynecol 2003; 22(4): 351–357.
G, Battaglia FC, Galan HL. Temporal sequence of abnormal 45. Del Rio M, Martínez JM, Figueras F, et al. Doppler assessment of
Doppler changes in the peripheral and central circulatory sys- the aortic isthmus and perinatal outcome in preterm fetuses with
tems of the severely growth-restricted fetus. Ultrasound Obstet severe intrauterine growth restriction. Ultrasound Obstet Gynecol
Gynecol 2002; 19(2): 140–146. 2008; 31(1): 41–47.
27. Cosmi E, Ambrosini G, D’Antona D, Saccardi C, Mari G. Doppler, 46. Fouron JC, Gosselin J, Raboisson MJ et al. The relationship
cardiotocography, and biophysical profile changes in growth- between an aortic isthmus blood flow velocity index and the
restricted fetuses. Obstet Gynecol 2005; 106(6): 1240–1245. postnatal neurodevelopmental status of fetuses with placental
28. Thornton JG, Hornbuckle J, Vail A, Spiegelhalter DJ, Levene circulatory insufficiency. Am J Obstet Gynecol 2005; 192(2):
M. Infant wellbeing at 2 years of age in the Growth Restriction 497–503.
Intervention Trial (GRIT): Multicentred randomised controlled 47. Figueras F, Benavides A, Del Rio M et al. Monitoring of fetuses
trial. Lancet 2004; 364(9433): 513–520. with intrauterine growth restriction: Longitudinal changes in ductus
29. Grit Study Group. A randomised trial of timed delivery for the venosus and aortic isthmus flow. Ultrasound Obstet Gynecol 2009;
compromised preterm fetus: Short term outcomes and Bayesian 33(1): 39–43.
interpretation. BJOG 2003; 110(1): 27–32. 48. Evertson LR, Gauthier RJ, Schifrin BS, Paul RH. Antepartum fetal
30. Eixarch E, Meler E, Iraola A, Illa M, Crispi F, Hernandez-Andrade E, heart rate testing. I. Evolution of the nonstress test. Am J Obstet
Gratacos E, Figueras F. Neurodevelopmental outcome in 2-year- Gynecol 1979; 133(1): 29–33.
old infants who were small-for-gestational age term fetuses with 49. Pattison N, McCowan L. Cardiotocography for antepartum fetal
cerebral blood flow redistribution. Ultrasound Obstet Gynecol assessment. Cochrane Database Syst Rev 2000; (2): CD001068.
2008; 32(7): 894–899. 50. Kaur S, Picconi JL, Chadha R, Kruger M, Mari G. Biophysical
31. Hershkovitz R, Kingdom JC, Geary M, Rodeck CH. Fetal cerebral profile in the treatment of intrauterine growth-restricted fetuses
blood flow redistribution in late gestation: Identification of com- who weigh <1000 g. Am J Obstet Gynecol 2008; 199(3):
promise in small fetuses with normal umbilical artery Doppler. 264.e1–264.e4.
Ultrasound Obstet Gynecol 2000; 15(3): 209–212. 51. Alfirevic Z, Neilson JP. Biophysical profile for fetal assessment
32. Cruz-Martinez R, Figueras F, Hernandez-Andrade E, Oros D, in high risk pregnancies. Cochrane Database Syst Rev 2000; (2):
Gratacos E. Fetal brain Doppler to predict cesarean delivery CD000038.
for nonreassuring fetal status in term small-for-gestational-age 52. Chauhan SP, Sanderson M, Hendrix NW, Magann EF, Devoe LD.
fetuses. Obstet Gynecol 2011; 117(3): 618–626. Perinatal outcome and amniotic fluid index in the antepartum and
33. Boers KE, van Wyk L, van der Post JA et al. Neonatal morbidity intrapartum periods: A meta-analysis. Am J Obstet Gynecol 1999;
after induction vs expectant monitoring in intrauterine growth 181: 1473–1478.
restriction at term: A subanalysis of the DIGITAT RCT. Am J 53. Gulmezoglu AM, Hofmeyr GJ. Plasma volume expansion for sus-
Obstet Gynecol 2012; 206(4): 344.e1–344.e7. pected impaired fetal growth. Cochrane Database Syst Rev 2000;
34. Boers KE, Vijgen SM, Bijlenga D et al. Induction versus expect- (2): CD000167.
ant monitoring for intrauterine growth restriction at term: 54. Gulmezoglu AM, Hofmeyr GJ. Betamimetics for suspected
Randomised equivalence trial (DIGITAT). Br Med J 2010; 341: impaired fetal growth. Cochrane Database Syst Rev 2001; (4):
c7087. CD000036.
35. Cruz-Lemini M, Crispi F, Van Mieghem T et al. Risk of perina- 55. Laurin J, Persson PH. The effect of bedrest in hospital on fetal
tal death in early-onset intrauterine growth restriction according outcome in pregnancies complicated by intra-uterine growth
to gestational age and cardiovascular Doppler indices: A multi- retardation. Acta Obstet Gynecol Scand 1987; 66(5): 407–411.
center study. Fetal Diagn Ther 2012; 32(1–2): 116–122. 56. Say L, Gulmezoglu AM, Hofmeyr GJ. Maternal nutrient supplemen-
36. Oros D, Figueras F, Cruz-Martinez R, Padilla N, Meler E, tation for suspected impaired fetal growth. Cochrane Database Syst
Hernandez-Andrade E, Gratacos E. Middle versus anterior cere- Rev 2003; (1): CD000148.
bral artery Doppler for the prediction of adverse outcome in term 57. Say L, Gulmezoglu AM, Hofmeyr GJ. Maternal oxygen administra-
small-for-gestational-age fetuses. Ultrasound Obstet Gynecol tion for suspected impaired fetal growth. Cochrane Database Syst
2010; 35: 456–461. Rev 2003; (1): CD000137.

4 5805688099241461709 PDF

Uploaded by

Copyright:

Available Formats

4 5805688099241461709 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

4 5805688099241461709 PDF

Uploaded by

Copyright:

Available Formats

Accessing the E-book edition

Using the VitalSource® ebook DOWNLOAD AND READ OFFLINE

1. Create a VitalSource Bookshelf account at On your iPhone/iPod Touch/iPad

On your Kindle Fire

Howard Carp, Recurrent Pregnancy Loss, ISBN 9780415421300

Howard Carp, Recurrent Pregnancy Loss: Causes, Controversies, and

Oded Langer MD PhD

No claim to original U.S. Government works

International Standard Book Number-13: 978-1-4822-1362-1 (eBook - PDF)

For their tolerance, patience, and love—they made it all possible

21. Gestational diabetes mellitus in multiple pregnancies 169

44. Insulin infusion pumps in pregnancy 368

Table P.1 Options for diagnosis of GDM based on resource settings

d Latin America Study Group.

given. It is important to note that there is no long-term Preconception care

Salvatore Alberico Ahmet A. Baschat Eulàlia Brugués

Maria Grazia Dalfrà Graziano Di Cianni Rinat Gabbay-Benziv

Emilio Herrera Michal Kovo Gianpaolo Maso

Delphine Mitanchez Michael J. Paidas Isabel Sánchez-Vera

Sara Ornaghi Lene Ringholm Pierre Singer

Rina Su Tuangsit Wataganara Yariv Yogev

Such long-term effects on the health of mothers and their

Pregnancy-induced Early and late pregnancy—

The great obstetrical syndromes contingent management

Assessment Integrated clinic at 11–13 weeks Intervention

Figure 1.2 The first trimester clinic and contingent management.

Community clinic Postnatal evaluation Community clinic

Prevention Annual follow-up Treatment

Planning new pregnancy

Figure 1.3 The postpregnancy clinic.

sugar,” and in early Chinese and Japanese writings “the urine

facilitated diffusion according to concentration-dependent

Carbohydrate metabolism Protein and amino acid metabolism

lipolysis released into the circulation, NEFA and glycerol, is

6 a quantitatively low,38 and therefore their main destination is

NEFA Ketone bodies Muscle

ATGL HSL MAGL

DAG TAG NEFA

intrauterine and early postnatal life, little is known about Vitamin A

triglycerides and cholesterol reach the highest levels.1,2,10,11

FFA Glycerol FA FFA Glycerol FA

(a) (b) (c)

These changes develop gradually during gestation. By the Plasma changes

preterm infants, indicating that the lack of marked fatty

Carbon (g/kg/day) Calories (kcal/kg/day)

Table 5.2 Fetal carbon substrate oxidation in relation to fetal oxygen

Carbon for oxidation

Uteroplacental glucose uptake

Glucose utilization rate (mg/min/kg fetus) Percent of total

is upregulated by hypoglycemia and hypoinsulinemia

Liver glycogen or total carbohydrate

The vertical line indicates both term and time of birth.

Maternal Placenta Fetal Fetal body

Percentage body fat

Maternal Placenta Fetal Fetal

Lactate Lactate Glucose

screening challenge tests. A detailed discussion of variations

first-degree relatives of patients with DM-1 (FDRs-DM-1).17

Table 7.1 Prevalence of diabetes-related antibodies in women with