Absence of Excess Body Fatness

IARC HANDBOOKS
ABSENCE OF EXCESS
BODY FATNESS
VOLUME 16
IARC HANDBOOKS OF
CANCER PREVENTION
IARC HANDBOOKS
ABSENCE OF EXCESS
BODY FATNESS
VOLUME 16
This publication represents the views and

expert opinions of an IARC Working Group on
the Evaluation of Cancer-Preventive Interventions,
which met in Lyon, 5–12 April 2016
LYON, FRANCE - 2018
IARC HANDBOOKS OF
CANCER PREVENTION
Published by the International Agency for Research on Cancer, 150 cours Albert Thomas, 69372
Lyon Cedex 08, France
©International Agency for Research on Cancer, 2018
Distributed by WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland
(tel.: +41 22 791 3264; fax: +41 22 791 4857; email: [email protected]).
Publications of the World Health Organization enjoy copyright protection in accordance with the provisions of Protocol 2
of the Universal Copyright Convention. All rights reserved.
The designations employed and the presentation of the material in this publication do not imply the expression of any
opinion whatsoever on the part of the Secretariat of the World Health Organization concerning the legal status of any country,
territory, city, or area or of its authorities, or concerning the delimitation of its frontiers or boundaries.
The boundaries and names shown and the designations used on the maps do not imply the expression of any opinion
whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city, or area
or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent
approximate border lines for which there may not yet be full agreement.
The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or
recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors and
omissions excepted, the names of proprietary products are distinguished by initial capital letters.
The IARC Handbooks Working Group alone is responsible for the views expressed in this publication.
The International Agency for Research on Cancer welcomes requests for permission to reproduce or translate its
publications, in part or in full. Requests for permission to reproduce or translate IARC publications – whether for sale or for
non-commercial distribution – should be addressed to the IARC Communications Group at [email protected].
Corrigenda to the IARC Handbooks are published online at http://publications.iarc.fr/.
To report an error, please contact [email protected].
How to cite this publication:

IARC (2018). Absence of excess body fatness. IARC Handb Cancer Prev. 16:1–646. Available from: http://publications.iarc.
fr/570.
IARC Library Cataloguing in Publication Data

Absence of excess body fatness / IARC Working Group on the Evaluation of Cancer-Preventive Interventions, 2016.
(IARC Handbooks of Cancer Prevention ; Volume 16)
1. Neoplasms – prevention & control 2. Overweight 3. Body Mass Index 4. Prevalence 5. Risk Factors
ISBN 978-92-832-3020-5 (NLM Classification: W1)

ISSN 1027-5622
International Agency for Research on Cancer
The International Agency for Research on Cancer (IARC) was established in 1965 by the World
Health Assembly, as an independently funded organization within the framework of the World Health
Organization. The headquarters of the Agency are in Lyon, France.
The Agency has as its mission to reduce the cancer burden worldwide through promoting international
collaboration in research. The Agency addresses this mission through conducting cancer research for
cancer prevention in three main areas: describing the occurrence of cancer; identifying the causes of
cancer, and evaluating preventive interventions and their implementation. Each of these areas is a vital
contribution to the spectrum of cancer prevention.
The publications of the Agency contribute to the dissemination of authoritative information on
different aspects of cancer research. Information about IARC publications, and how to order them, is
available at http://publications.iarc.fr/.
IARC Handbooks of Cancer Prevention
In 1969, the International Agency for Research on Cancer (IARC) initiated a programme on the
evaluation of the carcinogenic risk of chemicals to humans involving the production of monographs of
critical reviews and evaluations of individual chemicals.
The IARC Handbooks of Cancer Prevention complement the IARC Monographs’ evaluations of
carcinogenic hazards. The objective of the programme is to coordinate and publish critical reviews of data
on the cancer-preventive effects of primary or secondary interventions, and to evaluate these data in terms
of cancer prevention with the help of international working groups of experts in prevention and related
fields. The lists of evaluations are regularly updated and are available at http://handbooks.iarc.fr/.
This IARC Handbook of Cancer Prevention is partly funded by the American Cancer Society (contract
ACS #26531) and by the Grant or Cooperative Agreement Number DP004954-05, funded by the United
States Centers for Disease Control and Prevention. Its contents are solely the responsibility of the authors
and do not necessarily represent the official views of the Centers for Disease Control and Prevention or
the Department of Health and Human Services.
Cover image: Overweight people sitting on a bench. © Tony Alter CC-BY-2.0

CONTENTS
NOTE TO THE READER. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
LIST OF PARTICIPANTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
WORKING PROCEDURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
A. GENERAL PRINCIPLES AND PROCEDURES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2. Objective and scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3. Selection of interventions for review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4. Data for the IARC Handbooks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
5. Meeting participants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
B. SCIENTIFIC REVIEW AND EVALUATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1. Characteristics and occurrence of the intervention. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2. Studies of cancer prevention in humans. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3. Studies of cancer prevention in experimental animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4. Mechanistic and other relevant data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5. Summary of data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
6. Evaluation and rationale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
GENERAL REMARKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
LIST OF ABBREVIATIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1. BODY FATNESS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.1 Background and definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.1.1 Scientific definitions of obesity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.1.2 Cultural definitions of obesity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.1.3 Body fatness as a public health problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.2 Prevalence and trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.2.1 Prevalence and trends in adults. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.2.2 Prevalence and trends in children and adolescents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.2.3 Prevalence by age and sex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
1.2.4 Prevalence by ethnicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.2.5 Prevalence by social class and education. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
I
IARC HANDBOOKS OF CANCER PREVENTION – 16
1.3 Risk factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

1.3.1 Regulation of hunger and satiety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
1.3.2 Weight gain throughout the life-course . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
1.3.3 Excessive energy intake. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
1.3.4 Endocrine disruptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
1.3.5 Physical activity and sedentary behaviour. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
1.3.6 Built environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
1.3.7 Social determinants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
1.3.8 Microbiota and gastrointestinal environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
1.3.9 Genetic and epigenetic determinants of body fatness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
1.4 Assessment of anthropometric measures and body composition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
1.4.1 Weight and height . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
1.4.2 Indexes of adiposity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
1.4.3 Other measures of adiposity or body composition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
1.4.4 Change in weight. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
1.4.5 Assessment of paediatric adiposity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2. CANCER-PREVENTIVE EFFECTS IN HUMANS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

2.1 Methodological considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
2.1.1 Bias. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
2.1.2 Confounding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
2.1.3 Reverse causation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
2.1.4 Mendelian randomization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
2.2 Cancer-preventive effects by organ site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
2.2.1 Cancer of the colorectum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
2.2.2 Cancer of the oesophagus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
2.2.3 Cancer of the stomach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
2.2.4 Cancer of the liver (hepatocellular carcinoma) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
2.2.5 Cancer of the gall bladder. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
2.2.6 Cancers of the biliary tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
2.2.7 Cancer of the pancreas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
2.2.8 Cancer of the lung . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
2.2.9 Cancer of the breast in women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
2.2.10 Cancer of the breast in men. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
Reference. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
II
Contents
2.2.11 Cancer of the endometrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291

References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
2.2.12 Cancer of the cervix. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
2.2.13 Cancer of the ovary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
2.2.14 Cancer of the prostate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
2.2.15 Cancer of the testis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
2.2.16 Cancer of the kidney (renal cell carcinoma) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 396
2.2.17 Cancer of the urinary bladder. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
2.2.18 Primary tumours of the brain and central nervous system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410
2.2.19 Cancer of the thyroid. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 420
2.2.20 Haematopoietic malignancies of lymphoid origin. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
2.2.21 Other haematopoietic malignancies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454
2.2.22 Cancers of the head and neck. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 456
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
2.2.23 Malignant melanoma. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462
2.3. Excess body fatness in early life and subsequent cancer risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463
2.3.1 Weight and height measured in childhood. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463
2.3.2 Body shape in early adulthood determined by recall. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464
2.3.3 Trajectories of body shape determined from early life. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 475
2.4 Excess body fatness in cancer survivors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 477
2.4.1 Studies of weight at diagnosis and cancer outcomes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 477
2.4.2 Studies of weight change after cancer diagnosis and cancer outcomes. . . . . . . . . . . . . . . . . 478
2.4.3 Intervention trials of weight-loss intervention and dietary modification . . . . . . . . . . . . . . . 478
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
2.5 Sustained weight loss and cancer risk: illustrative examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481
2.5.1 Studies of weight loss and cancer risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481
2.5.2 Studies of bariatric surgery and cancer risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485
3. CANCER-PREVENTIVE EFFECTS IN EXPERIMENTAL ANIMALS. . . . . . . . . . . . . . . . . . . . . . 487

3.1 Methodological considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 487
3.1.1 Definition of dietary/calorie restriction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 487
3.1.2 Design issues in studies of dietary restriction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 487
3.2 Overview of the effects of excess body weight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489
III
3.2.1 Obesity models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489

3.2.2 Cancer of the mammary gland. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 490
3.2.3 Cancer of the colon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493
3.2.4 Cancer of the liver. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494
3.2.5 Cancer of the prostate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495
3.2.6 Cancer of the skin (melanoma). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497
3.2.8 Cancer of the endometrium of the uterus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
3.2.9 Leukaemia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
3.3 Preventive effects of dietary/calorie restriction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500
3.3.1 Cancer of the mammary gland. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500
3.3.2 Cancer of the colon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 514
3.3.3 Cancer of the liver. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 518
3.3.5 Cancer of the skin and subcutaneous tissue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526
3.3.6 Cancer of the pituitary gland . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 531
3.3.7 Cancer of the prostate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 537
3.3.8 Cancers of the haematopoietic system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 537
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 545
4. MECHANISTIC AND OTHER RELEVANT DATA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553
4.2 Intracellular factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 554
4.2.1 Cell proliferation, apoptosis, and angiogenesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 554
4.2.2 The mTOR network and other energy-sensor networks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557
4.2.3 Epigenetics, oxidative stress, DNA repair, and telomeres. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 559
4.3 Receptor-mediated effects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564
4.3.1 Sex hormones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564
4.3.2 Insulin resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 567
4.3.3 Insulin-like growth factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 570
4.3.4 Chronic inflammation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 571
4.3.5 Vitamin D. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 577
4.3.6 Other factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 579
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 611
5. SUMMARY OF DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 633

5.1 Exposure data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 633
5.2 Cancer-preventive effects in humans. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 634
5.2.1 Cancer of the colorectum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 634
5.2.2 Cancer of the oesophagus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 635
5.2.3 Cancer of the stomach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 635
5.2.4 Cancer of the liver (hepatocellular carcinoma) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 635
5.2.5 Cancer of the gall bladder. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 635
5.2.6 Cancers of the biliary tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 636
5.2.8 Cancer of the lung . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 636
IV
Contents
5.2.9 Cancer of the breast in women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 636

5.2.10 Cancer of the breast in men. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 636
5.2.11 Cancer of the endometrium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 636
5.2.12 Cancer of the cervix. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 637
5.2.13 Cancer of the ovary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 637
5.2.14 Cancer of the prostate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 637
5.2.15 Cancer of the testis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 637
5.2.16 Cancer of the kidney (renal cell carcinoma) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 637
5.2.17 Cancer of the urinary bladder. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 637
5.2.18 Primary tumours of the brain and central nervous system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 638
5.2.19 Cancer of the thyroid. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 638
5.2.20 Tumours of the haematopoietic system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 638
5.2.21 Cancers of the head and neck. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 638
5.2.22 Malignant melanoma. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639
5.2.23 Excess body fatness in early life and subsequent cancer risk . . . . . . . . . . . . . . . . . . . . . . . . . 639
5.2.24 Excess body fatness in cancer survivors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639
5.2.25 Sustained weight loss and cancer risk. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639
5.3 Cancer-preventive effects in experimental animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639
5.3.1 Excess body weight. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639
5.3.2 Dietary/calorie restriction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 640
5.4 Mechanistic and other relevant data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 643
5.4.1 Sex hormone metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 643
5.4.2 Inflammation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 643
5.4.3 Insulin and insulin-like growth factor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 644
5.4.4 Epigenetic alterations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 644
5.4.5 Oxidative stress. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 644
5.4.6 DNA repair. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 644
5.4.7 Telomeres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 644
5.4.8 Other mechanisms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 644
6. EVALUATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645
6.1 Cancer-preventive effects in humans. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645
6.2 Cancer-preventive effects in experimental animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645
6.3 Mechanistic and other relevant data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645
6.4 Overall evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645
V
NOTE TO THE READER
The IARC Handbooks of Cancer Prevention series was launched in 1995 to complement the IARC
Monographs’ evaluations of carcinogenic hazards. The IARC Handbooks of Cancer Prevention evalu-
ate the published scientific evidence of cancer-preventive interventions.
Inclusion of an intervention in the Handbooks does not imply that it is cancer-preventive, only
that the published data have been examined. Equally, the fact that an intervention has not yet been
evaluated in a Handbook does not mean that it may not prevent cancer. Similarly, identification of
organ sites with sufficient evidence or limited evidence of cancer-preventive activity in humans should
not be viewed as precluding the possibility that an intervention may prevent cancer at other sites.
The evaluations of cancer-preventive interventions are made by international Working Groups
of independent scientists and are qualitative in nature. No recommendation is given for regulation
or legislation.
Anyone who is aware of published data that may alter the evaluation of cancer-preventive interven-
tions is encouraged to make this information available to the IARC Handbooks Group, International
Agency for Research on Cancer, 150 cours Albert Thomas, 69372 Lyon Cedex 08, France, or by email
to [email protected], in order that these data may be considered for re-evaluation by a future Working
Group.
Although every effort is made to prepare the Handbooks as accurately as possible, mistakes may
occur. Readers are requested to communicate any errors to the IARC Handbooks Group at [email protected].
1
LIST OF PARTICIPANTS
Members 1
Annie S. Anderson (Subgroup Chair, João Breda

Characteristics of Body Weight, Nutrition, Physical Activity and Obesity
Prevalence and Risk Factors of Division of Noncommunicable Diseases
Overweight/Obesity) and Promoting Health through the
Centre for Public Health Nutrition Research Life-Course
Centre for Research into Cancer Prevention World Health Organization (WHO)
and Screening Regional Office for Europe
Division of Cancer Research Copenhagen
Medical Research Institute Denmark
University of Dundee
Dundee Tim Byers
United Kingdom
Colorado School of Public Health
Aurora, CO
Jennifer L. Baker USA
Institute of Preventive Medicine
Bispebjerg and Frederiksberg Hospital
Copenhagen
Denmark
1
Working Group Members and Invited Specialists serve in their individual capacities as scientists and not as representa-
tives of their government or any organization with which they are affiliated. Affiliations are provided for identification
purposes only.
Each participant was asked to disclose pertinent research, employment, and financial interests. Current financial
interests and research and employment interests during the past 4 years or anticipated in the future are identified here.
Minor pertinent interests are not listed and include stock valued at no more than US$ 1000 overall, grants that provide
no more than 5% of the research budget of the expert’s organization and that do not support the expert’s research or
position, and consulting or speaking on matters not before a court or government agency that does not exceed 2% of
total professional time or compensation. All grants that support the expert’s research or position and all consulting or
speaking on behalf of an interested party on matters before a court or government agency are listed as significant perti-
nent interests.
3
Margot P. Cleary Ronald A. Herbert (Subgroup Chair, Cancer-

Nutrition and Metabolism Preventive Effects in Experimental
The Hormel Institute Animals)
University of Minnesota Pathology Support Group
Austin, MN Cellular and Molecular Pathology Branch
USA National Toxicology Program Division
National Institute of Environmental Health
Sciences
Graham Colditz (Meeting Chair)
Research Triangle Park, NC
Washington University School of Medicine USA
Washington University in St. Louis
St. Louis, MO
USA Stephen D. Hursting
Department of Nutrition and Lineberger
Comprehensive Cancer Center
Mariachiara Di Cesare
University of North Carolina at Chapel Hill
Department of Natural Sciences Chapel Hill, NC
School of Science and Technology USA
Middlesex University
London
United Kingdom Rudolf Kaaks (Subgroup Chair, Cancer-
Preventive Effects in Humans)
Susan M. Gapstur Department of Cancer Epidemiology

German Cancer Research Center (DKFZ)
Epidemiology Research Program Heidelberg
American Cancer Society Germany
Atlanta, GA
USA
Michael Leitzmann
Marc Gunter University Hospital Regensburg

Institute of Epidemiology and Preventive
Nutritional Epidemiology Group Medicine
Section of Nutrition and Metabolism Regensburg
International Agency for Research on Cancer Germany
Lyon
France Jennifer Ligibel
Dana-Farber Cancer Institute
Harvard Medical School
Boston, MA
USA
4
Participants
Andrew G. Renehan 2 Cornelia Ulrich

Department of Surgery, The Christie NHS Population Sciences
Foundation Trust Huntsman Cancer Institute
Institute of Cancer Sciences University of Utah
University of Manchester Salt Lake City, UT
Manchester USA
United Kingdom
Kaitlin Wade
Isabelle Romieu Faculty of Health Sciences
Instituto Nacional de Salud Pública School of Social and Community Medicine
Cuernavaca University of Bristol
Mexico Bristol
United Kingdom
Isao Shimokawa
Elisabete Weiderpass
Department of Investigative Pathology
Unit of Basic Medical Science Research Department
Graduate School of Biomedical Sciences Etiological Research Unit
Nagasaki University Cancer Registry of Norway
Nagasaki Oslo
Japan Norway
Henry J. Thompson (Subgroup Chair,

Mechanistic and Other Relevant Data)
Observer 3
Cancer Prevention Laboratory
Colorado State University Mathilde His
Fort Collins, CO
USA Inserm U1018
Centre for Research in Epidemiology and
Population Health (CESP)
Generations and Health Team
Gustave Roussy
Villejuif
France
2
Andrew G. Renehan has received research support (ceased in 2013) from Novo Nordisk, Denmark, a global health-care
company and manufacturer of insulins.
3
Each Observer agreed to respect the Guidelines for Observers at IARC Handbooks meetings. Observers did not serve as
Meeting Chair or Subgroup Chair, draft or revise any part of the Handbook, or participate in the evaluations. They also
agreed not to contact participants before or after the meeting, not to lobby them at any time, not to send them written
materials, and not to offer them meals or other favours. IARC asked and reminded Working Group Members to report
any contact or attempt to influence that they may have encountered, either before or during the meeting.
5
IARC/WHO Secretariat Administrative Assistance
Melina Arnold, Section of Cancer Marieke Dusenberg

Surveillance Sandrine Egraz
Franca Bianchini, Visiting Scientist Michel Javin
(Rapporteur) Helene Lorenzen-Augros
Robert Carreras Torres, Section of Genetics
Jean-Marie Dangou, WHO Regional Office
for Africa
Production Team
Laure Dossus, Section of Nutrition and
Metabolism
Elisabeth Elbers
Eleonora Feletto, Section of Environment
Fiona Gould
and Radiation
Solène Quennehen
Inge Huybrechts, Section of Nutrition and
Metabolism
Béatrice Lauby-Secretan, Section of IARC
Monographs (Responsible Officer, Pre-Meeting Scientific Assistance
Rapporteur)
Dana Loomis, Section of IARC Monographs Yann Grosse
(Rapporteur)
Heidi Mattock, Section of IARC
Monographs (Editor)
Jason Montez, Unit of Nutrition for Health
Post-Meeting Scientific Assistance
and Development, WHO headquarters
Amy Mullee, Section of Nutrition and Karen Müller (Editor for Volume 16)
Metabolism Nadia Vilahur
Hwayoung Noh, Section of Nutrition and
Metabolism
Robin Ohannessian, Section of Cancer
Surveillance
Sabina Rinaldi, Section of Nutrition and
Metabolism
Chiara Scoccianti (Co-Responsible Officer,
Rapporteur)
Magdalena Stepien, Section of Nutrition and
Metabolism
Kurt Straif, Section of IARC Monographs
(Head of Programme)
Kayo Togawa, Section of Environment and
Radiation
6
WORKING PROCEDURES
The Working Procedures of the IARC Handbooks of Cancer Prevention describe the objective
and scope of the programme, the scientific principles and procedures used in developing
a Handbook, the types of evidence considered, and the scientific criteria that guide the
evaluations. The Working Procedures should be consulted when reading a Handbook or a
summary of evaluations made by the IARC Handbooks. These Working Procedures apply to
the review and evaluation of primary prevention.
A. GENERAL PRINCIPLES AND which are prepared by a Working Group of inter-

PROCEDURES national experts, are scientific judgements about
the available evidence on efficacy, effectiveness,
1. Background and safety of a wide range of cancer-preventive
interventions. No recommendation is given with
Prevention of cancer is one of the key objec- regard to national or international regulations
tives of the International Agency for Research on or legislation, which are the responsibility of
Cancer (IARC). The aim of the IARC Handbooks individual governments and/or other interna-
of Cancer Prevention series is to review and eval- tional authorities. The IARC Handbooks may
uate scientific information on interventions that assist national and international authorities in
may reduce the incidence of or mortality from devising programmes of health promotion and
cancer. As a result of the IARC Handbooks eval- cancer prevention, and in making benefit–risk
uations, national and international health agen- assessments.
cies have been able, on scientific grounds, to take In this document, the term “intervention”
measures to develop interventions or recommen- refers to any chemical, activity, or strategy that
dations that will reduce the risk of developing is subject to evaluation in a Handbook. Cancer-
cancer. preventive interventions encompass pharmaco-
The criteria guiding the evaluations were first logical, immunological, dietary, and behavioural
established in 1995 at the inception of the IARC interventions that may delay, block, or reverse
Handbooks series, and were revised in subse- carcinogenic processes, or reduce underlying
quent volumes. risk factors.
Preventive interventions can be applied
2. Objective and scope across a continuum of: (1) the general population;
The objective of the IARC Handbooks (2) subgroups with particular predisposing host
programme is the preparation of critical reviews or environmental risk factors, including genetic
and evaluations of the evidence that a particular susceptibility to cancer; (3) persons with precan-
intervention can prevent cancer. The evaluations, cerous lesions; and (4) cancer patients at risk of
7
developing second primary tumours. Use of the Working Group to be relevant to making an eval-
same interventions in the treatment of cancer uation are included.
patients to control the growth, metastasis, and With regard to intervention trials, epidemio-
recurrence of tumours is considered to be patient logical studies, cancer bioassays, and mechanistic
management and not prevention, although data and other relevant data, in the interests of trans-
from clinical trials of such interventions may be parency, only reports that have been published or
pertinent when reaching an evaluation. accepted for publication in the openly available
peer-reviewed scientific literature are reviewed.
3. Selection of interventions for review The same publication requirement applies
to studies originating from IARC, including
Interventions to be evaluated in the IARC meta-analyses or pooled analyses commissioned
Handbooks series are selected on the basis of one by IARC in advance of a meeting (see Part B,
or more of the following criteria: Section 2c). Data from government-agency
• The available evidence suggests potential for reports that are publicly available in final form
significantly reducing the incidence of cancer. are also considered. Exceptionally, doctoral
• There is a substantial body of human, experi- theses and other material that are in their final
mental, clinical and/or mechanistic data suit- form and publicly available may be reviewed.
able for evaluation. Data on exposure and other information
on an intervention under consideration are
• The intervention is in widespread use and of
also reviewed. In the sections on chemical and
putative protective value, but of uncertain
physical properties, on analysis, on production
efficacy and safety.
and use, and on occurrence and exposure, the
If significant new data become available on Working Group may consider published and
an intervention for which a Handbook exists, unpublished sources of information.
a re-evaluation may be made at a subsequent In some cases it may be appropriate to review
meeting of the Working Group. only the data published subsequent to a previous
evaluation; this can be useful for updating a data-
4. Data for the IARC Handbooks base, to resolve a previously open question, or to
identify new organ sites associated with a protec-
Each Handbook considers all pertinent inter-
tive effect of the intervention. Major changes (e.g.
vention trials and observational epidemiolog-
a large body of additional data that may lead to
ical studies, and all relevant cancer bioassays
a new classification; see Part B, Section 6) are
in experimental animals. Those studies that are
more appropriately addressed by a full review
judged by the Working Group to be uninforma-
and re-evaluation of the entire body of data.
tive for the evaluation (e.g. because of method-
Inclusion of a study does not imply accep-
ological limitations or small numbers) may be
tance of the adequacy of the study design or of
cited but not summarized. When such studies
the authors’ analysis and interpretation of the
are not reviewed, the reasons are indicated.
results; any limitations noted by the Working
Mechanistic and other relevant data are also
Group are clearly outlined in square brackets at
reviewed. A Handbook does not necessarily
the end of each study description (see Part B).
cite all the mechanistic literature concerning
The reasons for not giving further consideration
the intervention being evaluated (see Part B,
to an individual study also are indicated in the
Section 4). Only those data considered by the
square brackets.
8
Working procedures
5. Meeting participants They may also contribute text on non-influential

issues, for example for the general description
Five categories of participant can be present of the intervention or for the exposure (see Part
at meetings of the IARC Handbooks: B, Section 1). Invited Specialists do not serve as
(a) Member of the Working Group meeting chair or subgroup chair, do not draft text
that pertains to the description or interpretation
The Working Group is responsible for the of data directly relevant to the evaluations, and
critical reviews and evaluations that are devel- do not participate in the evaluations.
oped during the meeting. The tasks of members
of the Working Group are: (i) to ascertain that all (c) Representative
appropriate data have been collected; (ii) to select Representatives of national and international
the data relevant for the evaluation on the basis of health agencies may attend meetings because
scientific merit; (iii) to prepare accurate summa- such agencies sponsor the IARC Handbooks
ries of the data to enable the reader to follow the programme or are interested in the subject of a
reasoning of the Working Group; (iv) to evaluate meeting. Representatives do not serve as meeting
the results of epidemiological and experimental chair or subgroup chair, do not draft any part
studies on cancer-preventive effects; (v) to eval- of a Handbook, and do not participate in the
uate data relevant to the understanding of mech- evaluations.
anisms of cancer prevention; and (vi) to make
an overall evaluation of the cancer-preventive (d) Observer
effect of the intervention in humans. Members
Observers with relevant scientific credentials
of the Working Group are selected on the basis of
are admitted to an IARC Handbook meeting
(a) knowledge and experience; and (b) absence of
in limited numbers. Attention will be given to
real or perceived conflicts of interests. Members
achieving a balance of Observers from constit-
of the Working Group generally have published
uencies with differing perspectives. They are
significant research related to the cancer-preven-
invited to observe the meeting and should not
tive effects of the interventions being reviewed,
attempt to influence it. Observers do not serve
and have in most cases been identified as experts
as meeting chair or subgroup chair, do not draft
by IARC on the basis of literature searches.
any part of a Handbook, and do not participate
Consideration is also given to demographic diver-
in the evaluations. At the meeting, the meeting
sity and balance of scientific findings and views.
chair and subgroup chairs may grant Observers
Each member of the Working Group serves as an
an opportunity to speak, generally after a discus-
individual scientist and not as a representative of
sion has been completed by the Working Group.
any organization, government, or industry.
Observers agree to respect the Guidelines for
(b) Invited Specialist Observers at IARC Handbooks meetings (avail-
able from http://handbooks.iarc.fr).
Invited Specialists are experts who have
knowledge and experience that is critical to (e) The IARC Secretariat
consideration of the intervention being eval-
The Secretariat consists of IARC scientific
uated, but who also have a real or perceived
staff who have relevant expertise. They serve as
conflict of interests. These experts are invited
rapporteurs and participate in discussions. When
when necessary to assist the Working Group by
requested by the meeting chair or subgroup
contributing their unique knowledge and expe-
chair, they may also draft text or prepare tables
rience during subgroup and plenary discussions.
9
and analyses. Members of the Secretariat do not in the original report; units are converted when
participate in the evaluations. necessary for easier comparison. The Working
Group may conduct additional analyses of the
(f) Declaration of Interests published data and use them in their assessment
Before an invitation is extended, each poten- of the evidence; the results of such supplemen-
tial participant, including the IARC Secretariat, tary analyses are given in square brackets. When
completes the WHO Declaration of Interests an important aspect of a study that directly
to report financial interests, employment and impinges on its interpretation should be brought
consulting, and individual and institutional to the attention of the reader, a Working Group
research support related to the subject of the comment is given in square brackets.
meeting or any tobacco-related interests. IARC The IARC Handbooks evaluate a wide range of
assesses these interests to determine whether interventions for primary prevention, including
there is a conflict that warrants some limitation those involving chemical or pharmacological
on participation. The declarations are updated agents (e.g. drugs, vitamins, minerals, other
and reviewed again at the opening of the meeting. nutritional supplements), immunological agents
Interests related to the subject of the meeting are (vaccination), foods, behaviour changes (e.g.
disclosed to the meeting participants and in the weight control, physical activity), and public-
published volume (Cogliano et al., 2004). health policies (e.g. smoking restrictions). The
The names and principal affiliations of partic- structure of a Handbook typically comprises the
ipants are published on the website of the IARC following sections:
Handbooks programme (http://handbooks.iarc.fr) 1. Exposure data
approximately two months before each meeting.
2. Studies of cancer prevention in humans
It is not acceptable for Observers or third parties
to contact other participants before a meeting or 3. Studies of cancer prevention in experimental
to lobby them at any time. Meeting participants animals
are asked to report all such contacts to IARC 4. Mechanistic and other relevant data
(Cogliano et al., 2005). The names and principal 5. Summary
affiliations of all meeting participants are also 6. Evaluation and rationale
listed at the beginning of the corresponding
volume of the Handbooks. In addition, a section entitled “General
Remarks” at the front of the volume discusses the
reasons why the interventions were scheduled for
B. SCIENTIFIC REVIEW AND evaluation, and key issues the Working Group
EVALUATION encountered during the meeting.
The following part of the Working Procedures
A wide range of findings must be taken into discusses the types of evidence considered and
account before a particular intervention can be summarized in each section of a Handbook,
recognized as preventing cancer, and a systematic followed by the scientific criteria that guide the
approach to data presentation has been adopted evaluations.
for Handbooks evaluations.
The available studies are summarized by the
Working Group, with particular regard to the
qualitative aspects discussed below. In general,
numerical findings are indicated as they appear
10
Working procedures
1. Characteristics and occurrence of the descriptive epidemiological studies, case-series,

intervention and case reports are usually not reviewed. The
uncertainties that surround the interpretation
Each Handbook includes general information of such studies make them inadequate, except
identifying and describing the intervention. in exceptional circumstances, to form the basis
As preventive interventions can range from for inferring a preventive relationship. However,
community-based interventions to measures when considered together with experimental and
targeted to individuals (e.g. behavioural, dietary, analytical observational studies, these types of
pharmacological measures), this information study can sometimes contribute to the decision
may vary substantially between interventions. of the Working Group as to whether or not a
Depending on the intervention, this section causal relationship exists.
may include information on production and Intervention studies are experimental in
use, occurrence and exposure, prevalence, risk design – that is, the use of, or exposure to, the
factors, and regulations and guidelines. intervention is assigned by the investigator.
Given the wide variety of preventive interven- Experimental studies can provide the strongest
tions, this section will have an outline specific to and most direct evidence of a protective or
each Handbook. preventive effect; however, the use of such studies
is limited for practical and ethical reasons and
2. Studies of cancer prevention in the subjects are often drawn from select groups
humans that may not represent the population at large.
This section includes all pertinent experi- In exceptional cases, epidemiological studies
mental and observational epidemiological studies on advanced pre-neoplastic lesions and other
of cancer prevention in humans, with cancer as end-points thought to be relevant to cancer are
an outcome (see Part A, Section 4). Studies of also reviewed in this section. The results of such
biomarkers as indicators of the intervention are studies may strengthen inferences drawn from
included in Section 4 when they are relevant to other studies.
an evaluation of the cancer-preventive effect in (b) Quality of studies considered
humans.
In considering whether a particular study
(a) Types of study considered should contribute to the evaluation of an inter-
This section focuses on studies that assess the vention, the Working Group considers the
prevention of cancer as an outcome in humans. following aspects:
Relevant evidence is normally provided by • The relevance of the study;
experimental studies (for example, random- • The appropriateness of the design and analysis
ized clinical trials and community intervention to the question being asked;
trials), and analytical observational studies,
• The adequacy and completeness of the presen-
primarily cohort studies and case–control
tation of the data; and
studies. For certain interventions applied at
the population level, well-designed ecolog- • The degree to which chance, bias, and
ical studies (studies measuring both outcome confounding may have affected the results;
and exposure on the aggregate, or population, for drugs or other marketed products, this
level) or interrupted time-series studies may bias assessment should include review of the
also be informative. Cross-sectional studies, funding source.
11
Aspects that are particularly important in that was independent of the intervention of
evaluating randomized controlled trials are: the interest, and exposure to the intervention should
selection of participants, the nature and adequacy have been assessed in a way that was not related
of the randomization procedure, evidence that to disease status.
randomization achieved an adequate balance Secondly, the authors should have taken into
between the groups, exclusion criteria used account – in the study design and analysis – other
before and after randomization, compliance variables that could influence the risk of disease
with the intervention in the intervention group, and may have been related to the exposure of
and “contamination” of the control group with interest. Potential confounding by such variables
the intervention. Other considerations are the should have been dealt with either in the design
means by which the end-point was determined of the study (e.g. by matching) or in the analysis
and validated (either by screening or by other (by statistical adjustment). Internal compari-
means of detection of the disease), the length and sons of frequency of disease among individuals
completeness of follow-up of the groups, and the with different levels of exposure are desirable in
adequacy of the analysis. cohort studies, since they minimize the potential
It is necessary to take into account the for confounding related to the difference in risk
possible roles of bias, confounding, and chance factors between an external reference group and
in the interpretation of cohort and case–control the study population.
studies. Bias is the effect of factors in study design Thirdly, the authors should have reported the
or execution that leads erroneously to a stronger basic data on which the conclusions are founded,
or weaker association than in fact exists between even if sophisticated statistical analyses were
an intervention and outcome. Confounding is a employed. At the very least, they should have
form of bias that occurs when the relationship given the numbers of exposed and unexposed
with the outcome is made to appear stronger or cases and controls in a case–control study, and
weaker than it is in reality, due to an association the numbers of cases observed and expected in
between the apparent causal factor and another a cohort study. Further tabulations by duration
factor that is associated with either an increase of exposure and other temporal factors are also
or a decrease in the incidence of the disease. The important. In a cohort study, data on all cancer
role of chance is related to biological variability sites and all causes of death should have been
and the influence of sample size on the precision given, to reveal the possibility of reporting bias.
of estimates of effect. In a case–control study, the effects of investigated
In evaluating the extent to which these factors other than the exposure of interest should
factors have been minimized in an individual have been reported.
study, consideration is given to a number of Finally, the statistical methods used to obtain
aspects of design and analysis as described in the estimates of relative risk, absolute rates of cancer,
report of the study. Most of these considerations confidence intervals, and significance tests, and
apply equally to all types of study. Lack of clarity to adjust for confounding should have been
regarding any of these aspects in the reporting of clearly stated by the authors. These methods have
a study can decrease its credibility and the weight been reviewed for case–control studies (Breslow
given to it in the final evaluation. & Day, 1980) and for cohort studies (Breslow &
Firstly, the study population, target organ, Day, 1987).
and exposure should have been well defined
by the authors. Cancer occurrence in the study
population should have been identified in a way
12
Working procedures
(c) Quantitative aspects study methodology. The results of studies judged

The Working Group gives special attention to to be of high quality are given more weight. Note
quantitative assessment of the preventive effect is taken both of the applicability of preventive
of the intervention under study, by assessing action to several cancers and of possible differ-
data from studies investigating different doses ences in activity, including the possibility of
or levels of exposure. The Working Group also different findings between cancer sites.
addresses issues of timing and duration of use or
exposure. Such quantitative assessment is impor- 3. Studies of cancer prevention in
tant to clarify the circumstances under which a experimental animals
preventive effect can be achieved, as well as the (a) Types of study considered
dose or level of exposure at which a toxic effect
has been shown. Animal models are an important component
of research into cancer prevention. Models that
(d) Criteria for preventive effects permit evaluation of the effects of cancer-preven-
After summarizing and assessing the indi- tive interventions on the occurrence of cancer
vidual studies, the Working Group makes a judge- in most major organ sites are available. Animal
ment concerning the strength of the evidence that models for such studies include: those in which
the intervention in question prevents cancer in cancer is produced by the administration of a
humans. In making its judgement, the Working chemical or physical carcinogen; those involving
Group considers several criteria for each relevant genetically engineered animals; and those in
cancer site. which tumours develop spontaneously. Most
Evidence is frequently available from different cancer-preventive interventions investigated
types of study and is evaluated as a whole. in such studies can be placed into one of three
Findings that are replicated in several studies categories: interventions that prevent molecules
of the same design or in studies using different from reaching or reacting with critical target sites
approaches are more likely to provide evidence of (blocking agents); interventions that decrease
a true protective effect than are isolated observa- the sensitivity of target tissues to carcinogenic
tions from single studies. stimuli; and interventions that prevent evolution
Evidence of protection derived from inter- of the neoplastic process (suppressing agents).
vention studies of good quality is particularly There is increasing interest in the use of combi-
informative. Evidence of a substantial and nations of interventions as a means of improving
significant reduction in risk, including a “dose”– efficacy and minimizing toxicity; animal models
response relationship, is more likely to indicate are useful in evaluating such combinations. The
a true effect. Nevertheless, a small effect, or an development of optimal strategies for interven-
effect without a dose–response relationship, does tion trials in humans can be facilitated by the
not imply lack of real benefit and may be impor- use of animal models that mimic the neoplastic
tant for public health if the cancer is common. process in humans.
The Working Group evaluates possible Specific factors to be considered in such
explanations for inconsistencies across studies, experiments are: (1) the temporal requirements
including differences in use of, or exposure to, of administration of the cancer-preventive
the intervention, differences in the underlying interventions; (2) dose–response effects; (3) the
risk of cancer, and metabolism and genetic differ- site specificity of cancer-preventive activity;
ences in the population, as well as differences in and (4) the number and structural diversity of
13
carcinogens whose activity can be reduced by the of the study); (2) the consistency of the results, for
intervention being evaluated. example across species and target organ(s); (3) the
An important variable in the evaluation of stage or stages of the neoplastic process studied,
the cancer-preventive response is the time and from pre-neoplastic lesions and benign tumours
duration of administration of the intervention to malignant tumours; and (4) the possible role
in relation to any carcinogenic treatment, or in of modifying factors.
transgenic or other experimental models in which In the interpretation and evaluation of a
no carcinogen is administered. Furthermore, particular study, the Working Group takes into
concurrent administration of an intervention consideration: (1) how clearly the interven-
may result in a decreased incidence of tumours tion was defined and, in the case of mixtures,
in a given organ and an increase in incidence in how adequately the sample composition was
another organ of the same animal. Thus, in these reported; (2) the composition of the diet and
experiments it is important that multiple organs the stability of the intervention in the diet;
be examined. (3) whether the source, strain, and quality of the
For all these studies, the nature and extent animals was reported; (4) whether there were
of impurities or contaminants present in the adequate numbers of animals, of appropriate
cancer-preventive intervention or interven- age, per group; (5) whether males and females
tions being evaluated are given when available. were used, if appropriate; (6) whether animals
Also, consideration is given to the possibility of were allocated randomly to groups; (7) whether
changes in the physicochemical properties of the appropriate respective controls were used;
test substance during collection, storage, extrac- (8) whether the dose and schedule of treatment
tion, concentration, and delivery. Chemical and with the known carcinogen were appropriate in
toxicological interactions of the components of assays of combined treatment; (9) whether the
mixtures may result in non-linear dose–response doses of the cancer-preventive intervention were
relationships. adequately monitored; (10) whether the agent(s)
As certain components of commonly used was absorbed, as shown by blood concentrations;
diets of experimental animals are themselves (11) whether the survival of treated animals was
known to have cancer-preventive activity, similar to that of controls; (12) whether the body
particular consideration should be given to the and organ weights of treated animals were similar
interaction between the diet and the apparent to those of controls; (13) whether the duration
effect of the intervention being studied. Likewise, of the experiment was adequate; (14) whether
restriction of diet may be important. The appro- there was adequate statistical analysis; and
priateness of the diet given relative to the compo- (15) whether the data were adequately reported.
sition of human diets may be commented on by
the Working Group. (c) Quantitative aspects
The incidence of tumours may depend on
(b) Quality of studies considered the species, sex, strain, and age of the animals,
An assessment of the experimental preven- the dose of carcinogen (if any), the dose of the
tion of cancer involves several considerations agent, and the route and duration of exposure.
of qualitative importance, including: (1) the A decreased incidence and/or decreased multi-
experimental conditions under which the test plicity of tumours in adequately designed studies
was performed (route and schedule of expo- provide evidence of a cancer-preventive effect. A
sure, species, strain, sex and age of the animals dose-related decrease in incidence and/or multi-
studied, duration of the exposure, and duration plicity further strengthens this association.
14
Working procedures
The nature of the dose–response relationship The focus of this section is on studies in
can vary widely, depending on the agent and the humans, including intervention trials and epide-
target organ. Saturation of steps such as absorp- miological studies with cancer-relevant molec-
tion, activation, inactivation, and elimination ular biomarkers or intermediate end-points as
may produce non-linearity in the dose–response an outcome. Studies in experimental systems
relationship (Hoel et al., 1983; Gart et al., 1986), can strengthen the evidence for the poten-
as could saturation of the detoxication processes. tial cancer-preventive effect of an intervention
The dose–response relationship can also be observed in humans, and can elucidate the mech-
affected by differences in survival between the anism(s) of cancer prevention. A brief summary
treatment groups. of important findings in experimental systems is
therefore included.
(d) Statistical analyses Evaluation of the results of intervention
Factors considered in the statistical analysis studies in humans includes consideration of
by the Working Group include: (1) the adequacy quality, as described above. Study quality factors
of the data for each treatment group; (2) the generally consider the adequacy of the methods
initial and final effective numbers of animals and the reporting of results, addressing: (1) the
studied and the survival rate; (3) body weights; description of the methods; (2) the appropri-
and (4) tumour incidence and multiplicity. ateness of control populations; (3) whether
The statistical methods used should be clearly toxic effects were considered in the outcome;
stated and should be the generally accepted tech- (4) whether the data were appropriately compiled
niques defined for this purpose. In particular, the and analysed; (5) whether appropriate quality
statistical methods should be appropriate for the controls were used; (6) whether appropriate
characteristics of the expected data distribution concentration ranges were used; (7) whether
and should account for interactions in multi- adequate numbers of independent measurements
factorial studies. Consideration is given as to were made per group; and (8) the relevance of the
whether the appropriate adjustment was made end-points.
for differences in survival. The observation of effects on the occurrence
If available, recent data on the incidence of of lesions presumed to be pre-neoplastic, or the
specific tumours in historical controls, as well as emergence of benign or malignant tumours,
in concurrent controls, are taken into account in may aid in assessing the mode of action of the
the evaluation of tumour response. intervention being considered. Particular atten-
tion is given to assessing the reversibility of these
4. Mechanistic and other relevant data lesions and their predictive value in relation to
cancer development.
In evaluating an intervention, effects other than
cancer are described and weighed. Furthermore, (a) Toxicokinetics
information that facilitates an understanding of Information is given on absorption, distribu-
the applicability of findings to different species, tion (including placental transfer), metabolism,
or to different human populations is particularly and excretion in humans. If human data are
important; this includes metabolic, kinetic, and sparse, evidence from experimental animals may
genetic data. Whenever possible, quantitative be summarized. Studies in humans that indicate
data, including information on dose, duration, the metabolic pathways and fate of an interven-
and potency, are considered. tion are summarized. Data indicating long-term
accumulation in human tissues are included.
15
Observations are made on inter-individual vari- molecules, angiogenesis, interactions with the
ations and relevant metabolic polymorphisms. extracellular matrix, hormonal status, and the
Physiologically based pharmacokinetic models immune system.
and their parameter values are relevant and are Many cancer-preventive interventions are
included whenever they are available. known or suspected to act by several mech-
Information from experimental systems, anisms, which may operate in a coordinated
including on the fate of the compound within manner and allow them a broader spectrum of
tissues and cells (transport, role of cellular recep- anticarcinogenic activity. Therefore, a range of
tors, compartmentalization, binding to macro- possible mechanisms of action are taken into
molecules) may be briefly summarized. account in the evaluation of cancer prevention.
The metabolic consequences of interventions These can be conceptually organized to encom-
are described. pass impacts on one or more related key char-
acteristics of carcinogens (Smith et al., 2016),
(b) Mechanisms of cancer prevention particularly interference with: (1) metabolic acti-
For a rational implementation of cancer- vation of carcinogens; (2) mutagenesis; (3) DNA
preventive measures, it is essential not only to repair or genomic instability; (4) epigenetic
assess protective end-points but also to under- effects; (5) oxidative stress; (6) inflammation;
stand the mechanisms by which the interven- (7) immune function; (8) receptor-mediated
tion exerts its anticarcinogenic action. Data effects; (9) immortalization; or (10) cell prolifer-
on mechanisms will be primarily from studies ation, cell death, or nutrient supply.
in humans. Data from relevant experimental
models can also be summarized, including (c) Susceptible populations
studies of the inhibition of tumorigenesis in This section summarizes studies of cancer
vivo, studies of intermediate biomarkers in vivo, in humans that have addressed differential
analyses of interactions between agents and susceptibility due to toxicokinetics, mechanisms
specific molecular targets, and studies of specific of cancer prevention, and other factors. Such
end-points in vitro. Information on the mecha- studies may identify individuals, populations,
nisms of cancer-preventive activity inferred from and life-stages with greater or lesser suscep-
relationships between chemical structure and tibility. Examples of host and genetic factors
biological activity can also be included. that affect individual susceptibility include sex,
Cancer-preventive interventions may act genetic polymorphisms of genes involved in
at different levels: (1) extracellular, for example the metabolism of the intervention, differences
inhibiting the uptake or endogenous forma- in metabolic capacity due to life-stage or the
tion of carcinogens, or forming complexes presence of disease, differences in DNA repair
with, diluting, and/or deactivating carcinogens; capacity, competition for alteration of metabolic
(2) intracellular, for example trapping carcin- capacity by medications or other chemical expo-
ogens in nontarget cells, modifying trans- sures, a pre-existing hormonal imbalance that is
membrane transport, modulating metabolism, exacerbated by a chemical exposure, a suppressed
blocking reactive molecules, inhibiting cell repli- immune system, periods of higher-than-usual
cation, or modulating gene expression or DNA tissue growth or regeneration, and genetic poly-
metabolism; or (3) at the level of the cell, tissue, morphisms that lead to differences in behav-
or organism, for example affecting cell differen- iour (e.g. addiction). Genotyping is being used
tiation, intercellular communication, proteases, increasingly, not only to identify subpopulations
signal transduction, growth factors, cell adhesion at increased or decreased risk for cancers but also
16
Working procedures
to characterize variation in the biotransforma- relationships, dose–response, and other quanti-

tion of and response to cancer-preventive inter- tative data are also summarized.
ventions. Such data can substantially increase
the strength of the evidence from epidemiolog- (d) Mechanistic and other relevant data
ical data and enhance the linkage of in vivo and Human data relevant to the toxicokinetics
in vitro laboratory studies to humans. (absorption, distribution, metabolism, elimina-
tion) and the possible mechanism(s) of cancer
(d) Adverse effects prevention are summarized. In addition, human
Relevant clinical or other evidence that would studies on cancer susceptibility including on
impact any recommendations may be summa- genetic polymorphisms, susceptible populations
rized as appropriate. and life-stages are summarized. This section
also reports briefly on adverse effects as well as
5. Summary of data any additional relevant data from experimental
systems that are considered to be influential for
This section is a summary of data presented the evaluation of a cancer-preventive effect.
in the preceding sections.
(a) Exposure data 6. Evaluation and rationale

Data are summarized, as appropriate, on Evaluations of the strength of the evidence
elements such as characteristics and production for cancer-preventive effects from studies in
or implementation of the intervention, and humans and experimental animals are made
patterns of use or exposure in human popula using standard terms. Similarly, an evaluation of
tions. Quantitative data and time trends are given the strength of the mechanistic evidence is given.
to compare exposure, use, or implementation in It is recognized that the criteria for these
different regions and settings. evaluation categories, described below, cannot
encompass all factors that may be relevant to
(b) Cancer prevention in humans an evaluation of cancer-preventive effects. In
Results of epidemiological studies pertinent considering all the relevant scientific data, the
to an assessment of the cancer-preventive effect Working Group may assign the intervention to a
in humans are summarized. The target organ(s) higher or lower category than a strict interpreta-
or tissue(s) in which a decrease in cancer occur- tion of these criteria would indicate.
rence was observed is identified. Dose–response The evaluation categories refer only to the
and other quantitative data may be summarized strength of the evidence that an intervention
when available. prevents cancer, and not to the extent of its
cancer-preventive effects (potency). The evalua-
(c) Cancer in experimental animals tions may change as new information becomes
Data relevant to an evaluation of a cancer- available.
preventive effect in animals are summarized. For Evaluations are inevitably limited to the inter-
each animal species, study design, and route of vention as actually implemented and observed,
administration, it is stated whether decreased for example to the cancer sites, conditions, and
incidence, increased latency, or decreased duration of observation covered by the available
severity or multiplicity of tumours or pre- studies.
neoplastic lesions were observed, and the tumour
sites are indicated. Negative findings, positive
17
(a) Cancer-preventive effects in humans the observed results should be considered and
The evidence relevant to cancer prevention excluded with confidence.
in humans is classified into one of the following (b) Cancer-preventive effects in experimental
categories: animals
Sufficient evidence of cancer-preventive effects:
The Working Group considers that a preventive Cancer-preventive effects in experimental
relationship has been established between the animals can be evaluated using conventional
intervention and the risk of cancer in humans. bioassays, bioassays that employ genetically
That is, a preventive association has been modified animals, and other in vivo bioassays
observed in studies in which chance, bias, and that focus on one or more of the critical stages of
confounding could be ruled out with confidence. carcinogenesis.
A statement that there is sufficient evidence is Evidence for cancer prevention in exper-
followed by a sentence identifying the organ(s) imental animals is classified into one of the
or tissue(s) for which a preventive effect has been following categories:
observed in humans. Identification of preven- Sufficient evidence of cancer-preventive effects:
tive effects in a specific organ or tissue does not The Working Group considers that a causal
preclude the possibility that the intervention may relationship has been established between the
prevent cancer at other sites. intervention and a decreased incidence and/or
Limited evidence of cancer-preventive effects: multiplicity of spontaneous or chemically
A reduced risk of cancer is associated with the induced malignant neoplasms, or of an appro-
intervention for which a preventive effect is priate combination of benign and malignant
considered credible by the Working Group, but neoplasms in an adequate number (four or more)
chance, bias, or confounding could not be ruled of independent studies carried out at different
out with confidence. times, or in different laboratories, or under
Inadequate evidence of cancer-preventive different protocols.
effects: The available studies are not of sufficient Limited evidence of cancer-preventive effects:
quality, consistency, or statistical power to permit The data indicate a cancer-preventive effect,
a conclusion regarding the presence or absence of but are limited for making a definitive evalua-
a cancer-preventive effect of the intervention, or tion because, for example: (a) the evidence of a
no data on the prevention of cancer by this inter- cancer-preventive effect is restricted to a small
vention in humans are available. number (fewer than four) of experiments; or
Evidence suggesting lack of cancer-preventive (b) the intervention decreases the incidence and/
effects: When several epidemiological studies or multiplicity of benign neoplasms only.
show little or no indication of an association Inadequate evidence of cancer-preventive
between an intervention and a reduced risk effects: The studies cannot be interpreted as
of cancer, a judgement may be made that the showing either the presence or absence of a
studies, taken together, show evidence of lack of preventive effect because of major methodo-
a preventive effect. Such a judgement requires logical or quantitative limitations: unresolved
that the studies meet the standards of design and questions regarding the adequacy of the design,
analysis described above. Specifically, the possi- conduct, or interpretation of the study, or few or
bility that bias, confounding, or misclassification no data on cancer prevention in experimental
of the intervention or the outcome could explain animals are available.
18
Working procedures
Evidence suggesting lack of cancer-preventive The overall evaluation is described according

activity: Adequate evidence from conclusive to the wording of one of the following standard
studies in several models shows that, within the categories. The categorization of an intervention
limits of the tests used, the intervention has no is a matter of scientific judgement that reflects the
cancer-preventive effects. strength of the evidence derived from studies in
humans and in experimental animals, and from
(c) Mechanistic data on cancer-preventive mechanistic and other relevant data.
effects
(i) The intervention prevents cancer
Mechanistic and other evidence judged to be (Group A)
relevant to an evaluation of a cancer-preventive
This category is used for interventions for
effect and of sufficient importance to affect the
overall evaluation is brought forward to the which there is sufficient evidence of a cancer-
preventive effect in humans.
evaluation.
The sites on which the evidence in humans is
The strength of mechanistic evidence
based are given.
supporting the cancer-preventive effect is eval-
uated, using terms such as ‘weak’, ‘moderate’, or (ii) The intervention probably prevents
‘strong’. Indications that a particular mechanism cancer (Group B1)
operates in humans are strongest. The data may This category is used for interventions for
be considered to be especially relevant if they which there is limited evidence of a cancer-
show in humans that the intervention in ques- preventive effect in humans and sufficient
tion has caused suppression of effects that are on evidence in animals. An intervention may also
the pathway to cancer. The mechanistic evidence be classified in this category when there is limited
can be strengthened by findings of consistent evidence in humans, less than sufficient evidence
results in different experimental designs, by the in experimental animals, and strong supporting
demonstration of biological plausibility, and by evidence from mechanistic and other relevant
coherence of the overall database. data that the mechanism(s) of prevention also
The Working Group considers whether operates in humans.
multiple mechanisms might contribute to cancer The sites on which the evidence in humans is
prevention, whether different mechanisms might based are given.
operate in different dose ranges or at different
sites, or whether separate mechanisms might (iii) The intervention possibly prevents
operate in a susceptible group. cancer (Group B2)
For complex interventions, such as food This category is used for interventions for
categories, the chemical composition and the which there is inadequate evidence in humans
potential contribution of different nutrients and sufficient evidence in experimental animals.
known to be present may be considered by the An intervention may also be classified in this
Working Group in its overall evaluation of cancer category when there is inadequate evidence
prevention. in humans, limited evidence in experimental
animals, and strong supporting evidence from
(d) Overall evaluation mechanistic and other relevant data that the
Finally, the body of evidence is considered as mechanism(s) of prevention also operates in
a whole, and summary statements are made that humans.
encompass the effects of the intervention with
regard to cancer-preventive effects in humans.
19
(iv) The intervention is unclassifiable as to References

its cancer-preventive effects (Group C)
This category is used for interventions for Breslow NE, Day NE (1980). The analysis of case–control
which the evidence is inadequate in humans studies. Statistical methods in cancer research, Volume
I. Lyon, France: International Agency for Research on
and less than sufficient in experimental animals. Cancer (IARC Scientific Publication No. 32). Available
Interventions that do not fall into any other from: http://publications.iarc.fr/175.
group are also placed in this category. Breslow NE, Day NE (1987). The design and analysis of
cohort studies. Statistical methods in cancer research,
(v) The intervention probably does not Volume II. Lyon, France: International Agency for
prevent cancer (Group D) Research on Cancer (IARC Scientific Publication No.
82). Available from: http://publications.iarc.fr/225.
This category is used for interventions for Cogliano VJ, Baan RA, Straif K, Grosse Y, Secretan MB,
which there is evidence suggesting lack of a El Ghissassi F, et al. (2004). The science and practice
cancer-preventive effect both in humans and in of carcinogen identification and evaluation. Environ
Health Perspect, 112(13):1269–74. doi:10.1289/ehp.6950
experimental animals. PMID:15345338
Cogliano V, Baan R, Straif K, Grosse Y, Secretan B,
(e) Rationale El Ghissassi F, et al. (2005). Transparency in IARC
Monographs. Lancet Oncol, 6(10):747. doi:10.1016/
The reasoning that the Working Group used S1470-2045(05)70380-6
to reach its evaluation is presented and discussed. Gart JJ, Krewski D, Lee PN, Tarone RE, Wahrendorf J
This section integrates the major findings from (1986). The design and analysis of long-term animal
studies in humans, studies in experimental experiments. Statistical methods in cancer research,
Volume III. Lyon, France: International Agency for
animals, and mechanistic and other relevant data. Research on Cancer (IARC Scientific Publication No.
It includes concise statements of the principal 79). Available from: http://publications.iarc.fr/222.
line(s) of argument that emerged, the conclu- Hoel DG, Kaplan NL, Anderson MW (1983). Implication
of nonlinear kinetics on risk estimation in carcinogen-
sions of the Working Group on the strength of esis. Science, 219(4588):1032–7. PMID:6823565
the evidence for each group of studies, and an Smith MT, Guyton KZ, Gibbons CF, Fritz JM, Portier CJ,
explanation of the reasoning of the Working Rusyn I, et al. (2016). Key characteristics of carcino-
Group in weighing data and making evaluations. gens as a basis for organizing data on mechanisms of
carcinogenesis. Environ Health Perspect, 124(6):713–21.
The human populations that were the subject of doi:10.1289/ehp.1509912 PMID:26600562
study should be identified. Additionally, impor-
tant health concerns identified – such as adverse
effects, including cancer-causing properties –
should be clearly addressed.
When there are significant differences in
scientific interpretation among Working Group
members, a brief summary of the alternative
interpretations is provided, together with their
scientific rationale and an indication of the rela-
tive degree of support for each alternative.
20
GENERAL REMARKS
This sixteenth Volume of the IARC Hand Previous evaluations

books of Cancer Prevention series evaluates the
cancer-preventive effects of absence of excess In 2001, a Working Group of interna-
body fatness. It is the second Volume since the tional experts developed Volume 6 of the IARC
relaunch of the IARC Handbooks series in 2014, Handbooks, on weight control and physical
and the first Volume on primary prevention in activity (IARC, 2002). The resulting consensus
the new series. evaluations are presented in Table 1.
The IARC Handbooks provide the same
rigorous evaluation process as the IARC
Monographs. They serve national health agen- Rationale for a re-evaluation
cies to inform their preventive strategies for
cancer control. To support the World Health
A re-evaluation of the cancer-preventive
Organization Global Action Plan for the preven-
effects of avoidance of weight gain was highly
tion and control of noncommunicable diseases,
desired. The mean body mass index (BMI)
the availability of an international consensus
of the population has increased dramatically
from an independent, specialized agency within
worldwide during the past 40 years (NCD
the United Nations family provides an author-
Risk Factor Collaboration (NCD-RisC), 2016).
itative basis for national decision-making, and
The United Nations High-Level Meeting on
should facilitate national recommendations and
Noncommunicable Diseases in September 2011
communication with the population at risk.
identified obesity as one of the leading risk factors
For this Volume, the Working Procedures
for chronic diseases, including coronary heart
of the IARC Handbooks have been updated in
disease, diabetes, and cancer (Beaglehole et al.,
accordance with the current Preamble of the
2011). Overweight and obesity have been esti-
IARC Monographs (IARC, 2006), with defini-
mated to have accounted for 4.0 million deaths
tions of the different types of participants and
(95% uncertainty interval, 2.7–5.3 million)
guidelines for selection of experts and literature
worldwide in 2015, representing 7.1% (95%
searches. In addition, more detailed instructions
uncertainty interval, 4.9–9.6%) of total global
are given for the scientific review and evaluation
mortality (Afshin et al., 2017). In 2014, the
criteria. (See the Working Procedures in this
overall socioeconomic cost associated with
Volume.)
obesity was estimated at US$ 2 trillion globally
(Dobbs et al., 2014).
21
IARC HANDBOOKS OF CANCER PREVENTION - 16
Table 1 Evaluations of IARC Handbooks Volume 6 (2002)
Intervention Humans Experimental animals Overall evaluation

Strength of Organ site Strength of Organ site
evidence evidence
Avoidance of Sufficient Colon Limiting weight gain during adult
weight gain Breast life, thereby avoiding overweight
(postmenopausal) and obesity, reduces the risk of
Endometrium postmenopausal breast cancer
Kidney (renal cell) and cancers of the colon, uterus
Oesophagus (endometrium), kidney (renal cell),
(adenocarcinoma) and oesophagus (adenocarcinoma).
ESLE Breast
(premenopausal)
Intentional Inadequate Sufficient (Calorie/dietary Weight loss among overweight or
weight loss restriction) obese persons possibly reduces
Mammary gland risks of these cancers, but no
Liver firm conclusion can be drawn
Pituitary gland because of the sparsity of the
(adenoma) epidemiological evidence.
Colon
Skin (non-melanoma)
Lymphoma
Limited Prostate
Pancreas
ESLE, evidence suggesting a lack of effect
Worldwide, it has been estimated that 481 000 Content of this Handbook

new cancer cases (3.6% of all new cases) in adults
in 2012 could be attributed to high BMI; the In this Volume, in addition to the identifi-
attributable fraction was as high as 9% in women cation of target organs for excess body fatness,
in North America, Europe, and the Middle East the following topics have been reviewed when
(Arnold et al., 2015). This estimation was based available:
on evidence for an association of high BMI with
oesophageal adenocarcinoma and cancers of the • Sex specificity
colon, rectum, pancreas, gall bladder, kidney, • Anthropometric measures of body fatness
postmenopausal breast, corpus uteri, and ovary other than BMI: weight, waist circumference,
(Arnold et al., 2015). Taking into account the and waist-to-hip ratio
evaluations of this Volume, which indicate that • Effect of change in BMI or weight over the
excess body fatness increases cancer risk at addi- life-course
tional sites (a total of 13 cancer sites or subtypes), • Risk reduction after intentional weight loss
the fraction of cancer cases worldwide that are • Effect of excess body fatness on cancer survival
attributable to overweight and obesity is even in cancer patients, and on recurrence in
higher than previously estimated. cancer survivors
• Excess body fatness in children, adolescents,
and young adults (age ≤ 25 years) and subse-
quent cancer risk.
22
General remarks
Weight loss References
Few data are available on intentional weight Afshin A, Forouzanfar MH, Reitsma MB, Sur P, Estep
K, Lee A, et al.; GBD 2015 Obesity Collaborators
loss in humans. Therefore, data in experimental (2017). Health effects of overweight and obesity in 195
animals provide important information to assess countries over 25 years. N Engl J Med, 377(1):13–27.
the effect of intentional weight loss. Studies doi:10.1056/NEJMoa1614362 PMID:28604169
in animals use dietary or calorie restriction Arnold M, Pandeya N, Byrnes G, Renehan PAG, Stevens
GA, Ezzati PM, et al. (2015). Global burden of cancer
to induce a lower weight gain compared with attributable to high body-mass index in 2012: a
animals fed ad libitum, or to induce weight population-based study. Lancet Oncol, 16(1):36–46.
reduction in obese animals. doi:10.1016/S1470-2045(14)71123-4 PMID:25467404
Beaglehole R, Bonita R, Alleyne G, Horton R, Li L,
For humans, the Working Group considered
Lincoln P, et al.; Lancet NCD Action Group (2011). UN
the data on bariatric surgery as a proxy for the High-Level Meeting on Non-Communicable Diseases:
evidence on intentional weight loss. The clinical addressing four questions. Lancet, 378(9789):449–55.
effectiveness of bariatric surgery for weight loss doi:10.1016/S0140-6736(11)60879-9 PMID:21665266
Dobbs R, Sawers C, Thompson F, Manyika J, Woetzel J,
and improved health has been established (Picot Child P, et al. (2014). Overcoming obesity: an initial
et al., 2009), although risks of complications, economic analysis. London, UK: McKinsey Global
reoperation, and death exist. Institute, McKinsey & Company. Available from:
ht tp://w w w.mck insey.com/~/media/McK insey/
Business Functions/Economic Studies TEMP/Our
Insights/How the world could better fight obesity/
Impact of physical activity on MGI_Overcoming_obesity_Full_report.ashx.
IARC (2002). Weight control and physical activity. Lyon,
the assessment of the cancer- France: IARC Press (IARC Handbooks of Cancer
preventive effects of absence of Prevention, Vol. 6). Available from: http://publications.
iarc.fr/376.
excess body fatness IARC (2006). Preamble to the IARC Monographs.
Available from: http://monographs.iarc.fr/ENG/
The major contributors to weight gain are Preamble/index.php.
Lauby-Secretan B, Scoccianti C, Loomis D, Grosse
excess energy intake and insufficient levels of Y, Bianchini F, Straif K; International Agency for
physical activity, which both lead to chronic Research on Cancer Handbook Working Group
positive energy balance. In recent years, new (2016). Body fatness and cancer – viewpoint of the
evidence has accumulated on the different types IARC Working Group. N Engl J Med, 375(8):794–8.
doi:10.1056/NEJMsr1606602 PMID:27557308
of physical inactivity and on sedentary behav- NCD Risk Factor Collaboration (NCD-RisC) (2016).
iour as risk factors for cancer. In this Handbook, Trends in adult body-mass index in 200 countries
the cancer-preventive effects of absence of excess from 1975 to 2014: a pooled analysis of 1698 popula-
tion-based measurement studies with 19.2 million
body fatness were evaluated taking into account participants. Lancet, 387(10026):1377–96. doi:10.1016/
potential confounding and/or effect modifica- S0140-6736(16)30054-X PMID:27115820
tion by physical activity. Physical activity will be Picot J, Jones J, Colquitt JL, Gospodarevskaya E, Loveman
evaluated separately in a future Handbook. E, Baxter L, et al. (2009). The clinical effectiveness and
cost-effectiveness of bariatric (weight loss) surgery for
A summary of the findings of this Volume obesity: a systematic review and economic evaluation.
has appeared in The New England Journal of Health Technol Assess, 13(41):1–190, 215–357, iii–iv.
Medicine (Lauby-Secretan et al., 2016). doi:10.3310/hta13410 PMID:19726018
23
LIST OF ABBREVIATIONS
AKT protein kinase B

AL ad libitum
AML acute myeloid leukaemia
AMP adenosine monophosphate
AMPK AMP-activated protein kinase
AOM azoxymethane
ATP adenosine triphosphate
BIA bioelectrical impedance analysis
BMI body mass index
BOP N-nitrosobis(2-oxopropyl)amine
CDK cyclin-dependent kinase
cDNA complementary DNA
CDR chronic dietary restriction
CI confidence interval
CIMP CpG island methylator phenotype
CLL chronic lymphocytic leukaemia
CML chronic myeloid leukaemia
COX-2 cyclooxygenase-2
CR calorie restriction
CRC colorectal cancer
CRP C-reactive protein
CT computed tomography
DEN diethylnitrosamine
DHEA dehydroepiandrosterone
DIO diet-induced obesity
DLBCL diffuse large B-cell lymphoma
DMBA 7,12-dimethylbenz[a]anthracene
DR dietary restriction
DXA dual-energy X-ray absorptiometry
EDC endocrine-disrupting chemical
8-epi-PGF2α 8-epi-prostaglandin F2α
EPIC European Prospective Investigation into Cancer and Nutrition
ER estrogen receptor
F344 Fischer 344
25
GIANT Genetic Investigation of Anthropometric Traits

GRS genetic risk score
GTG gold thioglucose
HBV hepatitis B virus
HCC hepatocellular carcinoma
HCV hepatitis C virus
HER2 human epidermal growth factor receptor 2
HFD high-fat diet
HPV human papillomavirus
HR hazard ratio
HRT hormone replacement therapy
IARC International Agency for Research on Cancer
ICD International Classification of Diseases
IDR intermittent dietary restriction
IGF-1 insulin-like growth factor 1
IGF-1R IGF-1 receptor
IGFBP IGF binding protein
IL interleukin
IWL intentional weight loss
LFD low-fat diet
MAPK mitogen-activated protein kinase
METs metabolic equivalents
MMTV mouse mammary tumour virus
MNU N-methyl-N-nitrosourea
MRI magnetic resonance imaging
MSI microsatellite instability
MSS microsatellite-stable
mTOR mammalian target of rapamycin
NAFLD non-alcoholic fatty liver disease
NAFPD non-alcoholic fatty pancreatic disease
NASH non-alcoholic steatohepatitis
NCDs noncommunicable diseases
NF-κB nuclear factor kappa-light-chain-enhancer of activated B cells
NHANES National Health and Nutrition Examination Survey
NHL non-Hodgkin lymphoma
NIH-AARP National Institutes of Health–AARP Diet and Health Study
NOS not otherwise specified
25(OH)D 25-hydroxyvitamin D
OR odds ratio
8-oxo-dG 8-hydroxydeoxyguanosine
PAI-1 plasminogen activator inhibitor-1
PanIN pancreatic intraepithelial neoplasia
PCR polymerase chain reaction
PDAC pancreatic ductal adenocarcinoma
PI3K phosphoinositide 3-kinase
PIN prostatic intraepithelial neoplasia
PPAR peroxisome proliferator-activated receptor
PR progesterone receptor
PSA prostate-specific antigen
RCC renal cell carcinoma
26
Abbreviations
RCT randomized controlled trial

ROS reactive oxygen species
RR relative risk
SASP salicylazosulfapyridine
SD standard deviation
SEER Surveillance, Epidemiology, and End Results
sGC soluble guanylyl cyclase
SHBG sex hormone-binding globulin
SHT scopolamine hydrobromide trihydrate
SLL small lymphocytic lymphoma
SNP single nucleotide polymorphism
STAT signal transducer and activator of transcription
TDR total dietary restriction
TGF transforming growth factor
TNF-α tumour necrosis factor alpha
TPA 12-O-tetradecanoylphorbol-13-acetate
TRAMP transgenic adenocarcinoma of the mouse prostate
VEGF vascular endothelial growth factor
VO2max maximal oxygen uptake
WC waist circumference
WCRF World Cancer Research Fund
WHO World Health Organization
WHR waist-to-hip ratio
27
1. BODY FATNESS
1.1 Background and definitions 1.1.2 Cultural definitions of obesity

1.1.1 Scientific definitions of obesity Perceptions of overweight and obesity can
vary across different settings and populations.
Obesity is the abnormal or excessive accu- People who are overweight or obese may be
mulation of body fat that results from energy admired or may be stigmatized, depending on
imbalance, i.e. energy intake exceeding energy contextual, cultural, spiritual, and other relevant
expenditure, and presents a risk to health. factors.
Obesity is both a condition and an important It is well documented that in many popul-
risk factor for other noncommunicable diseases, ations, overweight and obese people are discrim-
including diabetes, cardiovascular disease, and inated against and stigmatized because of their
many types of cancer. physical appearance – in interpersonal settings,
Overweight and obesity in humans are often in the workplace, and in health-care settings
classified by the body mass index (BMI), which (Link & Phelan, 2001; Puhl & Brownell, 2001;
is obtained by dividing the body weight (in kilo- Sikorski et al., 2011; Spahlholz et al., 2016).
grams) by the square of the height (in metres). Stigmatization of children is also common,
In adults, overweight is defined as BMI leading to rejection and harassment, especially
≥ 25 kg ⁄m2 and obesity as BMI ≥ 30 kg ⁄m2 (WHO, in educational settings (Puhl & Brownell, 2001).
2000). Obesity can be further classified, by level In contrast, in cultures where overweight
of severity and the corresponding different and obesity in children are viewed favourably,
medical approaches for treatment, into class I an obese child may be seen as having a healthy
(30–34.9 kg ⁄m2), class II (35–39.9 kg ⁄m2), and body status (Peña et al., 2012) and as a reflection
class III (≥ 40 kg ⁄m2) obesity (Table 1.1). In chil- of good parenting (Brown, 1991). Favourable
dren younger than 5 years, overweight is defined perceptions of increased body weight can also
as a weight-for-height more than 2 standard lead to inaccurate interpretations of a “healthy”
deviations (SD) above the WHO Child Growth body weight when parents assess the weight of
Standards median, and obesity is often defined their children. In a systematic review based on 13
as a weight-for-height more than 3 SD above the studies, 13% to 100% of the parents interviewed
WHO Child Growth Standards median. In chil- underestimated the obesity status of their child,
dren and adolescents from age 5 years to younger and in six studies 70% or more of the parents
than 19 years, overweight and obesity are defined were unable to identify their child as overweight
as a BMI-for-age more than 1 SD and more than (Tompkins et al., 2015). In cultures that favour
2 SD, respectively, above the WHO Growth larger body size, women of reproductive age
Reference median (WHO, 2016, 2017a, b). who are overweight may be perceived as having
29
Table 1.1 Overweight and obesity cut-off values
Population Measure Overweight Obese Obese I Obese II Obese III

Adults BMI (kg/m )2 ≥ 25.0 ≥ 30.0 30.0–34.9 35.0–39.9 ≥ 40.0
Children
< 5 years Weight-for-height > 2 SDa > 3 SDa – – –
5–19 years BMI-for-age > 1 SDb > 2 SDb – – –
aAbove the WHO Child Growth Standards median
bAbove the WHO Growth Reference median
BMI, body mass index; SD, standard deviation.
Sources: WHO (2000, 2016, 2017a, b)
been well cared for and, by extension, will care 1.1.3 Body fatness as a public health problem
for their children in a similar manner. In such
cultures, increased body weight can be linked to Until recently, obesity was perceived as being
favourable qualities such as self-worth, health, an issue of an individual’s behaviour. As a result,
prosperity, and maternity. For men, overweight interventions for the prevention and treatment
and obesity can symbolize (in addition to good of obesity were focused on the individual level
health) economic success, political power, or (Caballero, 2007). This perception has changed
social status, and in some societies the power of with the increasing awareness of the influence
the elders or leaders (Brown, 1991). of external environmental factors on obesity
Positive perceptions of large body size in (e.g. the built environment, the marketing of
cancer survivors have been reported among food and beverages), and obesity is now seen as
cancer care clinicians. Until the past few a health problem that demands a social response
decades, cancer was frequently diagnosed at (Opalinski, 2013); this is recognized both by the
a late stage of the disease, in which weight loss public and by health-care professionals (Obesity
was a major diagnostic sign. However, with early Society, 2014; see Sections 1.3.6 and 1.3.7).
detection programmes, many cancer survivors
are overweight and obese and suffer obesity-re- 1.2 Prevalence and trends
lated comorbidities (e.g. diabetes). In a weight-
loss feasibility trial in overweight patients with 1.2.1 Prevalence and trends in adults
colorectal cancer, Anderson et al. (2010) reported Worldwide, in 2014 more than 640 million
that clinicians describe avoidance of weight loss adults (14% of adults) were obese, 6 times the
as desirable and express concern about reported number in 1975; of those, more than 18% lived
weight loss (even intentional weight loss). In in high-income English-speaking countries, and
a study of colorectal clinicians in the United 13.9% lived in the Middle East and North Africa
Kingdom and Ireland, Anderson et al. (2013) (NCD Risk Factor Collaboration, 2016).
reported that current opinion and practice are From 1975 to 2014, the average weight of
influenced by the lack of evidence for the impact the population increased in all world regions;
of weight management on health, and by a belief during those four decades, the global prevalence
that weight gain is good and weight loss is bad in of underweight decreased and the global prev-
the cancer setting. alence of obesity increased, so that the number
of obese adults surpassed the number of under-
weight adults (NCD Risk Factor Collaboration,
30
Absence of excess body fatness
2016). The global average BMI in the adult students in 22 countries in different regions of
population (≥ 18 years) in 1975 was 21.7 kg/m2 the world suggested that the prevalence of trying
in men and 22.1 kg/m2 in women, and by 2014 to lose weight increases with levels of BMI, and is
these averages had increased to 24.2 kg/m2 in higher in women than in men at any given BMI
men and 24.4 kg/m2 in women, according to level (Wardle et al., 2006). Similarly, figures from
the latest available estimates (NCD Risk Factor a survey of European Union citizens aged 15 years
Collaboration, 2016). This means that between and older showed that 34% of people reported
1975 and 2014, the average weight of a man with a changing their eating and drinking behaviour to
height of 170 cm increased by about 7 kg, and that lose weight, a percentage that increased to 48%
of a woman with a height of 160 cm increased by among those who perceived their weight as being
about 6 kg. This general increase in the average too high (European Commission, 2006).
BMI affected both high-income countries and With earlier diagnosis and improved cancer
low- and middle-income countries. treatments, many cancer survivors have similar
The overall increase in the average BMI BMI levels to those of the general population; in
corresponded to a general increase in the preva- addition, some cancer treatments may induce
lence of obesity. In 2014, the global prevalence of weight gain. Data on BMI distribution by cancer
obesity in the adult population was 10.8% in men diagnosis are not routinely reported. It was
and 14.9% in women, ranging from less than found that 47.3% of patients with breast cancer
1% (Burundi) to almost 50% (Cook Islands and had a BMI greater than 25 kg/m2 (Nichols et al.,
French Polynesia) in men, and from less than 2009), and in a cohort of colorectal cancer survi-
3% (Timor-Leste and Japan) to more than 58% vors, 29.1% were obese (Rohan et al., 2015). Gross
(American Samoa) in women (NCD Risk Factor et al. (2015) reported that breast cancer survivors
Collaboration, 2016; Fig. 1.1). In 2014, the preva- gained weight at a higher rate than their cancer-
lence of obesity was lowest in low-income coun- free peers. The implications of high body mass
tries, whereas lower-middle-, upper-middle-, for treatment dosing, subsequent morbidity,
and high-income countries were characterized and recurrence are not fully understood (see
by a high level of heterogeneity; in countries in Section 2.4).
any of these three income categories, the highest
prevalence of obesity was more than 40% in men 1.2.2 Prevalence and trends in children and
and about 55% in women. In general, the prev- adolescents
alence of obesity was higher in women than in
men, but trends over the four decades suggested Recent estimates have shown a rapid rise in
a greater increase in the prevalence of obesity in the prevalence of overweight and obesity in chil-
men. However, it should be noted that despite the dren and adolescents worldwide; however, trend
worldwide increase in the prevalence of obesity analyses in developed countries have shown a
and the fact that the number of obese adults is tendency for the prevalence to stabilize starting
now higher that the number of underweight from mid-2000 (Ng et al., 2014; UNICEF, WHO,
adults, the proportion of the population that is and World Bank Group, 2015). The long-term
underweight is still very high (> 20%) in some impact of obesity during childhood and adoles-
countries (India and Bangladesh) (NCD Risk cence is a higher risk of obesity during adulthood
Factor Collaboration, 2016). (Guo & Chumlea, 1999; Freedman et al., 2005;
In light of the obesity epidemic, there is an Singh et al., 2008).
increasing interest in understanding weight-loss The WHO European Childhood Obesity
strategies in the population. A study of university Surveillance Initiative (COSI), established in
31
Fig. 1.1 Prevalence of obesity (BMI ≥ 30 kg/m2) by sex and country in 2014
32

Reproduced from NCD Risk Factor Collaboration (2016). Creative Commons Attribution License (CC BY).
33
2007 to monitor changes in the prevalence According to estimates from the United
of overweight (including obesity) in primary Nations Children’s Fund (UNICEF), WHO, and
schoolchildren, reported that in 2009–2010 the the World Bank Group, the number of over-
prevalence of overweight ranged from 18% to 57% weight children younger than 5 years has been
in boys and from 18% to 50% in girls; 6–31% of increasing steadily everywhere in the world;
boys and 5–21% of girls were obese (Wijnhoven in 2014, almost 50% of all overweight children
et al., 2014). A recent position statement from the younger than 5 years lived in Asia, and about
European Association for the Study of Obesity 25% lived in Africa (UNICEF, WHO, and World
(EASO) Childhood Obesity Task Force (COTF) Bank Group, 2015).
classified obesity as a chronic disease in children Similarly to findings for adult cancer survi-
and adolescents (Farpour-Lambert et al., 2015). vors, a study of paediatric cancer survivors
In 2013, about 110 million children and reported that boys aged 6–11 years were more
adolescents worldwide aged 2–19 years (preva- likely to be overweight than the general popul-
lence of 4.7%) were obese (IHME, 2014), almost ation (Nathan et al., 2006).
twice the number in 1980. In 2013, 24% of boys and
22% of girls living in high-income countries were 1.2.3 Prevalence by age and sex
overweight or obese, whereas the corresponding
percentages observed in low- and middle-income Several studies in the USA and Europe,
countries were about 13% for boys and girls (Ng using cross-sectional data, have shown that BMI
et al., 2014). The highest prevalence of obesity in increases with age up to the sixth decade of life
children and adolescents was observed in North and then starts to decrease (Flegal et al., 2002;
America (Canada and the USA) for both boys Vasan et al., 2005; Ogden et al., 2006). However,
(12.1%) and girls (13.0%). The next-highest prev- results from cohort studies suggest a more
alences were seen in southern Latin America for modest decline in BMI at older ages (Grinker
boys (~10%) and in North Africa and the Middle et al., 1995). This difference in findings may be
East for girls (~10%). The lowest prevalence was due to higher mortality rates at younger ages in
observed in South Asia for both sexes. In 2013, the obese population, leading to a lower prev-
the two countries with the highest prevalence alence of obesity in the surviving older popul-
of obesity for both boys and girls were Kiribati ation. Ageing is also associated with a change
and Samoa, where 20% of boys and 33% of girls in body composition, with a decline in fat-free
were obese; similar prevalences were observed mass, which may cause a decrease in weight, and
in Qatar for boys and in the Federated States of therefore in the measured obesity (Villareal et al.,
Micronesia for girls, whereas prevalences were 2005). [The validity of assessing obesity by BMI is
less than 1.5% for boys and girls in Bangladesh limited in elderly people (see Section 1.4.2).]
and the Republic of Korea. The prevalence of obesity and its association
Independently of the level of income, sex with age vary widely across countries, showing
differences in the prevalence and trends of over- both similarities and differences regardless of
weight and obesity were small (Ng et al., 2014). the socioeconomic level of the country (IHME,
Nevertheless, the prevalence of obesity was 2014). For example, low prevalences of obesity
higher in boys than in girls in 70% of low-income are observed at all ages in China, Ethiopia, India,
countries and in 63% of high-income countries; and Japan (Fig. 1.2), whereas in Colombia, France,
the corresponding percentage in lower-middle- the Islamic Republic of Iran, New Zealand, the
and upper-middle-income countries was 33% Russian Federation, and the USA the prevalence
(IHME, 2014; Ng et al., 2014).
34
Fig. 1.2 Prevalence of obese population by age and sex in selected countries
China Colombia Ethiopia
40
● ●
30 ●
●
●
● ●
20 ●
● ●
● ●
10 ● ● ● ● ●
●
● ● ●
● ● ● ● ●
●
0
● ●
Prevalence of obese population (BMI ≥ 30 kg/m2)
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
France India Iran
40 ●
● ● ●
●
●
30 ● ●
●
● ● ● ● ● ●
● ●
20 ● ●
● ● ●
●
●
10 ●
●
●
● ● ● ● ● ● ● ●
● ● ● ● ● ● ●
●
0 ● ●
Japan Malawi New Zealand
40 ●
● ●
● ●
30 ● ●
●
●
● ●
●
20 ●
10 ● ● ●
● ● ● ● ● ●
● ●
●
● ● ● ● ●
● ● ● ● ● ● ● ● ●
0
● ● ● ● ● ●
Russia United States Global

● ● ●
●
40 ●
●
● ●
●
●
●
● ● ● ● ●
●
30 ● ●
●
● ● ●
20 ● ● ● ● ●
●
●
●
● ● ●
● ●
10
● ●
● ●
●
● ● ● ●
0
10 to 14 yrs
15 to 19 yrs
20 to 24 yrs
25 to 29 yrs
30 to 34 yrs
35 to 39 yrs
40 to 44 yrs
45 to 49 yrs
50 to 54 yrs
55 to 59 yrs
60 to 64 yrs
65 to 69 yrs
70 to 74 yrs
75 to 79 yrs
10 to 14 yrs
15 to 19 yrs
20 to 24 yrs
25 to 29 yrs
30 to 34 yrs
35 to 39 yrs
40 to 44 yrs
45 to 49 yrs
50 to 54 yrs
55 to 59 yrs
60 to 64 yrs
65 to 69 yrs
70 to 74 yrs
75 to 79 yrs
10 to 14 yrs
15 to 19 yrs
20 to 24 yrs
25 to 29 yrs
30 to 34 yrs
35 to 39 yrs
40 to 44 yrs
45 to 49 yrs
50 to 54 yrs
55 to 59 yrs
60 to 64 yrs
65 to 69 yrs
70 to 74 yrs
75 to 79 yrs
80+ yrs
80+ yrs
80+ yrs
Age
sex ● female male
From IHME (2014). Overweight and Obesity Viz. Seattle (WA), USA: Institute for Health Metrics and Evaluation. © 2017 University of
Washington. Available from: http://www.healthdata.org/data-visualization/overweight-and-obesity-viz.
increases with age, followed by a decrease later in China, Ethiopia, France, India, Japan, New
in life. Zealand, and the USA.
The difference between the prevalence of
obesity in men and in women is only weakly 1.2.4 Prevalence by ethnicity
associated with the level of socioeconomic devel-
opment of the country. The prevalence of obesity Several studies have examined the associ-
is markedly higher in women than in men in ation between ethnicity and obesity, mainly in
Colombia, the Islamic Republic of Iran, and the high-income countries with high levels of immi-
Russian Federation, whereas a smaller differ- gration (e.g. the United Kingdom and the USA),
ence or no difference is observed, for example, but very little evidence is available for low- and
middle-income countries.
35
Studies in the USA have clearly shown circumference, were higher among the Pakistani
differences in health between different ethnic and Sri Lankan groups, and these differences
groups (Murray et al., 2005; Caprio et al., 2008; persisted after adjustment for sociodemographic,
Ogden et al., 2014). The prevalence of obesity biological, and lifestyle factors.
in the adult population (≥ 20 years) was almost As a way of avoiding biases when comparing
50% in non-Hispanic Blacks (37.1% in men and data, it has been suggested to define different
56.6% in women), more than 42% in Hispanics obesity cut-off values when comparison between
(40.1% in men and 44.4% in women), about 32% ethnicities is performed (see Section 1.4.2(ii)).
in non-Hispanic Whites, and almost 11% in
non-Hispanic Asians (Ogden et al., 2014). The 1.2.5 Prevalence by social class and
prevalence of obesity in children and adolescents education
(2–19 years) was 22.4% in Hispanics, 20.2% in
non-Hispanic Blacks, 14.1% in non-Hispanic The association between socioeconomic
Whites, and 8.6% in non-Hispanic Asians (Ogden status and prevalence of overweight and obesity
et al., 2014). Similarly, important ethnic differ- changes according to the level of economic devel-
ences have been observed in England (National opment of the country (Dinsa et al., 2012). Lower
Obesity Observatory, 2011). According to data socioeconomic status is protective against obesity
from the Health Survey for England 2004, in in low-income countries, whereas it is a risk factor
women the prevalence of obesity was highest for for obesity in middle- and high-income coun-
Black African women (38%), followed by Black tries, mostly in women (Monteiro et al., 2004a,
Caribbean (32%) and Pakistani (28%) women, b; Dinsa et al., 2012). Similar patterns have been
and lowest in Chinese women (6%). The largest observed in children. Studies in high-income
gap between the sexes was observed among countries have shown an inverse association
the Black African, Pakistani, and Bangladeshi between socioeconomic status and child obesity
groups, in which the prevalence of obesity in (Lamerz et al., 2005; Shrewsbury & Wardle,
women was 1.9–2.8 times that in men. In addi- 2008; Due et al., 2009). For example, data from
tion, data from the National Child Measurement the National Child Measurement Programme in
Programme in England in the 2012–2013 school England in the 2012–2013 school year showed
year compared the prevalence of obesity in a positive association [increasing deprivation
children aged 4–5 years and 10–11 years. The leads to increasing prevalence] between level of
prevalence of obesity was highest among “Asian deprivation and prevalence of obesity in children
or Asian British”, “Black or Black British”, and aged 4–5 years and 10–11 years (Fig. 1.3). Using
“Other than White, Chinese, and Mixed” chil- the deciles of the Index of Multiple Deprivation
dren, with levels of 20–30% (Health and Social 2010 (Department for Communities and Local
Care Information Centre, 2013). [These differ- Government, 2011), children in the most deprived
ences may be partly due to differences in socio- decile had a prevalence of obesity twice that
economic status.] observed in the least deprived decile, with a prev-
A study of the ethnic differences in obesity alence of about 10% in children aged 4–5 years
among immigrants to Norway from low- and and about 20% in children aged 10–11 years
middle-income countries showed large vari- (National Obesity Observatory, 2012; Health
ability in general adiposity among different and Social Care Information Centre, 2013). In
ethnicities (Kumar et al., 2006). Whereas 50% contrast, in low-income countries the associa-
of Turkish women were obese, the levels of tion is inverted, and children in affluent groups
other anthropometric indicators, such as waist
36
Fig. 1.3 Prevalence of obesity by age and decile (e.g. palatability and taste) and opportunities
of Index of Multiple Deprivation in England for high energy intakes (hedonic response)
(Blundell & Finlayson, 2004). In addition, a
range of psychological factors can affect appe-
tite control, including stress, mood, and guilt.
Polymorphisms in the fat mass and obesity-
associated gene (FTO) are associated with
increased food intake, which is thought to be due
to the effect of the gene on ghrelin and its role
in appetite control (Eissing, 2013; Benedict et al.,
2014). Thus, it seems that alterations in appetite
control and subsequent food intake (rather than
metabolic effects) are important factors for posi-
tive energy balance in humans.
The hypothalamus is the main regulatory
organ for human appetite; peripheral signals from
adipose tissue and the gastrointestinal system are
delivered to the hypothalamus to influence short-
term food intake (Lean & Malkova, 2015) (see
Adapted from National Obesity Observatory (2012). NOO data
factsheet: child obesity and socioeconomic status. Available from: also Sections 4.3.4a and 4.3.6b). Of the several
http://webarchive.nationalarchives.gov.uk/20160805121933/http://
www.noo.org.uk/uploads/doc/vid_16967_ChildSocioeconSep2012.
adipokines produced by white adipose tissue,
pdf. only leptin appears to have a significant effect on
appetite suppression (Blundell et al., 2015a, b);
are more likely to be obese (Dinsa et al., 2012; in contrast, more than 30 gut-derived peptides,
Lissner et al., 2016). including hormones and neuropeptides, are
As is the case for the socioeconomic level, BMI known to affect appetite (Lean & Malkova,
has been shown to be positively associated with 2015). Meal ingestion results in production of
education in lower-middle-income countries hormones from enteroendocrine cells found at
in women, and for men also in middle-income multiple sites in the digestive system (stomach,
countries, whereas in high-income countries an proximal/distal small intestine, pancreas, and
inverse association was observed between level of colon). Gut hormones inhibit or stimulate food
education and BMI in women and no association intake (anorexigenic or orexigenic effects). Key
was seen in men (Di Cesare et al., 2013; Fig. 1.4; anorexigenic gut peptides include glucagon-like
see also Section 1.3.7). peptide-1 (GLP-1) and peptide YY (PYY), whereas
ghrelin is a gut hormone with potent orexigenic
effects. Interest in the effect of bariatric surgery
1.3 Risk factors on long-term appetite control has highlighted
1.3.1 Regulation of hunger and satiety the role of gut hormones (Chakravartty et al.,
2015). Lean & Malkova (2015) reported that
The current epidemic of obesity suggests diet-induced weight loss affects gut hormones
that the processes of satiety and body weight and favours increased appetite (and therefore
control are not tightly regulated, and physiolog- weight regain), whereas physical exercise favours
ical systems appear to permit fat storage when enhanced satiety (supporting maintenance of
humans are exposed to sensory stimulation weight loss). They postulated that the sustained
37
Fig. 1.4 Mean BMI by education level in adults aged 25–64 years in selected countries
Cambodia Colombia Iran Malawi South Korea USA

30·0
27·5
25·0
Men
22·5
Mean BMI (kg/m2)
20·0
0
30·0
27·5
Women
25·0
22·5
20·0
0
Lowest Highest Lowest Highest Lowest Highest Lowest Highest Lowest Highest Lowest Highest
Education level
Reprinted from Di Cesare et al. (2013). Inequalities in non-communicable diseases and effective responses. Lancet, 381(9866):585–97. Copyright
2013, with permission from Elsevier.
weight loss achieved by bariatric surgery may in Maternal obesity is associated with both
part be mediated via favourable changes to gut large-for-gestational-age babies and macrosomia
hormones. (birth weight > 90th percentile). It has been
There has been much speculation about the reported that in women with a pre-pregnancy
concept of a physiological “set point” that is auto- BMI greater than 30 kg/m2, the likelihood of
matically adjusted to maintain weight. Recent having a large-for-gestational-age baby is
evidence suggests that structural changes in the 3.75 times that in women with a BMI less than
hypothalamus may lead to a “reset”, and thus 30 kg/m2, and that large-for-gestational-age babies
maintain an increased body weight. have both higher insulin resistance and higher
free fatty acid levels (Liu et al., 2013). There is
1.3.2 Weight gain throughout the life-course global concern about the number of women in
developing countries who enter pregnancy with
The association of noncommunicable diseases excess body weight (WHO, 2011; Torloni et al.,
in adulthood with nutrient status during preg- 2012).
nancy was proposed in 1990 (Barker, 1990), and Systematic reviews of studies mainly in
there is substantial evidence supporting an early- developed countries have demonstrated a posi-
life origin of susceptibility to obesity in later life tive association between birth weight and BMI
(Scientific Advisory Committee on Nutrition, in later life. The association has been described
2011). Both prenatal and postnatal factors, influ- as strongest in children and young adults and
encing fetal and infant growth, development, weaker in middle-aged adults. It has been estim-
birth weight, and programming of metabolism, ated that a 1 kg increment in birth weight is asso-
have been implicated in the later development ciated with a rise of 0.5–0.7 kg/m2 in the BMI of
of obesity. Overall, it is difficult to disentangle young adults (Scientific Advisory Committee on
the impact of single early-life risk factors in the Nutrition, 2011). Some studies suggest that both
development of adult obesity. low and high birth weight are associated with
subsequent risk of obesity in children and young
38
adults; however, a meta-analysis of 66 studies most obese adults were not obese in childhood
from 26 countries shows that low birth weight (Simmonds et al., 2015). Observational studies in
(< 2500 g) is followed by a decreasing long-term industrialized countries have shown a consistent
risk of overweight, whereas high birth weight association of poor early growth followed by rapid
predisposes to being overweight in later life catch-up in infancy with later obesity in child-
(Schellong et al., 2012). hood and adulthood (Monasta et al., 2010). An
The role of maternal diet (both quantity early adiposity rebound (a period of increasing
and quality) during pregnancy is of particular BMI after the early childhood nadir), which
interest with respect to fetal developmental usually occurs at ages 4–8 years, is associated
programming (Liu et al., 2013), and it is plausible with high body weight in later life (Williams &
that epigenetic mechanisms are responsible for Goulding, 2009; Brisbois et al., 2012). Nutritional
the associations between birth weight and risk of stunting (usually caused by chronic undernutri-
overweight later in life (Dominguez-Salas et al., tion) during childhood is also positively asso-
2014; Tobi et al., 2014). ciated with overweight and obesity in later life
Weight trajectories during infancy are influ- (Popkin et al., 1996; Black et al., 2013).
enced by breastfeeding and complementary Several studies have reported an association
feeding practices, which in turn influence body between parental smoking and later obesity
fatness in childhood and growth patterns (Poskitt in the offspring, with the greatest effect from
& Breda, 2012; Bagci Bosi et al., 2016). There is maternal smoking (Power et al., 2010; Dior et al.,
growing evidence from recent reviews of obser- 2014; Wang et al., 2014; Han et al., 2015). Prenatal
vational studies that breastfeeding is associated maternal exposure to smoking (including
with a decreased risk of becoming overweight secondhand tobacco smoke) was associated with
later in life (Weng et al., 2012; Horta et al., 2015). increased risk of obesity in adolescents, inde-
Horta et al. (2015) reported an odds ratio (OR) pendent of birth weight (Wang et al., 2014).
of 0.74 (95% confidence interval [CI], 0.70–0.78) Throughout adult life, there are further
that people who were breastfed as babies were periods or key life events associated with weight
less likely to be classified as overweight or obese gain, including student life, the transition from
at ages 1–9 years, 10–19 years, and 20 years and single status to marriage or cohabitation, diet
older. and activity behaviours during pregnancy,
Feeding choices during infancy and child- postpartum weight retention, the demands of
hood are influenced by parental factors, food employment, unemployment and job loss, the
availability, and other socioeconomic factors, demands of parenting, and physiological changes
including cultural norms, peer effects, and the associated with the ageing process (including
marketing of food and beverages (Boyland et al., menopause) (WHO Regional Office for Europe,
2016). However, current research findings on 2016). Maternal weight gained during pregnancy
the type of complementary foods and the age at is also associated with the likelihood of weight
which complementary foods are introduced into retention in the mother in the postpartum
the infant diet and subsequent body fatness are period and beyond. It is estimated that in the
inconsistent (Langley-Evans, 2015). USA 50–67% of women gain more weight during
The development of overweight during child- pregnancy than is considered desirable by the
hood appears to track into adult life (Brisbois United States Institute of Medicine (now known
et al., 2012); however, a systematic review showed as the National Academy of Medicine) (Liu et al.,
that childhood BMI is not a good predictor of the 2013). It is likely that the metabolic load of exclu-
incidence of adult obesity or overweight, because sive breastfeeding (~500 kcal per day) can help
39
women in losing the weight gained during preg- (b) Macronutrient composition of diet
nancy, although results have been inconsistent (i) Total fat
(Bobrow et al., 2013; Jarlenski et al., 2014; Neville
et al., 2014; Palmer et al., 2015). There has been a lack of consensus about
Physiological changes associated with whether the macronutrient composition of the
the ageing process influence weight gain (see diet is a significant determinant of body fatness
Section 1.2.3). Comorbidities also have the poten- (Bray & Popkin, 1998; Willett, 2002). Two system-
tial to increase weight gain, for example through atic reviews of observational studies found no
the action of different treatment modalities significant evidence for an association between
(Leslie et al., 2007; Rummel-Kluge et al., 2010; percentage of energy from fat and body weight
McPheeters et al., 2011; Almandil et al., 2013). (Fogelholm et al., 2012; Hooper et al., 2015),
Unintentional weight loss has also been reported whereas two systematic reviews with meta-ana-
as a side-effect of several therapeutic agents, lyses of randomized controlled trials (RCTs) in
and this is usually dose-dependent (Leslie et al., people with no intention of losing weight suggest
2007). that the percentage of energy from fat is positively
associated with body weight (Hooper et al., 2015;
Tobias et al., 2015). However, in weight-loss trials,
1.3.3 Excessive energy intake no difference in weight loss was observed between
(a) Total energy intake diets with different percentages of energy from
Obesity is determined primarily by increased fat, whereas diets with a low percentage of energy
total energy intake (Vandevijvere et al., 2015), from carbohydrates led to greater weight loss
which itself is influenced by portion size, the than those with a low percentage of energy from
frequency of consumption, and the energy fat (Tobias et al., 2015). A third meta-analysis of
density of food and drinks consumed (Ello- RCTs reported that diets with a low percentage of
Martin et al., 2005). Assessing total energy energy from fat and those with a low percentage
intake is challenging in epidemiological studies, of energy from carbohydrates resulted in compa-
because of measurement error, including selective rable reductions in body weight (Hu et al., 2012).
underreporting by overweight and obese people [It is recognized that the success of any dietary
(Livingstone & Black, 2003) and the inability to restriction regimen in achieving weight loss is
quantify energy provided by the colonic micro- dependent on compliance with that regimen.]
biome (Rahat-Rozenbloom et al., 2014). Because
(ii) Sugars
objective measures of energy intake do not exist
(de Jonge et al., 2007), the best long-term indi- Two systematic reviews and meta-analyses of
cator (or life-course indicator) of positive energy RCTs in adults found a significant positive asso-
balance is weight gain (Hill et al., 2012). Weight ciation between intake of free sugars (defined as
gain cannot differentiate between the proportions monosaccharides and disaccharides added to
of fat and lean tissues gained, but in most adults it foods and beverages by the manufacturer, cook,
is primarily determined by gain in adipose tissue or consumer, and sugars naturally present in
(Okorodudu et al., 2010). The amount of adipose honey, syrups, fruit juices, and fruit juice concen-
tissue gained over time depends on the degree trates; WHO, 2015) and weight gain (Te Morenga
of positive energy balance, the duration of the et al., 2012) and energy intake (Scientific Advisory
energy surplus, and the body composition before Committee on Nutrition, 2015), respectively.
weight gain (Schutz et al., 2014). Overconsumption of sugar-sweetened bever-
ages in particular has been linked to weight
40
gain and obesity. Reviews and meta-analyses of (d) Fast food

prospective cohort studies and RCTs indicate Fast foods are energy-dense, micronutrient-
that consumption of sugar-sweetened bever- poor foods that are often high in fatty acids,
ages promotes weight gain and obesity (Hu & processed starches, and added sugars (Jaworowska
Malik, 2010; Te Morenga et al., 2012; Malik et al., et al., 2013); they have more recently been termed
2013; Fardet & Boirie, 2014; Scientific Advisory “ultra-processed foods” (Martínez Steele et al.,
Committee on Nutrition, 2015). In contrast, 2016). There is concern about the rising global
there are investigations that consistently show trend in consumption of fast food and its poten-
weak or null results for the association between tial impact on obesity (McCrory et al., 1999;
consumption of sugar-sweetened beverages and Paeratakul et al., 2003; Louzada et al., 2015).
obesity (Forshee et al., 2008; Gibson, 2008; Kaiser Most available studies are limited by their
et al., 2013; Trumbo & Rivers, 2014). [Concern cross-sectional designs, non-standardized defi-
has been expressed about potential bias of these nitions of fast food, and potential confounding
studies, because of sources of funding.] by unmeasured factors. A recent narrative review
(c) Dietary patterns reported that frequent consumption of fast foods
is accompanied by weight gain and obesity
Studies using dietary patterns as derived by (Bahadoran et al., 2016). In one observational
factor or cluster analysis to examine common cohort study, frequency of consumption of fast
eating behaviour in relation to weight control and food was positively related to changes in body
obesity found that diets characterized by high weight. Specifically, more frequent consumers
intakes of sugar-sweetened beverages, refined (more than twice per week) gained an extra 4.5 kg
grains, potatoes, and red and processed meat of body weight over a 15-year period compared
were related to greater gains in BMI and waist with less frequent consumers (less than once per
circumference, whereas diets typified by high week) of fast-food products (Pereira et al., 2005).
intakes of fruit, vegetables, whole grains, fish,
poultry, and reduced-fat dairy products were (e) Alcoholic beverages
associated with smaller weight gains (Newby Findings on the association between
et al., 2003; Schulze et al., 2006). Similarly, the consumption of alcoholic beverages and body
traditional Mediterranean diet – as character- weight have been inconsistent, with observa-
ized by high intakes of fruit, vegetables, legumes, tional studies showing positive, inverse, or null
nuts/seeds, and wholegrain cereals; regular associations (reviewed in Poppitt, 2015). In addi-
consumption of fish and seafood; moderate tion, evidence linking alcohol consumption to
intakes of dairy products, poultry, and eggs; low weight gain and obesity was ranked insufficient
consumption of red and processed meat; use of by international reports (WHO, 2003; WCRF/
olive oil as the main source of dietary fat; and AICR, 2007).
moderate intake of wine – has been consistently
found to be inversely related to the development (f) Sleep
of weight gain and obesity (Buckland et al., 2008;
The relationship of sleep duration to adiposity
García-Fernández et al., 2014; Gotsis et al., 2015).
has been examined in numerous epidemiological
It has been suggested that eating patterns, such as
studies. Results from studies in adults have been
eating breakfast, frequency of eating, snacking,
mixed, but most investigations support an asso-
and timing of meals, may also modulate the risk
ciation of short sleep duration with adiposity,
of overweight and obesity (van der Heijden et al.,
with the association appearing to wane with
2007; McCrory, 2014; Jiang & Turek, 2017).
41
increasing age (Marshall et al., 2008; Patel & 1.3.4 Endocrine disruptors
Hu, 2008). By comparison, studies in children
showed a clear pattern, with short sleep duration An endocrine-disrupting chemical (EDC) is
being positively related to obesity (Chen et al., an exogenous chemical, or mixture of chemicals,
2008; Marshall et al., 2008; Patel & Hu, 2008). that interferes with any aspect of the regulation
Results from systematic reviews and meta-ana- of hormone action (Zoeller et al., 2012). EDCs
lyses are in line with these findings (Cappuccio act directly by binding or interfering with recep-
et al., 2008; Magee & Hale, 2012; Capers et al., tors, or indirectly by disrupting hormone levels
2015). [One possible explanation for the diver- or by altering hormonal transport mechanisms
gent results between children and adults is that (Heindel et al., 2015). Because endocrine systems
the association between short sleep duration and exhibit tissue-, cell-, and receptor-specific
weight gain diminishes over time after the tran- actions during the life-cycle, EDCs can produce
sition to chronic short sleep duration (Magee & a complex mosaic of effects across the life span
Hale, 2012).] (Zoeller et al., 2012).
Sleep deprivation can lead to enhanced energy Some EDCs, referred to as “obesogens”
intake, possibly through a mechanism involving (Janesick & Blumberg, 2016), improperly inter-
greater propensity to overeat and altered levels of fere with lipid homeostasis, and promote adipo-
hormones involved in appetite regulation, such as genesis by perturbing various endocrine axes
ghrelin and leptin (Bayon et al., 2014; Kim et al., (Burgio et al., 2015). The developmental age at
2015). By comparison, there is little evidence to which EDC exposures occur is a critical consid-
support that insufficient sleep leads to decreased eration in understanding their effects. Exposure
energy expenditure (Chaput & Tremblay, 2012). to obesogenic EDCs during fetal life and early
Light exposure during the night decreases childhood appears to have an impact on obesity
melatonin synthesis (Haus & Smolensky, 2013; during childhood and adulthood (Grün &
McFadden et al., 2014), and melatonin deficiency Blumberg, 2009; Heindel et al., 2015).
has been shown to exert beneficial effects on The EDCs with obesogenic properties that
obesity and to normalize the expression and are most prevalently used are presented in
secretion patterns of adipokines such as leptin Table 1.2 (adapted from Heindel et al., 2015), and
and adiponectin (Szewczyk-Golec et al., 2015). the available data are summarized below. [The
[Limitations of study designs in the area of overall limitations of the data from humans lie
short sleep duration and obesity include a failure in the cross-sectional nature of the majority of
to account for potential confounding, mediating, the studies and in the use of a single measure
and moderating variables, such as sedentary of exposure. The inconsistences in results in the
behaviour, and measurement issues such as a small number of birth cohort studies is likely to
paucity of objective measures of sleep duration, be related to the small sample size, differences
lack of repeated assessments of both sleep dura- in population characteristics, and differences in
tion and body weight, and an insufficient number levels of exposure and timing of exposure.]
of experimental study designs that manipulate (a) Bisphenol A
sleep.]
Studies in animals have shown that expo-
sure to bisphenol A can disrupt many metabolic
pathways and could alter the hypothalamic
energy balance circuitry, resulting in increased
susceptibility to developing diet-induced obesity
42
Table 1.2 Endocrine-disrupting chemicals with obesogenic properties
Chemical Source Potential mechanism

Bisphenol A Plastic and epoxy resins Estrogenic
Inhibition of proliferation of neural progenitor cells
Phthalates Plasticizers, adhesives, and personal care Increase the rate of adipocyte differentiation
products
Tributyltin Fungicide in paints and component of Activation of peroxisome proliferator-activated receptor-γ
polyvinyl chlorides and increased fat cell differentiation
Polychlorinated Coolants, plasticizers, and flame retardants Altered thyroid function
biphenyls Altered metabolism
Bioaccumulation in fat cells
Polycyclic Incomplete combustion of fossil fuels Accumulation of visceral fat
aromatic Inflammation
hydrocarbons
Perfluorinated Components of lubricants, non-stick Increase serum levels of insulin
chemicals coatings, and stain-resistant compounds Increase serum level of leptin
Flame retardants Chemicals applied to furniture and Increase rate of adipogenesis
electronics Increase glucose intolerance
Adapted from Heindel et al. (2015) by permission from Springer Nature.
and metabolic impairment (Heindel et al., 2015). (c) Tributyltin

Some studies reported an association between In animals, prenatal exposure to tributyltin
urinary concentrations of bisphenol A and causes lipid accumulation in adipose tissue, with
increased obesity in children (OR, 2.53; 95% CI, effects persisting into adulthood and into future
1.72–3.74, between the highest and lowest quar- generations (Heindel et al., 2015). A recent study
tile of bisphenol A concentration; Trasande et al., showed a positive association between placenta
2012) and in adults (OR, 1.50; 95% CI, 1.15–1.97, tributyltin levels and weight gain in male infants
between the highest and lowest quartile) (Wang at age 3 months (Rantakokko et al., 2014).
et al., 2012), whereas other studies reported no
association (Lakind et al., 2014). Also, exposure to (d) Polychlorinated biphenyls and other
bisphenol A during fetal life increased adipokine persistent organic pollutants
levels in childhood (Volberg et al., 2013; Ashley-
Data from the National Health and Nutrition
Martin et al., 2014).
Examination Survey 1999–2002, in adults and
(b) Phthalates children, showed an association of detectable
levels of persistent organic pollutants with waist
In animal models, prenatal exposure to circumference and BMI (Elobeid et al., 2010).
phthalates led to increased body weight, increased Cohort studies in children have not confirmed
number and size of adipocytes, and activation the association with weight gain (Cupul-Uicab
of peroxisome proliferator-activated receptor et al., 2013; Tang-Péronard et al., 2015). However,
gamma in male offspring (Heindel et al., 2015). in a birth cohort in the Faroe Islands, higher
Also, cross-sectional studies suggest an associa- prenatal exposure to polychlorinated biphenyls
tion between exposure to phthalates and weight was associated with increased BMI at age 7 years
gain (Kim & Park, 2014). However, data from in daughters of overweight mothers (Tang-
human birth cohorts are inconsistent (Ashley- Péronard et al., 2014). In addition, higher levels of
Martin et al., 2014; Maresca et al., 2015). persistent organic pollutants have been observed
43
in visceral fat of obese subjects with larger meta- studies since the 1980s with measurements of
bolic dysfunction compared with those without energy expenditure by the doubly labelled water
metabolic dysfunction (Pestana et al., 2014). method indicate that levels of physical activity
have not declined over the time period in which
(e) Polycyclic aromatic hydrocarbons the prevalence of obesity has risen (Westerterp
Studies in animals have shown that the combi- & Speakman, 2008). Thus, it seems likely that
nation of exposure to diesel exhaust in utero and the recent rise in the average body weight of
a high-fat diet as an adult increases susceptibility the population globally is determined largely by
to diet-induced obesity and neuroinflammation increased energy intake rather than by decreased
in females but not in males (Bolton et al., 2012). energy expenditure (Swinburn et al., 2009).
In a birth cohort of women who were exposed
to air pollutants during pregnancy, higher expo- (a) Physical activity and prevention of weight
sure levels were associated with higher obesity in gain
their children at ages 5 years and 7 years (Rundle Evidence from long-term observational
et al., 2012). studies (Williamson et al., 1993; Di Pietro et al.,
2004; Gordon-Larsen et al., 2009; Lee et al., 2010;
(f) Other compounds Mozaffarian et al., 2011) suggests a positive rela-
Very limited data in humans are available for tionship between self-reported physical activity
perfluorinated chemicals and flame retardants, and prevention of weight gain. Data from RCTs
and the results are inconclusive (Erkin-Cakmak are sparse and are based on short-term interven-
et al., 2015; Heindel et al., 2015). tions and follow-up (Simkin-Silverman et al.,
2003). Expert consensus has estimated that the
1.3.5 Physical activity and sedentary amount of moderate-intensity physical activity
behaviour associated with prevention of weight gain is
150–250 minutes per week (Donnelly et al., 2009),
Physical activity is considered “any bodily or 45–60 minutes per day (Saris et al., 2003).
movement produced by skeletal muscles that
results in energy expenditure” (Caspersen et al., (b) Physical activity and weight loss
1985). The intensity of physical activity refers to In adults, moderate physical activity shows
the rate of energy expenditure brought about a dose–response association with weight loss.
by physical activity; in epidemiological studies, Based on evidence from observational studies
it is commonly expressed in metabolic equiv- of varying durations (spanning time periods of
alents (METs) of tasks. Specifically, activities several months to several years), the American
are grouped according to light (1.6–2.9 METs), College of Sports Medicine concluded that less
moderate (3.0–5.9 METs), and vigorous than 150 minutes per week results in minimal
(≥ 6 METs) levels of activity (Ainsworth et al., weight loss, more than 150 minutes per week
2011). Several methods for measuring physical yields weight loss of 2–3 kg, and 225–420 minutes
activity are available and have been used in epide- per week leads to weight loss of 5–7.5 kg (Donnelly
miological studies (Ndahimana & Kim, 2017). et al., 2009). Moreover, a positive relationship
Globally, there have been declines in levels was reported, for a similar duration, between the
of physical activity and increases in time spent intensity of the physical activity and the amount
sedentary in adults over the past decades (Ng et al., of weight lost (Shaw et al., 2006). A meta-analysis
2012; Dearth-Wesley et al., 2014). However, exper- of trials with durations of 12 weeks to 12 months
imental evidence from repeated cross-sectional reported that moderate-intensity physical activity
44
is only modestly effective in producing weight involving physical activity only. However, the
loss in overweight and obese adults, showing a combination of physical activity and dietary
small weighted mean difference in body weight energy restriction resulted in a difference of
of −1.6 kg (Thorogood et al., 2011). −1.56 kg in weight regain compared with controls
Physical activity may generate a more at 12 months (Dombrowski et al., 2014). Taken
pronounced effect on weight loss in children together, findings from observational studies and
and adolescents, as suggested by a meta-ana- RCTs show inconsistent results, and the volume
lysis of 20 intervention studies reporting that of physical activity needed to prevent weight
an exercise programme involving approxi- regain after weight loss remains poorly defined.
mately 150 minutes per week of physical activity Data about the influence of resistance training
for 13 weeks led to a 3.6% decrease in BMI in on prevention of weight gain and weight regain
overweight and obese children and adolescents after weight loss are sparse.
(Kelley et al., 2015).
A meta-analysis of 15 RCTs compared the (d) Sedentary behaviour
effects of 2.5–6 months of aerobic and resistance Sedentary behaviour is a behaviour with
exercise on weight loss in overweight and obese reduced light activity and is distinct from lack
subjects and concluded that the combination of of moderate to vigorous activity (Lynch et al.,
aerobic and resistance training may be the most 2010). Time spent in sedentary activities reduces
efficacious physical activity regimen for weight energy expenditure, potentially promoting
loss (Schwingshackl et al., 2013). weight gain over time (Owen et al., 2010). Long-
Physical activity and energy restriction term observational epidemiological studies in
should yield similar amounts of weight loss if adults reported that sedentary behaviour, such
they provide comparable degrees of negative as television viewing and computer gaming, is
energy balance, and they are both important for associated with enhanced risk of adiposity (Hu
achieving optimal weight loss. However, energy et al., 2003; Thomée et al., 2015). Also, television
restriction combined with physical activity viewing is associated with enhanced intake of
results in the combination of loss of fat mass sugar-sweetened beverages and sweets (Lipsky
and maintenance of lean mass, thereby leading & Iannotti, 2012). Moreover, a meta-analysis of
to a more desirable effect on body composition 25 RCTs showed a small but significant decrease
(Miller et al., 2013). in BMI with reducing sedentary behaviour in
children and adolescents (Liao et al., 2014).
(c) Physical activity and prevention of weight [Potential confounding by energy intake could
regain after weight loss not be ruled out.] Data on sedentary behaviour
An early systematic review of RCTs and and prevention of weight regain after weight loss
observational studies reported that people who are sparse.
engaged in physical activity experienced less
weight regain than their sedentary counterparts, 1.3.6 Built environment
but confounding by a healthy lifestyle or reverse
causation by better exercise adherence among The human body has evolved to transform
those with less weight regain could not be ruled excess energy into body fat during exceptional
out (Fogelholm & Kukkonen-Harjula, 2000). In times of food abundance for protection during
contrast, a more recent meta-analysis of RCTs periods of low food availability, but has not
on long-term maintenance of weight loss found adapted to meet the challenge of the supply
no evidence of effectiveness for interventions
45
provided by current food systems (Hill et al., balance (Egger & Swinburn, 1997). The built
1998). environment is thought to play a key role in
The built environment refers to aspects of a promoting both reductions in energy expend-
person’s surroundings that are human-made (or iture and increases in energy intake, but it
modified), compared with naturally occurring also has the potential to promote and sustain
features (Caballero, 2007; Papas et al., 2007). It healthy behaviours for significant periods of
is one part of the so-called obesogenic environ- time throughout the life-course (Ludwig et al.,
ment, which includes sociocultural, economic, 2011). The built environment affects a person’s
and marketing barriers to the achievement of ability to comply with interventions, including
healthy ways of life. educational interventions, that are focused on
Food availability has undergone many individual behaviour change (Sallis et al., 2012).
changes in recent decades, as a result of Most research has focused on the role of the
economic, social, and demographic changes built environment in providing opportunities for
(the nutrition transition) (Brown, 1991; Popkin, physical activity; with respect to the food envi-
1993), resulting in an increased dependence ronment, most work has focused on food availa-
on processed foods, a strong emphasis on the bility within communities.
marketing and promotion of energy-dense Several reviews have examined the relation-
foods and beverages, a decreased intake of fruit ship between the built environment and obesity,
and vegetables, and an increased intake of foods but they have had heterogeneous measurement
that are high in fat and sugar (Caballero, 2007). approaches. The lack of valid and reliable meas-
The food environment has changed as the oppor- urement tools and analytical approaches may
tunities for consumption have increased through have contributed to inconsistent findings (Feng
increased diversification of food outlets (e.g. et al., 2010).
vending and fast-food restaurants) and intensive Metrics of the food environment (relevant for
and creative marketing strategies. In addition, energy intake) include measures of the density,
changes in the built environment, including proximity, and diversity of and spatial access to
shifts in transportation patterns, labour-saving food establishments (supermarkets, convenience
devices, limited public space for recreational stores, farmers’ markets, and restaurants). It is
physical activity, and increased public safety difficult to compare studies, because of metrics
concerns, have promoted sedentary lifestyles used and the overall suboptimal quality of the
(Caballero, 2007; Parizkova et al., 2007) (see also studies (Ding & Gebel, 2012). A review (Cobb
Section 1.3.5). et al., 2015) that assessed the relationship of
It is important to note that although indi- obesity to local food environments reported few
vidual energy balance is strongly predicted by significant findings, although a trend towards
personal behaviours, the factors that influence an inverse association between availability of
these behaviours are a response to the stimulus supermarkets and obesity was noted. Metrics
of the obesogenic environment. Thus, societal that are relevant for physical activity (and thus
change that will have an impact both on public energy expenditure) include measures of popul-
health and on the health of the individual is ation density, diversity, connectivity, design of
required to counter the current obesity epidemic and spatial access to physical activity facilities,
(Mackenbach et al., 2014). walkability, and sprawl (Lopez & Hynes, 2003).
The ecological approach to obesity research Although a consistent relationship has been
embraces the biological, behavioural, and envi- seen between physical activity (in the form of
ronmental factors that have an impact on energy active transport and recreational walking) and
46
neighbourhood walkability in adults, these to be obese than their urban counterparts (even
studies have not always reported measures of after adjustment for individual-level behaviours).
obesity. Where obesity has been reported, there A review by Hansen et al. (2015) suggests that
seems to be a lower than average BMI in neigh- limited “active living” built environments are a
bourhoods with higher perceived mixed land contributory factor.
use, improved walkability, and better access to
sports facilities (Black & Macinko, 2008). In a 1.3.7 Social determinants
review by Papas et al. (2007), significant associa-
tions between some objectively measured aspects Socioeconomic status is an aggregate concept
of the built environment (e.g. residential density, that consists of numerous indicators that reflect
street connectivity, greenery, and access to desti- a person’s position in society, including educa-
nations) and obesity were observed in 84% of tion level, occupation, income, wealth, poverty,
the studies. In a study on environmental attrib- and deprivation. Because levels of BMI vary
utes and adult weight status in 12 countries, De according to levels of economic development,
Bourdeaudhuij et al. (2015) reported that safety sociocultural factors, and characteristics of the
from traffic had the most robust correlation with health-care system of a country, the relationship
BMI. between social determinants and obesity cannot
Several studies have evaluated accessibility to be explained by any individual factor; rather,
recreation or exercise spaces and facilities (such numerous interrelated factors are involved
as parks, playgrounds, cycle routes, and sports (WHO, 2008).
facilities) using adiposity as an outcome. In a It has been widely shown that in high-income
review of cross-sectional studies that directly countries, the prevalence of obesity is higher
measured body weight in adults or children, among people with lower socioeconomic status
two out of three studies in adults observed that (WHO, 2014); this inverse association is driven
shorter distance to (or greater density of) fitness mainly by the fact that socially disadvantaged
facilities was associated with lower BMI and people experience a more obesogenic envi-
lower prevalence of overweight (Papas et al., ronment, which favours access to cheap, ener-
2007). In their review of population approaches gy-dense foods, lower levels of physical activity,
to improving health behaviours, Mozaffarian and poor education. In contrast, in low-income
et al. (2012) concluded that greater access to countries a positive association is seen between
recreation and exercise spaces and facilities is socioeconomic status and the prevalence of
relatively consistently linked to higher levels of obesity (Monteiro et al., 2004a, b; Blakely et al.,
physical activity and lower adiposity or other 2005; Dinsa et al., 2012; Pampel et al., 2012).
metabolic risk factors. However, they noted that Possible reasons for this are food scarcity among
nearly all the evidence is cross-sectional, which the poor in low-income countries, leading to low
limits inferences about causality. or moderate food intake, and greater engage-
In a review by Dunton et al. (2009) of child- ment in manual labour, requiring higher energy
hood obesity, associations between physical envi- expenditure. By comparison, the rich in low-in-
ronmental variables and obesity differed by sex, come countries may be particularly susceptible
age, socioeconomic status, population density, to obesity because of easy access to excess food
and whether reports were made by the parent or and lower levels of occupational physical activity.
the child. Also, in some low-income countries, a large body
Research in the USA shows that children size may be perceived as an indicator of socio-
and adults living in rural areas are more likely economic status (for additional cultural aspects,
47
see also Section 1.1). Inequalities in childhood intake) across the population, which may have
obesity generally mirror those seen in adults differential health effects according to socio-
(Due et al., 2009; Dinsa et al., 2012) (see also economic status (McKinnon et al., 2014). Such
Section 1.2.5). factors include perception of sensory attrib-
The relationship of socioeconomic status utes, such as a liking for sweetness (Deglaire
to obesity appears to be affected by the choice et al., 2015); psychological factors, such as mood
of socioeconomic indicator (income vs educa- (Singh, 2014); mental health and well-being,
tion level) in approximately 20–30% of studies including stress (Moore & Cunningham, 2012);
conducted in low-income countries. This may be and the food marketing environment, notably
due to the weak correlation observed between with marketing of foods high in sugars, fats,
wealth and education level in some low-income and salt being targeted at children (Cairns et al.,
countries where educational investment has not 2013).
yet translated into a higher income (Dinsa et al.,
2012). One literature review of studies from 1.3.8 Microbiota and gastrointestinal
high-income countries examined various indica- environment
tors of socioeconomic status and found inverse
relationships of education level and occupation The adult human gut hosts a complex
to weight gain, with less consistent findings when community of microorganisms (micro-
income was used as the measure of socioeconomic biota), including about 1.5 kg of bacteria,
status (Ball & Crawford, 2005). Another review most of which belong to four major phyla:
of studies conducted in industrialized coun- Firmicutes, Bacteroidetes, Actinobacteria, and
tries reported that women who changed social Proteobacteria. The amount, composition,
class after childhood took on the prevalence of and ratio of these vary according to diet and
adiposity of the class they joined, a relation- energy intake, and it is clear that variations in
ship that was not evident in men (Parsons et al., eating habits can lead to selective promotion of
1999). Data about other social determinants, particular species. It is estimated that 57% of
such as family size, number of parents at home, the bacterial variation in the gut is due to diet,
or availability of childcare and their relationship and there are particular concerns that high-fat,
to obesity are sparse. high-sugar diets may alter the functionality
Education level has been one of the most of the microbiota (Harris et al., 2012). There is
frequently used measures of socioeconomic accumulating evidence to support a role for the
status (Krieger et al., 1997) and has been shown microbiota in obesity and its downstream meta-
to be associated with obesity. This has led to an bolic sequelae.
emphasis on nutrition education and practical For example, the gut microbiota has been
food skills as a route to improving dietary habits. found to be significantly altered in animal
However, recent data from the United Kingdom models of obesity, with a reduction in bacte-
(Adams & White, 2015) show an inconsistent rial diversity as well as a change in bacterial
relationship between sociodemographic vari- composition. It is generally agreed that a lower
ables and all markers of cooking skills, high- diversity of microorganisms and the presence of
lighting the need to explore the wider social more tolerant bacteria are also observed in obese
factors that educational background reflects (see, people. In many studies, the pattern associated
for example, Lawrence & Barker, 2009). with obesity is a relative decrease in Bacteroidetes
It is recognized that numerous factors poten- (gram-negative) and a corresponding relative
tially influence food choice (and thus energy increase in Firmicutes (gram-positive), although
48
there is some inconsistency in results in relation demonstrating weight loss (Graham et al., 2015;
to ratios of different phyla (Sanmiguel et al., Sanmiguel et al., 2015). It seems likely that in
2015). Reduction in body weight results in a shift humans, differences in the gut microbiota accom-
in the composition of the gut microbiota, with a pany obesity rather than causing the problem.
significant relative increase in Bacteroidetes and Further research in experimental systems and in
relative decrease in Firmicutes as weight loss humans is needed to understand the link between
progresses. Exercise alone is also associated with the microbiome and obesity and whether varia-
an alteration in the ratio of the major bacteria, tion in the microbiome is a direct cause of obesity
with a higher proportion of butyrate-producing or is a consequence of it.
bacteria (Bacteroidetes) (Harris et al., 2012).
Both animals and humans experience major 1.3.9 Genetic and epigenetic determinants of
changes in the gut microbiota after gastric bypass, body fatness
when there is a restoration to a normal ratio of
Firmicutes to Bacteroidetes. It is postulated that There is growing evidence that gene–envi-
this occurs as a result of major changes in the ronment interactions play a major role in obesity
composition of the diet after surgery, a decrease (Reddon et al., 2016). A hereditary component
in production of bile acid, and a raised pH, which to obesity has long been recognized, but the
encourages the growth of bacteria associated with underlying genetic variants have only recently
weight loss (Harris et al., 2012; Sanmiguel et al., been identified. The first single nucleotide poly-
2015). The changes in the microbiota appear to morphism robustly associated with increased
be independent of energy intake, suggesting that BMI was identified in 2007 and mapped to a
weight loss may also be the result of an interplay gene now known as FTO (Frayling et al., 2007).
between the microbiota and host biology. Since then, genome-wide association studies and
It is recognized that along with an impact meta-analyses conducted through large-scale
on metabolic state, obesity-associated changes consortium efforts (e.g. the Genetic Investigation
in bacterial diversity, as well as specific shifts of Anthropometric Traits [GIANT] consortium)
in gut bacteria such as Faecalibacterium praus- have identified almost 100 gene variants robustly
nitzii (phylum Firmicutes), are associated linked to BMI, including genes involved in appe-
with alterations in the inflammatory state (Le tite regulation, neural networking, glutamate
Chatelier, et al., 2013; Marchesi et al., 2016) (see receptor activity, insulin function, and energy
also Section 4.4.1). In studies in rats, it has been metabolism (Speliotes et al., 2010; Locke et al.,
demonstrated that inflammation alone can cause 2015). In addition, BMI-associated variants in
weight gain in normal rats and that the absence genes linked to immune function, such as TLR4,
of inflammation protects rats against weight gain may confer susceptibility to obesity through
from a high-fat diet (Harris et al., 2012). interactions with the microbiome. Genome-wide
Studies in animal models have demonstrated association studies for other anthropometric
that transfer of the gut microbiota (faecal micro- parameters, such as central adiposity (as deter-
biota transplantation) from obese mice to lean mined by waist circumference), have identified
mice results in weight gain despite decreases additional variants (Heard-Costa et al., 2009). As
in food intake. Similarly, the introduction of a even larger studies of common genetic variation
“lean” microbiota will result in weight loss in and studies of rare variants are undertaken, and
obese animals. However, in humans similar more sophisticated computational tools become
changes in the gut microbiota created by probio- available, it is likely that additional genetic vari-
tics and prebiotics have not been successful in ants associated with obesity will be identified.
49
Nevertheless, the variants identified to date, composition and fat distribution, but their use
combined, explain only a modest fraction (~3%) is limited by the lack of practicality and the cost;
of the variation in BMI (Locke et al., 2015). they are mostly used to validate simpler methods
Epigenetic mechanisms, such as DNA or in small clinical and epidemiological studies
methylation and histone modifications, may (Willett & Hu, 2013).
also play a role in obesity and represent an inte-
grated measure of both genetic and environ- 1.4.1 Weight and height
mental factors (see also Section 4.2.3). Emerging
data from population-based studies indicate that Standardized methods are available to
specific regions of the genome are differentially measure weight, height, and other anthro-
methylated in obese people compared with those pometric variables (Gibson, 2005). These are
of normal weight (Dick et al., 2014; Aslibekyan currently used in nutritional surveys (CDC,
et al., 2015). These studies demonstrate that 2016) and analytical epidemiological studies.
increased adiposity is associated with specific These measurements should be obtained by
epigenetic changes and may provide clues about trained health technicians, with routine cali-
the biology of obesity. Maternal exposure to bration of equipment and strict quality control.
environmental risk factors during pregnancy can Body weight is measured in kilograms (to the
alter the metabolic phenotype of offspring (e.g. second decimal place) using a self-zeroing digital
“body size”) by means of epigenetic regulation of scale, while the subject is wearing light clothing
specific genes, and the epigenetic modifications and without shoes, preferably in the fasting state.
could be passed on to future generations (Cutfield Height is measured to the nearest millimetre
et al., 2007; Chamorro-García & Blumberg, with a stadiometer, while the subject is without
2014). Although developmental programming shoes, with the back square against a metal wall
of body weight regulation occurs in humans, it tape, and with the eyes looking straight ahead
is still unclear whether it occurs via epigenetic and a set square resting on the scalp (Willett &
mechanisms (Institute of Medicine, 2015). Hu, 2013).
Among various anthropometric variables,
weight and height are measured with the highest
1.4 Assessment of anthropometric precision (reproducibility) and accuracy (little
measures and body composition deviance from the true value) and with the least
amount of technical error (Ulijaszek & Kerr,
The assessment of body composition is essen- 1999). In many large cohort studies, weight and
tial in obesity research, and several approaches height are based on self-reports. Self-reported
have been proposed. The simpler and less expen- measures are more feasible for large samples
sive approach is to use weight and height – the and are necessary when considering past weight
most commonly used measures of nutritional (Ulijaszek & Kerr, 1999). When self-reported
status in epidemiological studies – and waist measures are validated with objective measures,
circumferences as a measure of abdominal it has been shown that subjects tend to understate
obesity. More technologically advanced methods weight and overstate height (Krul et al., 2011);
include bioelectrical impedance analysis (BIA), in general, the achieved degree of accuracy is
dual-energy X-ray absorptiometry (DXA), sufficient to rank individuals in epidemiological
computed tomography (CT), and magnetic reso- studies, but use of self-reported measures will
nance imaging (MRI). These methods provide lead to an underestimation of the prevalence of
more precise and accurate estimates of body obesity (Willett & Hu, 2013). Self-reported weight
50
and height obtained by telephone interviews have 1.4.2 Indexes of adiposity

been shown to be less reliable than those obtained
by in-person interviews (Ezzati et al., 2006). (a) BMI
Adult height is a complex variable, deter- The most commonly used index of obesity
mined primarily by genetics but also by nutri- is the BMI (see Section 1.1). The cut-off value
tional factors, especially intakes of energy and for underweight (18.5 kg/m2) is based largely
proteins during the pre-adult period (Cole, 2000). on health-related problems associated with
Thus, height may reflect energy balance during malnutrition in developing countries (IARC,
childhood and adolescence (Willett & Hu, 2013). 2002).
Recalled weight from many years in the past BMI is easy to assess and has high precision
appears to be valid, although the error is greater and accuracy. It has been shown to correlate
than for self-reported current weight. In women, strongly with both absolute body fat and body
the correlation between weight, height, and BMI fat percentage in different age, sex, and racial
measured at age 18 years and reported at ages groups when compared with more sophisticated
71–76 years was found to be 0.84, 0.92, and 0.83, measures of fat distribution (DXA) (Gallagher
respectively (Must et al., 1993). et al., 1996; Blew et al., 2002; Evans et al., 2006).
In addition, useful information can be BMI is a useful indicator of body fatness, but it is
obtained by asking subjects to describe their an imperfect measure of obesity, because it does
body profile (body silhouette) currently and at not differentiate between lean mass and fat mass,
different earlier ages using pictograms that range the relative proportions of which vary between
from very thin to massively obese (Sørensen individuals and with age, sex, and race/ethnicity
et al., 1983). Recalled body silhouette at ages 20, (Garn et al., 1986; Prentice & Jebb, 2001).
15, and 10 years has been found to correlate well (i) Age and sex differences
with measured BMI at ages 71–76 years (Pearson
correlation ranging from 0.53 to 0.75 in men and For biological reasons, women have higher
women) (Must et al., 1993). The use of these picto- body fat percentages than men for the same
grams is particularly helpful in populations in BMI (Gallagher et al., 1996). For children, BMI
which it is challenging to obtain accurate anthro- centiles and Z-scores [measures of standard devi-
pometric measurements and to obtain informa- ation] should be used to determine BMI status
tion about body fatness at earlier ages (Romieu for ages up to 18 years, because these indicators
et al., 2012). Similar pictograms have been used to are age- and sex-specific in children and adoles-
study personal perception of current body silhou- cents (Cole, 2002; WHO 2016, 2017a, b). Among
ette (Thompson & Gray, 1995) and perception of elderly people, the validity of BMI as a measure
body silhouette in relation to age and health (Han of body fatness appears to be reduced because
et al., 1999). of changes in body composition, such as loss of
Self-reported birth weight also appears to be lean mass (sarcopenia) and increase in abdom-
reasonably valid. In the Nurses’ Health Study II inal fat mass associated with ageing (Gallagher
cohort, the correlation between recorded birth et al., 1996: Hu, 2008). This lower validity of BMI
weight and self-reported birth weight was 0.74 in elderly people may explain why the relation-
(Troy et al., 1996). ship of BMI with mortality is less pronounced in
elderly people than in younger adults; in elderly
people, waist circumference is a better marker
of adiposity, in particular of abdominal fatness
(Seidell & Visscher, 2000, Janssen et al., 2005).
51
(ii) Ethnic differences standardized metric than BMI. Correlation of

The interpretation of BMI in epidemiolog- waist circumference with predicted total abdom-
ical studies is further complicated by the ethnic inal fat and abdominal visceral fat using CT
variation in body composition (Hu, 2008). For scan was 0.87–0.93 and 0.84–0.93, respectively,
the same BMI, Blacks have a lower adiposity and in men and women, and waist circumference
body fat percentage than Whites, on average, performed as well as DXA (Clasey et al., 1999).
whereas Asians have a higher body fat percentage However, in non-obese women, DXA appears to
than Whites (Wagner & Heyward, 2000; Rush be superior to waist circumference (Kamel et al.,
et al., 2009; Liu et al., 2011). On the basis of these 1999). Despite the fact that waist circumference
observations, lower cut-off values have been has greater accuracy than BMI for measuring
proposed for overweight (23 kg/m2 to 27.4 kg/m2) abdominal fat, it still does not differentiate the
and obesity (≥ 27.5 kg/m2) in Asian populations subcutaneous fat from the visceral fat.
(WHO Expert Consultation, 2004). However, the Hip circumference is typically measured at
available data do not allow a clear cut-off value the maximal circumference over the buttocks
to be established for all Asians, given the hetero- (Hu, 2008). Hip circumference is more difficult
geneity in the observed risks related to BMI to interpret than waist circumference, because it
in different Asian populations (WHO Expert can reflect more accumulation of subcutaneous
Consultation, 2004). fat, greater gluteal muscle mass, or larger bone
structure (pelvic width) (Willett, 1998).
(b) Waist circumference and hip circumference Waist-to-hip ratio is used as an indirect
There is increasing recognition that body measure of abdominal and central obesity, but
fat distribution contributes to obesity-related there is some evidence that waist circumfer-
disease risk, independently of overall adiposity ence may be superior to waist-to-hip ratio as a
(Eckel et al., 2005). Distribution of body fat has surrogate marker of central obesity, in particular
been used to delineate two body shapes: gyne- among elderly people, for whom waist-to-hip
coid (or pear shape), with fat accumulation in ratio may be an indicator of visceral obesity
the lower part of the body, such as the hips and combined with muscle loss (Clasey et al., 1999;
the thighs), and android (or apple shape), with Snijder et al., 2006).
fat accumulation in the upper part of the body, Waist circumference-to-height ratio has been
such as in the abdomen. Consistent evidence used as an alternative to waist-to-hip ratio as a
has linked android obesity with more metabolic measure of central obesity, but it does not seem to
alterations than gynecoid obesity (Hu, 2008). be a better predictor of total mortality than waist
Waist circumference is widely used as an circumference or waist-to-hip ratio (Hu, 2008).
indirect measure of abdominal and central Other more complex measures, such as
obesity in epidemiological studies. Waist circum- sagittal abdominal diameter (i.e. the horizontal
ference is measured at the natural waist (midway distance between the anterior and the posterior
between the lowest rib margin and the iliac crest) of the abdomen), waist-to-thigh ratio, or conicity
at the level of the umbilicus (Hu, 2008). Because index (waist circumference (m)/{0.109 × square
a “natural waist” may be difficult to locate for root of weight (kg)/height (m)}) have been
obese subjects, the umbilicus site is preferred, proposed; however, the difference with other
although it may introduce substantial variations measures is small (IARC, 2002).
in defining the measurement site for very obese Waist circumference has been recommended
patients. Therefore, waist circumference is a less as a measure of central obesity over waist-to-hip
ratio (NHLBI Obesity Education Initiative,
52
1.4.3 Other measures of adiposity or body

Table 1.3 Recommendations of body mass
index and waist circumference cut-off values
composition
made for overweight or obesity (a) Skinfold thickness
Category Body mass Obesity Waist Skinfold thickness measurements are used
index (kg/m2) class circumference as an indirect assessment of body fat distribu-
(cm)
tion using a special caliper (skinfold caliper) to
Overweight 25.0–29.9 measure the thickness of a double layer of skin
Obesity 30.0–34.9 Class I Men: > 102
Women: > 88 and the fat beneath it at predetermined sites,
35.0–39.9 Class II
≥ 40.0 Class III such as the triceps, the biceps, the subscapular
Adapted from WHO (2000).
region, the abdomen, and the thigh. The meas-
urements are repeated most commonly at four
2000). Recommended cut-off values for waist body sites that are high in fat mass (usually the
circumference were 102 cm for men and 88 cm chest, arm, abdomen, and thigh), and in some
for women (Table 1.3), and for waist-to-hip ratio cases at seven body sites. Skinfold measurement
were 0.95 for men and 0.88 for women (Hu, requires great expertise, is prone to inter-ob-
2008). However, as for BMI, these cut-off values server variations, and is less reproducible than
are arbitrary, given that metabolic risk appears to other anthropometric methods. Skinfold thick-
increase linearly. In addition, the association of ness measurements and the use of standardized
central obesity with chronic disease varies across equations have been validated as a good measure
different ages, sexes, and ethnicities. Recently, of body fat percentage in adults and in children,
the International Diabetes Federation proposed but this is highly dependent on the validity of the
a new definition of metabolic syndrome diag- prediction equation, which is population-specific
nosis that includes central obesity measured by (Hu, 2008; Horan et al., 2015). However, because
waist circumference (Zimmet et al., 2005; IDF, skinfold thickness measurements are measures
2006). Table 1.4 presents the recommended waist of subcutaneous fat, they are unable to quantify
circumference cut-off values for Caucasian and intra-abdominal or visceral fat (Steinberger et al.,
Asian populations. 2005).
In large cohort studies, waist circumfer-
(b) Bioelectrical impedance (BIA)
ence and hip circumference are self-reported.
Although obese participants tend to under BIA estimates body composition by meas-
estimate waist circumference, validation studies uring the impedance or resistance to a small elec-
have shown high correlations between self- trical current (typically 800 µA; 50 kHz) passing
reported and technician-measured waist circum- across the body tissues. A simple version is foot-
ferences (r = 0.95 for men and 0.89 for women), to-foot BIA assessment, where an individual
whereas for waist-to-hip ratio the correlation steps onto scales with electrode foot plates; this is
was slightly lower (r = 0.69 for men and 0.70 for particularly advantageous for children. The BIA
women) (Hu, 2008). [It is important to note that method is based on the principle that resistance
the validation studies on self-reported height, to an applied alternating electrical current is a
weight, waist circumference, and hip circum- function of tissue composition; because fat is a
ference have been done in educated populations non-conductor of electricity, the greater the lean
in the USA and that, depending on the cultural mass or water content of a person, the faster the
and social context, self-reported anthropometric current will pass through, and the greater the
measures may have less validity.] fatty tissue content, the greater the resistance to
53
of the body and are detected by transducers. This

Table 1.4 International Diabetes Federation method is non-invasive, portable, and quick, and
criteria for waist circumference cut-off by
ultrasound is a readily available technique in the
ethnicity
clinical setting. It can also distinguish between
Ethnicity Sex Waist visceral and subcutaneous fat (Wagner, 2013).
circumference However, interpretation of ultrasound images
(cm)
requires technical skill and practice. Ultrasound
Caucasian Men ≥ 94 measures subcutaneous and pre-peritoneal fat,
Women ≥ 80
which give a good approximation of visceral fat
Asian (including South Asian, Men ≥ 90
Chinese, and Japanese) Women ≥ 80 in adults. Among children, results are conflicting
Adapted from IDF (2006) with permission.
(Horan et al., 2015).
(d) Dual-energy X-ray absorptiometry (DXA)
the current (Hu, 2008). BIA measures total body
water content and enables the calculation of lean DXA involves the use of X-ray beams with
mass and fat mass (Kyle et al., 2004). Numerous different photon energies to determine body
prediction equations have been used to determine composition. High-density material (i.e. bone)
fat-free mass and body fat percentage. Models attenuates the X-ray beam the most, whereas
typically involved height, age, sex, race, weight, lower-density material allows more photons to
and other anthropometric measures. As with pass through. DXA enables the determination
other prediction equations for body fat, those for of three components of the whole body (fat-free
BIA tend to be population-specific. [The limita- mass, fat mass, and bone mineral density) as well
tions of this method include the hydration levels as of specific regions, such as the arms, legs, and
of the individual, the composition of the latest trunk (Hu, 2008). The procedure is relatively
meal, whether a workout took place before the simple, and the measurements are highly repro-
measurement, and the time of the day when the ducible and accurate for lean mass and fat mass.
measurement took place.] DXA has the inconvenience of involving ionizing
Comparison of estimates of body fat radiation, but the exposure is very low.
percentage has shown a good correlation between DXA measures of body composition have
BIA and DXA in subjects within a normal range shown high reproducibility over several months
of body fat (r = 0.88 for the whole population; (Cordero-MacIntyre et al., 2002) and high validity
0.78 for men and 0.85 for women). However, (Lohman & Chen, 2005). DXA estimates of trunk
BIA tended to overestimate body fat percentage and abdominal fat mass were strongly correlated
in lean subjects and to underestimate it in over- with CT scan estimates of total abdominal fat
weight and obese subjects (by 3–4%) (Sun et al., (correlations, 0.94–0.97) and abdominal visceral
2005). fat (correlations, 0.86–0.90) (Clasey et al., 1999;
Because BIA equipment is portable, it can be Glickman et al., 2004).
used in large surveys or epidemiological studies Although the DXA instrument is expensive
(Kyle et al., 2003). and immobile, it has been used in several epide-
miological studies either as a references method
(c) Ultrasound or in an association study with metabolic disor-
Ultrasound, or ultrasonography, involves ders or chronic outcomes. DXA has been used
exposure of the body to high-frequency sound as a reference method to determine the body
waves which reflect off the structures and tissues fat percentage corresponding to BMI cut-off
values stratified by age, sex, and race/ethnicity
54
in a large sample from the United States popul- simpler anthropometric indices (Neamat-Allah
ation (n = 12 906) (Heo et al., 2012). The oldest et al., 2014). Measured BMI, waist circumfer-
age group had the highest cut-off values of body ence, and hip circumference correlated well
fat percentage. Non-Hispanic Blacks had lower with MRI measures of total body volume, total
cut-off values, whereas Mexicans had the highest. adipose tissue, and subcutaneous adipose tissue
Cut-off values of body fat percentage were higher in a cohort of German men (n = 598) and women
in women than in men. (n = 594) aged 51–81 years (see Table 1.5).
DXA is becoming a frequent measure of
human body composition in small clinical and 1.4.4 Change in weight
epidemiological studies to obtain a reliable esti-
mate of total body fat mass and fat distribution, Change in weight is also of interest when
alone or together with ultrasound. considering the association between weight and
risk of cancer, and change in weight has been
(e) Computed tomography (CT) and magnetic widely used in epidemiological studies as both
resonance imaging (MRI) exposure and outcome. Several definitions have
been used across studies, such as the highest
CT and MRI are considered to be the most
attained weight or current weight, weight gain,
accurate methods for assessing body compo-
weight loss, weight cycling (repeated loss and
sition and ascertaining fat distribution at the
regain of body weight over time), or body weight
levels of tissues and organs. Both CT and MRI
variability. Body weight gain during the period
provide high-resolution cross-sectional scans
from young adulthood (18–20 years) to middle
of selected tissues or organs and can be used to
age (30–55 years) is of great interest.
measure the volume and distribution of subcu-
Change in weight may rely on recall of
taneous versus visceral fat, muscle mass, and
previous weights by participants. The Nurses’
organ composition. Unlike CT, MRI does not
Health Study found a high correlation between
expose subjects to ionizing radiation and can
recalled weight at age 18 years and measured
be used in children and pregnant women (Hu,
weight from medical records, and this finding
2008). The measurements are highly reproduc-
has been confirmed in additional studies (Troy
ible and accurate (Ross & Janssen, 2005). These
et al., 1995). When considering weight loss, it is
methods are expensive and not readily accessible
important to distinguish between intentional
and are used mostly for calibrating or validating
and unintentional weight loss, because uninten-
simpler and less expensive methods for meas-
tional weight loss is associated with increased
uring body fat distribution. CT and MRI have
morbidity and mortality (Wannamethee et al.,
been used in small clinical and epidemiological
2005).
studies to measure total adipose tissue, subcuta-
When epidemiological evidence is evaluated
neous adipose tissue, visceral adipose tissue, and
for a relationship between weight change and risk
hepatic and intramuscular triglyceride content.
of cancer, the timing of weight change measure-
Significant correlations have been found between
ments must be considered, because evaluation of
visceral adipose tissue (but not subcutaneous
weight change at certain life stages may not be
adipose tissue), insulin resistance, and metabolic
adequate to fully address potential associations.
syndrome (Lebovitz & Banerji, 2005).
Analysis to relate weight change to cancer
Recent epidemiological studies have imple-
risk should control for baseline BMI. As demon-
mented CT or MRI measurements on large
strated in cardiovascular disease, to assess
subsamples to investigate biomarkers of fat
the contribution of waist circumference and
distribution (Shah et al., 2016) or to validate
55
Table 1.5 Prediction of body compartments by anthropometric indices in multiple linear

regression analysis (partial correlation coefficient adjusted for age and height) in a cohort of
German men (n = 598) and women (n = 594) aged 51–81 years
Anthropometric variable Total body volume Total adipose Subcutaneous adipose Visceral adipose
tissue tissue tissue
Men
BMI (kg/m2) 0.72 0.25 0.21 0.15
Waist circumference (cm) 0.32 0.48 0.32 0.44
Hip circumference (m) 0.28 0.27 0.39 −0.13
Women
BMI (kg/m2) 0.80 0.49 0.46 0.22
Waist circumference (cm) 0.14 0.20 0.0 0.42
Hip circumference (m) 0.34 0.45 0.52 −0.13
BMI, body mass index.
Adapted from Neamat-Allah et al. (2014).
waist-to-hip ratio separately from BMI, it is weight have been shown to underestimate the
helpful to present results in strata of baseline prevalence of overweight, with a 25–50% under-
BMI (Wormser et al., 2011). In particular, the diagnosis of overweight (Sherry et al., 2007).
initial value of BMI before the weight gain should
be included in the model to directly assess the
independent effect of the subsequent weight gain References
on risk; then, the subsequent weight change is
predominantly change in adiposity. At present, Adams J, White M (2015). Characterisation of UK diets
analysis and reporting of weight gain is incon- according to degree of food processing and associ-
ations with socio-demographics and obesity: cross-
sistent in the cancer epidemiology literature. sectional analysis of UK National Diet and Nutrition
Survey (2008-12). Int J Behav Nutr Phys Act, 12(1):160.
1.4.5 Assessment of paediatric adiposity doi:10.1186/s12966-015-0317-y PMID:26684833
Ainsworth BE, Haskell WL, Herrmann SD, Meckes
The main anthropometric indicators of N, Bassett DR Jr, Tudor-Locke C, et al. (2011). 2011
Compendium of Physical Activities: a second update
adiposity in children are BMI and waist circum- of codes and MET values. Med Sci Sports Exerc,
ference, which provide sufficient information 43(8):1575–81. doi:10.1249/MSS.0b013e31821ece12
to enable classification of overweight or obesity PMID:21681120
when growth centile charts and ratios are used Almandil NB, Liu Y, Murray ML, Besag FM, Aitchison
KJ, Wong IC (2013). Weight gain and other metabolic
(Horan et al., 2015). There is little evidence that adverse effects associated with atypical antipsychotic
alternative measures, including skinfold thick- treatment of children and adolescents: a systematic
ness and foot-to-foot BIA assessment, provide review and meta-analysis. Paediatr Drugs, 15(2):139–
better estimates than BMI when measurements 50. doi:10.1007/s40272-013-0016-6 PMID:23519708
Anderson AS, Caswell S, MacAskill S, Steele RJC, Wells
of weight and height are taken accurately and M (2010). LiveWell, a feasibility study of a personal-
compared with appropriate growth charts and/ ized lifestyle programme for colorectal cancer survi-
or cut-off values. Indirect proxy reporting, such vors. Support Care Cancer, 18(4):409–15. doi:10.1007/
s00520-009-0677-4 PMID:19554354
as parental reports, should only be used with Anderson AS, Caswell S, Wells M, Steele RJC (2013). Obesity
caution (Himes, 2009; Van Cauwenberghe et al., and lifestyle advice in colorectal cancer survivors – how
2014). In adolescents, self-reported height and
56
well are clinicians prepared? Colorectal Dis, 15(8):949– Obes Res, 10(8):799–808. doi:10.1038/oby.2002.108
57. doi:10.1111/codi.12203 PMID:23480570 PMID:12181389
Ashley-Martin J, Dodds L, Arbuckle TE, Ettinger AS, Blundell JE, Finlayson G (2004). Is susceptibility to weight
Shapiro GD, Fisher M, et al. (2014). A birth cohort gain characterized by homeostatic or hedonic risk
study to investigate the association between prenatal factors for overconsumption? Physiol Behav, 82(1):21–5.
phthalate and bisphenol A exposures and fetal markers doi:10.1016/j.physbeh.2004.04.021 PMID:15234585
of metabolic dysfunction. Environ Health, 13(84):84. Blundell JE, Finlayson G, Gibbons C, Caudwell P,
doi:10.1186/1476-069X-13-84 PMID:25336252 Hopkins M (2015b). The biology of appetite control: do
Aslibekyan S, Demerath EW, Mendelson M, Zhi D, Guan resting metabolic rate and fat-free mass drive energy
W, Liang L, et al. (2015). Epigenome-wide study iden- intake? Physiol Behav, 152(Pt B):473–8. doi:10.1016/j.
tifies novel methylation loci associated with body mass physbeh.2015.05.031 PMID:26037633
index and waist circumference. Obesity (Silver Spring), Blundell JE, Gibbons C, Caudwell P, Finlayson G, Hopkins
23(7):1493–501. doi:10.1002/oby.21111 PMID:26110892 M (2015a). Appetite control and energy balance: impact
Bagci Bosi AT, Eriksen KG, Sobko T, Wijnhoven TM, of exercise. Obes Rev, 16(Suppl 1):67–76. doi:10.1111/
Breda J (2016). Breastfeeding practices and policies in obr.12257 PMID:25614205
WHO European Region Member States. Public Health Bobrow KL, Quigley MA, Green J, Reeves GK, Beral V;
Nutr, 19(4):753–64. doi:10.1017/S1368980015001767 Million Women Study Collaborators (2013). Persistent
PMID:26096540 effects of women’s parity and breastfeeding patterns on
Bahadoran Z, Mirmiran P, Azizi F (2016). Fast food their body mass index: results from the Million Women
pattern and cardiometabolic disorders: a review of Study. Int J Obes (Lond), 37(5):712–7. doi:10.1038/
current studies. Health Promot Perspect, 5(4):231–40. ijo.2012.76 PMID:22777544
doi:10.15171/hpp.2015.028 PMID:26933642 Bolton JL, Smith SH, Huff NC, Gilmour MI, Foster WM,
Ball K, Crawford D (2005). Socioeconomic status and Auten RL, et al. (2012). Prenatal air pollution exposure
weight change in adults: a review. Soc Sci Med, induces neuroinflammation and predisposes offspring
60(9):1987–2010. doi:10.1016/j.socscimed.2004.08.056 to weight gain in adulthood in a sex-specific manner.
PMID:15743649 FASEB J, 26(11):4743–54. doi:10.1096/fj.12-210989
Barker DJ (1990). The fetal and infant origins of PMID:22815382
adult disease. BMJ, 301(6761):1111. doi:10.1136/ Boyland EJ, Nolan S, Kelly B, Tudur-Smith C, Jones
bmj.301.6761.1111 PMID:2252919 A, Halford JC, et al. (2016). Advertising as a cue to
Bayon V, Leger D, Gomez-Merino D, Vecchierini MF, consume: a systematic review and meta-analysis of
Chennaoui M (2014). Sleep debt and obesity. Ann the effects of acute exposure to unhealthy food and
Med, 46(5):264–72. doi:10.3109/07853890.2014.931103 nonalcoholic beverage advertising on intake in children
PMID:25012962 and adults. Am J Clin Nutr, 103(2):519–33. doi:10.3945/
Benedict C, Axelsson T, Söderberg S, Larsson A, Ingelsson ajcn.115.120022 PMID:26791177
E, Lind L, et al. (2014). Fat mass and obesity-associ- Bray GA, Popkin BM (1998). Dietary fat intake does affect
ated gene (FTO) is linked to higher plasma levels of obesity! Am J Clin Nutr, 68(6):1157–73. PMID:9846842
the hunger hormone ghrelin and lower serum levels Brisbois TD, Farmer AP, McCargar LJ (2012). Early
of the satiety hormone leptin in older adults. Diabetes, markers of adult obesity: a review. Obes Rev,
63(11):3955–9. doi:10.2337/db14-0470 PMID:24898142 13(4):347–67. doi:10.1111/j.1467-789X.2011.00965.x
Black JL, Macinko J (2008). Neighborhoods and PMID:22171945
obesity. Nutr Rev, 66(1):2–20. doi:10.1111/j.1753- Brown PJ (1991). Culture and the evolution of obesity.
4887.2007.00001.x PMID:18254880 Hum Nat, 2(1):31–57. doi:10.1007/BF02692180
Black RE, Victora CG, Walker SP, Bhutta ZA, Christian P, PMID:24222189
de Onis M, et al.; Maternal and Child Nutrition Study Buckland G, Bach A, Serra-Majem L (2008). Obesity and
Group (2013). Maternal and child undernutrition and the Mediterranean diet: a systematic review of observa-
overweight in low-income and middle-income coun- tional and intervention studies. Obes Rev, 9(6):582–93.
tries. Lancet, 382(9890):427–51. doi:10.1016/S0140- doi:10.1111/j.1467-789X.2008.00503.x PMID:18547378
6736(13)60937-X PMID:23746772 Burgio E, Lopomo A, Migliore L (2015). Obesity and
Blakely T, Hales S, Kieft C, Wilson N, Woodward A diabetes: from genetics to epigenetics. Mol Biol
(2005). The global distribution of risk factors by Rep, 42(4):799–818. doi:10.1007/s11033-014-3751-z
poverty level. Bull World Health Organ, 83(2):118–26. PMID:25253098
PMID:15744404 Caballero B (2007). The global epidemic of obesity: an
Blew RM, Sardinha LB, Milliken LA, Teixeira PJ, Going SB, overview. Epidemiol Rev, 29(1):1–5. doi:10.1093/epirev/
Ferreira DL, et al. (2002). Assessing the validity of body mxm012 PMID:17569676
mass index standards in early postmenopausal women. Cairns G, Angus K, Hastings G, Caraher M (2013).
Systematic reviews of the evidence on the nature, extent
57
and effects of food marketing to children. A retro- (Silver Spring), 23(7):1331–44. doi:10.1002/oby.21118
spective summary. Appetite, 62:209–15. doi:10.1016/j. PMID:26096983
appet.2012.04.017 PMID:22561190 Cole TJ (2000). Secular trends in growth. Proc Nutr
Capers PL, Fobian AD, Kaiser KA, Borah R, Allison Soc, 59(2):317–24. doi:10.1017/S0029665100000355
DB (2015). A systematic review and meta-analysis of PMID:10946801
randomized controlled trials of the impact of sleep Cole TJ (2002). A chart to link child centiles of body mass
duration on adiposity and components of energy index, weight and height. Eur J Clin Nutr, 56(12):1194–9.
balance. Obes Rev, 16(9):771–82. doi:10.1111/obr.12296 doi:10.1038/sj.ejcn.1601473 PMID:12494304
PMID:26098388 Cordero-MacIntyre ZR, Peters W, Libanati CR, España
Cappuccio FP, Taggart FM, Kandala NB, Currie A, Peile RC, Abila SO, Howell WH, et al. (2002). Reproducibility
E, Stranges S, et al. (2008). Meta-analysis of short sleep of DXA in obese women. J Clin Densitom, 5(1):35–44.
duration and obesity in children and adults. Sleep, doi:10.1385/JCD:5:1:035 PMID:11940727
31(5):619–26. PMID:18517032 Cupul-Uicab LA, Klebanoff MA, Brock JW, Longnecker
Caprio S, Daniels SR, Drewnowski A, Kaufman FR, MP (2013). Prenatal exposure to persistent organochlo-
Palinkas LA, Rosenbloom AL, et al. (2008). Influence rines and childhood obesity in the US Collaborative
of race, ethnicity, and culture on childhood obesity: Perinatal Project. Environ Health Perspect,
implications for prevention and treatment: a consensus 121(9):1103–9. PMID:23799652
statement of Shaping America’s Health and the Obesity Cutfield WS, Hofman PL, Mitchell M, Morison IM
Society. Diabetes Care, 31(11):2211–21. doi:10.2337/ (2007). Could epigenetics play a role in the develop-
dc08-9024 PMID:18955718 mental origins of health and disease? Pediatr Res,
Caspersen CJ, Powell KE, Christenson GM (1985). Physical 61(5 Pt 2):68R–75R. doi:10.1203/pdr.0b013e318045764c
activity, exercise, and physical fitness: definitions and PMID:17413843
distinctions for health-related research. Public Health De Bourdeaudhuij I, Van Dyck D, Salvo D, Davey R, Reis
Rep, 100(2):126–31. PMID:3920711 RS, Schofield G, et al. (2015). International study of
CDC (2016). NHANES III standardized anthropo- perceived neighbourhood environmental attributes
metric procedures, videos. Atlanta (GA), USA: and body mass index: IPEN Adult study in 12 coun-
Centers for Disease Control and Prevention. Available tries. Int J Behav Nutr Phys Act, 12(1):62. doi:10.1186/
from: https://wwwn.cdc.gov/nchs/nhanes/nhanes3/ s12966-015-0228-y PMID:25982214
anthropometricvideos.aspx. Department for Communities and Local Government
Chakravartty S, Tassinari D, Salerno A, Giorgakis E, (2011). The English Indices of Deprivation 2010.
Rubino F (2015). What is the mechanism behind Available from: https://www.gov.uk/government/
weight loss maintenance with gastric bypass? Curr u p l o a d s /s y s t e m / u p l o a d s /a t t a c h m e n t _ d a t a /
Obes Rep, 4(2):262–8. doi:10.1007/s13679-015-0158-7 file/6871/1871208.pdf.
PMID:26627220 de Jonge L, DeLany JP, Nguyen T, Howard J, Hadley EC,
Chamorro-García R, Blumberg B (2014). Transgener Redman LM, et al. (2007). Validation study of energy
ational effects of obesogens and the obesity epidemic. expenditure and intake during calorie restriction using
Curr Opin Pharmacol, 19:153–8. doi:10.1016/j. doubly labeled water and changes in body composition.
coph.2014.10.010 PMID:25460228 Am J Clin Nutr, 85(1):73–9. PMID:17209180
Chaput J-P, Tremblay A (2012). Insufficient sleep as a Dearth-Wesley T, Popkin BM, Ng SW (2014). Estimated
contributor to weight gain: an update. Curr Obes Rep, and forecasted trends in domain specific time-use and
1(4):245–56. doi:10.1007/s13679-012-0026-7 energy expenditure among adults in Russia. Int J Behav
Chen X, Beydoun MA, Wang Y (2008). Is sleep duration Nutr Phys Act, 11(1):11. doi:10.1186/1479-5868-11-11
associated with childhood obesity? A systematic review PMID:24475868
and meta-analysis. Obesity (Silver Spring), 16(2):265– Deglaire A, Méjean C, Castetbon K, Kesse-Guyot E,
74. doi:10.1038/oby.2007.63 PMID:18239632 Hercberg S, Schlich P (2015). Associations between
Clasey JL, Bouchard C, Teates CD, Riblett JE, Thorner weight status and liking scores for sweet, salt and
MO, Hartman ML, et al. (1999). The use of anthro- fat according to the gender in adults (the Nutrinet-
pometric and dual-energy X-ray absorptiometry Santé study). Eur J Clin Nutr, 69(1):40–6. doi:10.1038/
(DXA) measures to estimate total abdominal and ejcn.2014.139 PMID:25074389
abdominal visceral fat in men and women. Obes Res, Di Cesare M, Khang YH, Asaria P, Blakely T, Cowan MJ,
7(3):256–64. doi:10.1002/j.1550-8528.1999.tb00404.x Farzadfar F, et al.; Lancet NCD Action Group (2013).
PMID:10348496 Inequalities in non-communicable diseases and effec-
Cobb LK, Appel LJ, Franco M, Jones-Smith JC, Nur A, tive responses. Lancet, 381(9866):585–97. doi:10.1016/
Anderson CA (2015). The relationship of the local S0140-6736(12)61851-0 PMID:23410608
food environment with obesity: a systematic review
of methods, study quality, and results. Obesity
58
Di Pietro L, Dziura J, Blair SN (2004). Estimated change Egger G, Swinburn B (1997). An “ecological” approach
in physical activity level (PAL) and prediction of 5-year to the obesity pandemic. BMJ, 315(7106):477–80.
weight change in men: the Aerobics Center Longitudinal doi:10.1136/bmj.315.7106.477 PMID:9284671
Study. Int J Obes Relat Metab Disord, 28(12):1541–7. Eissing L (2013). Metabolism: FTO-associated obesity
doi:10.1038/sj.ijo.0802821 PMID:15543159 risk is linked to brain food responses via modula-
Dick KJ, Nelson CP, Tsaprouni L, Sandling JK, Aïssi tion of ghrelin levels. Nat Rev Endocrinol, 9(10):564.
D, Wahl S, et al. (2014). DNA methylation and doi:10.1038/nrendo.2013.155 PMID:23917583
body-mass index: a genome-wide analysis. Lancet, Ello-Martin JA, Ledikwe JH, Rolls BJ (2005). The influ-
383(9933):1990–8. doi:10.1016/S0140-6736(13)62674-4 ence of food portion size and energy density on energy
PMID:24630777 intake: implications for weight management. Am J Clin
Ding D, Gebel K (2012). Built environment, physical Nutr, 82(1 Suppl):236S–41S. PMID:16002828
activity, and obesity: what have we learned from Elobeid MA, Padilla MA, Brock DW, Ruden DM, Allison
reviewing the literature? Health Place, 18(1):100–5. DB (2010). Endocrine disruptors and obesity: an exam-
doi:10.1016/j.healthplace.2011.08.021 PMID:21983062 ination of selected persistent organic pollutants in the
Dinsa GD, Goryakin Y, Fumagalli E, Suhrcke M (2012). NHANES 1999–2002 data. Int J Environ Res Public
Obesity and socioeconomic status in developing coun- Health, 7(7):2988–3005. doi:10.3390/ijerph7072988
tries: a systematic review. Obes Rev, 13(11):1067–79. PMID:20717554
doi:10.1111/j.1467-789X.2012.01017.x PMID:22764734 Erkin-Cakmak A, Harley KG, Chevrier J, Bradman A,
Dior UP, Lawrence GM, Sitlani C, Enquobahrie D, Kogut K, Huen K, et al. (2015). In utero and childhood
Manor O, Siscovick DS, et al. (2014). Parental smoking polybrominated diphenyl ether exposures and body
during pregnancy and offspring cardio-meta- mass at age 7 years: the CHAMACOS study. Environ
bolic risk factors at ages 17 and 32. Atherosclerosis, Health Perspect, 123(6):636–42. PMID:25738596
235(2):430–7. doi:10.1016/j.atherosclerosis.2014.05.937 European Commission (2006). Health and food. Special
PMID:24937467 Eurobarometer 246/Wave 64.3 – TNS Opinion & Social.
Dombrowski SU, Knittle K, Avenell A, Araújo-Soares V, Available from: http://ec.europa.eu/public_opinion/
Sniehotta FF (2014). Long term maintenance of weight archives/ebs/ebs_246_en.pdf.
loss with non-surgical interventions in obese adults: Evans EM, Rowe DA, Racette SB, Ross KM, McAuley E
systematic review and meta-analyses of randomised (2006). Is the current BMI obesity classification appro-
controlled trials. BMJ, 348:g2646. doi:10.1136/bmj. priate for black and white postmenopausal women? Int
g2646 PMID:25134100 J Obes (Lond), 30(5):837–43. doi:10.1038/sj.ijo.0803208
Dominguez-Salas P, Moore SE, Baker MS, Bergen PMID:16418761
AW, Cox SE, Dyer RA, et al. (2014). Maternal nutri- Ezzati M, Martin H, Skjold S, Vander Hoorn S, Murray
tion at conception modulates DNA methylation of CJ (2006). Trends in national and state-level obesity in
human metastable epialleles. Nat Commun, 5:3746. the USA after correction for self-report bias: analysis of
doi:10.1038/ncomms4746 PMID:24781383 health surveys. J R Soc Med, 99(5):250–7. doi:10.1258/
Donnelly JE, Blair SN, Jakicic JM, Manore MM, Rankin jrsm.99.5.250 PMID:16672759
JW, Smith BK; American College of Sports Medicine Fardet A, Boirie Y (2014). Associations between food
(2009). American College of Sports Medicine Position and beverage groups and major diet-related chronic
Stand. Appropriate physical activity intervention strat- diseases: an exhaustive review of pooled/meta-ana-
egies for weight loss and prevention of weight regain for lyses and systematic reviews. Nutr Rev, 72(12):741–62.
adults. Med Sci Sports Exerc, 41(2):459–71. doi:10.1249/ doi:10.1111/nure.12153 PMID:25406801
MSS.0b013e3181949333 PMID:19127177 Farpour-Lambert NJ, Baker JL, Hassapidou M, Holm
Due P, Damsgaard MT, Rasmussen M, Holstein BE, JC, Nowicka P, O’Malley G, et al. (2015). Childhood
Wardle J, Merlo J, et al.; HBSC obesity writing group obesity is a chronic disease demanding specific health
(2009). Socioeconomic position, macroeconomic care – a position statement from the Childhood Obesity
environment and overweight among adolescents in Task Force (COTF) of the European Association for
35 countries. Int J Obes, 33(10):1084–93. doi:10.1038/ the Study of Obesity (EASO). Obes Facts, 8(5):342–9.
ijo.2009.128 PMID:19621018 doi:10.1159/000441483 PMID:26469067
Dunton GF, Kaplan J, Wolch J, Jerrett M, Reynolds KD Feng J, Glass TA, Curriero FC, Stewart WF, Schwartz BS
(2009). Physical environmental correlates of childhood (2010). The built environment and obesity: a system-
obesity: a systematic review. Obes Rev, 10(4):393–402. atic review of the epidemiologic evidence. Health Place,
doi:10.1111/j.1467-789X.2009.00572.x PMID:19389058 16(2):175–90. doi:10.1016/j.healthplace.2009.09.008
Eckel RH, Grundy SM, Zimmet PZ (2005). The metabolic PMID:19880341
syndrome. Lancet, 365(9468):1415–28. doi:10.1016/
S0140-6736(05)66378-7 PMID:15836891
59
Flegal KM, Carroll MD, Ogden CL, Johnson CL (2002). Gotsis E, Anagnostis P, Mariolis A, Vlachou A,
Prevalence and trends in obesity among US adults, Katsiki N, Karagiannis A (2015). Health benefits
1999–2000. JAMA, 288(14):1723–7. doi:10.1001/ of the Mediterranean Diet: an update of research
jama.288.14.1723 PMID:12365955 over the last 5 years. Angiology, 66(4):304–18.
Fogelholm M, Anderssen S, Gunnarsdottir I, Lahti-Koski doi:10.1177/0003319714532169 PMID:24778424
M (2012). Dietary macronutrients and food consump- Graham C, Mullen A, Whelan K (2015). Obesity and the
tion as determinants of long-term weight change in gastrointestinal microbiota: a review of associations
adult populations: a systematic literature review. and mechanisms. Nutr Rev, 73(6):376–85. doi:10.1093/
Food Nutr Res, 56(0). doi:10.3402/fnr.v56i0.19103 nutrit/nuv004 PMID:26011912
PMID:22893781 Grinker JA, Tucker K, Vokonas PS, Rush D (1995).
Fogelholm M, Kukkonen-Harjula K (2000). Does physical Body habitus changes among adult males from the
activity prevent weight gain – a systematic review. Obes normative aging study: relations to aging, smoking
Rev, 1(2):95–111. doi:10.1046/j.1467-789x.2000.00016.x history and alcohol intake. Obes Res, 3(5):435–46.
PMID:12119991 doi:10.1002/j.1550-8528.1995.tb00173.x PMID:8521163
Forshee RA, Anderson PA, Storey ML (2008). Sugar- Gross AL, May BJ, Axilbund JE, Armstrong DK, Roden RB,
sweetened beverages and body mass index in children Visvanathan K (2015). Weight change in breast cancer
and adolescents: a meta-analysis. Am J Clin Nutr, survivors compared to cancer-free women: a prospec-
87(6):1662–71. PMID:18541554 tive study in women at familial risk of breast cancer.
Frayling TM, Timpson NJ, Weedon MN, Zeggini E, Cancer Epidemiol Biomarkers Prev, 24(8):1262–9.
Freathy RM, Lindgren CM, et al. (2007). A common doi:10.1158/1055-9965.EPI-15-0212 PMID:26178712
variant in the FTO gene is associated with body mass Grün F, Blumberg B (2009). Endocrine disrupters as
index and predisposes to childhood and adult obesity. obesogens. Mol Cell Endocrinol, 304(1–2):19–29.
Science, 316(5826):889–94. doi:10.1126/science.1141634 doi:10.1016/j.mce.2009.02.018 PMID:19433244
PMID:17434869 Guo SS, Chumlea WC (1999). Tracking of body mass index
Freedman DS, Khan LK, Serdula MK, Dietz WH, in children in relation to overweight in adulthood. Am
Srinivasan SR, Berenson GS (2005). The relation of J Clin Nutr, 70:145S–8S.
childhood BMI to adult adiposity: the Bogalusa Heart Han TS, Hart CL, Haig C, Logue J, Upton MN, Watt GC,
Study. Pediatrics, 115(1):22–7. PMID:15629977 et al. (2015). Contribution of maternal and paternal
Gallagher D, Visser M, Sepúlveda D, Pierson RN, Harris adiposity and smoking to adult offspring adiposity
T, Heymsfield SB (1996). How useful is body mass and CVD risk. BMJ Open, 5(11):e007682. doi:10.1136/
index for comparison of body fatness across age, sex, bmjopen-2015-007682 PMID:26525718
and ethnic groups? Am J Epidemiol, 143(3):228–39. Han TS, Morrison CE, Lean ME (1999). Age and health
doi:10.1093/oxfordjournals.aje.a008733 PMID:8561156 indications assessed by silhouette photographs. Eur
García-Fernández E, Rico-Cabanas L, Rosgaard N, J Clin Nutr, 53(8):606–11. doi:10.1038/sj.ejcn.1600816
Estruch R, Bach-Faig A (2014). Mediterranean diet and PMID:10477246
cardiodiabesity: a review. Nutrients, 6(9):3474–500. Hansen AY, Umstattd Meyer MR, Lenardson JD, Hartley
doi:10.3390/nu6093474 PMID:25192027 D (2015). Built environments and active living in
Garn SM, Leonard WR, Hawthorne VM (1986). Three rural and remote areas: a review of the literature. Curr
limitations of the body mass index. Am J Clin Nutr, Obes Rep, 4(4):484–93. doi:10.1007/s13679-015-0180-9
44(6):996–7. PMID:3788846 PMID:26364307
Gibson RS (2005). Principles of nutritional assessment. Harris K, Kassis A, Major G, Chou CJ (2012). Is the
2nd edition. Oxford, UK: Oxford University Press. gut microbiota a new factor contributing to obesity
Gibson S (2008). Sugar-sweetened soft drinks and obesity: and its metabolic disorders? J Obes, 2012:879151.
a systematic review of the evidence from observational doi:10.1155/2012/879151 PMID:22315672
studies and interventions. Nutr Res Rev, 21(2):134–47. Haus EL, Smolensky MH (2013). Shift work and cancer
doi:10.1017/S0954422408110976 PMID:19087367 risk: potential mechanistic roles of circadian disrup-
Glickman SG, Marn CS, Supiano MA, Dengel DR (2004). tion, light at night, and sleep deprivation. Sleep Med
Validity and reliability of dual-energy X-ray absorp- Rev, 17(4):273–84. doi:10.1016/j.smrv.2012.08.003
tiometry for the assessment of abdominal adiposity. PMID:23137527
J Appl Physiol (1985), 97(2):509–14. doi:10.1152/ Health and Social Care Information Centre (2013).
japplphysiol.01234.2003 PMID:15075304 National Child Measurement Programme: England,
Gordon-Larsen P, Hou N, Sidney S, Sternfeld B, Lewis CE, 2012-13 school year. Available from: http://content.
Jacobs DR Jr, et al. (2009). Fifteen-year longitudinal digital.nhs.uk/catalogue/PUB13115/nati-chil-meas-
trends in walking patterns and their impact on weight prog-eng-2012-2013-rep.pdf.
change. Am J Clin Nutr, 89(1):19–26. doi:10.3945/
ajcn.2008.26147 PMID:19056560
60
Heard-Costa NL, Zillikens MC, Monda KL, Johansson A, Am J Epidemiol, 176(Suppl 7):S44–54. doi:10.1093/aje/
Harris TB, Fu M, et al. (2009). NRXN3 is a novel locus kws264 PMID:23035144
for waist circumference: a genome-wide association IARC (2002). Weight control and physical activity. Lyon,
study from the CHARGE Consortium. PLoS Genet, France: IARC Press (IARC Handbooks of Cancer
5(6):e1000539. doi:10.1371/journal.pgen.1000539 Prevention, Vol. 6). Available from: http://publications.
PMID:19557197 iarc.fr/376.
Heindel JJ, Newbold R, Schug TT (2015). Endocrine IDF (2006). The IDF consensus worldwide definition of the
disruptors and obesity. Nat Rev Endocrinol, 11(11):653– metabolic syndrome. Brussels, Belgium: International
61. doi:10.1038/nrendo.2015.163 PMID:26391979 Diabetes Federation. Available from: https://www.idf.
Heo M, Faith MS, Pietrobelli A, Heymsfield SB (2012). org/e-library/consensus-statements/60-idfconsensus-
Percentage of body fat cutoffs by sex, age, and race- worldwide-definitionof-the-metabolic-syndrome.
ethnicity in the US adult population from NHANES IHME (2014). Overweight and Obesity Viz. Seattle (WA),
1999–2004. Am J Clin Nutr, 95(3):594–602. doi:10.3945/ USA: Institute for Health Metrics and Evaluation.
ajcn.111.025171 PMID:22301924 Available from: http://www.healthdata.org/
Hill JO, Peters JC, Peters JC (1998). Environmental data-visualization/overweight-and-obesity-viz.
contributions to the obesity epidemic. Science, Institute of Medicine (2015). Examining a developmental
280(5368):1371–4. doi:10.1126/science.280.5368.1371 approach to childhood obesity: the fetal and early
PMID:9603719 childhood years. Workshop summary. Washington
Hill JO, Wyatt HR, Peters JC (2012). Energy balance (DC), USA: National Academies Press. Available from:
and obesity. Circulation, 126(1):126–32. doi:10.1161/ https://www.nap.edu/catalog/21782.
CIRCULATIONAHA.111.087213 PMID:22753534 Janesick AS, Blumberg B (2016). Obesogens: an emerging
Himes JH (2009). Challenges of accurately measuring and threat to public health. Am J Obstet Gynecol, 214(5):559–
using BMI and other indicators of obesity in children. 65. doi:10.1016/j.ajog.2016.01.182 PMID:26829510
Pediatrics, 124(Suppl 1):S3–22. doi:10.1542/peds.2008- Janssen I, Katzmarzyk PT, Ross R (2005). Body mass index
3586D PMID:19720665 is inversely related to mortality in older people after
Hooper L, Abdelhamid A, Bunn D, Brown T, Summerbell adjustment for waist circumference. J Am Geriatr Soc,
CD, Skeaff CM (2015). Effects of total fat intake on 53(12):2112–8. doi:10.1111/j.1532-5415.2005.00505.x
body weight. Cochrane Database Syst Rev, 8:CD011834. PMID:16398895
doi:10.1002/14651858.CD011834 PMID:26250104 Jarlenski MP, Bennett WL, Bleich SN, Barry CL, Stuart
Horan M, Gibney E, Molloy E, McAuliffe F (2015). EA (2014). Effects of breastfeeding on postpartum
Methodologies to assess paediatric adiposity. Ir J weight loss among U.S. women. Prev Med, 69:146–50.
Med Sci, 184(1):53–68. doi:10.1007/s11845-014-1124-1 doi:10.1016/j.ypmed.2014.09.018 PMID:25284261
PMID:24791970 Jaworowska A, Blackham T, Davies IG, Stevenson L (2013).
Horta BL, Loret de Mola C, Victora CG (2015). Long- Nutritional challenges and health implications of take-
term consequences of breastfeeding on cholesterol, away and fast food. Nutr Rev, 71(5):310–8. doi:10.1111/
obesity, systolic blood pressure and type 2 diabetes: nure.12031 PMID:23590707
a systematic review and meta-analysis. Acta Paediatr, Jiang P, Turek FW (2017). Timing of meals: when is
104(467):30–7. doi:10.1111/apa.13133 PMID:26192560 as critical as what and how much. Am J Physiol
Hu FB (2008). Measurements of adiposity and body Endocrinol Metab, 312(5):E369–80. doi:10.1152/
composition. In: Hu FB, editor. Obesity epidemiology. ajpendo.00295.2016 PMID:28143856
Oxford, UK: Oxford University Press. doi:10.1093/acpr Kaiser KA, Shikany JM, Keating KD, Allison DB (2013).
of:oso/9780195312911.003.0005 Will reducing sugar-sweetened beverage consump-
Hu FB, Li TY, Colditz GA, Willett WC, Manson JE (2003). tion reduce obesity? Evidence supporting conjec-
Television watching and other sedentary behaviors in ture is strong, but evidence when testing effect is
relation to risk of obesity and type 2 diabetes mellitus weak. Obes Rev, 14(8):620–33. doi:10.1111/obr.12048
in women. JAMA, 289(14):1785–91. doi:10.1001/ PMID:23742715
jama.289.14.1785 PMID:12684356 Kamel EG, McNeill G, Han TS, Smith FW, Avenell A,
Hu FB, Malik VS (2010). Sugar-sweetened beverages Davidson L, et al. (1999). Measurement of abdominal
and risk of obesity and type 2 diabetes: epidemiologic fat by magnetic resonance imaging, dual-energy X-ray
evidence. Physiol Behav, 100(1):47–54. doi:10.1016/j. absorptiometry and anthropometry in non-obese men
physbeh.2010.01.036 PMID:20138901 and women. Int J Obes Relat Metab Disord, 23(7):686–
Hu T, Mills KT, Yao L, Demanelis K, Eloustaz M, Yancy 92. doi:10.1038/sj.ijo.0800904 PMID:10454101
WS Jr, et al. (2012). Effects of low-carbohydrate Kelley GA, Kelley KS, Pate RR (2015). Exercise and BMI
diets versus low-fat diets on metabolic risk factors: a in overweight and obese children and adolescents: a
meta-analysis of randomized controlled clinical trials. systematic review and trial sequential meta-analysis.
61
Biomed Res Int, 2015:704539. doi:10.1155/2015/704539 Le Chatelier E, Nielsen T, Qin J, Prifti E, Hildebrand F,
PMID:26579538 Falony G, et al.; MetaHIT consortium (2013). Richness
Kim SH, Park MJ (2014). Phthalate exposure and child- of human gut microbiome correlates with meta-
hood obesity. Ann Pediatr Endocrinol Metab, 19(2):69– bolic markers. Nature, 500(7464):541–6. doi:10.1038/
75. doi:10.6065/apem.2014.19.2.69 PMID:25077088 nature12506 PMID:23985870
Kim TW, Jeong JH, Hong SC (2015). The impact of sleep and Lean MEJ, Malkova D (2015). Altered gut and adipose
circadian disturbance on hormones and metabolism. tissue hormones in overweight and obese individ-
Int J Endocrinol, 2015:591729. doi:10.1155/2015/591729 uals: cause or consequence? Int J Obes (Lond), 1:1–10.
PMID:25861266 PMID:26499438
Krieger N, Williams DR, Moss NE (1997). Measuring Lebovitz HE, Banerji MA (2005). Point: visceral adiposity
social class in US public health research: concepts, is causally related to insulin resistance. Diabetes
methodologies, and guidelines. Annu Rev Public Health, Care, 28(9):2322–5. doi:10.2337/diacare.28.9.2322
18(1):341–78. doi:10.1146/annurev.publhealth.18.1.341 PMID:16123512
PMID:9143723 Lee IM, Djoussé L, Sesso HD, Wang L, Buring JE
Krul AJ, Daanen HA, Choi H (2011). Self-reported and (2010). Physical activity and weight gain prevention.
measured weight, height and body mass index (BMI) JAMA, 303(12):1173–9. doi:10.1001/jama.2010.312
in Italy, the Netherlands and North America. Eur J PMID:20332403
Public Health, 21(4):414–9. doi:10.1093/eurpub/ckp228 Leslie WS, Hankey CR, Lean ME (2007). Weight gain as
PMID:20089678 an adverse effect of some commonly prescribed drugs:
Kumar BN, Meyer HE, Wandel M, Dalen I, Holmboe- a systematic review. QJM, 100(7):395–404. doi:10.1093/
Ottesen G (2006). Ethnic differences in obesity among qjmed/hcm044 PMID:17566010
immigrants from developing countries, in Oslo, Liao Y, Liao J, Durand CP, Dunton GF (2014). Which type
Norway. Int J Obes (Lond), 30(4):684–90. doi:10.1038/ of sedentary behaviour intervention is more effective at
sj.ijo.0803051 PMID:16130029 reducing body mass index in children? A meta-analytic
Kyle UG, Bosaeus I, De Lorenzo AD, Deurenberg P, review. Obes Rev, 15(3):159–68. doi:10.1111/obr.12112
Elia M, Gómez JM, et al.; Composition of the ESPEN PMID:24588966
Working Group (2004). Bioelectrical impedance Link BG, Phelan JC (2001). Conceptualizing stigma.
analysis – part I: review of principles and methods. Clin Annu Rev Sociol, 27(1):363–85. doi:10.1146/annurev.
Nutr, 23(5):1226–43. doi:10.1016/j.clnu.2004.06.004 soc.27.1.363
PMID:15380917 Lipsky LM, Iannotti RJ (2012). Associations of tele-
Kyle UG, Pirlich M, Schuetz T, Luebke HJ, Lochs H, vision viewing with eating behaviors in the 2009
Pichard C (2003). Prevalence of malnutrition in 1760 Health Behaviour in School-aged Children Study.
patients at hospital admission: a controlled population Arch Pediatr Adolesc Med, 166(5):465–72. doi:10.1001/
study of body composition. Clin Nutr, 22(5):473–81. archpediatrics.2011.1407 PMID:22566548
doi:10.1016/S0261-5614(03)00049-9 PMID:14512035 Lissner L, Wijnhoven TMA, Mehlig K, Sjöberg A, Kunesova
Lakind JS, Goodman M, Mattison DR (2014). Bisphenol M, Yngve A, et al. (2016). Socioeconomic inequalities
A and indicators of obesity, glucose metabolism/type in childhood overweight: heterogeneity across five
2 diabetes and cardiovascular disease: a systematic countries in the WHO European Childhood Obesity
review of epidemiologic research. Crit Rev Toxicol, Surveillance Initiative (COSI-2008). Int J Obes (Lond),
44(2):121–50. doi:10.3109/10408444.2013.860075 40(5):796–802. doi:10.1038/ijo.2016.12 PMID:27136760
PMID:24392816 Liu A, Byrne NM, Kagawa M, Ma G, Poh BK, Ismail MN,
Lamerz A, Kuepper-Nybelen J, Wehle C, Bruning N, et al. (2011). Ethnic differences in the relationship
Trost-Brinkhues G, Brenner H, et al. (2005). Social between body mass index and percentage body fat
class, parental education, and obesity prevalence among Asian children from different backgrounds. Br
in a study of six-year-old children in Germany. Int J J Nutr, 106(9):1390–7. doi:10.1017/S0007114511001681
Obes (Lond), 29(4):373–80. doi:10.1038/sj.ijo.0802914 PMID:21736824
PMID:15768043 Liu KY, Chow JM, Sherry C (2013). Early life obesity and
Langley-Evans SC (2015). Nutrition in early life and diabetes: origins in pregnancy. Open J Endocr Metab
the programming of adult disease: a review. J Hum Dis, 3:1–12. doi:10.1007/s11154-013-9238-8
Nutr Diet, 28(Suppl 1):1–14. doi:10.1111/jhn.12212 Livingstone MB, Black AE (2003). Markers of the validity
PMID:24479490 of reported energy intake. J Nutr, 133(Suppl 3):895S–
Lawrence W, Barker M (2009). A review of factors affecting 920S. PMID:12612176
the food choices of disadvantaged women. Proc Nutr Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH,
Soc, 68(2):189–94. doi:10.1017/S0029665109001013 Day FR, et al.; LifeLines Cohort Study; ADIPOGen
PMID:19208268 Consortium; AGEN-BMI Working Group;
CARDIOGRAMplusC4D Consortium; CKDGen
62
Consortium; GLGC; ICBP; MAGIC Investigators; Marshall NS, Glozier N, Grunstein RR (2008). Is sleep
MuTHER Consortium; MIGen Consortium; PAGE duration related to obesity? A critical review of the
Consortium; ReproGen Consortium; GENIE epidemiological evidence. Sleep Med Rev, 12(4):289–98.
Consortium; International Endogene Consortium doi:10.1016/j.smrv.2008.03.001 PMID:18485764
(2015). Genetic studies of body mass index yield new Martínez Steele E, Baraldi LG, Louzada ML, Moubarac JC,
insights for obesity biology. Nature, 518(7538):197–206. Mozaffarian D, Monteiro CA (2016). Ultra-processed
doi:10.1038/nature14177 PMID:25673413 foods and added sugars in the US diet: evidence from
Lohman TG, Chen Z (2005). Dual-energy x-ray absorp- a nationally representative cross-sectional study. BMJ
tiometry. In: Heymsfield SB, Lohman TG, Wang Z, Open, 6(3):e009892. doi:10.1136/bmjopen-2015-009892
Going SB, editors. Human body composition. 2nd PMID:26962035
edition. Champaign (IL), USA; Leeds, UK: Human McCrory MA (2014). Meal skipping and variables related
Kinetics Press; pp. 63–77. to energy balance in adults: a brief review, with emphasis
Lopez R, Hynes HP (2003). Sprawl in the 1990s: on the breakfast meal. Physiol Behav, 134:51–4.
measurement, distribution, and trends. Urban doi:10.1016/j.physbeh.2014.05.005 PMID:24825781
Aff Rev Thousand Oaks Calif, 38(3):325–55. McCrory MA, Fuss PJ, Hays NP, Vinken AG, Greenberg
doi:10.1177/1078087402238805 PMID:16990927 AS, Roberts SB (1999). Overeating in America: associ-
Louzada ML, Baraldi LG, Steele EM, Martins AP, Canella ation between restaurant food consumption and body
DS, Moubarac JC, et al. (2015). Consumption of fatness in healthy adult men and women ages 19 to 80.
ultra-processed foods and obesity in Brazilian adoles- Obes Res, 7(6):564–71. doi:10.1002/j.1550-8528.1999.
cents and adults. Prev Med, 81:9–15. doi:10.1016/j. tb00715.x PMID:10574515
ypmed.2015.07.018 PMID:26231112 McFadden E, Jones ME, Schoemaker MJ, Ashworth
Ludwig J, Sanbonmatsu L, Gennetian L, Adam E, Duncan A, Swerdlow AJ (2014). The relationship between
GJ, Katz LF, et al. (2011). Neighborhoods, obesity, and obesity and exposure to light at night: cross-sectional
diabetes – a randomized social experiment. N Engl J analyses of over 100,000 women in the Breakthrough
Med, 365(16):1509–19. doi:10.1056/NEJMsa1103216 Generations Study. Am J Epidemiol, 180(3):245–50.
PMID:22010917 doi:10.1093/aje/kwu117 PMID:24875371
Lynch BM, Healy GN, Dunstan DW, Owen N (2010). McKinnon L, Giskes K, Turrell G (2014). The contrib-
Sedentary versus inactive: distinctions for disease ution of three components of nutrition knowledge
prevention. Nat Rev Cardiol, 7(11):–68c1, author to socio-economic differences in food purchasing
reply, 68-c2. doi:10.1038/nrcardio.2010.68-c1 choices. Public Health Nutr, 17(8):1814–24. doi:10.1017/
PMID:21080592 S1368980013002036 PMID:23920283
Mackenbach JD, Rutter H, Compernolle S, Glonti K, Oppert McPheeters ML, Warren Z, Sathe N, Bruzek JL,
JM, Charreire H, et al. (2014). Obesogenic environ- Krishnaswami S, Jerome RN, et al. (2011). A systematic
ments: a systematic review of the association between review of medical treatments for children with autism
the physical environment and adult weight status, the spectrum disorders. Pediatrics, 127(5):e1312–21.
SPOTLIGHT project. BMC Public Health, 14(1):233. doi:10.1542/peds.2011-0427 PMID:21464191
doi:10.1186/1471-2458-14-233 PMID:24602291 Miller CT, Fraser SF, Levinger I, Straznicky NE, Dixon JB,
Magee L, Hale L (2012). Longitudinal associations between Reynolds J, et al. (2013). The effects of exercise training
sleep duration and subsequent weight gain: a system- in addition to energy restriction on functional capac-
atic review. Sleep Med Rev, 16(3):231–41. doi:10.1016/j. ities and body composition in obese adults during
smrv.2011.05.005 PMID:21784678 weight loss: a systematic review. PLoS One, 8(11):e81692.
Malik VS, Pan A, Willett WC, Hu FB (2013). Sugar- doi:10.1371/journal.pone.0081692 PMID:24409219
sweetened beverages and weight gain in children and Monasta L, Batty GD, Cattaneo A, Lutje V, Ronfani L,
adults: a systematic review and meta-analysis. Am J Van Lenthe FJ, et al. (2010). Early-life determinants
Clin Nutr, 98(4):1084–102. doi:10.3945/ajcn.113.058362 of overweight and obesity: a review of systematic
PMID:23966427 reviews. Obes Rev, 11(10):695–708. doi:10.1111/j.1467-
Marchesi JR, Adams DH, Fava F, Hermes GD, Hirschfield 789X.2010.00735.x PMID:20331509
GM, Hold G, et al. (2016). The gut microbiota and Monteiro CA, Conde WL, Lu B, Popkin BM (2004a).
host health: a new clinical frontier. Gut, 65(2):330–9. Obesity and inequities in health in the developing
doi:10.1136/gutjnl-2015-309990 PMID:26338727 world. Int J Obes Relat Metab Disord, 28(9):1181–6.
Maresca MM, Hoepner LA, Hassoun A, Oberfield SE, doi:10.1038/sj.ijo.0802716 PMID:15211362
Mooney SJ, Calafat AM, et al. (2015). Prenatal expo- Monteiro CA, Moura EC, Conde WL, Popkin BM (2004b).
sure to phthalates and childhood body size in an urban Socioeconomic status and obesity in adult populations
cohort. Environ Health Perspect, 124(4). doi:10.1289/ of developing countries: a review. Bull World Health
ehp.1408750 PMID:26069025 Organ, 82(12):940–6. PMID:15654409
63
Moore CJ, Cunningham SA (2012). Social position, psycho- resonance imaging – a study within the German
logical stress, and obesity: a systematic review. J Acad European Prospective Investigation into Cancer and
Nutr Diet, 112(4):518–26. doi:10.1016/j.jand.2011.12.001 Nutrition (EPIC) cohorts. PLoS One, 9(3):e91586.
PMID:22709702 doi:10.1371/journal.pone.0091586 PMID:24626110
Mozaffarian D, Afshin A, Benowitz NL, Bittner V, Daniels Neville CE, McKinley MC, Holmes VA, Spence D,
SR, Franch HA, et al.; American Heart Association Woodside JV (2014). The relationship between breast-
Council on Epidemiology and Prevention, Council on feeding and postpartum weight change – a system-
Nutrition, Physical Activity and Metabolism, Council atic review and critical evaluation. Int J Obes (Lond),
on Clinical Cardiology, Council on Cardiovascular 38(4):577–90. doi:10.1038/ijo.2013.132 PMID:23892523
Disease in the Young, Council on the Kidney in Newby PK, Muller D, Hallfrisch J, Qiao N, Andres R,
Cardiovascular Disease, Council on Peripheral Tucker KL (2003). Dietary patterns and changes in
Vascular Disease, and the Advocacy Coordinating body mass index and waist circumference in adults.
Committee (2012). Population approaches to Am J Clin Nutr, 77(6):1417–25. PMID:12791618
improve diet, physical activity, and smoking habits: Ng M, Fleming T, Robinson M, Thomson B, Graetz N,
a scientific statement from the American Heart Margono C, et al. (2014). Global, regional, and national
Association. Circulation, 126(12):1514–63. doi:10.1161/ prevalence of overweight and obesity in children and
CIR.0b013e318260a20b PMID:22907934 adults during 1980–2013: a systematic analysis for
Mozaffarian D, Hao T, Rimm EB, Willett WC, Hu FB (2011). the Global Burden of Disease Study 2013. Lancet,
Changes in diet and lifestyle and long-term weight gain 384(9945):766–81. doi:10.1016/S0140-6736(14)60460-8
in women and men. N Engl J Med, 364(25):2392–404. PMID:24880830
doi:10.1056/NEJMoa1014296 PMID:21696306 Ng SW, Popkin BM, Popkin BM (2012). Time use and
Murray CJ, Kulkarni S, Ezzati M (2005). Eight physical activity: a shift away from movement across
Americas: new perspectives on U.S. health dispari- the globe. Obes Rev, 13(8):659–80. doi:10.1111/j.1467-
ties. Am J Prev Med, 29(5 Suppl 1):4–10. doi:10.1016/j. 789X.2011.00982.x PMID:22694051
amepre.2005.07.031 PMID:16389119 NHLBI Obesity Education Initiative (2000). The prac-
Must A, Willett WC, Dietz WH (1993). Remote recall of tical guide: identification, evaluation, and treatment of
childhood height, weight, and body build by elderly overweight and obesity in adults. Bethesda (MD), USA:
subjects. Am J Epidemiol, 138(1):56–64. PMID:8333427 National Heart, Lung, and Blood Institute, National
Nathan PC, Jovcevska V, Ness KK, Mammone D’Agostino Institutes of Health. Available from: https://www.
N, Staneland P, Urbach SL, et al. (2006). The preva- nhlbi.nih.gov/files/docs/guidelines/prctgd_c.pdf.
lence of overweight and obesity in pediatric survi- Nichols HB, Trentham-Dietz A, Egan KM, Titus-Ernstoff
vors of cancer. J Pediatr, 149(4):518–25. doi:10.1016/j. L, Holmes MD, Bersch AJ, et al. (2009). Body mass
jpeds.2006.06.039 PMID:17011325 index before and after breast cancer diagnosis: asso-
National Obesity Observatory (2011). Obesity and ciations with all-cause, breast cancer, and cardiovas-
ethnicity. Available from: http://webarchive. cular disease mortality. Cancer Epidemiol Biomarkers
nationalarchives.gov.uk/20160805121933/http://www. Prev, 18(5):1403–9. doi:10.1158/1055-9965.EPI-08-1094
noo.org.uk/uploads/doc/vid_9851_Obesity_ethnicity. PMID:19366908
pdf. Obesity Society (2014). Americans’ view on obesity is
National Obesity Observatory (2012). NOO data factsheet: changing: fewer adults see it as a personal problem
child obesity and socioeconomic status. Available of bad choices. Science Daily, 5 November 2014.
from: http://webarchive.nationalarchives.gov. Available from: https://www.sciencedaily.com/
uk/20160805121933/http://www.noo.org.uk/uploads/ releases/2014/11/141105112514.htm.
doc/vid_16967_ChildSocioeconSep2012.pdf. Ogden CL, Carroll MD, Curtin LR, McDowell MA,
NCD Risk Factor Collaboration (2016). Trends in adult Tabak CJ, Flegal KM (2006). Prevalence of over-
body-mass index in 200 countries from 1975 to weight and obesity in the United States, 1999–2004.
2014: a pooled analysis of 1698 population-based JAMA, 295(13):1549–55. doi:10.1001/jama.295.13.1549
measurement studies with 19.2 million participants PMID:16595758
Lancet, 387(10026):1377–96. doi:10.1016/S0140- Ogden CL, Carroll MD, Kit BK, Flegal KM (2014).
6736(16)30054-X PMID:27115820 Prevalence of childhood and adult obesity in the United
Ndahimana D, Kim EK (2017). Measurement methods States, 2011–2012. JAMA, 311(8):806–14. doi:10.1001/
for physical activity and energy expenditure: a review. jama.2014.732 PMID:24570244
Clin Nutr Res, 6(2):68–80. doi:10.7762/cnr.2017.6.2.68 Okorodudu DO, Jumean MF, Montori VM, Romero-
PMID:28503503 Corral A, Somers VK, Erwin PJ, et al. (2010). Diagnostic
Neamat-Allah J, Wald D, Hüsing A, Teucher B, Wendt A, performance of body mass index to identify obesity as
Delorme S, et al. (2014). Validation of anthropometric defined by body adiposity: a systematic review and
indices of adiposity against whole-body magnetic
64
meta-analysis. Int J Obes, 34(5):791–9. doi:10.1038/ Popkin BM, Richards MK, Montiero CA (1996). Stunting
ijo.2010.5 PMID:20125098 is associated with overweight in children of four
Opalinski A (2013). Cultural perspectives of childhood nations that are undergoing the nutrition transition.
obesity in Haitan and Hispanic children. J Nurse Pract, J Nutr, 126(12):3009–16. PMID:9001368
9(3):149–54. doi:10.1016/j.nurpra.2012.10.009 Poppitt SD (2015). Beverage consumption: are alco-
Owen N, Sparling PB, Healy GN, Dunstan DW, Matthews holic and sugary drinks tipping the balance towards
CE (2010). Sedentary behavior: emerging evidence for overweight and obesity? Nutrients, 7(8):6700–18.
a new health risk. Mayo Clin Proc, 85(12):1138–41. doi:10.3390/nu7085304 PMID:26270675
doi:10.4065/mcp.2010.0444 PMID:21123641 Poskitt EME, Breda J (2012). Complementary feeding
Paeratakul S, Ferdinand DP, Champagne CM, Ryan DH, and non communicable diseases: current knowledge
Bray GA (2003). Fast-food consumption among US and future research needs. Nutr Metab Cardiovasc
adults and children: dietary and nutrient intake profile. Dis, 22(10):819–22. doi:10.1016/j.numecd.2012.08.005
J Am Diet Assoc, 103(10):1332–8. doi:10.1016/S0002- PMID:22917600
8223(03)01086-1 PMID:14520253 Power C, Atherton K, Thomas C (2010). Maternal
Palmer JR, Kipping-Ruane K, Wise LA, Yu J, Rosenberg smoking in pregnancy, adult adiposity and other risk
L (2015). Lactation in relation to long-term maternal factors for cardiovascular disease. Atherosclerosis,
weight gain in African-American women. Am J 211(2):643–8. doi:10.1016/j.atherosclerosis.2010.03.015
Epidemiol, 181(12):932–9. doi:10.1093/aje/kwv027 PMID:20400081
PMID:25944887 Prentice AM, Jebb SA (2001). Beyond body mass
Pampel FC, Denney JT, Krueger PM (2012). Obesity, index. Obes Rev, 2(3):141–7. doi:10.1046/j.1467-
SES, and economic development: a test of the reversal 789x.2001.00031.x PMID:12120099
hypothesis. Soc Sci Med, 74(7):1073–81. doi:10.1016/j. Puhl R, Brownell KD (2001). Bias, discrimination,
socscimed.2011.12.028 PMID:22341204 and obesity. Obes Res, 9(12):788–805. doi:10.1038/
Papas MA, Alberg AJ, Ewing R, Helzlsouer KJ, Gary TL, oby.2001.108 PMID:11743063
Klassen AC (2007). The built environment and obesity. Rahat-Rozenbloom S, Fernandes J, Gloor GB, Wolever
Epidemiol Rev, 29(1):129–43. doi:10.1093/epirev/ TM (2014). Evidence for greater production of colonic
mxm009 PMID:17533172 short-chain fatty acids in overweight than lean humans.
Parizkova J, Chin MK, Chia M, Yang J (2007). An inter- Int J Obes, 38(12):1525–31. doi:10.1038/ijo.2014.46
national perspective on obesity, health and physical PMID:24642959
activity: current trends and challenges in China and Rantakokko P, Main KM, Wohlfart-Veje C, Kiviranta H,
Asia. J Exerc Sci Fit, 5(1):7–23. Airaksinen R, Vartiainen T, et al. (2014). Association
Parsons TJ, Power C, Logan S, Summerbell CD (1999). of placenta organotin concentrations with growth and
Childhood predictors of adult obesity: a systematic ponderal index in 110 newborn boys from Finland
review. Int J Obes Relat Metab Disord, 23(Suppl 8):S1– during the first 18 months of life: a cohort study.
107. PMID:10641588 Environ Health, 13(1):45. doi:10.1186/1476-069X-13-45
Patel SR, Hu FB (2008). Short sleep duration and weight PMID:24899383
gain: a systematic review. Obesity (Silver Spring), Reddon H, Guéant JL, Meyre D (2016). The importance of
16(3):643–53. doi:10.1038/oby.2007.118 PMID:18239586 gene-environment interactions in human obesity. Clin
Peña MM, Dixon B, Taveras EM (2012). Are you talking to Sci (Lond), 130(18):1571–97. doi:10.1042/CS20160221
ME? The importance of ethnicity and culture in child- PMID:27503943
hood obesity prevention and management. Child Obes, Rohan EA, Townsend JS, Fairley TL, Stewart SL (2015).
8(1):23–7. PMID:22799474 Health behaviors and quality of life among colorectal
Pereira MA, Kartashov AI, Ebbeling CB, Van Horn cancer survivors. J Natl Compr Canc Netw, 13(3):297–
L, Slattery ML, Jacobs DR Jr, et al. (2005). Fast-food 302. PMID:25736006
habits, weight gain, and insulin resistance (the Romieu I, Escamilla-Núñez MC, Sánchez-Zamorano
CARDIA study): 15-year prospective analysis. Lancet, LM, Lopez-Ridaura R, Torres-Mejía G, Yunes EM,
365(9453):36–42. doi:10.1016/S0140-6736(04)17663-0 et al. (2012). The association between body shape
PMID:15639678 silhouette and dietary pattern among Mexican
Pestana D, Faria G, Sá C, Fernandes VC, Teixeira D, women. Public Health Nutr, 15(1):116–25. doi:10.1017/
Norberto S, et al. (2014). Persistent organic pollutant S1368980011001182 PMID:21875454
levels in human visceral and subcutaneous adipose Ross R, Janssen I (2005). Computed tomography and
tissue in obese individuals – depot differences and magnetic resonance imaging. In: Heymsfield SB,
dysmetabolism implications. Environ Res, 133:170–7. Lohman TG, Wang Z, Going SB, editors. Human
doi:10.1016/j.envres.2014.05.026 PMID:24949816 body composition. 2nd edition. Champaign (IL), USA;
Popkin BM (1993). Nutritional patterns and transitions. Leeds, UK: Human Kinetics Press; pp. 89–108.
Popul Dev Rev, 19(1):138–57. doi:10.2307/2938388
65
Rummel-Kluge C, Komossa K, Schwarz S, Hunger H, Scientific Advisory Committee on Nutrition (2011).

Schmid F, Lobos CA, et al. (2010). Head-to-head The influence of maternal, fetal and child nutri-
comparisons of metabolic side effects of second gener- tion on the development of chronic disease
ation antipsychotics in the treatment of schizophrenia: later in life. London, UK: TSO. Available from:
a systematic review and meta-analysis. Schizophr ht t ps://w w w.gov.u k /government /publ icat ions/
Res, 123(2–3):225–33. doi:10.1016/j.schres.2010.07.012 sacn-early-life-nutrition-report.
PMID:20692814 Scientific Advisory Committee on Nutrition (2015).
Rundle A, Hoepner L, Hassoun A, Oberfield S, Freyer Carbohydrates and health. London, UK: TSO.
G, Holmes D, et al. (2012). Association of childhood Available from: https://www.gov.uk/government/
obesity with maternal exposure to ambient air poly- uploads/system/uploads/attachment_data/file/445503/
cyclic aromatic hydrocarbons during pregnancy. Am SACN_Carbohydrates_and_Health.pdf.
J Epidemiol, 175(11):1163–72. doi:10.1093/aje/kwr455 Seidell JC, Visscher TL (2000). Body weight and weight
PMID:22505764 change and their health implications for the elderly.
Rush EC, Freitas I, Plank LD (2009). Body size, body Eur J Clin Nutr, 54(Suppl 3):S33–9. doi:10.1038/
composition and fat distribution: comparative sj.ejcn.1601023 PMID:11041073
analysis of European, Maori, Pacific Island and Asian Shah RV, Allison MA, Lima JA, Abbasi SA, Eisman A, Lai
Indian adults. Br J Nutr, 102(4):632–41. doi:10.1017/ C, et al. (2016). Abdominal fat radiodensity, quantity
S0007114508207221 PMID:19203416 and cardiometabolic risk: the Multi-Ethnic Study of
Sallis JF, Floyd MF, Rodríguez DA, Saelens BE (2012). Atherosclerosis. Nutr Metab Cardiovasc Dis, 26(2):114–
Role of built environments in physical activity, obesity, 22. PMID:26817938
and cardiovascular disease. Circulation, 125(5):729– Shaw K, Gennat H, O’Rourke P, Del Mar C (2006). Exercise
37. doi:10.1161/CIRCULATIONAHA.110.969022 for overweight or obesity. Cochrane Database Syst Rev,
PMID:22311885 (4):CD003817. PMID:17054187
Sanmiguel C, Gupta A, Mayer EA (2015). Gut microbiome Sherry B, Jefferds ME, Grummer-Strawn LM (2007).
and obesity: a plausible explanation for obesity. Curr Accuracy of adolescent self-report of height and weight
Obes Rep, 4(2):250–61. doi:10.1007/s13679-015-0152-0 in assessing overweight status: a literature review. Arch
PMID:26029487 Pediatr Adolesc Med, 161(12):1154–61. doi:10.1001/
Saris WH, Blair SN, van Baak MA, Eaton SB, Davies PS, archpedi.161.12.1154 PMID:18056560
Di Pietro L, et al. (2003). How much physical activity Shrewsbury V, Wardle J (2008). Socioeconomic status and
is enough to prevent unhealthy weight gain? Outcome adiposity in childhood: a systematic review of cross-
of the IASO 1st Stock Conference and consensus sectional studies 1990–2005. Obesity (Silver Spring),
statement. Obes Rev, 4(2):101–14. doi:10.1046/j.1467- 16(2):275–84. doi:10.1038/oby.2007.35 PMID:18239633
789X.2003.00101.x PMID:12760445 Sikorski C, Luppa M, Kaiser M, Glaesmer H, Schomerus
Schellong K, Schulz S, Harder T, Plagemann A (2012). G, König HH, et al. (2011). The stigma of obesity in the
Birth weight and long-term overweight risk: systematic general public and its implications for public health
review and a meta-analysis including 643,902 persons – a systematic review. BMC Public Health, 11(1):661.
from 66 studies and 26 countries globally. PLoS doi:10.1186/1471-2458-11-661 PMID:21859493
One, 7(10):e47776. doi:10.1371/journal.pone.0047776 Simkin-Silverman LR, Wing RR, Boraz MA, Kuller LH
PMID:23082214 (2003). Lifestyle intervention can prevent weight gain
Schulze MB, Fung TT, Manson JE, Willett WC, Hu FB during menopause: results from a 5-year randomized
(2006). Dietary patterns and changes in body weight clinical trial. Ann Behav Med, 26(3):212–20. doi:10.1207/
in women. Obesity (Silver Spring), 14(8):1444–53. S15324796ABM2603_06 PMID:14644697
doi:10.1038/oby.2006.164 PMID:16988088 Simmonds M, Burch J, Llewellyn A, Griffiths C, Yang H,
Schutz Y, Byrne NM, Dulloo A, Hills AP (2014). Energy Owen C, et al. (2015). The use of measures of obesity in
gap in the aetiology of body weight gain and obesity: childhood for predicting obesity and the development
a challenging concept with a complex evaluation and of obesity-related diseases in adulthood: a system-
pitfalls. Obes Facts, 7(1):15–25. doi:10.1159/000357846 atic review and meta-analysis. Health Technol Assess,
PMID:24457473 19(43):1–336. doi:10.3310/hta19430 PMID:26108433
Schwingshackl L, Dias S, Strasser B, Hoffmann G (2013). Singh AS, Mulder C, Twisk JW, van Mechelen W, Chinapaw
Impact of different training modalities on anthropo- MJ (2008). Tracking of childhood overweight into
metric and metabolic characteristics in overweight/ adulthood: a systematic review of the literature. Obes
obese subjects: a systematic review and network Rev, 9(5):474–88. doi:10.1111/j.1467-789X.2008.00475.x
meta-analysis. PLoS One, 8(12):e82853. doi:10.1371/ PMID:18331423
journal.pone.0082853 PMID:24358230 Singh M (2014). Mood, food, and obesity. Front Psychol,
5:925. doi:10.3389/fpsyg.2014.00925 PMID:25225489
66
Snijder MB, van Dam RM, Visser M, Seidell JC (2006). Thomée S, Lissner L, Hagberg M, Grimby-Ekman A
What aspects of body fat are particularly hazardous and (2015). Leisure time computer use and overweight
how do we measure them? Int J Epidemiol, 35(1):83–92. development in young adults – a prospective study.
doi:10.1093/ije/dyi253 PMID:16339600 BMC Public Health, 15(1):839. doi:10.1186/s12889-015-
Sørensen TI, Stunkard AJ, Teasdale TW, Higgins MW 2131-5 PMID:26329006
(1983). The accuracy of reports of weight: children’s Thompson MA, Gray JJ (1995). Development and vali-
recall of their parents’ weights 15 years earlier. Int J dation of a new body-image assessment scale. J Pers
Obes, 7(2):115–22. PMID:6862758 Assess, 64(2):258–69. doi:10.1207/s15327752jpa6402_6
Spahlholz J, Baer N, König HH, Riedel-Heller SG, Luck- PMID:7722852
Sikorski C (2016). Obesity and discrimination – a Thorogood A, Mottillo S, Shimony A, Filion KB, Joseph
systematic review and meta-analysis of observational L, Genest J, et al. (2011). Isolated aerobic exercise and
studies. Obes Rev, 17(1):43–55. doi:10.1111/obr.12343 weight loss: a systematic review and meta-analysis of
PMID:26596238 randomized controlled trials. Am J Med, 124(8):747–55.
Speliotes EK, Willer CJ, Berndt SI, Monda KL, Thorleifsson doi:10.1016/j.amjmed.2011.02.037 PMID:21787904
G, Jackson AU, et al.; MAGIC; Procardis Consortium Tobi EW, Goeman JJ, Monajemi R, Gu H, Putter H,
(2010). Association analyses of 249,796 individ- Zhang Y, et al. (2014). DNA methylation signatures
uals reveal 18 new loci associated with body mass link prenatal famine exposure to growth and metab-
index. Nat Genet, 42(11):937–48. doi:10.1038/ng.686 olism. Nat Commun, 5:5592. doi:10.1038/ncomms6592
PMID:20935630 PMID:25424739
Steinberger J, Jacobs DR, Raatz S, Moran A, Hong CP, Tobias DK, Chen M, Manson JE, Ludwig DS, Willett W,
Sinaiko AR (2005). Comparison of body fatness meas- Hu FB (2015). Effect of low-fat diet interventions versus
urements by BMI and skinfolds vs dual energy X-ray other diet interventions on long-term weight change in
absorptiometry and their relation to cardiovascular risk adults: a systematic review and meta-analysis. Lancet
factors in adolescents. Int J Obes (Lond), 29(11):1346– Diabetes Endocrinol, 3(12):968–79. doi:10.1016/S2213-
52. doi:10.1038/sj.ijo.0803026 PMID:16044176 8587(15)00367-8 PMID:26527511
Sun G, French CR, Martin GR, Younghusband B, Green Tompkins CL, Seablom M, Brock DW (2015). Parental
RC, Xie YG, et al. (2005). Comparison of multifre- perception of child’s body weight: a systematic
quency bioelectrical impedance analysis with dual- review. J Child Fam Stud, 24(5):1384–91. doi:10.1007/
energy X-ray absorptiometry for assessment of s10826-014-9945-0
percentage body fat in a large, healthy population. Am Torloni MR, Betrán AP, Merialdi M (2012). Demography
J Clin Nutr, 81(1):74–8. PMID:15640463 of obesity. In: Gillman MW, Poston L, editors. Maternal
Swinburn B, Sacks G, Ravussin E (2009). Increased food obesity. Cambridge, UK: Cambridge University Press;
energy supply is more than sufficient to explain the pp. 1–7.
US epidemic of obesity. Am J Clin Nutr, 90(6):1453–6. Trasande L, Attina TM, Blustein J (2012). Association
doi:10.3945/ajcn.2009.28595 PMID:19828708 between urinary bisphenol A concentration and
Szewczyk-Golec K, Woźniak A, Reiter RJ (2015). Inter- obesity prevalence in children and adolescents.
relationships of the chronobiotic, melatonin, with leptin JAMA, 308(11):1113–21. doi:10.1001/2012.jama.11461
and adiponectin: implications for obesity. J Pineal Res, PMID:22990270
59(3):277–91. doi:10.1111/jpi.12257 PMID:26103557 Troy LM, Hunter DJ, Manson JE, Colditz GA, Stampfer
Tang-Péronard JL, Heitmann BL, Andersen HR, MJ, Willett WC (1995). The validity of recalled weight
Steuerwald U, Grandjean P, Weihe P, et al. (2014). among younger women. Int J Obes Relat Metab Disord,
Association between prenatal polychlorinated biphenyl 19(8):570–2. PMID:7489028
exposure and obesity development at ages 5 and 7 y: Troy LM, Michels KB, Hunter DJ, Spiegelman D, Manson
a prospective cohort study of 656 children from the JE, Colditz GA, et al. (1996). Self-reported birthweight
Faroe Islands. Am J Clin Nutr, 99(1):5–13. doi:10.3945/ and history of having been breastfed among younger
ajcn.113.066720 PMID:24153349 women: an assessment of validity. Int J Epidemiol,
Tang-Péronard JL, Jensen TK, Andersen HR, Ried-Larsen 25(1):122–7. doi:10.1093/ije/25.1.122 PMID:8666479
M, Grøntved A, Andersen LB, et al. (2015). Associations Trumbo PR, Rivers CR (2014). Systematic review of the
between exposure to persistent organic pollutants in evidence for an association between sugar-sweetened
childhood and overweight up to 12 years later in a low beverage consumption and risk of obesity. Nutr Rev,
exposed Danish population. Obes Facts, 8(4):282–92. 72(9):566–74. doi:10.1111/nure.12128 PMID:25091794
doi:10.1159/000438834 PMID:26228100 Ulijaszek SJ, Kerr DA (1999). Anthropometric measure-
Te Morenga L, Mallard S, Mann J (2012). Dietary sugars ment error and the assessment of nutritional status. Br
and body weight: systematic review and meta-analyses J Nutr, 82(3):165–77. doi:10.1017/S0007114599001348
of randomised controlled trials and cohort studies. PMID:10655963
BMJ, 346:e7492. doi:10.1136/bmj.e7492 PMID:23321486
67
UNICEF, WHO, and World Bank Group (2015). Levels Endocrinol Metab, 97(2):E223–7. doi:10.1210/jc.2011-
and trends in child malnutrition: UNICEF – WHO – 1989 PMID:22090277
World Bank Group joint child malnutrition estimates: Wannamethee SG, Shaper AG, Lennon L (2005).
key findings of the 2015 edition. New York, USA: Reasons for intentional weight loss, unintentional
United Nations Children’s Fund; Geneva, Switzerland: weight loss, and mortality in older men. Arch Intern
World Health Organization; Washington (DC), USA: Med, 165(9):1035–40. doi:10.1001/archinte.165.9.1035
The World Bank. Available from: http://www.who.int/ PMID:15883243
nutgrowthdb/jme_brochure2015.pdf?ua=1. Wardle J, Haase AM, Steptoe A (2006). Body image and
Van Cauwenberghe J, Delvaux I, Michels N, Den Hond E, weight control in young adults: international compar-
Schoeters G, Nelen V, et al. (2014). Validity of parentally isons in university students from 22 countries. Int J
reported versus measured weight, length and waist in Obes, 30(4):644–51. Available from: http://www.nature.
7- to 9-year-old children for use in follow-up studies. com/ijo/journal/v30/n4/pdf/0803050a.pdf doi:10.1038/
Eur J Pediatr, 173(7):921–8. doi:10.1007/s00431-014- sj.ijo.0803050 PMID:16151414
2274-x PMID:24497184 WCRF/AICR (2007). Food, nutrition, physical activity,
van der Heijden AA, Hu FB, Rimm EB, van Dam RM and the prevention of cancer: a global perspective.
(2007). A prospective study of breakfast consump- Washington (DC), USA: American Institute for Cancer
tion and weight gain among U.S. men. Obesity (Silver Research. Available from: www.aicr.org/assets/docs/
Spring), 15(10):2463–9. doi:10.1038/oby.2007.292 pdf/reports/Second_Expert_Report.pdf.
PMID:17925472 Weng SF, Redsell SA, Swift JA, Yang M, Glazebrook CP
Vandevijvere S, Chow CC, Hall KD, Umali E, Swinburn (2012). Systematic review and meta-analyses of risk
BA (2015). Increased food energy supply as a major factors for childhood overweight identifiable during
driver of the obesity epidemic: a global analysis. Bull infancy. Arch Dis Child, 97(12):1019–26. doi:10.1136/
World Health Organ, 93(7):446–56. doi:10.2471/ archdischild-2012-302263 PMID:23109090
BLT.14.150565 PMID:26170502 Westerterp KR, Speakman JR (2008). Physical activity
Vasan RS, Pencina MJ, Cobain M, Freiberg MS, energy expenditure has not declined since the 1980s
D’Agostino RB (2005). Estimated risks for developing and matches energy expenditures of wild mammals.
obesity in the Framingham Heart Study. Ann Intern Int J Obes, 32(8):1256–63. doi:10.1038/ijo.2008.74
Med, 143(7):473–80. doi:10.7326/0003-4819-143-7- PMID:18504442
200510040-00005 PMID:16204159 WHO (2000). Obesity: preventing and managing the
Villareal DT, Apovian CM, Kushner RF, Klein S; global epidemic: report of a WHO Consultation.
American Society for Nutrition; NAASO, The Obesity Geneva, Switzerland: World Health Organization
Society (2005). Obesity in older adults: technical review (WHO Technical Report Series, No. 894). Available
and position statement of the American Society for from: http://www.who.int/nutrition/publications/
Nutrition and NAASO, The Obesity Society. Am J Clin obesity/WHO_TRS_894/en/.
Nutr, 82(5):923–34. PMID:16280421 WHO (2003). Diet, nutrition and the prevention of
Volberg V, Harley K, Calafat AM, Davé V, McFadden J, chronic diseases. Report of a joint WHO/FAO
Eskenazi B, et al. (2013). Maternal bisphenol A exposure Expert Consultation. Geneva, Switzerland: World
during pregnancy and its association with adipokines Health Organization (WHO Technical Report
in Mexican-American children. Environ Mol Mutagen, Series, No. 916). Available from: http://www.who.int/
54(8):621–8. doi:10.1002/em.21803 PMID:23908009 dietphysicalactivity/publications/trs916/en/.
Wagner DR (2013). Ultrasound as a tool to assess WHO (2008). Closing the gap in a generation: health
body fat. J Obes, 2013:1–9. doi:10.1155/2013/280713 equity through action on the social determinants
PMID:24062944 of health. Final report of the Commission on Social
Wagner DR, Heyward VH (2000). Measures of body Determinants of Health. Geneva, Switzerland: World
composition in blacks and whites: a comparative review. Health Organization. Available from: http://www.who.
Am J Clin Nutr, 71(6):1392–402. PMID:10837277 int/social_determinants/thecommission/finalreport/
Wang L, Mamudu HM, Alamian A, Anderson JL, Brooks en/.
B (2014). Independent and joint effects of prenatal WHO (2011). Global database on body mass index: an
maternal smoking and maternal exposure to second- interactive surveillance tool for monitoring nutrition
hand smoke on the development of adolescent transition. Available from: www.who.int/bmi/index.
obesity: a longitudinal study. J Paediatr Child Health, jsp.
50(11):908–15. doi:10.1111/jpc.12667 PMID:24920104 WHO (2014). Obesity and inequities: guidance for address-
Wang T, Li M, Chen B, Xu M, Xu Y, Huang Y, et al. ing inequities in overweight and obesity. Available from:
(2012). Urinary bisphenol A (BPA) concentration http://www.euro.who.int/en/publications/abstracts/
associates with obesity and insulin resistance. J Clin obesity-and-inequities.-guidance-for-addressing-
inequities-in-overweight-and-obesity-2014.
68
WHO (2015). Information note about intake of Wormser D, Kaptoge S, Di Angelantonio E, Wood AM,
sugars recommended in the WHO guideline Pennells L, Thompson A, et al.; Emerging Risk Factors
for adults and children. Available from: http:// Collaboration (2011). Separate and combined associ-
w w w.who.int/nutrition/publications/guidelines/ ations of body-mass index and abdominal adiposity
sugar_intake_information_note_en.pdf. with cardiovascular disease: collaborative analysis
WHO (2016). Obesity and overweight: fact sheet. Available of 58 prospective studies. Lancet, 377(9771):1085–95.
from: http://www.who.int/mediacentre/factsheets/ doi:10.1016/S0140-6736(11)60105-0 PMID:21397319
fs311/en/. Zimmet P, Magliano D, Matsuzawa Y, Alberti G, Shaw J
WHO (2017a). WHO child growth standards. Geneva, (2005). The metabolic syndrome: a global public health
Switzerland: World Health Organization. Available problem and a new definition. J Atheroscler Thromb,
from: http://www.who.int/childgrowth/standards/ 12(6):295–300. doi:10.5551/jat.12.295 PMID:16394610
Technical_report.pdf. Zoeller RT, Brown TR, Doan LL, Gore AC, Skakkebaek
WHO (2017b). WHO growth reference data for 5–19 years. NE, Soto AM, et al. (2012). Endocrine-disrupting
Geneva, Switzerland: World Health Organization. chemicals and public health protection: a state-
Available from: http://www.who.int/nutrition/ ment of principles from The Endocrine Society.
publications/growthref_who_bulletin/en/. Endocrinology, 153(9):4097–110. doi:10.1210/en.2012-
WHO Expert Consultation (2004). Appropriate body- 1422 PMID:22733974
mass index for Asian populations and its implica-
tions for policy and intervention strategies. Lancet,
363(9403):157–63. doi:10.1016/S0140-6736(03)15268-3
PMID:14726171
WHO Regional Office for Europe (2016). Good
maternal nutrition: the best start in life.
Copenhagen, Denmark: World Health Organization
Regional Office for Europe. Available from:
h t t p : // w w w. e u r o .w h o . i n t /e n / h e a l t h - t o p i c s /
disease-prevention/nutrition/publications/2016/
good-maternal-nutrition.-the-best-start-in-life-2016.
Wijnhoven TM, van Raaij JM, Spinelli A, Starc G,
Hassapidou M, Spiroski I, et al. (2014). WHO
European Childhood Obesity Surveillance Initiative:
body mass index and level of overweight among
6-9-year-old children from school year 2007/2008 to
school year 2009/2010. BMC Public Health, 14(1):806.
doi:10.1186/1471-2458-14-806 PMID:25099430
Willett W (1998). Nutritional epidemiology. 2nd edition.
Oxford, UK: Oxford University Press. doi:10.1093/acpr
of:oso/9780195122978.001.0001
Willett W, Hu F (2013). Anthropometric measures and
body composition. In: Willett W, editor. Nutritional
epidemiology. 3rd edition. New York (NY), USA:
Oxford University Press; pp. 213–40.
Willett WC (2002). Dietary fat plays a major role in
obesity: no. Obes Rev, 3(2):59–68. doi:10.1046/j.1467-
789X.2002.00060.x PMID:12120421
Williams SM, Goulding A (2009). Patterns of growth asso-
ciated with the timing of adiposity rebound. Obesity
(Silver Spring), 17(2):335–41. doi:10.1038/oby.2008.547
PMID:19057527
Williamson DF, Madans J, Anda RF, Kleinman JC, Kahn
HS, Byers T (1993). Recreational physical activity and
ten-year weight change in a US national cohort. Int J
Obes Relat Metab Disord, 17(5):279–86. PMID:8389337
69
2. CANCER-PREVENTIVE
EFFECTS IN HUMANS
2.1 Methodological considerations (b) Selection bias

Non-randomized studies are at risk of selec-
Randomized trials addressing body fatness
tion bias, because subjects are not allocated to
and risk of cancer are rare and are often not
groups at random, and instead are generally
feasible. Hence, observational epidemiological
selected based on their disease or exposure status.
studies on various weight parameters are relied on
Therefore, if cases and controls, or exposed and
to provide evidence. Body fatness can be a reflec-
unexposed individuals, are selected systemati-
tion of genetic, metabolic, lifestyle, dietary, envi-
cally in a different way, estimation of the associ-
ronmental, and psychosocial factors. Therefore,
ation between exposure and risk can be affected,
it is important that epidemiological studies are
depending on the study design.
designed appropriately to control for the many
For example, in case–control studies, those
potential confounders. This section reviews some
who agree to participate as controls may be
of the methodological issues in epidemiological
more likely to have a history of being at a healthy
studies that must be carefully considered when
weight, and may be more likely to engage in other
evaluating the body of evidence on the associa-
healthy behaviours, than those who do not agree
tion between body fatness and risk of cancer.
to participate. They may not be representative
of the larger population from which they are
2.1.1 Bias selected, and this can result in an overestimation
(a) Recall bias or underestimation of the association between
body fatness and risk of cancer.
Retrospective studies addressing body fatness
and risk of cancer may rely on participants’ recol- (c) Detection bias
lections of their past weight or other measures.
If there is differential recall between cases and Detection bias refers to systematic differences
controls, or between overweight people and lean between groups in the detection of outcomes of
people, this is considered recall bias. This imbal- interest. Studies of cancers that can be detected
ance can have an impact on estimates of effect, by screening are at higher risk of this bias,
particularly in case–control studies. affecting their estimate of effect. Individuals
who are likely to engage in healthy behaviours,
such as behaviours that lead to maintaining a
healthy weight, may also be more likely to seek
the recommended screening tests. They may
71
therefore be more likely to receive early diagnosis explored as potential confounders when inves-
and to have access to early treatment, which can tigating the association of body fatness with risk
affect their prognosis. If the outcome of interest of cancer. In high-income countries, people with
is mortality, individuals who receive early diag- lower socioeconomic status are more likely to
nosis may be less likely to die from the disease, be overweight or obese. Race and other factors
because of earlier treatment. If these individuals may also be related to body fatness and to risk of
are also more likely to have a lower weight, this cancer, and when the results of epidemiological
could result in an overestimation of the impact studies are evaluated, it is important to consider
of these behaviours. whether such confounders have been adjusted
Similarly, individuals who are less likely to for appropriately.
engage in healthy behaviours, and may be less Tobacco use is strongly associated with a
likely to be at a healthy weight, may also be less higher risk of many cancers. However, smoking
likely to participate in screening and therefore is more common among lean individuals than
will be less likely to receive early diagnosis and among overweight or obese individuals; one
to have access to early treatment. The estimated mechanism that could explain this association
effect of body fatness on the poorer outcomes is that smoking can have an anorectic effect.
in such individuals can be affected by their Smoking must therefore be properly adjusted for
behaviour. This type of bias is of less concern for to ensure that it is not confounding the relation-
cancers that are more likely to be fatal, because ship between body fatness and risk of cancer.
early detection or screening may not have as Current smoking modifies weight gain trajec-
large an effect if the outcome of interest is death. tory; therefore, among former smokers, time
since quitting must be considered when strati-
2.1.2 Confounding fying by smoking status.
Confounding is the result of an association
2.1.3 Reverse causation
between exposures, resulting in the conclusion
that the effect on the risk of disease is due to one Reverse causation occurs when the exposure
variable rather than another. Although the expo- is affected by the outcome, whereas it is usually
sure and the risk of disease are linked, this is due assumed that the outcome is affected by the expo-
to their joint relationship with the confounding sure. The direction of causality must be consid-
variable, rather than due to a direct relationship. ered when evaluating associations between body
Potential confounders can be addressed fatness and risk of cancer. Weight may affect risk
either in the design of studies or in the analysis of cancer, but preclinical cancer can also cause
of the data. In case–control studies, suspected weight loss. Additional chronic diseases that may
confounders can be controlled for by matching affect risk of cancer may also result in weight
on those variables. Similarly, in cohort studies, loss. The timing of measurement must also be
unexposed and exposed groups can be selected considered, because closer to the time of diag-
to be as similar as possible with respect to the nosis, body fatness is more likely to be affected
potential confounders. In the analysis of the by disease.
data, stratification or statistical adjustment can
be used to control for potential confounders.
Individuals who maintain a healthy weight
may be more likely to engage in other healthy
behaviours, so these associations should be
72
2.1.4 Mendelian randomization References

In the absence of large-scale and long-
Lawlor DA, Harbord RM, Sterne JAC, Timpson N, Davey
term randomized controlled trials (RCTs) on Smith G (2008). Mendelian randomization: using
body fatness and risk of cancer, the concept of genes as instruments for making causal inferences in
Mendelian randomization can provide insights epidemiology. Stat Med, 27(8):1133–63. doi:10.1002/
into whether observed associations are causal, sim.3034 PMID:17886233
Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH,
by leveraging the properties of genetic variation Day FR, et al.; LifeLines Cohort Study; ADIPOGen
to overcome limitations present in observational Consortium; AGEN-BMI Working Group;
epidemiological studies. Mendelian randomiz- CARDIOGRAMplusC4D Consortium; CKDGen
Consortium; GLGC; ICBP; MAGIC Investigators;
ation exploits the random allocation of alleles MuTHER Consortium; MIGen Consortium; PAGE
between parents and offspring at conception as Consortium; ReproGen Consortium; GENIE
the basis of natural experiment to strengthen Consortium; International Endogene Consortium
causal inference within the association between (2015). Genetic studies of body mass index yield new
insights for obesity biology. Nature, 518(7538):197–206.
a modifiable exposure and an outcome of interest doi:10.1038/nature14177 PMID:25673413
(Smith & Ebrahim, 2003, 2004; Lawlor et al., Smith GD, Ebrahim S (2003). ‘Mendelian randomization’:
2008). can genetic epidemiology contribute to understanding
The method relies on three main assump- environmental determinants of disease? Int J Epidemiol,
32(1):1–22. doi:10.1093/ije/dyg070 PMID:12689998
tions: the genetic variant (i) is a valid instrument, Smith GD, Ebrahim S (2004). Mendelian randomization:
in that it is reliably associated with the exposure prospects, potentials, and limitations. Int J Epidemiol,
of interest, (ii) is not independently associated 33(1):30–42. doi:10.1093/ije/dyh132 PMID:15075143
with the outcome, except through the exposure
(known as the exclusion restriction criterion), and
(iii) is not associated with any of the confounding
factors that would otherwise distort the observa-
tional association between the exposure and the
outcome. There are several general limitations
to this methodology (for reviews, see Smith &
Ebrahim, 2003, 2004). Importantly, effects of
common genetic variants on the exposure are
small and prone to weak instrumentation if
used alone, which can bias estimates (Smith &
Ebrahim, 2003, 2004). Therefore, using a greater
number of variants included within Mendelian
randomization analyses increases the variance
explained in a given trait and can thus improve
the precision of the causal estimate (Locke et al.,
2015).
73
2.2 Cancer-preventive effects by their results for body mass index (BMI) at base-
organ site line, with comments on findings according
to other measures of body fatness, such as
2.2.1 Cancer of the colorectum weight change over the life-course and waist
circumference.
Colorectal cancer (CRC) accounts for about
10% of all cancer diagnoses and 8.5% of all (i) Body mass index
cancer deaths worldwide (Ferlay et al., 2013). Although findings vary across studies, there
CRC is more common in high-income countries is a general observation of a positive associa-
than in low- and middle-income countries and is tion between BMI and colon cancer risk across
more prevalent in men than in women. It is well most studies, and a much weaker (but still posi-
established that the risk of CRC changes within tive) association between BMI and rectal cancer
one generation after migration from low-inci- risk. In the studies that included both colon
dence areas to high-incidence areas and thus has cancer and rectal cancer, the association with
a strong environmental component. Cancers of BMI for colon cancer was almost always either
the colon and of the rectum, although similar in stronger or of the same magnitude as that for
many ways, have important differences in their rectal cancer. For both colon cancer and rectal
risk factor profiles. Cancers of the rectum seem to cancer, the association with BMI is stronger in
be less associated with dietary factors and more men than in women. The association between
associated with consumption of alcohol (particu- BMI and colon cancer is approximately linear
larly beer). Cancers of the colon arise most often with increasing BMI levels. In a meta-analysis
from colorectal adenomas, and cancers in the of prospective studies (Table 2.2.1c), the relative
proximal colon tend to have a worse prognosis risk per 5 kg/m2 increase in BMI was estimated
than cancers in the distal colon. to be 1.24 in men and 1.09 in women for colon
In 2001, the Working Group of the IARC cancer, and 1.09 in men and 1.02 in women for
Handbook on weight control and physical activity rectal cancer (all P < 0.05, except for rectal cancer
(IARC, 2002) concluded that there was sufficient in women, with P = 0.26) (Renehan et al., 2008).
evidence for a cancer-preventive effect of avoidance Another meta-analysis reported a relative risk
of weight gain for cancer of the colon. The 2007 of CRC for obesity relative to normal weight of
World Cancer Research Fund (WCRF) review 1.53 (95% confidence interval [CI], 1.44–1.62) in
concluded that there was convincing evidence men and 1.25 (95% CI, 1.14–1.37) in women, and
that both body fatness and waist circumfer- an overall increase in CRC risk of 18% (95% CI,
ence were associated with increased risk of CRC 14–21%) per 5 kg/m2 increase in BMI (Ning et al.,
(WCRF/AICR, 2007). The 2007 conclusions were 2010). The most recent meta-analysis of CRC, by
reaffirmed in 2011 (WCRF/AICR, 2011). Results Ma et al. (2013), based on 43 cohorts, estimated
from studies published since 2001 are summa- the relative risk for obesity relative to normal
rized here and in Table 2.2.1a, Table 2.2.1b, and weight to be 1.33 (95% CI, 1.25–1.42).
Table 2.2.1c. In women, an interaction between use of
hormone replacement therapy (HRT) and the
(a) Cohort studies BMI–CRC association has not been found
A total of 39 cohort studies have been consistently in the identified cohort studies that
published since 2001 (excluding analyses that have investigated this (Lin et al., 2004; Adams
were later updated and analyses based on fewer et al., 2007; Wang et al., 2007; Aleksandrova et al.,
than 100 incident cases). Table 2.2.1a summarizes 2013; Kabat et al., 2015). There is not a consistent
74
set of evidence pointing to a differential of the reported on the association of BMI with CRC
BMI association for proximal versus distal colon risk (Table 2.2.1b). In most of the studies, BMI
subsites (Lin et al., 2004; Larsson et al., 2006; was calculated from body height and self-re-
Bassett et al., 2010; Laake et al., 2010; Oxentenko ported body weight for a recent period before
et al., 2010; Hughes et al., 2011; Matsuo et al., 2012; cancer diagnosis; in some of the studies, body
Kitahara et al., 2013). BMI is also associated with weight was measured after diagnosis. Most
risk of colorectal adenomas (Keum et al., 2015). studies showed an increase in risk of cancers of
The BMI–CRC association is observed consist- both the colon and the rectum with increasing
ently in diverse parts of the world (Renehan et al., BMI, and in some studies the association of BMI
2008; Ma et al., 2013). with risk was stronger for colon cancer than for
Several investigators have assessed the asso- rectal cancer. Some, but not all, studies showed
ciation between BMI at different ages or weight more pronounced BMI-associated increases in
gain over the life-course and later colon cancer risk in men than in women, although globally
risk and/or rectal cancer risk. BMI at earlier ages the evidence indicated increases in risk in both
seems to also be related to colon cancer risk (see sexes. A meta-analysis of 12 case–control studies
Section 2.3), but BMI closer to the time of diag- (Ning et al., 2010) found a relative risk of 1.23 for
nosis is more consistently and strongly associated colon and rectal cancers combined, per 5 kg/m2
with risk than is BMI earlier in life (Bassett et al., increase in BMI.
2010; Hughes et al., 2011). Weight gain since age The frequent observation of stronger associa-
18 years has been found to be associated with tions of BMI with colon cancer risk in men than
colon cancer risk in several studies (Thygesen in women has led to the hypothesis that high
et al., 2008; Bassett et al., 2010; Renehan et al., blood levels of estrogens might confer protec-
2012), but it is difficult to separate the effects of tion against colon cancer. To address this issue, a
long-term weight gain from those of the resultant few studies provided results in women stratified
excess adiposity. by estrogen status (determined by menopausal
status and use of HRT). In a study in Germany in
(ii) Waist circumference
postmenopausal women only, a stratified analysis
Several cohorts have included measurements by users and non-users of postmenopausal HRT
of waist circumference. Waist circumference showed a strong association between BMI and
at baseline is about as strongly associated with CRC risk in the non-users only (odds ratio [OR],
risk as is BMI in those studies that used iden- 3.30; 95% CI, 1.25–8.72 for BMI ≥ 30 kg/m2
tical quantile cut-off points for both measures compared with BMI < 23 kg/m2, based on 31 cases
(Table 2.2.1a). The meta-analysis of CRC and in the highest BMI category) and no association
waist circumference by Ma et al. (2013), based on in the ever-users (OR, 0.89; 95% CI, 0.29–2.75)
13 prospective cohort studies, estimated the rela- (Hoffmeister et al., 2007). These findings were
tive risk for the highest versus lowest categories opposite to those from a previous large study
of waist circumference across studies to be 1.46 in the USA, which showed an increase in colon
(95% CI, 1.33–1.60), and no heterogeneity among cancer risk only in estrogen-positive women (i.e.
studies was found (P = 0.323). women who were premenopausal or who were
(b) Case–control studies users of postmenopausal HRT; OR, 2.38; 95%
CI, 1.50–3.77 for BMI > 30 kg/m2 compared
Since 2002, a total of 11 case–control studies, with BMI < 23 kg/m2, based on 77 cases in the
in Australia, Canada, China, Europe, the highest BMI category) compared with no asso-
Republic of Korea, Thailand, and the USA, have ciation in estrogen-negative women (i.e. women
75
who were postmenopausal and were non-users from conventional covariate-adjusted analysis
of HRT; OR, 1.02; 95% CI, 0.71–1.46 for BMI (minimally adjusted OR per 5 kg/m2 increase in
> 30 kg/m2 compared with BMI < 23 kg/m2, BMI, 1.18; 95% CI, 1.15–1.22); however, the 95%
based on 134 cases in the highest BMI category) confidence intervals overlapped and they were
(Slattery et al., 2003). Another study, conducted not statistically significantly different from one
in Shanghai, China, in a relatively lean popul- another (Pdifference = 0.10). In addition, there was a
ation, showed a direct association of BMI with positive association between BMI and CRC risk
colon cancer risk in premenopausal women in women (GRS-related OR per 5 kg/m2 increase
(OR, 2.9; 95% CI, 1.7–8.6 for BMI > 23.6 kg/m2 in BMI, 1.82; 95% CI, 1.26–2.61), and this esti-
compared with BMI < 19.0 kg/m2, based on 62 mate was much greater than that obtained from
cases in the highest BMI category) and an inverse conventional observational analyses (OR, 1.14;
association in postmenopausal women (OR, 0.6; 95% CI, 1.10–1.18; Pdifference = 0.01); although
95% CI, 0.3–0.9 for BMI > 23.6 kg/m2 compared there was no strong evidence from Mendelian
with BMI < 19.0 kg/m2, based on 50 cases in randomization analyses for an association in
the highest BMI category) (Hou et al., 2006). A men (GRS-related OR per 5 kg/m2 increase in
fourth study, in Canada, found an absence of BMI, 1.18 (95% CI, 0.73–1.92), the results were in
association both in “estrogen-positive” women the same direction as in the observational results
and in “estrogen-negative” women (Campbell in the same sample (Pdifference = 0.70). [This discrep-
et al., 2007). ancy between the sexes may be due to sex-specific
With regard to molecular tumour subtypes, residual confounding or measurement error in
Campbell et al. (2010) showed a BMI-associated observational analyses. Alternatively, the distri-
increase in risk for tumours that have a micro- bution of body fat, rather than total body fatness
satellite-stable phenotype (recent BMI, OR (reflected by BMI), may be a more important
per 5 kg/m2 increase, 1.38; 95% CI, 1.24–1.54), predictor of CRC risk for men than for women.]
whereas no association was observed for tumours In the second study, Gao et al. used 15 SNPs
characterized by microsatellite instability (OR, reliably associated with childhood BMI (Felix
1.05; 95% CI, 0.84–1.31) (see Section 4.2.3c). et al., 2016) and 77 SNPs reliably associated with
adult BMI (Locke et al., 2015) as Mendelian
(c) Mendelian randomization studies randomization instruments and assessed their
Two recent studies have applied Mendelian association with CRC risk (Gao et al., 2016).
randomization to assess the association between Mendelian randomization analyses showed an
BMI and CRC risk (Table 2.2.1d). In the first study, 8% increase in risk of CRC with each increase of
Thrift et al. (2015) used a genetic risk score (GRS) 1 kg/m2 in adult BMI [assuming that a standard
derived from 77 single nucleotide polymorphisms deviation was equivalent to 4.5 kg/m2]. There was
(SNPs) associated with higher BMI, identified by no evidence of an association with childhood
the Genetic Investigation of Anthropometric BMI.
Traits (GIANT) consortium, which involved
more than 300 000 individuals of European
descent. In their analysis, higher BMI was associ-
ated with an increased risk of CRC (GRS-related
OR per 5 kg/m2 increase in BMI, 1.50; 95% CI,
1.13–2.01). The point estimate obtained using the
Mendelian randomization approach was greater
in magnitude than the point estimate obtained
76
Table 2.2.1a Cohort studies of measures of body fatness and cancer of the colorectum
Reference Total number of Organ site Exposure Exposed Relative risk Covariates Comments
Cohort subjects categories cases (95% CI)
Location Sex
Follow-up period Incidence/mortality
Terry et al. (2001) 61 463 Colon BMI 291 total Age, education level, Stronger risk
Women in Swedish Women < 22 1.0 alcohol consumption, within the women
mammography Incidence 22–24.2 1.05 (0.72–1.51) diet in age group 40–
programme (ages 24.2–26.7 1.09 (0.77–1.56) 54 yr (Ptrend = 0.06)
40–76 yr) > 26.7 1.21 (0.86–1.70)
Sweden [Ptrend] [0.25]
1987–1998 61 463 Rectum BMI 159 total Age, education level,
Women < 22 1.0 alcohol consumption,
Incidence 22–24.2 0.92 (0.56–1.54) diet
24.2–26.7 1.14 (0.71–1.83)
> 26.7 1.32 (0.83–2.08)
[Ptrend] [0.13]
Terry et al. (2002) 89 835 Colon and BMI 527 total Age, smoking, Association
Women in Women rectum < 18.5–24.9 1.0 education level, stronger in
Canadian Incidence 25–29.9 1.03 (0.84–1.26) physical activity, OC premenopausal
mammography ≥ 30 1.08 (0.82–1.41) use, HRT use, parity ages than
programme (ages [Ptrend] [0.57] postmenopausal
40–59 yr) ages
Canada (Pinteraction = 0.01)
1980–1993
Calle et al. (2003) 404 576 Colon and BMI Age, education level,
Population-based Men rectum 18.5–24.9 1292 1.00 smoking, physical
cohort Mortality 25–29.9 1811 1.20 (1.12–1.30) activity, alcohol

USA 30–34.9 337 1.47 (1.30–1.66) consumption, marital
1982–1998 35–39.9 54 1.84 (1.39–2.41) status, race, aspirin
[Ptrend] [< 0.001] use, fat intake,
vegetable intake
495 477 BMI Additionally adjusted
Women 18.5–24.9 1706 1.00 for HRT use
Mortality 25–29.9 906 1.10 (1.01–1.19)
30–34.9 312 1.33 (1.17–1.51)
35–39.9 67 1.36 (1.06–1.74)
≥ 40 21 1.46 (0.94–2.24)
[Ptrend] [< 0.001]
77
78

Table 2.2.1a (continued)
Location Sex
Lin et al. (2004) 39 876 Colon and BMI Age, study group, Stronger
Women’s Health Women rectum < 23 44 1.0 family history, association with
Study Incidence 23–24.9 45 1.45 (0.96–2.20) history polyps, proximal colon.
USA 25–26.9 31 1.28 (0.81–2.04) physical activity, Similar findings
1993–2002 27–29.9 40 1.72 (1.12–2.66) smoking, aspirin by HRT status
≥ 30 42 1.67 (1.08–2.59) use, consumption in never-users of
[Ptrend] [0.018] of red meat, alcohol HRT. Proximal
consumption, HRT and distal subsites
use similar
MacInnis et al. 16 556 Colon BMI Age, education level,
(2004) Men < 24.8 26 1.0 country of birth
Population-based Incidence 24.8–26.9 37 1.3 (0.8–2.2)
cohort 27–29.2 39 1.4 (0.8–2.3)
Australia ≥ 29.2 51 1.7 (1.1–2.8)
1990–2003 [Ptrend] [0.02]
16 556 WC Age, education level,
Men < 87 22 1.0 country of birth
Incidence 87–93 19 10.8 (0.4–1.4)
93–99.3 48 1.7 (1.0–2.8)
≥ 99.3 64 2.1 (1.3–3.5)
[Ptrend] [< 0.001]
Moore et al. (2004) 3764 Colon BMI Age, sex, education
Framingham Study Men and women 18.5–24.9 67 1.0 level, height, alcohol
cohort aged 30–54 yr at 25–29.9 69 1.3 (0.91–1.8) consumption,
USA baseline ≥ 30 21 1.5 (0.92–2.5) smoking, physical
1948–1999 Incidence activity
WC Age, sex, education Additional
Small 17 1.0 level, height, alcohol adjustment for
Medium 61 1.1 (0.66–2.0) consumption, BMI has no effect
Large 46 1.6 (0.91–2.9) smoking, physical on estimates
Extra large 33 2.0 (1.1–3.7) activity
Location Sex
Moore et al. (2004) 3802 Colon BMI Age, sex, education Associations more
(cont.) Men and women 18.5–24.9 39 1.0 level, height, alcohol evident in men
aged 55–79 yr at 25–29.9 79 1.8 (1.2–2.6) consumption, than in women,
baseline ≥ 30 31 2.4 (1.5–3.9) smoking, physical and stronger in the
Incidence activity proximal site
WC Age, sex, education Adjustment for
Small 11 1.0 level, height, alcohol BMI has no effect
Medium 53 1.4 (0.74–2.7) consumption, on estimates
Large 47 2.1 (1.1–4.0) smoking, physical
Extra large 38 2.6 (1.3–5.2) activity
Samanic et al. 4 500 700 Colon Obesity Age, calendar year Obesity defined
(2004) Men White men: as discharge
United States Incidence Non-obese 16 704 1.00 diagnosis of
Veterans cohort Obese 1420 1.47 (1.39–1.55) obesity: ICD-8:
USA Black men: 277; ICD-9: 278.0
1969–1996 Non-obese 3830 1.00 No significant
Obese 262 1.45 (1.28–1.64) differences in risk
observed between
White and Black
veterans
4 500 700 Rectum Obesity Age, calendar year No significant
Men White men: differences in risk
Incidence Non-obese 9849 1.00 observed between

Obese 719 1.23 (1.14–1.33) White and Black
Black men: veterans
Non-obese 1773 1.00
Obese 93 1.11 (0.90–1.37)
Wei et al. (2004) 46 632 Colon BMI Age, family history,
Nurses’ Health Men < 23 57 1.0 physical activity,
Study Incidence 23–24.9 119 1.33 (0.97–1.83) smoking, diet,
USA 25–29.9 225 1.54 (1.15–2.07) screening history.
1976–2000 ≥ 30 51 1.85 (1.26–2.72) alcohol consumption,
[Ptrend] [0.001] height
79
80

Location Sex
Wei et al. (2004) 87 733 BMI
(cont.) Women < 23 210 1.0
Incidence 23–24.9 141 1.10 (0.88–1.36)
25–29.9 207 1.11 (0.91–1.35)
≥ 30 113 1.28 (1.10–1.62)
[Ptrend] [0.05]
46 632 Rectum BMI Age, family history,
Men < 23 24 1.0 physical activity,
Incidence 23–24.9 42 1.16 (0.70–1.94) smoking, diet,
25–29.9 55 0.93 (0.57–1.53) screening history,
≥ 30 11 1.03 (0.49–2.14) alcohol consumption,
87 733 BMI
Women < 23 56 1.0
Incidence 23–24.9 46 1.37 (0.92–2.02)
25–29.9 68 1.40 (0.98–2.01)
≥ 30 34 1.56 (1.01–2.42)
[Ptrend] [0.04]
Engeland et al. 963 709 Colon and BMI Age at BMI Relationships
(2005) Men rectum < 18.5 90 0.84 (0.68–1.03) measurement, birth similar for colon vs
Population-based Incidence 18.5–24.9 11 432 1.0 cohort rectum
cohort 25–29.9 9953 1.15 (1.12–1.18)
Norway ≥ 30 1512 1.40 (1.32–1.48)
1963–2001 [Ptrend] [< 0.001]
1 038 010 BMI Age at BMI Relationships
Women < 18.5 298 1.04 (0.93–1.17) measurement, birth similar for colon vs
Incidence 18.5–24.9 11 136 1.0 cohort rectum. In women,
25–29.9 8780 1.02 (0.99–1.05) associations
≥ 30 3916 1.06 (1.02–1.10) stronger for colon
[Ptrend] [0.01]
Kuriyama et al. 12 485 Colon and BMI Age, smoking, alcohol
(2005) Men rectum < 18.5–24.9 114 1.00 consumption, diet,
Population-based Incidence 25–27.5 25 1.04 (0.67–1.60) health insurance
prospective cohort 27.5–29.9 11 1.58 (0.85–2.94)
Japan ≥ 30 5 1.78 (0.73–4.38)
1984–1992 [Ptrend] [0.3710]
Location Sex
Kuriyama et al. 15 052 BMI
(2005) Women < 18.5–24.9 73 1.00
(cont.) Incidence 25–27.5 22 1.11 (0.69–1.80)
27.5–29.9 11 1.28 (0.68–2.43)
≥ 30 9 2.06 (1.03–4.13)
[Ptrend] [0.06]
Oh et al. (2005) 781 283 Colon BMI Age, smoking, alcohol
Civil servants and Men (excluding < 18.5 14 1.00 (0.62–1.63) consumption, physical
private school Incidence rectosigmoid) 18.5–22.9 359 1.00 activity, family
workers cohort 23.0–24.9 316 1.24 (1.07–1.43) history, residence area
Republic of Korea 25.0–26.7 190 1.33 (1.13–1.57)
1992–2001 27.0–29.9 63 1.07 (0.83–1.38)
≥ 30 11 1.92 (1.15–3.22)
[Ptrend] [0.001]
781 283 Rectosigmoid BMI Age, smoking, alcohol
Men < 18.5 20 0.64 (0.36–1.13) consumption, physical
Incidence 18.5–22.9 606 1.00 activity, family
23.0–24.9 480 1.06 (0.92–1.22) history, residence area
25.0–26.7 326 1.29 (1.10–1.52)
27.0–29.9 117 1.15 (0.91–1.46)
≥ 30 14 1.08 (0.56–2.10)
[Ptrend] [0.003]
Rapp et al. (2005) 67 447 Colon BMI Age, smoking status,

VHM&PP Men 18.5–24.9 86 1.00 occupational group
(population-based Incidence 25–29.9 128 1.14 (0.86–1.50)
cohort) 30–34.9 39 1.56 (1.06–2.30)
Austria ≥ 35 7 2.48 (1.15–5.39)
1985–2002 [Ptrend] [0.005]
78 484 BMI Age, smoking status,
Women 18.5–24.9 122 1.00 occupational group
Incidence 25–29.9 106 1.13 (0.86–1.47)
30–34.9 35 1.11 (0.76–1.62)
≥ 35 8 0.88 (0.43–1.81)
[Ptrend] [0.73]
81
82

Location Sex
Rapp et al. (2005) 67 447 Rectum BMI Age, smoking status, All obese
(cont.) Men 18.5–24.9 45 1.00 occupational group categories were
Incidence 25–29.9 69 1.20 (0.82–1.75) merged (from BMI
30–34.9 24 1.66 (1.01–2.73) 30 kg/m2 onwards)
≥ 35 – – to ensure at least 5
[Ptrend] [0.05] cases
78 484 BMI Age, smoking status,
Women 18.5–24.9 68 1.00 occupational group
Incidence 25–29.9 48 0.90 (0.62–1.31)
30–34.9 12 0.66 (0.36–1.23)
≥ 35 5 0.96 (0.38–2.39)
[Ptrend] [0.32]
Bowers et al. (2006) 29 133 Colon BMI Age, number of Cohort of smokers
ATBC cohort Men < 18.5 2 1.47 (0.36–5.98) cigarettes smoked per
Finland Incidence 18.5–24.9 77 1.00 day, total cholesterol,
1985–2002 25–29.9 98 1.07 (0.79–1.44) height, type 2 diabetes
≥ 30 50 1.78 (1.25–2.55)
29 133 Rectum BMI
Men < 18.5 1 0.96 (0.13–6.96)
Incidence 18.5–24.9 61 1.0
25–29.9 87 1.18 (0.85–1.64)
≥ 30 34 1.51 (0.99–2.29)
29 133 Colon and BMI
Men rectum < 18.5 3 1.25 (0.40–3.93)
Incidence 18.5–24.9 138 1.0
25–29.9 185 1.12 (0.90–1.39)
≥ 30 84 1.66 (1.27–2.18)
Larsson et al. (2006) 45 906 Colon BMI Age, education Proximal and
Population-based Men < 23 47 1.00 level, family history, distal subsites
cohort Incidence 23–24.9 72 1.11 (0.77–1.61) diabetes, smoking, similar. WC
Sweden 25–26.9 65 1.07 (0.73–1.56) aspirin use, physical also positively
1997–2005 27–29.9 61 1.15 (0.78–1.70) activity associated
≥ 30 39 1.60 (1.03–2.48)
[Ptrend] [0.08]
Location Sex
Larsson et al. (2006) 45 906 Rectum BMI Age, education
(cont.) Men < 23 25 1.00 level, family history,
Incidence 23–24.9 39 1.08 (0.65–1.80) diabetes, smoking,
25–26.9 49 1.35 (0.83–2.19) aspirin use, physical
27–29.9 46 1.53 (0.93–2.51) activity
≥ 30 21 1.44 (0.79–2.61)
[Ptrend] [0.06]
45 906 Colon and WC Age, education
Men rectum < 88 47 1.00 level, family history,
Incidence 88–92 67 1.06 (0.73–1.55) diabetes, smoking,
93–97 95 1.32 (0.92–1.88) aspirin use, physical
98–103 96 1.37 (0.96–1.96) activity
≥ 104 102 1.29 (0.90–1.85)
[Ptrend] [0.03]
Lukanova et al. 33 424 Colon and BMI Age, calendar year, Association with
(2006) Men rectum < 18.5–24.9 45 1.0 smoking obesity significant
Population-based Incidence 25–29.9 69 1.17 (0.80–1.71) only when
cohort ≥ 30 22 1.61 (0.95–2.65) excluding cases
Sweden [Ptrend] [0.08] diagnosed within
1985–2003 35 362 BMI Age, calendar year, 1 yr of recruitment
Women < 18.5–24.9 43 1.0 smoking
Incidence 25–29.9 39 1.27 (0.82–1.97)
≥ 30 26 2.01 (1.22–3.27)

[Ptrend] [0.005]
MacInnis et al. 24 072 Colon BMI, tertiles 212 total Age, education level, No differences
(2006a) Women T1 (< 25) 1.0 country of birth, HRT between proximal
Melbourne Incidence T2 (25–29) 0.8 (0.6–1.1) use and distal, or by
Collaborative T3 (≥ 30) 1.0 (0.7–1.4) disease stage (early
Cohort Study [Ptrend] [0.59] vs late)
Australia WC, tertiles 212 total Age, education level, No differences
1990–2003 T1 (< 75) 1.0 country of birth, HRT between proximal
T2 (75–79) 1.4 (1.0–1.9) use and distal, or by
T3 (≥ 80) 1.4 (1.0–1.9) disease stage (early
[Ptrend] [0.02] vs late)
83
84

Location Sex
MacInnis et al. 16 867 Rectum BMI, tertiles Age, country of birth, Similar results
(2006b) Men < 25 24 1.0 SES, height in women
Population-based Incidence 25–29.9 86 1.7 (1.1–2.7) (n = 24 247), no sex
cohort ≥ 30 24 1.3 (0.8–2.4) interaction
Australia [Ptrend] [0.48]
1990–2003 WC Age, country of birth Similar results
< 94 57 1.0 in women
94–101.9 43 1.3 (0.9–1.9) (n = 24 247), no sex
≥ 102 34 1.4 (0.9–2.2) interaction
[Ptrend] [0.11]
Pischon et al. (2006) 129 731 Colon BMI Age, centre, smoking,
EPIC cohort Men < 23.6 64 1.0 education level,
Europe Incidence 23.6–25.3 85 1.18 (0.85–1.63) alcohol consumption,
1992–2003 25.4–27 74 1.00 (0.71–1.41) physical activity, diet
27.1–29.3 88 1.19 (0.85–1.66)
≥ 29.4 110 1.55 (1.12–2.15)
[Ptrend] [0.006]
WC Age, centre, smoking,
< 86 63 1.00 education level,
86–91.8 57 0.73 (0.50–1.04) alcohol consumption,
91.9–96.5 78 0.97 (0.69–1.36) physical activity, diet,
96.6–102.9 95 1.10 (0.79–1.53) height
≥ 103 125 1.39 (1.01–1.93)
[Ptrend] [0.001]
Location Sex
Pischon et al. (2006) 238 546 Colon BMI Age, centre, smoking,
(cont.) Women < 23.6 87 1.0 education level,
Incidence 23.6–25.3 96 0.92 (0.68–1.23) alcohol consumption,
25.4–27 120 1.02 (0.77–1.35) physical activity, diet
27.1–29.3 137 1.09 (0.83–1.45)
≥ 29.4 135 123 1.06 (0.79–1.42)
[Ptrend] [0.40]
WC Age, centre, smoking,
< 70.2 62 1.0 education level,
70.2–75.8 91 1.10 (0.80–1.52) alcohol consumption,
75.9–80.9 125 1.23 (0.90–1.68) physical activity, diet,
81–88.9 135 1.25 (0.91–1.70) height
≥ 89 149 1.48 (1.08–2.03)
[Ptrend] [0.008]
129 731 Rectum BMI Age, centre, smoking, WC, null
Men < 23.6 52 1.0 education level, association
27.1–29.3 69 1.11 (0.77–1.62)
≥ 29.4 64 1.05 (0.72–1.55)
[Ptrend] [0.47]
238 546 BMI Age, centre, smoking, WC, null
Women < 23.6 47 1.0 education level, association

27.1–29.3 63 0.95 (0.64–1.41)
≥ 29.4 65 1.06 (0.71–1.58)
[Ptrend] [0.51]
Samanic et al. 362 552 Colon BMI Age, year, smoking
(2006) Men 18.5–24.9 763 1.00 status
Swedish Incidence 25–29.9 842 1.24 (1.12–1.37)
Construction ≥ 30 190 1.74 (1.48–2.04)
Worker Cohort [Ptrend] [< 0.001]
Sweden
1971–1999
85
86

Location Sex
Samanic et al. 362 552 Rectum BMI Age, year, smoking
(2006) Men 18.5–24.9 626 1.00 status
(cont.) Incidence 25–29.9 610 1.08 (0.96–1.21)
≥ 30 126 1.36 (1.13–1.66)
[Ptrend] [< 0.01]
Adams et al. (2007) 307 708 Colon BMI Age, alcohol
NIH-AARP cohort Men 18.5–22.9 136 1.0 consumption,
USA Incidence 23–24.9 260 1.11 (0.90–1.37) smoking,
1995–2000 25–27.4 479 1.22 (1.01–1.48) supplemental calcium
27.5–29.9 367 1.44 (1.18–1.76) intake, consumption
30–32.5 219 1.53 (1.23–1.90) of red meat
32.5–34.9 110 1.57 (1.22–2.03)
35–39.9 76 1.71 (1.29–2.27)
≥ 40 29 2.39 (1.59–3.58)
[Ptrend] [< 0.0005]
209 436 BMI Additionally adjusted Similar findings by
Women 18.5–22.9 151 1.0 for HRT use HRT status
Incidence 23–24.9 141 1.20 (0.95–1.51)
25–27.4 172 1.29 (1.03–1.60)
27.5–29.9 106 1.31 (1.01–1.68)
30–32.5 77 1.28 (0.97–1.69)
32.5–34.9 42 1.13 (0.80–1.60)
35–39.9 52 1.46 (1.06–2.02)
≥ 40 28 1.49 (0.98–2.25)
[Ptrend] [0.02]
307 708 Rectum BMI Age, alcohol
Men 18.5–22.9 74 1.0 consumption,
Incidence 23–24.9 101 0.78 (0.58–1.06) smoking,
25–27.4 218 1.01 (0.77–1.31) supplemental calcium
27.5–29.9 135 0.96 (0.72–1.28) intake, consumption
30–32.5 74 0.94 (0.68–1.30) of red meat
32.5–34.9 42 1.10 (0.75–1.61)
≥ 35 33 1.0 (0.68–1.58)
[Ptrend] [0.31]
Location Sex
Adams et al. (2007) 209 436 BMI Additionally adjusted Similar findings by
(cont.) Women 18.5–22.9 60 1.0 for HRT use HRT status
Incidence 23–24.9 49 1.05 (0.72–1.53)
25–27.4 60 1.13 (0.79–1.63)
27.5–29.9 37 1.16 (0.76–1.76)
30–32.5 26 1.09 (0.68–1.75)
32.5–34.9 14 0.95 (0.52–1.71)
≥ 35 32 1.44 (0.92–2.25)
[Ptrend] [0.20]
Driver et al. (2007) 22 071 Colon and BMI Age, smoking,
Physicians’ Health Men rectum < 25 190 1.0 alcohol consumption,
Study Incidence 25–29.9 171 1.26 (1.05–1.52) diabetes, exercise
USA ≥ 30 20 1.62 (1.09–2.42)
1982–2004 [Ptrend] [Ptrend]
Fujino et al. (2007) 46 465 Colon BMI Age, study area Weight at age 20 yr
JACC cohort Men < 18.5 12 0.86 (0.48–1.57) also positively
Japan Incidence 18.5–24.9 155 1.0 associated with
1988–1997 25–29.9 36 1.14 (0.79–1.65) risk
≥ 30 1 0.54 (0.07–3.90)
64 327 BMI Age, study area Weight at age 20 yr
Women < 18.5 14 0.98 (0.56–1.71) also positively
Incidence 18.5–24.9 128 1.0 associated with
25–29.9 42 1.09 (0.77–1.56) risk

≥ 30 8 1.94 (0.94–3.98)
46 465 Rectum BMI Age, study area Weight at age 20 yr
Men < 18.5 6 0.57 (0.25–1.30) also positively
Incidence 18.5–24.9 128 1.0 associated with
25–29.9 21 0.78 (0.49–1.24) risk
≥ 30 2 1.27 (0.31–5.17)
64 321 BMI Age, study area
Women < 18.5 2 0.36 (0.08–1.48)
Incidence 18.5–24.9 58 1.0
25–29.9 19 1.04 (0.62–1.76)
≥ 30 2 1.00 (0.24–4.12)
87
88

Location Sex
Lundqvist et al. 24 821 older twins Colon and BMI Smoking, sex, country, No association
(2007) (mean baseline age, rectum < 18.5 7 1.0 (0.5–2.1) physical activity, with rectal cancer
Twin cohorts 56 yr) 18.5–24.9 274 1.0 education level,
Sweden and Finland 10 804 men and 25–29.9 196 1.1 (0.9–1.3) diabetes
1961–2004 14 017 women ≥ 30 36 1.3 (0.9–1.8)
Incidence [Ptrend] [0.12]
43 328 younger twins BMI Smoking, physical No association
(mean baseline age, < 18.5 4 0.6 (0.2–1.7) activity, education with rectal cancer
30 yr) 18.5–24.9 146 1.0 level, diabetes
20 992 men and 25–29.9 47 1.0 (0.7–1.4)
22 336 women ≥ 30 7 1.1 (0.5–2.5)
Incidence [Ptrend] [0.53]
Reeves et al. (2007) 1.2 million Colon and BMI Age, region, SES,
Population-based Women rectum < 22.5 789 1.02 (0.95–1.10) reproductive history,
cohort Incidence 22.5–24.9 1034 1.00 smoking, alcohol
United Kingdom 25.0–27.4 913 1.04 (0.97–1.11) consumption, physical
1996–2001 27.5–29.9 555 1.01 (0.93–1.10) activity, time since
≥ 30 717 1.01 (0.94–1.09) menopause, HRT use
per 10 kg/m2 1.00 (0.92–1.08)
Wang et al. (2007) 73 842 Colon and BMI Age, education level, Cohort of
Cancer Prevention Women rectum < 18.5–24.9 399 1.0 endoscopy history, postmenopausal
Study II (CPS II) 25–29.9 274 1.08 (0.93–1.27) baseline HRT women
Nutrition Cohort ≥ 30 141 1.19 (0.97–1.45) use, NSAID use, Similar findings by
USA [Ptrend] [0.04] multivitamin use, HRT status (never,
1992–2003 smoking, physical former, current
activity, diabetes use)
Song et al. (2008) 107 481 Colon (above BMI Age, height, smoking, Cohort of
Korean medical Women rectosigmoid < 18.5 11 0.94 (0.37–2.39) alcohol consumption, postmenopausal
insurance cohort Incidence junction) 18.5–20.9 46 1.03 (0.63–1.70) exercise, pay level at women (age
Republic of Korea 21–22.9 86 1.00 study entry 40–64 yr)
1994–2003 23.0–24.9 141 1.69 (1.17–2.44) Results presented
25.0–26.9 129 1.73 (1.18–2.53) are those after
27.0–29.9 64 1.21 (0.77–1.90) excluding patients
≥ 30 32 2.43 (1.40–4.23) diagnosed within
[risk per 1 kg/m2] [1.05 (1.02–1.09)] the first 5 yr of
follow-up
Location Sex
Song et al. (2008) 107 481 Rectum BMI Age, height, smoking,
(cont.) Women (below < 18.5 10 1.00 (0.43–2.33) alcohol consumption,
Incidence rectosigmoid 18.5–20.9 69 1.06 (0.67–1.67) exercise, pay level at
junction) 21–22.9 110 1.00 study entry
23.0–24.9 140 1.26 (0.88–1.81)
25.0–26.9 102 0.94 (0.63–1.40)
27.0–29.9 85 1.62 (1.10–2.38)
≥ 30 20 1.13 (0.57–2.24)
[risk per 1 kg/m2] [1.03 (0.99–1.06)]
Thygesen et al. 46 349 Colon BMI Age, physical activity, Weight gain since
(2008) Men < 20 9 1.69 (0.83–3.44) alcohol consumption, age 21 yr positively
Health Professionals Incidence 20.1–22.5 50 1.0 diet, smoking, aspirin associated
Follow-Up Study 22.6–25 205 1.40 (1.03–1.92) use, family history, with risk. The
USA 25.1–30 341 1.64 (1.21–2.22) prior screening. All association became
1986–2004 30.1–35 75 2.29 (1.58–3.31) confounders were stronger when
> 35 13 2.29 (1.23–4.26) lagged 2 yr 2–4 yr of lag time
for weight change
was allowed
Wang et al. (2008) 44 068 Colon BMI Height, education
Cancer Prevention Men < 18.5–24.9 143 1.0 level, physical activity,
Study II (CPS II) Incidence 25–29.9 179 0.93 (0.75–1.17) smoking, alcohol
Nutrition Cohort 30–34.9 64 1.34 (0.99–1.82) consumption, NSAID
USA ≥ 35 16 1.93 (1.14–3.28) use, multivitamin use,

1997–2005 [Ptrend] [0.01] screening history
WC
< 95 165 1.0
95–105 195 0.95 (0.77–1.17)
105–120 157 1.21 (0.96–1.52)
≥ 120 29 1.68 (1.12–2.53)
[Ptrend] [< 0.006]
89
90

Location Sex
Wang et al. (2008) 51 083 Colon BMI Height, education
(cont.) Women < 18.5–24.9 156 1.0 level, physical activity,
Incidence 25–29.9 97 0.92 (0.71–1.19) smoking, alcohol
30–34.9 44 1.25 (0.88–1.76) consumption, NSAID
≥ 35 17 1.40 (0.84–2.36) use, multivitamin
[Ptrend] [0.18] use, screening history,
HRT use
WC Height, education
< 85 158 1.0 level, physical activity,
85–95 109 1.01 (0.79–1.29) smoking, alcohol
95–110 104 1.27 (0.98–1.64) consumption, NSAID
≥ 110 36 1.75 (1.20–2.54) use, multivitamin
[Ptrend] [0.003] use, screening history,
HRT use
44 068 Rectum BMI Height, education WC, also null
Men < 18.5–24.9 50 1.0 level, physical activity, association
Incidence 25–29.9 63 0.80 (0.55–1.16) smoking, alcohol
30–34.9 23 1.01 (0.61–1.68) consumption, NSAID
≥ 35 6 1.38 (0.58–3.28) use, multivitamin use,
[Ptrend] [0.70] screening history; for
women, also adjusted
for HRT use
51 083 BMI Similar association
Women < 18.5–24.9 37 1.0 with WC
Incidence 25–29.9 31 1.34 (0.82–2.17)
30–34.9 19 2.62 (1.48–4.66)
≥ 35 6 2.67 (1.09–6.54)
[Ptrend] [0.001]
Andreotti et al. 39 628 Colon BMI Race, education level,
(2010) Men < 18.5 1 – family history of colon
Agricultural Incidence 18.5–24.9 44 1.0 cancer
workers 25.0–29.9 112 1.26 (0.86–1.86)
USA 30–34.9 58 1.88 (1.23–2.91)
1993–2005 ≥ 35 15 2.03 (1.05–3.93)
per 1 kg/m2 1.05 (1.02–1.09)
Location Sex
Andreotti et al. 28 319 BMI
(2010) Women < 18.5 1 –
(cont.) Incidence 18.5–24.9 40 1.0
25.0–29.9 49 1.48 (0.97–2.26)
30–34.9 19 1.36 (0.79–2.36)
≥ 35 4 –
per 1 kg/m2 1.00 (0.96–1.04)
[Ptrend] [0.92]
39 628 Rectum BMI Additionally adjusted Results in women
Men < 18.5 0 – for meat consumption not presented due
Incidence 18.5–24.9 23 1.0 to too few incident
25.0–29.9 53 0.96 (0.51–1.82) cases
30–34.9 16 0.60 (0.24–1.50)
≥ 35 10 3.21 (1.34–7.71)
per 1 kg/m2 1.06 (1.00–1.12)
[Ptrend] [0.06]
Bassett et al. (2010) 16 188 Colon BMI Place of birth, BMI at age 18 yr,
Population-based Men < 23 13 0.60 (0.32–1.13) education level, diet, null association.
cohort Incidence 23–24.9 38 1.0 smoking, alcohol Positive
Australia 25.0–29.9 160 1.31 (0.91–1.87) consumption association with
1990–2007 ≥ 30 66 1.51 (1.00–2.28) weight gain
[Ptrend] [< 0.01] since age 18 yr.
Association

stronger for
proximal colon
23 438 BMI BMI at age 18 yr,
Women < 23 64 0.95 (0.67–1.36) null association.
Incidence 23–24.9 59 1.0 Weight gain since
25.0–29.9 102 0.84 (0.61–1.17) age 18 yr, also null
≥ 30 67 1.00 (0.70–1.44) association
[Ptrend] [0.90]
91
92

Location Sex
Laake et al. (2010) 38 822 Colon BMI Age, physical activity, Association
Population-based Men < 18.5–22.9 695 1.0 height, energy intake, stronger for
cohort Incidence 23–24.9 112 1.16 (0.86–1.56) smoking, education distal colon than
Norway 25–27.4 140 1.19 (0.89–1.60) level, county proximal
1974–2005 27.5–29.9 75 1.20 (0.86–1.68)
≥ 30 54 1.80 (1.25–2.59)
[Ptrend] [0.004]
37 357 BMI Age, physical activity, Association
Women < 18.5–22.9 115 1.0 height, energy intake, stronger for distal
Incidence 23–24.9 95 1.05 (0.80–1.38) smoking, education colon
25–27.4 81 1.03 (0.77–1.38) level, county
27.5–29.9 57 1.27 (0.92–1.76)
≥ 30 71 1.48 (1.09–2.02)
[Ptrend] [0.01]
Oxentenko et al. 36 941 Colon and BMI Age, HRT use, OC Proximal and
(2010) Women rectum < 18.5 19 1.62 (0.98–2.66) use, smoking, physical distal subsites
Iowa Women’s Incidence after age 18.5–24.9 495 1.0 activity, diabetes, similar.
Health Study 55 yr 25–29.9 548 1.12 (0.99–1.28) alcohol consumption, Association
USA 30–34.9 272 1.31 (1.12–1.54) diet, calcium intake, stronger for distal
1986–2005 35–39.9 93 1.32 (1.03–1.68) folate intake, vitamin site
≥ 40 37 1.56 (1.10–2.22) E intake
[Ptrend] [< 0.001]
WC, quartiles Age, HRT use, OC Proximal and
Q1 292 1.0 use, smoking, physical distal subsites
Q2 351 1.18 (1.00–1.39) activity, diabetes, similar
Q3 431 1.34 (1.14–1.576) alcohol consumption,
Q4 390 1.32 (1.11–1.56) diet, calcium intake,
[Ptrend] [< 0.001] folate intake, vitamin
E intake
Location Sex
Parr et al. (2010) 424 519 Colon BMI 429 total Age, sex, tobacco use Stronger positive
Pooled analysis of Men and women < 12–18.4 0.63 (0.26–1.56) association in
39 cohort studies Incidence 18.5–24.9 1.0 obese men
Asia, Australia, and 25–29.9 1.13 (0.94–1.36)
New Zealand ≥ 30 1.50 (1.13–1.99)
1961–1999, median [Ptrend] [0.02]
follow-up 4 yr 424 519 Rectum BMI 233 total Age, sex, tobacco use
Men and women < 12–18.4 0.86 (0.37–2.02)
Mortality 18.5–24.9 1.0
25–29.9 1.44 (1.11–1.86)
≥ 30 1.68 (1.06–2.67)
[Ptrend] [0.03]
Hughes et al. (2011) 58 297 Colon and BMI, quintiles Age, diet, occupation, Rectal cancer not
Population-based Men rectum Q1 232 1.0 physical activity, associated with
cohort Incidence Q2 238 0.95 (0.74–1.24) education level, BMI. Proximal
The Netherlands Q3 240 0.99 (0.77–1.28) family history, and distal sites
1986–2002 Q4 247 1.05 (0.81–1.36) alcohol consumption, similar. Stronger
Q5 254 1.25 (0.96–1.62) smoking associations with
[Ptrend] [0.08] distal sites, Ptrend
significant. BMI at
age 20 yr weakly
associated
62 573 BMI, quintiles Age, diet, occupation, BMI at age 20 yr,

Women Q1 228 1.0 physical activity, null association
Incidence Q2 211 0.88 (0.69–1.13) education level, Rectal cancer also
Q3 223 0.94 (0.73–1.20) family history, not associated with
Q4 222 0.91 (0.71–1.16) alcohol consumption, BMI
Q5 222 0.97 (0.76–1.24) smoking
[Ptrend] [0.90]
93
94

Location Sex
Odegaard et al. 51 251 Colon BMI Age, sex, year Significant
(2011) Men and women < 18.5 51 1.23 (0.90–1.68) enrolment, dialect, U-shaped
Singapore Chinese Incidence 18.5–21.4 162 1.17 (0.95–1.45) education level, quadratic
Health Study cohort 21.5–24.4 181 1.0 diabetes, family association
Shanghai, China 24.5–27.4 123 1.12 (0.89–1.43) history, smoking, (Ptrend = 0.014).
1993–2007 ≥ 27.5 79 1.48 (1.13–1.92) alcohol consumption, Stronger
[Ptrend] [0.44] diet, physical activity, association in
sleep duration older subjects
51 251 Rectum BMI Age, sex, year of (> 65 yr) and non-
Men and women < 18.5 25 0.77 (0.50–1.19) enrolment, dialect, smokers
Incidence 18.5–21.4 111 1.04 (0.81–1.34) education level,
21.5–24.4 137 1.0 diabetes, family
24.5–27.4 76 0.95 (0.71–1.25) history, smoking,
≥ 27.5 35 0.93 (0.64–1.36) alcohol consumption,
[Ptrend] [0.92] diet, physical activity,
sleep duration
Matsuo et al. (2012) 157 927 Colon BMI Age, area, smoking, Association
8 population-based Men < 19 98 0.91 (0.70–1.17) alcohol consumption, stronger for
cohorts (pooled) Incidence 19–20.9 317 1.0 (0.85–1.16) diet, physical activity proximal colon
Japan 21–22.9 473 0.87 (0.75–1.00)
1984–2006 23–24.9 512 1.0
25–26.9 319 1.17 (1.01–1.36)
27–29.9 168 1.31 (1.09–1.58)
≥ 30 32 1.47 (0.99–2.18)
[Ptrend] [< 0.001]
183 457 BMI Age, area, smoking, Association
Women < 19 76 0.71 (0.52–0.97) alcohol consumption, stronger for
Incidence 19–20.9 215 0.87 (0.71–1.07) diet, physical activity proximal colon
21–22.9 330 1.00 (0.84–1.19)
23–24.9 512 1.0
25–26.9 217 1.21 (1.02–1.44)
27–29.9 136 1.11 (0.88–1.39)
≥ 30 48 1.18 (0.83–1.68)
[Ptrend] [0.003]
Location Sex
Matsuo et al. (2012) 157 927 Rectum BMI Age, area, smoking,
(cont.) Men < 19 59 0.91 (0.65–1.27) alcohol consumption,
Incidence 19–20.9 179 0.98 (0.80–1.21) diet, physical activity
21–22.9 325 1.12 (0.94–1.33)
23–24.9 284 1.0
25–26.9 158 1.12 (0.91–1.37)
27–29.9 80 1.20 (0.91–1.58)
≥ 30 26 1.57 (0.97–2.53)
[Ptrend] [0.20]
183 457 BMI Age, area, smoking,
Women < 19 53 1.44 (0.99–2.08) alcohol consumption,
Incidence 19–20.9 97 1.12 (0.84–1.50) diet, physical activity
21–22.9 147 1.05 (0.81–1.35)
23–24.9 284 1.0
25–26.9 80 0.88 (0.64–1.20)
27–29.9 54 0.99 (0.70–1.39)
≥ 30 20 1.39 (0.81–2.39)
[Ptrend] [0.785]
Park et al. (2012) 11 166 Colon and BMI Age, sex, smoking, WC, also null
EPIC-Norfolk study Men rectum < 23.9 67 1.00 alcohol consumption, association
cohort Incidence 23.9–25.5 41 0.75 (0.50–1.12) education level,
England 25.5–26.9 30 0.74 (0.48–1.14) exercise, family
1993–2006 27–28.8 32 0.90 (0.58–1.38) history, diet

≥ 28.9 27 0.97 (0.61–1.54)
[Ptrend] [0.85]
13 078 BMI Age, sex, smoking, WC, null
Women < 23.9 34 1.00 alcohol consumption, association
Incidence 23.9–25.5 31 1.20 (0.72–1.98) education level,
25.5–26.9 44 1.87 (1.17–2.99) exercise, family
27–28.8 21 1.10 (0.62–1.93) history, diet
≥ 28.9 30 1.97 (1.18–3.30)
[Ptrend] [0.02]
95
96

Location Sex
Park et al. (2012) 13 078 WC Age, sex, smoking,
(cont.) Women < 73 20 1.00 alcohol consumption,
Incidence 73–78 22 0.86 (0.46–1.62) education level,
78–83.3 30 1.16 (0.65–2.06) exercise, family
83.4–90.4 41 1.52 (0.88–2.62) history, diet
≥ 90.5 47 1.65 (0.97–2.86)
[Ptrend] [0.001]
Renehan et al. 168 294 Colon BMI Age, race, education BMI at ages 18,
(2012) Men < 18.5 6 0.89 (0.39–2.02) level, physical activity, 35, and 50 yr
NIH-AARP cohort Incidence 18.5–21.9 98 1.0 smoking, alcohol shows similar
USA 22.0–22.9 93 0.91 (0.68–1.22) consumption associations as
1995–2006 23.0–24.9 349 1.01 (0.80–1.27) baseline BMI
25.0–27.4 600 1.07 (0.86–1.34) (mean baseline
27.5–29.9 438 1.26 (1.01–1.58) age, 62.8 yr)
30.0–32.4 249 1.29 (1.01–1.64)
32.5–34.9 124 1.33 (1.01–1.75)
≥ 35 113 1.53 (1.16–2.03)
[Ptrend] [< 0.0001]
105 385 BMI Age, race, education BMI at ages
Women < 18.5 14 1.33 (0.76–2.30) level, physical activity, 35 yr and 50 yr
Incidence 18.5–21.9 148 1.0 smoking, alcohol shows similar
22.0–22.9 68 1.00 (0.75–1.34) consumption, HRT associations as
23.0–24.9 176 1.08 (0.87–1.35) use baseline BMI, but
25.0–27.4 207 1.11 (0.89–1.38) BMI at age 18 yr
27.5–29.9 127 1.15 (0.90–1.47) null association
30.0–32.4 82 1.00 (0.76–1.32)
32.5–34.9 54 1.07 (0.78–1.48)
≥ 35 86 1.23 (0.93–1.64)
[Ptrend] [0.20]
Location Sex
Renehan et al. 168 294 Rectum BMI Age, race, education BMI at ages 18,
(2012) Men < 18.5 4 1.63 (0.58–4.59) level, physical activity, 35, and 50 yr
(cont.) Incidence 18.5–21.9 37 1.0 smoking, alcohol shows similar
22.0–22.9 45 1.22 (0.78–1.91) consumption associations as
23.0–24.9 150 1.20 (0.82–1.74) baseline BMI
25.0–27.4 215 1.06 (0.74–1.53) (mean baseline
27.5–29.9 149 1.15 (0.79–1.67) age, 62.8 yr)
30.0–32.4 78 0.99 (0.65–1.49)
32.5–34.9 44 1.22 (0.77–1.92)
≥ 35 40 1.43 (0.90–2.28)
[Ptrend] [0.51]
105 385 BMI Age, race, education BMI at ages 18, 35,
Women < 18.5 6 1.94 (0.82–4.58) level, physical activity, and 50 yr also null
Incidence 18.5–21.9 43 1.0 smoking, alcohol association
22.0–22.9 22 1.15 (0.68–1.93) consumption, HRT
23.0–24.9 50 1.07 (0.71–1.63) use
25.0–27.4 64 1.21 (0.82–1.81)
27.5–29.9 32 1.01 (0.63–1.61)
30.0–32.4 20 0.85 (0.49–1.47)
32.5–34.9 20 1.45 (0.84–2.51)
≥ 35 25 1.28 (0.76–2.16)
[Ptrend] [0.45]

97
98

Location Sex
Aleksandrova et al. 74 091 Colon Weight change from age 20 yr Age, weight at age
(2013) Men Loss 37 0.84 (0.43–1.64) 20 yr, smoking,
EPIC cohort Incidence Stable 67 1.0 education level,
(6 centres) 2–5 kg gain 65 1.20 (0.67–2.14) alcohol consumption,
Europe 5–10 kg gain 122 0.97 (0.58–1.63) physical activity,
1992–2010 10–15 kg gain 127 0.88 (0.53–1.48) consumption of red
15–20 kg gain 114 1.09 (0.65–1.84) meat, fish and shellfish
≥ 20 kg gain 165 1.31 (0.78–2.19) intake, intake of fruits
[Ptrend] [0.13] and vegetables, fibre
127 605 Weight change from age 20 yr intake Similar findings by
Women Loss 70 0.97 (0.56–1.68) HRT status
Incidence Stable 66 1.0
2–5 kg gain 87 1.34 (0.81–2.23)
5–10 kg gain 158 1.07 (0.68–1.69)
10–15 kg gain 139 1.05 (0.65–1.69)
15–20 kg gain 112 1.36 (0.83–2.23)
≥ 20 kg gain 141 1.49 (0.92–2.42)
[Ptrend] [0.05]
74 091 Rectum Weight change from age 20 yr
Men Loss 31 1.15 (0.53–2.49)
2–5 kg gain 48 0.64 (0.30–1.35)
5–10 kg gain 107 1.37 (0.74–2.52)
10–15 kg gain 103 1.28 (0.69–2.35)
15–20 kg gain 72 1.22 (0.65–2.30)
≥ 20 kg gain 91 1.36 (0.73–2.52)
[Ptrend] [0.16]
127 605 Weight change from age 20 yr
Women Loss 32 1.77 (0.84–3.76)
2–5 kg gain 50 2.15 (1.12–4.11)
5–10 kg gain 84 1.34 (0.78–2.31)
10–15 kg gain 88 1.65 (0.93–2.93)
15–20 kg gain 53 1.82 (0.94–3.51)
≥ 20 kg gain 71 1.45 (0.79–2.66)
[Ptrend] [0.96]
Location Sex
Kitahara et al. 36 912 Colon and BMI Age, study centre, Proximal, distal,
(2013) Men rectum < 18.5–24.9 128 1.0 screening history, and rectal
PLCO trial subjects Incidence 25–29.9 270 1.19 (0.96–1.48) race/ethnicity, tobacco associations with
(screening arm) ≥ 30 148 1.48 (1.16–1.89) use, HRT use BMI all similar,
USA [Ptrend] [0.002] but only proximal
1993–2001 significant
37 562 BMI Age, study centre, All subsites null for
Women < 18.5–24.9 156 1.0 screening history, BMI associations
Incidence 25–29.9 154 1.07 (0.86–1.34) race/ethnicity, tobacco
≥ 30 106 1.03 (0.80–1.33) use, HRT use
[Ptrend] [0.74]
Bhaskaran et al. 5 243 978 Colon per 5 kg/m2 13 465 1.10 (1.07–1.13) Age, sex, year, Similar association
(2014) Men and women diabetes, alcohol in never-smokers.
Health system Incidence consumption, Significant sex
clinical database smoking, SES interaction above
United Kingdom 22 kg/m2 (stronger
1987–2012 association in
men)
5 243 978 Rectum per 5 kg/m2 6123 1.04 (1.00–1.08) Similar association
Men and women in never-smokers
Incidence

99
100

Location Sex
Kabat et al. (2015) 143 901 Colon and BMI, quintiles 1908 total Age, alcohol Associations
Women’s Health Women rectum Q1 1.0 consumption, stronger in ever-
Initiative cohort Incidence Q2 1.18 (1.01–1.38) smoking, physical users of HRT
USA Q3 1.15 (0.98–1.38) activity, age at
1992–2013 Q4 1.27 (1.09–1.48) menarche, age at first
Q5 1.44 (1.23–1.68) birth, parity, HRT
[Ptrend] [< 0.0001] use, family history,
WC, quintiles 1908 total ethnicity, education Similar findings by
Q1 1.0 level, aspirin use, HRT status
Q2 1.49 (1.26–1.75) diabetes, treatment
Q3 1.36 (1.15–1.61) allocation
Q4 1.67 (1.41–1.96)
Q5 1.90 (1.61–2.25)
[Ptrend] [< 0.0001]
ATBC, Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study; BMI, body mass index (in kg/m 2); CI, confidence interval; CRC, colorectal cancer; EPIC, European Prospective
Investigation into Cancer and Nutrition; HRT, hormone replacement therapy; JACC, Japan Collaborative Cohort Study for Evaluation of Cancer Risk; NIH-AARP, National Institutes
of Health–AARP Diet and Health Study; NSAID, non-steroidal anti-inflammatory drug; OC, oral contraceptive; PLCO, Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial;
SES, socioeconomic status; VHM&PP, Vorarlberg Health Monitoring and Prevention Program; WC, waist circumference (in cm); yr, year or years
Table 2.2.1b Case–control studies of measures of body fatness and cancer of the colorectum
Reference Total number of Exposure categories Exposed Relative risk Adjustment for Comments
Study location cases cases (95% CI) confounding
Period Source of
controls
Boutron-Ruault et CRC: BMI, quintiles (sex-specific) Age
al. (2001) Men: 109 Men: Women:
France Women: 62 < 22.9 < 20.3 29 1.0
(Burgundy) Population 23–24.4 20.4–22.6 45 1.7 (0.9–3.0)
Period NR 25–25.9 22.7–23.9 23 0.8 (0.4–1.6)
26–28.7 24–26.1 40 1.4 (0.8–2.6)
> 28.7 > 26.1 34 1.1 (0.6–2.1)
[Ptrend] [0.92]
Slattery et al. Colon cancer: BMI Men: Age Additional adjustment for
(2003) Men: 1095 < 23 56 1.00 dietary factors, NSAID
USA (Northern Women: 1286 23–24 119 0.06 (0.64–1.44) use, physical activity level,
California, Utah, Population 25–27 320 1.13 (0.79–1.63) and family history of CRC
Minnesota) 28–30 305 1.54 (1.06–2.23) did not significantly alter
1991–1994 > 30 295 1.88 (1.29–2.74) associations
BMI Women:
< 23 144 1.00
23–24 146 1.22 (0.90–1.65)
25–27 224 1.27 (0.96–1.67)
28–30 152 1.30 (0.96–1.76)
> 30 211 1.45 (1.09–1.92)
BMI in estrogen-positive women
< 23 56 1.00
23–24 60 1.28 (0.81–2.02)

25–27 59 1.09 (0.69–1.73)
28–30 49 1.56 (0.95–2.56)
> 30 77 2.38 (1.50–3.77)
BMI in estrogen-negative women
< 23 88 1.00
23–24 86 1.21 (0.80–1.82)
25–27 165 1.28 (0.90–1.82)
28–30 103 1.10 (0.75–1.62)
> 30 134 1.02 (0.71–1.46)
101
102

Table 2.2.1b (continued)
Period Source of
controls
Pan et al. (2004) Colon cancer: BMI 5-yr age group,
Canada (eight Men: 959 Men: NR province, education
Canadian Women: 768 < 25 1.00 level, smoking, alcohol
provinces), Population 25– < 30 1.54 (1.27–1.86) consumption, total
NECSS study ≥ 30 2.16 (1.68–2.78) energy intake, diet,
1994–1997 [Ptrend] [< 0.0001] recreational physical
Women: NR activity
< 25 1.00 Women only:
25– < 30 1.22 (0.98–1.52) menopausal status,
≥ 30 1.77 (1.35–2.32) number of live births,
[Ptrend] [< 0.0001] age at menarche, age at
end of first pregnancy
Rectal cancer: Men: NR
Men: 858 < 25 1.00
Women: 589 25– < 30 1.41 (1.15–1.71)
Population ≥ 30 1.75 (1.35–2.28)
[Ptrend] [0.0001]
Women: NR
< 25 1.00
25– < 30 1.28 (1.02–1.61)
≥ 30 1.50 (1.11–2.02)
[Ptrend] [0.0045]
Chung et al. CRC: BMI Age, sex, glucose,
(2006) 105 < 22.9 37 1.0 triglycerides,
Republic of Korea Hospital 23.0–24.9 32 1.4 (0.6–3.3) cholesterol
2002–2004 ≥ 25.0 36 2.3 (0.9–5.8)
Hou et al. (2006) Colon cancer: BMI, quintiles Men: Age, education level, In women, a significant
China (Shanghai) Men: 461 < 19.2 80 1.0 family income, marital interaction was observed
1990–1993 Women: 465 19.2–20.3 85 1.0 (0.7–1.4) status, total energy by menopausal status
Population 20.4–21.3 68 1.0 (0.7–1.4) intake, diet (Pinteraction = 0.03)
21.4–22.8 109 1.2 (0.9–1.8) Women only: number
> 22.8 119 1.7 (1.1–2.4) of pregnancies, years of
[Ptrend] [0.005] menstruation
Period Source of
controls
Hou et al. (2006) BMI, quintiles Women:
(cont.) < 19 86 1.0
19.1–20.5 91 1.2 (0.8–1.7)
20.6–21.9 80 0.9 (0.6–1.3)
22.0–23.6 92 1.1 (0.8–1.7)
> 23.6 116 1.4 (1.0–2.1)
[Ptrend] [0.08]
BMI in premenopausal women
< 19 15 1.0
19.1–20.5 19 1.2 (0.6–2.8)
20.6–21.9 20 1.2 (0.3–3.1)
22.0–23.6 24 1.3 (0.6–3.2)
> 23.6 62 2.9 (1.7–8.6)
[Ptrend] [0.01]
BMI in postmenopausal women
< 19 66 1.0
19.1–20.5 72 1.1 (0.6–1.5)
20.6–21.9 58 0.8 (0.5–1.2)
22.0–23.6 71 0.8 (0.6–1.4)
> 23.6 50 0.6 (0.3–0.9)
[Ptrend] [0.03]
Campbell et al. CRC: BMI Men: Age, education level, Associations were
(2007) Men: 1292 18.5–24.99 298 1.0 consumption of red moderately stronger for

Canada Women: 1404 25–29.99 627 1.29 (1.07–1.56) meat, physical activity, colon than rectum.
(Ontario and Population ≥ 30 322 1.80 (1.43–2.27) province of residence, Significant associations
Newfoundland) BMI Women: CRC screening with weight gain since
1997–2003 18.5–24.99 616 1.0 endoscopy, history age 20 yr were observed
25–29.99 443 0.99 (0.83–1.20) of high cholesterol/ in men only (≥ 20 kg vs
≥ 30 260 0.94 (0.75–1.18) triglycerides reference 1–5 kg)
BMI in estrogen-positive women Women only:
menopausal status,
18.5–24.99 260 1.0
use of postmenopausal
25–29.99 148 0.89 (0.66–1.21)
HRT
≥ 30 80 0.67 (0.45–0.98)
103
104

Period Source of
controls
Campbell et al. BMI in estrogen-negative women
(2007) 18.5–24.99 356 1.0
(cont.) 25–29.99 295 1.08 (0.85–1.37)
≥ 30 180 1.05 (0.79–1.40)
Hoffmeister et al. CRC: BMI Age, county of Cohort of postmenopausal
(2007) Women: 208 < 23 51 1.00 residence, history of women
Germany Population 23– < 25 39 0.80 (0.42–1.53) rheumatic disease,
2003–2004 25– < 27 25 0.78 (0.39–1.58) hyperlipidaemia,
27– < 30 46 1.71 (0.89–3.31) former health check-
≥ 30 40 1.82 (0.92–3.62) up, former colorectal
[Ptrend] [0.02] endoscopy, smoking,
BMI in never-users of HRT alcohol consumption,
regular NSAID use, use
< 23 24 1.00
of statins, OC use
23– < 25 31 1.31 (0.55–3.12)
25– < 27 18 1.60 (0.58–4.44)
27– < 30 33 2.76 (1.07–7.12)
≥ 30 31 3.30 (1.25–8.72)
[Ptrend] [0.01]
BMI in ever-users of HRT
< 23 27 1.00
23– < 25 8 0.49 (0.16–1.48)
25– < 27 7 0.36 (0.11–1.13)
27– < 30 13 1.18 (0.40–3.48)
≥ 30 9 0.89 (0.29–2.75)
[Ptrend] [0.96]
Sriamporn et al. CRC: BMI Age, sex, place of
(2007) 253 < 25 34 1 residence
North-eastern Hospital ≥ 25 0.5 (0.3–0.8)
Thailand [Ptrend] [< 0.5]
2002–2006
Period Source of
controls
Campbell et al. CRC: BMI Age, endoscopy Only microsatellite stable
(2010) Men: 877 Women: screening, smoking tumours showed increased
Canada Women: 917 < 18.5 24 1.77 (0.91–3.45) Women only: risk at higher BMI
(Ontario and Sibling controls 18.5–24.99 404 1.00 postmenopausal HRT
Newfoundland) 25–29.99 252 1.00 (0.80–1.25) use
1997–2003 ≥ 30 212 1.34 (1.03–1.75)
per 5 kg/m2 1.20 (1.10–1.32)
[Ptrend] [< 0.001]
Men:
< 18.5 2 0.51 (0.09–2.89)
18.5–24.99 223 1.00
25–29.99 408 1.33 (1.06–1.68)
≥ 30 222 1.79 (1.33–2.40)
per 5 kg/m2 1.30 (1.15–1.47)
[Ptrend] [< 0.001]
Adult weight change
Women:
Loss 94 0.70 (0.049–1.00)
0–5 kg gain 158 1.00
6–10 kg gain 155 0.88 (0.64–1.20)
11–20 kg gain 249 0.93 (0.70–1.23)
≥ 21 kg gain 229 1.08 (0.80–1.47)
per 5 kg 1.06 (1.01–1.12)

[Ptrend] [< 0.01]
Men:
Loss 104 1.40 (0.95–2.06)
0–5 kg gain 93 1.00
6–10 kg gain 143 1.47 (1.05–2.07)
11–20 kg gain 257 1.72 (1.25–2.36)
≥ 21 kg gain 233 2.23 (1.58–3.14)
per 5 kg 1.08 (1.03–1.14)
[Ptrend] [0.003]
105
106

Period Source of
controls
Choe et al. (2013) CRC: BMI, quartiles NR Current smoking No significant associations
Republic of Korea 153 (stage I) Q1 1.0 status, alcohol were observed when
(Seoul) Hospital Q2 0.81 (0.48–1.38) consumption comparing CRC risk vs
2004–2008 Q3 1.32 (0.80–2.19) colorectal adenoma (554
Q4 1.58 (0.95–2.63) cases in total) across
quartiles of BMI
Boyle et al. (2014) CRC: BMI at age 20 yr NR Age group, sex, SES, No differences in
Australia 918 Normal 1.00 energy intake, lifetime associations were observed
2005–2007 Population Overweight 1.25 (0.92–1.71) vigorous recreational with BMI at age 40 yr
Obese 0.89 (0.44–1.77) physical activity,
[Ptrend] [0.401] alcohol consumption,
tobacco use, diabetes
BMI, body mass index (in kg/m 2); CI, confidence interval; CRC, colorectal cancer; HRT, hormone replacement therapy; NECSS, National Enhanced Cancer Surveillance System; NR, not
reported; NSAID, non-steroidal anti-inflammatory drug; OC, oral contraceptive; SES, socioeconomic status; yr, year or years
Table 2.2.1c Meta-analyses of measures of body fatness and cancer of the colorectum
Reference Total number of studies Organ Exposure categories Relative risk Adjustment for confounding
Total number of cases site (95% CI)
Moghaddam et al. 31 studies (23 cohort Colon and ΒΜΙ Age (all studies) and other factors (not
(2007) studies, 8 case–control rectum ≥ 30 vs < 25 1.35 (1.24–1.46) in all studies): sex, diabetes, smoking,
studies) alcohol consumption, hypertension,
70 906 cases (49% women) hypercholesterolaemia, medication, race, family
8 cohort studies Colon and WC history, physical activity, diet, education level,
N/A rectum Highest vs lowest 1.50 (1.35–1.67) SES, pregnancy (for women), menstruation (for
category women), study centre
Renehan et al. (2008) 22 prospective studies in Colon BMI Age (all studies) and other factors (not in all
men per 5 kg/m2 increase 1.24 (1.20–1.28) studies): family history, inflammatory bowel
22 440 incident cases disease, Western diet, increased weight,
19 prospective studies in Colon BMI alcohol consumption, previous CRC, medical
women per 5 kg/m2 increase 1.09 (1.05–1.12) conditions (e.g. type 2 diabetes, acromegaly),
20 975 incident cases intake of fruits and vegetables, fat intake,
vitamin D and calcium intake, physical activity,
aspirin use, HRT use
18 prospective studies in Rectum BMI Age (all studies) and other factors (not in all
men per 5 kg/m2 increase 1.09 (1.06–1.12) studies): family history, inflammatory bowel
14 894 incident cases disease, Western diet, increased weight,
14 prospective studies in Rectum BMI alcohol consumption, previous CRC, medical
women per 5 kg/m2 increase 1.02 (1.00–1.05) conditions (e.g. type 2 diabetes, acromegaly),
9052 incident cases intake of fruits and vegetables, fat intake,
vitamin D and calcium intake, physical activity,
aspirin use, HRT use
Ning et al. (2010) 51 studies (39 prospective Colon and BMI Cancer site, sex, menopausal status (for women),
and 12 retrospective) rectum per 5 kg/m2 increase 1.18 (1.14–1.21) directly measured BMI or self-reported BMI,

93 812 cases and adjustment for physical activity
Ma et al. (2013) 41 prospective studies Colon and ΒΜΙ Age (36 studies), smoking (32 studies), physical
85 935 cases rectum ≥ 30 vs < 25 1.33 (1.25–1.42) activity (23 studies), alcohol consumption (23
13 prospective studies Colon and WC studies). Fewer adjusted for energy intake (9
6546 cases rectum Highest vs lowest 1.46 (1.33–1.60) studies), NSAID/aspirin use (8 studies), folate
category intake (7 studies), calcium intake (6 studies),
diabetes (6 studies)
BMI, body mass index (in kg/m 2); CI, confidence interval; CRC, colorectal cancer; HRT, hormone replacement therapy; N/A, not applicable; NSAID, non-steroidal anti-inflammatory
drug; SES, socioeconomic status; WC, waist circumference
107
108

Table 2.2.1d Mendelian randomization studies of measures of body fatness and cancer of the colorectum
Reference Characteristics of study Sample size Exposure (unit) Odds ratio Adjustment for Comments
Study population (95% CI) confounding
Thrift et al. (2015) 11 studies of individuals 20 512 Weighted All: 1.50 (1.13–2.01) Study, and the top
Genetics and of European descent (10 226 cases genetic risk score Men: 1.18 (0.73–1.92) three principal
Epidemiology of (6 cohort and 5 case– and 10 286 representing Women: 1.82 (1.26–2.61) components of
Colorectal Cancer control) controls) an increase of ancestry
Consortium (GECCO) 5 kg/m2 in BMI
Gao et al. (2016) 6 studies of individuals of 9931 (5100 Increase of 1 SD Childhood BMI: N/A Waist-to-hip ratio,
Genetic Associations European ancestry cases and 4831 in genetically 1.20 (0.90–1.59) null association:
and Mechanisms in controls) predicted Adult BMI: 1.29 (0.75–2.22)
Oncology (GAME-ON) childhood BMI or 1.39 (1.06–1.82)
Consortium adult BMI
BMI, body mass index (in kg/m 2); CI, confidence interval; N/A, not applicable; SD, standard deviation
References Cancer Epidemiol Biomarkers Prev, 16(9):1735–44.

doi:10.1158/1055-9965.EPI-06-1059 PMID:17855691
Campbell PT, Jacobs ET, Ulrich CM, Figueiredo JC,
Adams KF, Leitzmann MF, Albanes D, Kipnis V, Mouw T, Poynter JN, McLaughlin JR, et al.; Colon Cancer
Hollenbeck A, et al. (2007). Body mass and colorectal Family Registry (2010). Case-control study of over-
cancer risk in the NIH-AARP cohort. Am J Epidemiol, weight, obesity, and colorectal cancer risk, overall
166(1):36–45. doi:10.1093/aje/kwm049 PMID:17449892 and by tumor microsatellite instability status. J Natl
Aleksandrova K, Pischon T, Buijsse B, May AM, Peeters Cancer Inst, 102(6):391–400. doi:10.1093/jnci/djq011
PH, Bueno-de-Mesquita HB, et al. (2013). Adult PMID:20208017
weight change and risk of colorectal cancer in the Choe EK, Kim D, Kim HJ, Park KJ (2013). Association of
European Prospective Investigation into Cancer and visceral obesity and early colorectal neoplasia. World
Nutrition. Eur J Cancer, 49(16):3526–36. doi:10.1016/j. J Gastroenterol, 19(45):8349–56. doi:10.3748/wjg.v19.
ejca.2013.06.021 PMID:23867126 i45.8349 PMID:24363527
Andreotti G, Hou L, Beane Freeman LE, Mahajan R, Chung YW, Han DS, Park YK, Son BK, Paik CH, Lee HL,
Koutros S, Coble J, et al. (2010). Body mass index, et al. (2006). Association of obesity, serum glucose and
agricultural pesticide use, and cancer incidence in lipids with the risk of advanced colorectal adenoma
the Agricultural Health Study cohort. Cancer Causes and cancer: a case-control study in Korea. Dig Liver
Control, 21(11):1759–75. doi:10.1007/s10552-010-9603-9 Dis, 38(9):668–72. doi:10.1016/j.dld.2006.05.014
PMID:20730623 PMID:16790371
Bassett JK, Severi G, English DR, Baglietto L, Krishnan Driver JA, Gaziano JM, Gelber RP, Lee IM, Buring
K, Hopper JL, et al. (2010). Body size, weight change, JE, Kurth T (2007). Development of a risk score for
and risk of colon cancer. Cancer Epidemiol Biomarkers colorectal cancer in men. Am J Med, 120(3):257–63.
Prev, 19(11):2978–86. doi:10.1158/1055-9965.EPI-10- doi:10.1016/j.amjmed.2006.05.055 PMID:17349449
0543 PMID:20870733 Engeland A, Tretli S, Austad G, Bjørge T (2005). Height
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, and body mass index in relation to colorectal and
Leon DA, Smeeth L (2014). Body-mass index and risk gallbladder cancer in two million Norwegian men
of 22 specific cancers: a population-based cohort study and women. Cancer Causes Control, 16(8):987–96.
of 5.24 million UK adults. Lancet, 384(9945):755–65. doi:10.1007/s10552-005-3638-3 PMID:16132807
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 Felix JF, Bradfield JP, Monnereau C, van der Valk RJ,
Boutron-Ruault MC, Senesse P, Méance S, Belghiti C, Stergiakouli E, Chesi A, et al.; Bone Mineral Density
Faivre J (2001). Energy intake, body mass index, in Childhood Study (BMDCS); Early Genetics and
physical activity, and the colorectal adenoma-carci- Lifecourse Epidemiology (EAGLE) consortium; Early
noma sequence. Nutr Cancer, 39(1):50–7. doi:10.1207/ Growth Genetics (EGG) Consortium; Bone Mineral
S15327914nc391_7 PMID:11588902 Density in Childhood Study BMDCS (2016). Genome-
Bowers K, Albanes D, Limburg P, Pietinen P, Taylor PR, wide association analysis identifies three new suscep-
Virtamo J, et al. (2006). A prospective study of anthro- tibility loci for childhood body mass index. Hum
pometric and clinical measurements associated with Mol Genet, 25(2):389–403. doi:10.1093/hmg/ddv472
insulin resistance syndrome and colorectal cancer PMID:26604143
in male smokers. Am J Epidemiol, 164(7):652–64. Ferlay J, Soerjomataram I, Ervik M, Dikshit R, Eser S,
doi:10.1093/aje/kwj253 PMID:16877536 Mathers C, et al. (2013). GLOBOCAN 2012 v1.0, cancer
Boyle T, Fritschi L, Tabatabaei SM, Ringwald K, Heyworth incidence and mortality worldwide: IARC CancerBase
JS (2014). Smoking, alcohol, diabetes, obesity, socio- No. 11 [Internet]. Lyon, France: International Agency
economic status, and the risk of colorectal cancer in for Research on Cancer. Available from: http://
a population-based case-control study. Cancer Causes globocan.iarc.fr.
Control, 25(12):1659–68. doi:10.1007/s10552-014- Fujino Y; Japan Collaborative Cohort Study for Evaluation
0470-7 PMID:25301194 of Cancer (2007). Anthropometry, development history
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ and mortality in the Japan Collaborative Cohort Study
(2003). Overweight, obesity, and mortality from cancer for Evaluation of Cancer (JACC). Asian Pac J Cancer
in a prospectively studied cohort of U.S. adults. N Engl Prev, 8(Suppl):105–12. PMID:18260709
J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423 Gao C, Patel CJ, Michailidou K, Peters U, Gong J,
PMID:12711737 Schildkraut J, et al.; Colorectal Transdisciplinary Study
Campbell PT, Cotterchio M, Dicks E, Parfrey P, Gallinger (CORECT); Discovery, Biology and Risk of Inherited
S, McLaughlin JR (2007). Excess body weight and Variants in Breast Cancer (DRIVE); Elucidating Loci
colorectal cancer risk in Canada: associations in Involved in Prostate Cancer Susceptibility (ELLIPSE);
subgroups of clinically defined familial risk of cancer. Follow-up of Ovarian Cancer Genetic Association and
Interaction Studies (FOCI); and Transdisciplinary
109
Research in Cancer of the Lung (TRICL) (2016). a cohort of Swedish men. Eur J Cancer, 42(15):2590–7.
Mendelian randomization study of adiposity-related doi:10.1016/j.ejca.2006.04.015 PMID:16914307
traits and risk of breast, ovarian, prostate, lung and Lin J, Zhang SM, Cook NR, Rexrode KM, Lee
colorectal cancer. Int J Epidemiol, 45(3):896–908. I-M, Buring JE (2004). Body mass index and risk
doi:10.1093/ije/dyw129 PMID:27427428 of colorectal cancer in women (United States).
Hoffmeister M, Raum E, Winter J, Chang-Claude J, Cancer Causes Control, 15(6):581–9. doi:10.1023/B:-
Brenner H (2007). Hormone replacement therapy, CACO.0000036168.23351.f1 PMID:15280637
body mass, and the risk of colorectal cancer among Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH,
postmenopausal women from Germany. Br J Day FR, et al.; LifeLines Cohort Study; ADIPOGen
Cancer, 97(11):1486–92. doi:10.1038/sj.bjc.6604066 Consortium; AGEN-BMI Working Group;
PMID:17987040 CARDIOGRAMplusC4D Consortium; CKDGen
Hou L, Ji BT, Blair A, Dai Q, Gao YT, Potter JD, et al. Consortium; GLGC; ICBP; MAGIC Investigators;
(2006). Body mass index and colon cancer risk in MuTHER Consortium; MIGen Consortium; PAGE
Chinese people: menopause as an effect modifier. Eur Consortium; ReproGen Consortium; GENIE
J Cancer, 42(1):84–90. doi:10.1016/j.ejca.2005.09.014 Consortium; International Endogene Consortium
PMID:16321519 (2015). Genetic studies of body mass index yield new
Hughes LAE, Simons CCJM, van den Brandt PA, insights for obesity biology. Nature, 518(7538):197–206.
Goldbohm RA, van Engeland M, Weijenberg MP (2011). doi:10.1038/nature14177 PMID:25673413
Body size and colorectal cancer risk after 16.3 years of Lukanova A, Björ O, Kaaks R, Lenner P, Lindahl B,
follow-up: an analysis from the Netherlands Cohort Hallmans G, et al. (2006). Body mass index and cancer:
Study. Am J Epidemiol, 174(10):1127–39. doi:10.1093/ results from the Northern Sweden Health and Disease
aje/kwr247 PMID:21984660 Cohort. Int J Cancer, 118(2):458–66. doi:10.1002/
IARC (2002). Weight control and physical activity. Lyon, ijc.21354 PMID:16049963
France: IARC Press (IARC Handbooks of Cancer Lundqvist E, Kaprio J, Verkasalo PK, Pukkala E,
Prevention, Vol. 6). Available from: http://publications. Koskenvuo M, Söderberg KC, et al. (2007). Co-twin
iarc.fr/376. control and cohort analyses of body mass index and
Kabat GC, Xue X, Kamensky V, Lane D, Bea JW, Chen height in relation to breast, prostate, ovarian, corpus
C, et al. (2015). Risk of breast, endometrial, colorectal, uteri, colon and rectal cancer among Swedish and
and renal cancers in postmenopausal women in asso- Finnish twins. Int J Cancer, 121(4):810–8. doi:10.1002/
ciation with a body shape index and other anthropo- ijc.22746 PMID:17455257
metric measures. Cancer Causes Control, 26(2):219–29. Ma Y, Yang Y, Wang F, Zhang P, Shi C, Zou Y, et al. (2013).
doi:10.1007/s10552-014-0501-4 PMID:25430815 Obesity and risk of colorectal cancer: a systematic
Keum N, Lee DH, Kim R, Greenwood DC, Giovannucci review of prospective studies. PLoS One, 8(1):e53916.
EL (2015). Visceral adiposity and colorectal adenomas: doi:10.1371/journal.pone.0053916 PMID:23349764
dose-response meta-analysis of observational studies. MacInnis RJ, English DR, Haydon AM, Hopper JL, Gertig
Ann Oncol, 26(6):1101–9. doi:10.1093/annonc/mdu563 DM, Giles GG (2006b). Body size and composition
PMID:25480876 and risk of rectal cancer (Australia). Cancer Causes
Kitahara CM, Berndt SI, de González AB, Coleman Control, 17(10):1291–7. doi:10.1007/s10552-006-0074-y
HG, Schoen RE, Hayes RB, et al. (2013). Prospective PMID:17111261
investigation of body mass index, colorectal MacInnis RJ, English DR, Hopper JL, Gertig DM, Haydon
adenoma, and colorectal cancer in the prostate, lung, AM, Giles GG (2006a). Body size and composition and
colorectal, and ovarian cancer screening trial. J Clin colon cancer risk in women. Int J Cancer, 118(6):1496–
Oncol, 31(19):2450–9. doi:10.1200/JCO.2012.48.4691 500. doi:10.1002/ijc.21508 PMID:16187280
PMID:23715565 MacInnis RJ, English DR, Hopper JL, Haydon AM, Gertig
Kuriyama S, Tsubono Y, Hozawa A, Shimazu T, Suzuki Y, DM, Giles GG (2004). Body size and composition and
Koizumi Y, et al. (2005). Obesity and risk of cancer in colon cancer risk in men. Cancer Epidemiol Biomarkers
Japan. Int J Cancer, 113(1):148–57. doi:10.1002/ijc.20529 Prev, 13(4):553–9. PMID:15066919
PMID:15386435 Matsuo K, Mizoue T, Tanaka K, Tsuji I, Sugawara Y,
Laake I, Thune I, Selmer R, Tretli S, Slattery ML, Veierød Sasazuki S, et al.; Development and Evaluation
MB (2010). A prospective study of body mass index, of Cancer Prevention Strategies in Japan (2012).
weight change, and risk of cancer in the proximal Association between body mass index and the
and distal colon. Cancer Epidemiol Biomarkers Prev, colorectal cancer risk in Japan: pooled analysis
19(6):1511–22. doi:10.1158/1055-9965.EPI-09-0813 of population-based cohort studies in Japan. Ann
PMID:20501754 Oncol, 23(2):479–90. doi:10.1093/annonc/mdr143
Larsson SC, Rutegård J, Bergkvist L, Wolk A (2006). Physical PMID:21597097
activity, obesity, and risk of colon and rectal cancer in
110
Moghaddam AA, Woodward M, Huxley R (2007). large cohort study of over 145,000 adults in Austria.
Obesity and risk of colorectal cancer: a meta-analysis Br J Cancer, 93(9):1062–7. doi:10.1038/sj.bjc.6602819
of 31 studies with 70,000 events. Cancer Epidemiol PMID:16234822
Biomarkers Prev, 16(12):2533–47. doi:10.1158/1055- Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D;
9965.EPI-07-0708 PMID:18086756 Million Women Study Collaboration (2007). Cancer
Moore LL, Bradlee ML, Singer MR, Splansky GL, Proctor incidence and mortality in relation to body mass
MH, Ellison RC, et al. (2004). BMI and waist circum- index in the Million Women Study: cohort study. BMJ,
ference as predictors of lifetime colon cancer risk in 335(7630):1134. doi:10.1136/bmj.39367.495995.AE
Framingham Study adults. Int J Obes Relat Metab PMID:17986716
Disord, 28(4):559–67. doi:10.1038/sj.ijo.0802606 Renehan AG, Flood A, Adams KF, Olden M, Hollenbeck
PMID:14770200 AR, Cross AJ, et al. (2012). Body mass index at different
Ning Y, Wang L, Giovannucci EL (2010). A quantitative adult ages, weight change, and colorectal cancer risk in
analysis of body mass index and colorectal cancer: the National Institutes of Health-AARP Cohort. Am
findings from 56 observational studies. Obes Rev, J Epidemiol, 176(12):1130–40. doi:10.1093/aje/kws192
11(1):19–30. doi:10.1111/j.1467-789X.2009.00613.x PMID:23186750
PMID:19538439 Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen
Odegaard AO, Koh W-P, Yu MC, Yuan J-M (2011). Body M (2008). Body-mass index and incidence of cancer:
mass index and risk of colorectal cancer in Chinese a systematic review and meta-analysis of prospec-
Singaporeans: the Singapore Chinese Health Study. tive observational studies. Lancet, 371(9612):569–78.
Cancer, 117(16):3841–9. doi:10.1002/cncr.25936 doi:10.1016/S0140-6736(08)60269-X PMID:18280327
PMID:21815135 Samanic C, Chow WH, Gridley G, Jarvholm B,
Oh SW, Yoon YS, Shin SA (2005). Effects of excess Fraumeni JF Jr (2006). Relation of body mass index
weight on cancer incidences depending on cancer to cancer risk in 362,552 Swedish men. Cancer Causes
sites and histologic findings among men: Korea Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9
National Health Insurance Corporation Study. J Clin PMID:16841257
Oncol, 23(21):4742–54. doi:10.1200/JCO.2005.11.726 Samanic C, Gridley G, Chow W-H, Lubin J, Hoover
PMID:16034050 RN, Fraumeni JF Jr (2004). Obesity and cancer
Oxentenko AS, Bardia A, Vierkant RA, Wang AH, risk among white and black United States veterans.
Anderson KE, Campbell PT, et al. (2010). Body size Cancer Causes Control, 15(1):35–43. doi:10.1023/B:-
and incident colorectal cancer: a prospective study of CACO.0000016573.79453.ba PMID:14970733
older women. Cancer Prev Res (Phila), 3(12):1608–20. Slattery ML, Ballard-Barbash R, Edwards S, Caan BJ,
doi:10.1158/1940-6207.CAPR-10-0116 PMID:20719902 Potter JD (2003). Body mass index and colon cancer:
Pan SY, Johnson KC, Ugnat AM, Wen SW, Mao Y; an evaluation of the modifying effects of estrogen
Canadian Cancer Registries Epidemiology Research (United States). Cancer Causes Control, 14(1):75–84.
Group(2004). Association of obesity and cancer risk in doi:10.1023/A:1022545017867 PMID:12708728
Canada. Am J Epidemiol, 159(3):259–68. doi:10.1093/ Song Y-M, Sung J, Ha M (2008). Obesity and risk of
aje/kwh041 PMID:14742286 cancer in postmenopausal Korean women. J Clin
Park JY, Mitrou PN, Keogh RH, Luben RN, Wareham NJ, Oncol, 26(20):3395–402. doi:10.1200/JCO.2007.15.7867
Khaw K-T (2012). Self-reported and measured anthro- PMID:18612154
pometric data and risk of colorectal cancer in the EPIC- Sriamporn S, Wiangnon S, Suwanrungruang K,
Norfolk study. Int J Obes, 36(1):107–18. doi:10.1038/ Rungsrikaji D, Sukprasert A, Thipsuntornsak N, et al.
ijo.2011.61 PMID:21427695 (2007). Risk factors for colorectal cancer in northeast
Parr CL, Batty GD, Lam TH, Barzi F, Fang X, Ho SC, et al.; Thailand: lifestyle related. Asian Pac J Cancer Prev,
Asia-Pacific Cohort Studies Collaboration (2010). 8(4):573–7. PMID:18260731
Body-mass index and cancer mortality in the Asia- Terry P, Giovannucci E, Bergkvist L, Holmberg L, Wolk
Pacific Cohort Studies Collaboration: pooled analyses A (2001). Body weight and colorectal cancer risk in a
of 424,519 participants. Lancet Oncol, 11(8):741–52. cohort of Swedish women: relation varies by age and
doi:10.1016/S1470-2045(10)70141-8 PMID:20594911 cancer site. Br J Cancer, 85(3):346–9. doi:10.1054/
Pischon T, Lahmann PH, Boeing H, Friedenreich C, Norat bjoc.2001.1894 PMID:11487263
T, Tjønneland A, et al. (2006). Body size and risk of Terry PD, Miller AB, Rohan TE (2002). Obesity and
colon and rectal cancer in the European Prospective colorectal cancer risk in women. Gut, 51(2):191–4.
Investigation Into Cancer and Nutrition (EPIC). J Natl doi:10.1136/gut.51.2.191 PMID:12117878
Cancer Inst, 98(13):920–31. doi:10.1093/jnci/djj246 Thrift AP, Gong J, Peters U, Chang-Claude J, Rudolph A,
PMID:16818856 Slattery ML, et al. (2015). Mendelian randomization
Rapp K, Schroeder J, Klenk J, Stoehr S, Ulmer H, Concin study of body mass index and colorectal cancer risk.
H, et al. (2005). Obesity and incidence of cancer: a
111
Cancer Epidemiol Biomarkers Prev, 24(7):1024–31.

doi:10.1158/1055-9965.EPI-14-1309 PMID:25976416
Thygesen LC, Grønbaek M, Johansen C, Fuchs CS, Willett
WC, Giovannucci E (2008). Prospective weight change
and colon cancer risk in male US health professionals.
Int J Cancer, 123(5):1160–5. doi:10.1002/ijc.23612
PMID:18546286
Wang Y, Jacobs EJ, Patel AV, Rodríguez C, McCullough
ML, Thun MJ, et al. (2008). A prospective study of
waist circumference and body mass index in rela-
tion to colorectal cancer incidence. Cancer Causes
Control, 19(7):783–92. doi:10.1007/s10552-008-9141-x
PMID:18322811
Wang Y, Jacobs EJ, Teras LR, Pavluck AL, Rodriguez C,
Thun MJ, et al. (2007). Lack of evidence for effect modi-
fication by estrogen of association between body mass
index and colorectal cancer risk among postmeno-
pausal women. Cancer Causes Control, 18(8):793–9.
doi:10.1007/s10552-007-9009-5 PMID:17619155
WCRF/AICR (2007). Food, nutrition, physical activity,
and the prevention of cancer: a global perspective.
Washington (DC), USA: American Institute for Cancer
Research. Available from: http://www.aicr.org/assets/
docs/pdf/reports/Second_Expert_Report.pdf.
WCRF/AICR (2011). Continuous Update Project Report.
Food, nutrition, physical activity, and the prevention of
colorectal cancer. Washington (DC), USA: American
Institute for Cancer Research. Available from: https://
www.wcrf.org/sites/default/files/Colorectal-Cancer-
2011-Report.pdf.
Wei EK, Giovannucci E, Wu K, Rosner B, Fuchs CS,
Willett WC, et al. (2004). Comparison of risk factors
for colon and rectal cancer. Int J Cancer, 108(3):433–42.
doi:10.1002/ijc.11540 PMID:14648711
112
2.2.2 Cancer of the oesophagus Associations were similar across follow-up

periods in one study (Engeland et al., 2004) and
There are two main histological subtypes of in another study that excluded the first 5 years
cancer of the oesophagus: adenocarcinoma and of follow-up (Abnet et al., 2008). There did not
squamous cell carcinoma. Oesophageal squa- appear to be any meaningful differences in
mous cell carcinoma arises from epithelial cells associations when stratifying by smoking status
that line the oesophagus and typically occurs in (O’Doherty et al., 2012; Lindkvist et al., 2014) or
the upper and middle parts of the oesophagus. when limiting results to non-smokers or never-
Oesophageal adenocarcinoma originates from smokers only (Reeves et al., 2007; Abnet et al.,
glandular cells; it occurs in the lower portion of 2008).
the oesophagus and can spread into the gastric In a meta-analysis including five prospec-
cardia. tive studies (Renehan et al., 2008), a relative risk
In 2001, the Working Group of the IARC of 1.5 for a 5 kg/m2 increase in BMI at baseline
Handbook on weight control and physical was reported, with similar values in men and in
activity (IARC, 2002) concluded that there was women.
sufficient evidence for a cancer-preventive effect Few studies have examined the association
of avoidance of weight gain for oesophageal between BMI measured at younger ages and
adenocarcinoma. Although recent pathological subsequent risk of oesophageal adenocarcinoma.
classification recognizes the histological simi- In the Netherlands Cohort Study, there was
larity between oesophageal adenocarcinoma evidence of a positive association between high
and gastric cardia cancer, most epidemiological BMI at age 20 years and risk, although the rela-
studies classify gastric cardia cancer with tive risk estimate was not statistically significant
stomach cancer, and therefore these studies (Merry et al., 2007).
are considered in Section 2.2.3. Also, because The association between BMI change and
evidence to date strongly suggests differences in incidence of oesophageal adenocarcinoma was
etiological factors between oesophageal adeno- examined in two prospective studies (Samanic
carcinoma and squamous cell carcinoma, the et al., 2006; Merry et al., 2007). The first study,
results are presented separately for each histo- which considered BMI change during a period of
logical subtype, and no results are presented for 6 years, did not find evidence for a positive associ-
oesophageal cancer overall. ation [the analysis was based on only 28 incident
(a) Cohort studies cases] (Samanic et al., 2006). The second study,
which included 113 cases, found that a 1 kg/m2
See Table 2.2.2a. increase in BMI from age 20 years to baseline was
(i) Adenocarcinoma of the oesophagus significantly associated with a 14% higher risk
Several cohort studies (with at least 75 (95% CI, 1.06–1.23) (Merry et al., 2007).
incident cases) have been published since the There have been few prospective studies of
previous IARC evaluation (IARC, 2002). In abdominal fatness in relation to risk of oesopha-
all of those studies, BMI and/or weight were geal adenocarcinoma. A study nested within the
positively associated with risk (Engeland et al., Multiphasic Health Check-up cohort of Kaiser
2004; Lindblad et al., 2005; Samanic et al., 2006; Permanente Northern California members
Merry et al., 2007; Reeves et al., 2007; Abnet observed a positive association between sagittal
et al., 2008; Corley et al., 2008; O’Doherty et al., abdominal diameter [distance from the anterior
2012; Lindkvist et al., 2014; Steffen et al., 2015). to the posterior of the abdomen] and incidence
113
of oesophageal adenocarcinoma (Corley et al., 2007). An inverse association was also observed
2008). Similarly, strong positive associations in the only study in Asia, which included 1958
were reported of both waist circumference and incident cases in China (Tran et al., 2005). There
waist-to-hip ratio with incidence of oesopha- was no evidence of differences in associations
geal adenocarcinoma in the National Institutes based on follow-up time (Engeland et al., 2004).
of Health–AARP Diet and Health Study A meta-analysis of five prospective studies
(NIH-AARP) cohort (Ptrend ≤ 0.01 for both) by Renehan et al. (2008) reported a relative risk
(O’Doherty et al., 2012) and with oesophageal per 5 kg/m2 increase in BMI of 0.71 (95% CI,
adenocarcinoma incidence/mortality in the 0.60–0.85) in men and 0.57 (95% CI, 0.47–0.69)
European Prospective Investigation into Cancer in women.
and Nutrition (EPIC) study (Ptrend ≤ 0.0001) The association between BMI measured at
(Steffen et al., 2015). age 20 years and risk of oesophageal squamous
cell carcinoma was examined in the Netherlands
(ii) Squamous cell carcinoma of the
Cohort Study (Merry et al., 2007). The relative
oesophagus
risk for BMI ≥ 25 kg/m2 compared with BMI
Since 2001, the association between BMI 20–21.4 kg/m2 was 2.49 (95% CI, 1.15–5.40),
and/or weight assessed at baseline and the inci- but there was no evidence of dose–response
dence and/or mortality of oesophageal squamous [Ptrend = 0.58]. In that study, weight loss from age
cell carcinoma has been examined in at least nine 20 years to baseline was associated with a statis-
individual prospective studies (Engeland et al., tically significant increased risk, with a relative
2004; Lindblad et al., 2005; Tran et al., 2005; risk of 2.57, but there was no evidence that weight
Samanic et al., 2006; Merry et al., 2007; Reeves gain was associated with risk.
et al., 2007; Corley et al., 2008; Steffen et al., 2009; Only two prospective studies examined
Lindkvist et al., 2014) and in one meta-analysis measures of abdominal fatness in relation to risk
(Renehan et al., 2008). In all of the studies, BMI of oesophageal squamous cell carcinoma. In the
and/or weight were inversely associated with Kaiser Permanente Multiphasic Health Check-up
risk. Notably, higher risks were found in the nested case–control study, there was no associa-
lowest BMI categories (i.e. BMI < 20 kg/m2) tion between sagittal abdominal diameter and
compared with categories within the normal risk (Corley et al., 2008), whereas in the EPIC
range of BMI, whereas lower risks were observed study, there was some evidence of a weak inverse
in the overweight and obese categories. Although trend of waist circumference with incidence/
most studies adjusted for tobacco use, not all mortality (Ptrend = 0.08) (Steffen et al., 2009).
studies included alcohol consumption, another
strong risk factor for oesophageal squamous cell (b) Case–control studies
carcinoma in their model. Furthermore, in two See Table 2.2.2b.
studies that stratified by smoking status, there
was an inverse association in current smokers (i) Adenocarcinoma of the oesophagus
but no association in non-smokers [supporting a Of the case–control studies reporting on
possible confounding effect of tobacco smoking] oesophageal adenocarcinoma, most studies
(Steffen et al., 2009; Lindkvist et al., 2014). In showed increases of 2.5-fold and higher in risk of
contrast, in the Million Women Study, an inverse oesophageal adenocarcinoma when comparing
association with both incidence and mortality of the highest and lowest BMI categories, although
oesophageal squamous cell carcinoma was noted in a few studies these associations were not statis-
even in the never-smokers group (Reeves et al., tically significant. When assessed, adjustments
114
for self-reported frequency or severity, or strat- (c) Mendelian randomization studies

ification by presence or absence of gastric reflux See Table 2.2.2c.
symptoms did not substantially alter the relative One Mendelian randomization study estim-
risk estimates (Chow et al., 1998; Lagergren et al., ated the causal association between BMI and risk
1999; de Jonge et al., 2006; Anderson et al., 2007; of oesophageal adenocarcinoma (Thrift et al.,
Löfdahl et al., 2008; Whiteman et al., 2008; Olsen 2014). Using a genetic risk score based on 29 SNPs
et al., 2011). previously shown to be associated with BMI
A pooled analysis of data from 10 case– (Speliotes et al., 2010), this Mendelian random-
control studies and 2 cohort studies (Hoyo et al., ization study showed that each 1 kg/m2 increase
2012), including a total of 3719 adenocarcinoma in BMI was associated with a 23% increase in
cases and 10 481 controls, showed significant risk (95% CI, 6–43%; P = 0.01), compared with
trends of increasing adenocarcinoma risk with a 6% increase in risk (95% CI, 5–8%; P < 0.001)
increasing BMI, up to odds ratios of 4.76 (95% observed in the same sample by conventional
CI, 2.96–7.66) for oesophageal adenocarcinoma epidemiological analyses.
and 3.07 (95% CI, 1.89–4.99) for oesophagogas-
tric junction adenocarcinoma when comparing
BMI ≥ 40 kg/m2 with BMI < 25 kg/m2. Subset
analyses showed similar increases in risk of
adenocarcinoma when stratifying by symptoms
of gastro-oesophageal reflux disease. No differ-
ences in associations were observed by sex.
(ii) Squamous cell carcinoma of the
oesophagus
For oesophageal squamous cell carci-
noma, several case–control studies reported
an inverse association between risk and recent
BMI (Vaughan et al., 1995; Chow et al., 1998;
Lahmann et al., 2012), and this inverse associ-
ation was observed within both smokers and
never-smokers (Lahmann et al., 2012). Of the
two studies that investigated the association of
risk of oesophageal squamous cell carcinoma
with recalled BMI at age 20 years, one found a
non-significant decrease in risk in relation to
higher BMI (Lahmann et al., 2012), whereas
the other study, based on a total of 167 cases in
Sweden, showed an increase in risk with higher
BMI (Lagergren et al., 1999).
115
116

Table 2.2.2a Cohort studies of measures of body fatness and cancer of the oesophagus
Reference Total Organ site or Exposure categories Exposed cases Relative risk Covariates Comments
Cohort number of cancer type (95% CI)
Location subjects (ICD code)
Follow-up Sex
period Incidence/
mortality
Adenocarcinoma
Engeland et al. Men: Oesophageal BMI 448 total Men: Age at measurement,
(2004) 963 709 adenocarcinoma < 18.5 – height, birth cohort
Population- Women: ICD-7: 150 18.5–24.9 1.00
based 1 038 010 25–29.9 1.80 (1.48–2.19)
Norwegian Incidence ≥ 30 2.58 (1.81–3.68)
cohort [Ptrend] [< 0.001]
Norway BMI 127 total Women:
1963–2002 < 18.5 4.07 (1.44–11)
18.5–24.9 1.00
25–29.9 1.64 (1.08–2.49)
≥ 30 2.06 (1.25–3.39)
[Ptrend] [0.002]
Lindblad et al. 10 287 Oesophageal BMI Age, sex, calendar
(2005) Men and adenocarcinoma < 20 8 1.44 (0.67–3.10) year, smoking,
Case–control women 20–24 49 1.00 alcohol consumption,
study nested Incidence 25–29 94 1.68 (1.18–2.40) reflux
in General ≥ 30 36 1.93 (1.24–3.01)
Practitioner [Ptrend] [0.005]
Research
Database
United
Kingdom
1994–2001
Samanic et al. 362 552 Oesophageal BMI Attained age (10-yr
(2006) Men adenocarcinoma 18.5–24.9 34 1.00 interval), calendar
Swedish Incidence 25–29.9 38 1.58 (0.98–2.53) year, smoking
Construction ≥ 30 10 2.72 (1.33–5.55)
Sweden
1958–1999
Follow-up Sex
period Incidence/
mortality
Samanic et al. BMI, 6-yr change
(2006) −4% to +4.9% 19 1.00
(cont.) 5–9.9% 3 0.44 (0.13–1.49)
10–14.9% 5 2.24 (0.81–6.21)
> 15% 1 1.21 (0.16–9.45)
[Ptrend] [> 0.5]
Merry et al. 4774 (case– Oesophageal BMI at baseline Age, sex First year of
(2007) cohort adenocarcinoma < 20 3 1.29 (0.40–4.16) For BMI change only: follow-up excluded
Netherlands sample from ICD-10: C15 20–24.9 51 1.00 adjustment for BMI at from the analyses
Cohort Study 120 852 Histology: 8140– 25–29.9 60 1.40 (0.95–2.04) age 20 yr
The main 8141, 8190–8231, ≥ 30 19 3.96 (2.27–6.88)
Netherlands cohort) 8260–8263, 8310, [Ptrend] [0.001]
1986–1999 Men and 8430, 8480–8490, per 1 kg/m2 1.14 (1.08–1.21)
women 8560, 8570–8572 BMI at age 20 yr
Incidence < 20 21 1.07 (0.59–1.94)
20–21.4 24 1.00
21.5–22.9 37 1.61 (0.95–2.72)
23.0–24.9 18 1.02 (0.55–1.90)
≥ 25 13 1.97 (0.99–3.94)
[Ptrend] [0.17]
per 1 kg/m2 1.04 (0.95–1.14)

BMI change, age 20 yr to baseline
< 0 8 0.75 (0.34–1.64)
0–3.9 51 1.00
4–7.9 37 1.34 (0.86–2.08)
≥ 8 17 3.41 (1.88–6.18)
[Ptrend] [0.001]
per 1 kg/m2 1.14 (1.06–1.23)
Reeves et al. 1 222 630 Oesophageal BMI Incidence: Age, geographical Results remained
(2007) Women adenocarcinoma < 22.5 22 1.06 (0.70–1.62) region, SES, significant after
Million Women Incidence ICD-10: C15 22.5–24.9 27 1.00 (0.68–1.46) reproductive history, excluding never-
Study and 25–27.4 30 1.28 (0.90–1.83) smoking status, smokers and
United mortality 27.5–29.9 23 1.57 (1.04–2.36) alcohol consumption, excluding the first
Kingdom ≥ 30 48 2.54 (1.89–3.41) physical activity 2 yr of follow-up
117
1996–2005 per 10 kg/m2 2.38 (1.59–3.56)

118

Follow-up Sex
period Incidence/
mortality
Reeves et al. BMI Mortality:
(2007) < 22.5 20 1.35 (0.87–2.11)
(cont.) 22.5–24.9 19 1.00 (0.64–1.57)
25–27.4 20 1.21 (0.78–1.87)
27.5–29.9 15 1.44 (0.87–2.39)
≥ 30 37 2.75 (1.97–3.85)
per 10 kg/m2 2.24 (1.40–3.58)
Abnet et al. 480 475 Oesophageal BMI Age, sex, cigarette Results were stable
(2008) Men and adenocarcinoma < 18.5 2 1.61 (0.39–6.55) smoking, alcohol after excluding
NIH-AARP women ICD-10: 18.5–24.9 71 1.00 consumption, the first 5 yr of
cohort Incidence C15.0–15.9 25–29.9 194 1.65 (1.26–2.18) education level, follow-up
USA Histology: 30–34.9 77 1.91 (1.38–2.66) physical activity
1995–2003 “adenocarcinoma” ≥ 35 27 2.27 (1.44–3.59)
Corley et al. 3150 Oesophageal BMI Age, sex, year of
(2008) Men and adenocarcinoma < 18.5 1 1.36 (0.12–15.52) health check-up
Nested women ICD-10: 18.5–24.9 28 1.00 BMI results also
case–control Incidence C15.0–15.9 25–29.9 51 2.20 (1.31–3.67) adjusted for ethnicity
of Kaiser Histology: ≥ 30 14 3.17 (1.43–7.04)
Permanente 8140–8573 per 1 kg/m2 increase 1.10 (1.04–1.17)
Multiphasic Sagittal abdominal diameter (cm)
Health Check- < 20 8 1.00
up cohort 20–22.4 13 0.92 (0.31–2.74)
USA 22.5–25 12 2.35 (0.78–7.12)
1964–1973 ≥ 25 22 3.47 (1.29–9.33)
per 1 cm increase 1.10 (1.03–1.17)
Renehan et al. 4 673 213 Oesophageal BMI Men: Geographical region,
(2008) Men and adenocarcinoma per 5 kg/m2 increase 1315 total 1.52 (1.33–1.74) age (all studies), and
Meta-analysis women BMI Women: other factors (not
1966–2007 Incidence per 5 kg/m2 increase 735 total 1.51 (1.31–1.74) in all studies) such
as Western diet,
alcohol consumption,
medical conditions
(e.g. type 2 diabetes,
acromegaly), or
physical activity
Follow-up Sex
period Incidence/
mortality
O’Doherty et al. 218 854 Oesophageal BMI Age, sex, total energy Waist-to-hip ratio
(2012) Men and adenocarcinoma < 18.5 0 – intake, antacid use, also significantly
NIH-AARP women ICD-10: 18.5–24.9 59 1.00 aspirin use, NSAID associated with
cohort Incidence C15.0–15.9 25–29.9 119 1.30 (0.94–1.78) use, marital status, risk (Q3 and Q4)
USA Histology: 30–34.9 64 2.28 (1.57–3.30) diabetes, cigarette
1995–2006 “adenocarcinoma” ≥ 35 11 2.11 (1.09–4.09) smoking, education
[Ptrend] [< 0.01] level, ethnicity,
Weight, quartiles (sex-specific) alcohol consumption,
physical activity,
Q1 41 1.00
intake of red and
Q2 58 1.49 (0.99–2.23)
white meat, intake of
Q3 53 1.37 (0.89–2.10)
fruits and vegetables;
Q4 101 2.66 (1.76–4.02)
for weight, also
[Ptrend] [< 0.01]
adjusted for height
WC, quartiles (sex-specific)
Q1 37 1.00
Q2 49 1.36 (0.89–2.09)
Q3 79 1.51 (1.02–2.25)
Q4 88 2.01 (1.35–3.00)
[Ptrend] [< 0.01]
Lindkvist et al. 587 700 Oesophageal BMI, quintiles Sex, age, study cohort,
(2014) Men and adenocarcinoma Q1 5 1.00 smoking status

Me-Can cohort women ICD-7: 150 Q2 18 3.37 (1.25–9.10)
(prospective Incidence Q3 18 3.17 (1.17–8.57)
cohorts) Q4 31 5.19 (2.00–13.42)
Austria, Q5 42 7.34 (2.88–18.68)
Norway, and [Ptrend] [< 0.0001]
Sweden per 5 kg/m2 1.78 (1.45–2.17)
1972–2006
119
120

Follow-up Sex
period Incidence/
mortality
Steffen et al. 391 456 Oesophageal BMI, quintiles Age at recruitment, Sex-specific
(2015) Men and adenocarcinoma Q1 15 1.00 centre, sex, education quintiles for
EPIC cohort women ICD-10: C15 Q2 22 1.30 (0.67–2.52) level, smoking, weight, BMI, and
10 European Incidence/ Q3 24 1.36 (0.71–2.62) alcohol consumption, WC. Cut-off points
countries mortality Q4 30 1.76 (0.93–3.31) physical activity, diet, not provided, only
1992–2008 Q5 33 2.15 (1.14–4.05) height the median values
[Ptrend] [0.004] for each
Weight, quintiles Positive
Q1 17 1.00 associations with
Q2 25 1.54 (0.82–2.88) waist-to-hip ratio
Q3 23 1.41 (0.74–2.70) (Q4 and Q5)
Q4 26 1.57 (0.82–3.01)
Q5 33 2.19 (1.14–4.21)
[Ptrend] [0.03]
WC, quintiles
Q1 7 1.00
Q2 22 2.78 (1.18–6.54)
Q3 20 2.47 (1.03–5.92)
Q4 26 3.19 (1.36–7.49)
Q5 39 5.08 (2.21–11.7)
[Ptrend] [< 0.0001]
Squamous cell carcinoma
Engeland et al. Men: Oesophageal BMI 1023 total Age at measurement,
(2004) 963 709 squamous cell < 18.5 2.80 (1.73–4.54) height, birth cohort
Population- Incidence carcinoma 18.5–24.9 1.00
based ICD-7: 150 25–29.9 0.72 (0.63–0.82)
Norwegian ≥ 30 0.68 (0.50–0.93)
cohort [Ptrend] [< 0.001]
Norway Women: Oesophageal BMI 472 total Age at measurement,
1963–2002 1 038 010 squamous cell < 18.5 2.11 (1.23–3.62) height, birth cohort
Incidence carcinoma 18.5–24.9 1.00
ICD-7: 150 25–29.9 0.52 (0.42–0.65)
≥ 30 0.43 (0.32–0.59)
[Ptrend] [< 0.001]
Follow-up Sex
period Incidence/
mortality
Lindblad et al. 10 140 Oesophageal BMI Age, sex, calendar
(2005) Men and squamous cell < 20 9 1.93 (0.90–4.11) year, smoking,
Case–control women carcinoma 20–24 34 1.00 alcohol consumption,
study nested Incidence 25–29 39 1.13 (0.71–1.80) reflux
in General ≥ 30 4 0.28 (0.10–0.79)
Practitioner [Ptrend] [0.01]
Research
Database
United
Kingdom
1994–2001
Tran et al. 29 584 Oesophageal BMI 1958 total Age, sex
(2005) Men and squamous cell < 20 1.00
Linxian General women carcinoma 20–21 0.96 (0.85–1.08)
Population Incidence 22 0.80 (0.71–0.91)
Trial ≥ 23 0.81 (0.72–0.92)
China [Ptrend] [< 0.001]
1986–2001
Samanic et al. 362 552 Oesophageal BMI Attained age,
(2006) Men squamous cell 18.5–24.9 134 1.00 calendar year,
Swedish Incidence carcinoma 25–29.9 57 0.53 (0.39–0.72) smoking

Construction ≥ 30 13 0.77 (0.43–1.36)
Sweden
1958–1999
121
122

Follow-up Sex
period Incidence/
mortality
Merry et al. 4774 (case– Oesophageal BMI at baseline Age, sex, current
(2007) cohort squamous cell < 20 9 2.21 (0.99–4.92) smoking, cigarettes
Netherlands sample from carcinoma 20–24.9 51 1.00 per day, number of
Cohort Study 120 852 ICD-10: C15 25–29.9 26 0.63 (0.39–1.02) years of smoking
The main Histology: ≥ 30 6 0.93 (0.38–2.26) For BMI change only:
Netherlands cohort) 8050–8076 [Ptrend] [0.04] adjustment for BMI at
1986–1999 Men and per 1 kg/m2 0.90 (0.82–0.98) age 20 yr
women BMI at age 20 yr
Incidence < 20 22 1.35 (0.70–2.62)
20–21.4 16 1.00
21.5–22.9 11 0.72 (0.33–1.57)
23.0–24.9 13 1.03 (0.48–2.21)
≥ 25 12 2.49 (1.15–5.40)
[Ptrend] [0.58]
per 1 kg/m2 1.07 (0.96–1.20)
< 0 18 2.57 (1.40–4.72)
0–3.9 32 1.00
4–7.9 16 0.73 (0.39–1.36)
≥ 8 8 1.39 (0.62–3.15)
[Ptrend] [0.10]
per 1 kg/m2 0.90 (0.81–1.00)
Reeves et al. 1 222 630 Oesophageal BMI Incidence: Age, geographical Negative
(2007) Women squamous cell < 22.5 106 2.04 (1.67–2.48) region, SES, associations
Million Women Incidence carcinoma 22.5–24.9 63 1.00 (0.78–1.28) reproductive history, remained stable in
Study and ICD-10: C15 25–27.4 52 0.96 (0.73–1.26) smoking status, non-smokers and
United mortality 27.5–29.9 21 0.61 (0.40–0.94) alcohol consumption, excluding the first
Kingdom ≥ 30 21 0.47 (0.31–0.73) physical activity 2 yr of follow-up
1996–2005 per 10 kg/m2 0.26 (0.18–0.38)
Follow-up Sex
period Incidence/
mortality
Reeves et al. BMI Mortality:
(2007) < 22.5 75 2.10 (1.66–2.65)
(cont.) 22.5–24.9 44 1.00 (0.74–1.35)
25–27.4 39 1.02 (0.75–1.40)
27.5–29.9 11 0.45 (0.25–0.82)
≥ 30 13 0.42 (0.24–0.73)
per 10 kg/m2 0.22 (0.14–0.35)
Corley et al. 3150 Oesophageal BMI Matched for age,
(2008) Men and squamous cell < 18.5 3 0.91 (0.19–4.29) sex, year of health
Nested women carcinoma 18.5–24.9 78 1.00 check-up
case–control Incidence ICD-10: 25–29.9 46 0.66 (0.44–1.00) BMI results also
of Kaiser C15.0–15.9 ≥ 30 9 0.30 (0.13–0.72) adjusted for ethnicity
Permanente Histology per 1 kg/m2 increase 0.89 (0.84–0.94)
Multiphasic 8050–8082 Sagittal abdominal diameter (cm)
Health Check- < 20 19 1.00
up cohort 20–22.4 24 0.91 (0.43–1.94)
USA 22.5–25 14 0.89 (0.35–2.24)
1964–1973 ≥ 25 15 0.78 (0.32–1.92)
per 1 cm increase 1.00 (0.94–1.06)
Renehan et al. 4 673 213 Oesophageal BMI Men: Geographical region,
(2008) Men and squamous cell per 5 kg/m2 increase 6201 total 0.71 (0.60–0.85) age (all studies), and

Meta-analysis women carcinoma other factors (not
BMI Women:
1966–2007 Incidence in all studies) such
per 5 kg/m2 increase 1114 total 0.57 (0.47–0.69)
as Western diet,
medical conditions
(e.g. type 2 diabetes,
acromegaly), or
physical activity
123
124

Follow-up Sex
period Incidence/
mortality
Steffen et al. 346 554 Oesophageal BMI, quintiles (sex-specific) Age, study centre, BMI and WC
(2009) Men and squamous cell Men: Women: education level, were significantly
EPIC cohort women carcinoma < 23.4 < 21.7 42 1.00 smoking, alcohol inversely related
10 European Incidence/ ICD-10: C15 23.4–25.2 21.7–23.6 22 0.47 (0.27–0.79) consumption, to oesophageal
countries mortality 25.2–26.9 23.6–25.6 15 0.31 (0.17–0.57) physical activity, squamous cell
1992–2007 26.9–29.1 25.6–28.7 14 0.27 (0.14–0.51) consumption of carcinoma only in
≥ 29.2 ≥ 28.8 17 0.26 (0.14–0.51) fruits/vegetables/meat smokers
[Ptrend] [< 0.0001]
Weight, quintiles
Q1 41 1.00
Q2 28 0.61 (0.37–1.01)
Q3 14 0.30 (0.16–0.57)
Q4 10 0.19 (0.09–0.40)
Q5 17 0.33 (0.18–0.60)
[Ptrend] [< 0.0001]
WC, quintiles
Q1 23 1.00
Q2 19 0.76 (0.41–1.43)
Q3 23 0.78 (0.43–1.43)
Q4 16 0.51 (0.26–1.00)
Q5 22 0.62 (0.32–1.20)
[Ptrend] [0.08]
Lindkvist et al. 587 700 Oesophageal BMI, quintiles Sex, age, study cohort,
(2014) Men and squamous cell Q1 55 1.00 smoking status
Me-Can cohort women carcinoma Q2 29 0.50 (0.32–0.79)
(prospective Incidence ICD-7: 150 Q3 46 0.76 (0.51–1.12)
cohorts) Q4 30 0.46 (0.30–0.72)
Austria, Q5 24 0.38 (0.23–0.62)
Norway, and [Ptrend] [< 0.0001]
Sweden per 5 kg/m2 0.62 (0.50–0.79)
1972–2006
BMI, body mass index (in kg/m 2); CI, confidence interval; EPIC, European Prospective Investigation into Cancer and Nutrition; ICD, International Classification of Diseases; NIH-
AARP, National Institutes of Health–AARP Diet and Health Study; NSAID, non-steroidal anti-inflammatory drug; SES, socioeconomic status; WC, waist circumference; yr, year or
years
Table 2.2.2b Case–control studies of measures of body fatness and cancer of the oesophagus
Reference Total number Exposure categories Exposed cases Relative risk Adjustment for Comments
Study location of cases (95% CI) confounding
Period Source of
controls
Vaughan et al. EAC: BMI, percentiles Age, sex, education BMI percentiles
(1995) Men and 1–10% 12 1.6 (0.7–3.6) level, race, cigarette based on sex-
USA (13 counties women: 133 10–49% 43 1.0 use, alcohol specific distribution
of Western Population 50–89% 50 1.2 (0.7–2.1) consumption in controls (1 yr
Washington State) 90–100% 26 2.5 (1.2–5.0) before diagnosis
1993–1990 ESCC: BMI, percentiles in cases, 1 yr
Men and 1–10% 34 3.2 (1.4–7.1) before interview in
women: 106 10–49% 41 1.0 controls)
Population 50–89% 24 0.7 (0.3–1.4)
90–100% 6 0.2 (0.1–1.0)
Chow et al. (1998) EAC: BMI up to 1 yr before diagnosis (sex-specific) Geographical No effect
USA Men and Men: Women: location, age, sex, modification
1993–1995 women: 292 < 23.12 < 21.95 45 1.0 race, cigarette was observed by
Population 23.12–25.08 21.95–24.12 63 1.3 (0.8–2.2) smoking, respondent history of gastro-
25.09–27.31 24.13–27.43 85 2.0 (1.3–3.3) status oesophageal reflux
≥ 27.32 ≥ 27.44 99 2.9 (1.8–4.7) disease
[Ptrend] [< 0.0001]
ESCC: BMI up to 1 yr before diagnosis (sex-specific)
Men and Men: Women:
women: 220 < 23.12 < 21.95 79 1.0
Population 23.12–25.08 21.95–24.12 50 0.5 (0.3–0.9)
25.09–27.31 24.13–27.43 53 0.8 (0.5–1.3)
≥ 27.32 ≥ 27.44 38 0.6 (0.3–1.0)

[Ptrend] [< 0.11]
Lagergren et al. EAC: BMI 20 yr before interview Age, sex, tobacco No differences
(1999) Men and < 22 10 1.0 smoking, alcohol were observed in
Sweden women: 189 22–24.9 68 3.2 (1.6–6.7) consumption, SES, the associations for
1995–1997 Population 25–30 89 6.9 (3.3–14.4) reflux symptoms, both cancer types
> 30 22 16.2 (6.3–41.4) intake of fruits and when stratifying by
[Ptrend] [< 0.001] vegetables, energy presence of reflux
intake, physical symptoms
activity
125
126

Period Source of
controls
Lagergren et al. BMI at age 20 yr, quartiles (sex-specific)
(1999) Men: Women:
(cont.) < 20.7 < 19.3 28 1.0
20.7–22.1 19.3–20.4 29 0.9 (0.5–1.6)
22.2–23.7 20.5–22.1 51 1.6 (0.9–2.8)
> 23.7 > 22.1 81 2.7 (1.6–4.6)
[Ptrend] [< 0.001]
ESCC: BMI 20 yr before interview
Men and < 22 48 1.0
women: 820 22–24.9 67 1.0 (0.6–1.7)
Population 25–30 42 1.3 (0.8–2.3)
> 30 10 2.0 (0.8–4.9)
[Ptrend] [0.12]
BMI at age 20 yr, quartiles (sex-specific)
Men: Women:
< 20.7 < 19.3 36 1.0
20.7–22.1 19.3–20.4 38 1.2 (0.7–2.1)
22.2–23.7 20.5–22.1 40 1.4 (0.8–2.4)
> 23.7 > 22.1 53 1.8 (1.1–3.1)
[Ptrend] [0.03]
Wu et al. (2001) EAC: BMI at age 40 yr, quartiles (sex-specific) Smoking, sex, race,
USA Men and Men: Women: 202 total birthplace, education
1992–1997 women: 222 ≤ 22 ≤ 21 1.00 level
Population > 22–25 > 21–23 1.13 (0.7–1.7)
(proxy control) > 25– ≤ 27 > 23– ≤ 25 1.76 (1.1–2.9)
> 27 > 25 2.78 (1.7–4.4)
[Ptrend] [< 0.0001]
Men: Women: 207 total
≤ 20 ≤ 18 1.00
> 20–22 > 18–20 1.23 (0.8–1.9)
> 22– ≤ 24 > 20– ≤ 22 1.34 (0.9–2.1)
> 24 > 22 1.77 (1.1–2.7)
[Ptrend] [0.011]
Period Source of
controls
de Jonge et al. EAC: BMI 10 yr before questionnaire Age, sex, education Controls were
(2006) Men and < 25 29 1.0 level, smoking patients with
The Netherlands women: 91 > 25 58 1.8 (1.1–3.3) status, alcohol Barrett oesophagus
2003–2005 Hospital BMI at age 20 yr consumption, reflux
< 25 63 1.0 symptoms
> 25 20 2.6 (1.2–5.5)
Anderson et al. EAC: Current BMI, tertiles Sex, age at interview
(2007) 227 (192 men < 25.8 115 1.00 date, smoking
Ireland and 35 women) 25.8–29.0 54 0.35 (0.21–0.58) status, alcohol
2002–2004 Population > 29.0 50 0.33 (0.20–0.56) consumption,
BMI 5 yr before, tertiles years of full-time
< 25.0 51 1.00 education, job type,
25.0–28.1 55 1.74 (0.66–1.97) gastro-oesophageal
> 28.1 120 2.69 (1.62–4.46) reflux
BMI at age 21 yr
< 22.1 55 1.00
22.1–24.1 64 1.10 (0.65–1.25)
> 24.1 96 1.81 (1.08–3.02)
Löfdahl et al. EAC + EJAC: BMI 20 yr before interview Age, education The associations
(2008) Men: 388 Men: level, alcohol for maximum
Sweden Women: 63 < 22 45 1.0 consumption, adult BMI and for
1995–1997 Population 22–24.9 143 1.5 (1.0–2.3) cigarette smoking, minimum adult
25–29.9 164 2.7 (1.8–4.1) intake of fruits BMI were weaker,

≥ 30 36 5.4 (2.6–10.8) and vegetables, but also showed a
Women: Helicobacter pylori stronger association
< 22 12 1.0 infection in women than in
22–24.9 25 2.4 (0.9–6.0) Maximum and men
25–29.9 16 4.3 (1.4–13.1) minimum adult
≥ 30 10 10.3 (2.6–42.3) BMI, also adjusted
for gastro-
oesophageal reflux
127
128

Period Source of
controls
Whiteman et al. EAC: BMI in the last year Age, sex, state, Results did not
(2008) Men and < 18.5 1 0.3 (0.0–2.6) household income, significantly change
Australia women: 367 18.5–24.9 71 1.0 cumulative when additionally
2001–2005 Population 25.0–29.9 150 1.4 (1.0–1.9) smoking history, adjusted for
30.0–34.9 89 2.7 (1.8–3.9) mean alcohol gastro-oesophageal
35.0–39.9 25 3.1 (1.8–5.5) consumption, reflux; significantly
≥ 40 16 7.0 (3.3–15.0) frequency of aspirin higher risk in men
[Ptrend] [< 0.001] use in the 5 yr before than in women;
Maximum BMI diagnosis no significant
< 18.5 1 0.9 (0.1–8.7) associations or
18.5–24.9 39 1.0 trend between
25.0–29.9 136 1.4 (0.9–2.0) change in BMI and
30.0–34.9 114 2.5 (1.6–3.7) risk of EAC or EJAC
35.0–39.9 43 4.1 (2.4–6.8)
≥ 40 24 5.2 (2.7–9.9)
[Ptrend] [< 0.001]
BMI at age 20 yr
< 18.5 14 0.8 (0.4–1.4)
18.5–24.9 227 1.0
25.0–29.9 81 1.7 (1.2–2.3)
30.0–34.9 13 2.6 (1.3–5.2)
35.0–39.9 5 3.6 (1.0–13.0)
[Ptrend] [< 0.001]
EJAC: BMI in the last year
Men and < 18.5 1 0.2 (0.0–1.7)
women: 426 18.5–24.9 107 1.0
Population 25.0–29.9 168 1.1 (0.8–1.4)
30.0–34.9 98 1.9 (1.3–2.6)
35.0–39.9 27 2.0 (1.2–3.4)
≥ 40 9 2.6 (1.1–6.2)
[Ptrend] [< 0.001]
Period Source of
controls
Whiteman et al. Maximum BMI
(2008) < 18.5 0 –
(cont.) 18.5–24.9 55 1.0
25.0–29.9 178 1.3 (0.9–1.8)
30.0–34.9 122 1.9 (1.3–2.7)
35.0–39.9 47 2.9 (1.8–4.6)
≥ 40 13 2.1 (1.1–4.2)
[Ptrend] [< 0.001]
BMI at age 20 yr
< 18.5 9 0.4 (0.2–0.8)
18.5–24.9 282 1.0
25.0–29.9 97 1.6 (1.2–2.1)
30.0–34.9 13 2.1 (1.0–4.1)
≥ 35.0 2 1.1 (0.2–5.9)
[Ptrend] [< 0.001]
Olsen et al. (2011) EAC: BMI 1 yr before Age, sex, education
Australia Men and 18–24.9 71 1.0 level, NSAID use,
2002–2005 women: 364 25–29.9 149 1.4 (1.0–2.0) smoking status,
Population 30–34.9 89 2.5 (1.7–3.6) heartburn/acid
≥ 35 40 3.7 (2.2–6.2) reflux in the past
Overweight or obese 1.8 (1.3–2.5) 10 yr
EJAC: BMI 1 yr before
Men and 18–24.9 107 1.0

women: 425 25–29.9 168 1.1 (0.8–1.5)
Population 30–34.9 98 2.0 (1.4–2.9)
≥ 35 36 2.5 (1.5–4.1)
Overweight or obese 1.8 (1.3–2.5)
129
130

Period Source of
controls
Hoyo et al. (2012) EAC: BMI Age, sex, smoking, In stratified
International Men and < 25 577 1.00 education level, and analyses, results
Barrett’s and women: 1997 25.0–29.9 862 1.54 (1.26–1.88) other study-specific were independent
Esophageal Population 30.0–34.9 331 2.39 (1.86–3.06) adjustment variables of the presence
Adenocarcinoma 35.0–39.9 86 2.79 (1.89–4.12) (e.g. study centre) of symptoms of
Consortium ≥ 40 41 4.76 (2.96–7.66) gastro-oesophageal
(BEACON) Pooled continuous 1897 1.09 (1.06–1.12) reflux
analysis of 10 EJAC: BMI No differences in
case–control and Men and < 25 663 1.00 associations by sex
2 cohort studies women: 1900 25.0–29.9 742 1.28 (1.13–1.45)
from Australia, Population 30.0–34.9 304 2.08 (1.75–2.47)
Europe, and USA 35.0–39.9 85 2.36 (1.75–3.17)
≥ 40 28 3.07 (1.89–4.99)
continuous 1822 1.07 (1.05–1.09)
Lahmann et al. ESCC: BMI in the last year, quintiles (sex-specific) Age, sex, education
(2012) Men and Men: Women: level, alcohol
Australia women: 287 < 22.1 < 23.7 108 1.00 consumption,
2002–2005 Population 22.1– ≤ 24.6 23.7– < 25.6 65 0.61 (0.42–0.90) smoking status,
24.6– ≤ 27.0 25.6– ≤ 27.2 35 0.32 (0.20–0.50) NSAID/aspirin use,
27.0– ≤ 31.9 27.2– ≤ 29.7 41 0.40 (0.26–0.61) physical activity
> 31.9 > 29.7 38 0.36 (0.23–0.57) BMI at age 20 yr
[Ptrend] [< 0.001] (only for BMI in the
Maximum BMI, quintiles (sex-specific) last year)
Men: Women:
≤ 23.5 < 25.1 90 1.00
23.5– ≤ 26.0 25.1– ≤ 27.0 73 0.78 (0.53–1.15)
26.0– ≤ 28.7 27.0– ≤ 28.9 42 0.49 (0.32–0.76)
28.7– ≤ 33.9 28.9– ≤ 31.7 43 0.45 (0.29–0.69)
> 33.9 > 31.7 39 0.44 (0.28–0.69)
[Ptrend] [< 0.001]
BMI at age 20 yr
< 25 233 1.00
≥ 25 42 0.85 (0.57–1.25)
[Ptrend] [< 0.40]
BMI, body mass index (in kg/m 2); CI, confidence interval; EAC, oesophageal adenocarcinoma; EJAC, oesophagogastric junction adenocarcinoma; ESCC, oesophageal squamous cell
carcinoma; GCAC, gastric cardia adenocarcinoma; NSAID, non-steroidal anti-inflammatory drug; SES, socioeconomic status; yr, year or years
Table 2.2.2c Mendelian randomization studies of measures of body fatness and cancer of the oesophagus
Reference Characteristics of study Sample size Exposure (unit) Odds ratio Adjustment Comments
Study population (95% CI) for
confounding
Thrift et al. (2014) Subset of ethnically 5229 (999 1 kg/m2 increase 1.23 (1.06–1.43) NR Similar associations in men
Barrett’s and homogenous individuals EAC cases based on a genetic and women. Associations
Esophageal from 14 studies in and 2169 risk score of 29 SNPs with the genetic instrument
Adenocarcinoma Australia, North America, controls) were stronger than those of
Genetic and western Europe conventional epidemiological
Susceptibility Study analyses in the same sample
(BEAGESS)
CI, confidence interval; EAC, oesophageal adenocarcinoma; NR, not reported; SNP, single nucleotide polymorphism

131
References gastric cardia, and gastric non-cardia adenocarcinoma

among men and women in a nested case-control study.
Cancer Causes Control, 16(3):285–94. doi:10.1007/
Abnet CC, Freedman ND, Hollenbeck AR, Fraumeni JF Jr, s10552-004-3485-7 PMID:15947880
Leitzmann M, Schatzkin A (2008). A prospective study Lindkvist B, Johansen D, Stocks T, Concin H, Bjørge
of BMI and risk of oesophageal and gastric adeno- T, Almquist M, et al. (2014). Metabolic risk factors
carcinoma. Eur J Cancer, 44(3):465–71. doi:10.1016/j. for esophageal squamous cell carcinoma and adeno-
ejca.2007.12.009 PMID:18221867 carcinoma: a prospective study of 580,000 subjects
Anderson LA, Watson RG, Murphy SJ, Johnston BT, within the Me-Can project. BMC Cancer, 14(1):103.
Comber H, Mc Guigan J, et al. (2007). Risk factors doi:10.1186/1471-2407-14-103 PMID:24548688
for Barrett’s oesophagus and oesophageal adeno- Löfdahl HE, Lu Y, Lagergren J (2008). Sex-specific risk
carcinoma: results from the FINBAR study. World J factor profile in oesophageal adenocarcinoma. Br
Gastroenterol, 13(10):1585–94. doi:10.3748/wjg.v13. J Cancer, 99(9):1506–10. doi:10.1038/sj.bjc.6604701
i10.1585 PMID:17461453 PMID:18841152
Chow WH, Blot WJ, Vaughan TL, Risch HA, Gammon Merry AH, Schouten LJ, Goldbohm RA, van den Brandt PA
MD, Stanford JL, et al. (1998). Body mass index and (2007). Body mass index, height and risk of adenocarci-
risk of adenocarcinomas of the esophagus and gastric noma of the oesophagus and gastric cardia: a prospec-
cardia. J Natl Cancer Inst, 90(2):150–5. doi:10.1093/ tive cohort study. Gut, 56(11):1503–11. doi:10.1136/
jnci/90.2.150 PMID:9450576 gut.2006.116665 PMID:17337464
Corley DA, Kubo A, Zhao W (2008). Abdominal obesity O’Doherty MG, Freedman ND, Hollenbeck AR, Schatzkin
and the risk of esophageal and gastric cardia carci- A, Abnet CC (2012). A prospective cohort study of
nomas. Cancer Epidemiol Biomarkers Prev, 17(2):352–8. obesity and risk of oesophageal and gastric adeno-
doi:10.1158/1055-9965.EPI-07-0748 PMID:18268119 carcinoma in the NIH-AARP Diet and Health Study.
de Jonge PJ, Steyerberg EW, Kuipers EJ, Honkoop P, Gut, 61(9):1261–8. doi:10.1136/gutjnl-2011-300551
Wolters LM, Kerkhof M, et al. (2006). Risk factors for PMID:22174193
the development of esophageal adenocarcinoma in Olsen CM, Pandeya N, Green AC, Webb PM, Whiteman
Barrett’s esophagus. Am J Gastroenterol, 101(7):1421–9. DC; Australian Cancer Study (2011). Population
doi:10.1111/j.1572-0241.2006.00626.x PMID:16863542 attributable fractions of adenocarcinoma of the esoph-
Engeland A, Tretli S, Bjørge T (2004). Height and body agus and gastroesophageal junction. Am J Epidemiol,
mass index in relation to esophageal cancer; 23-year 174(5):582–90. doi:10.1093/aje/kwr117 PMID:21719746
follow-up of two million Norwegian men and women. Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D;
Cancer Causes Control, 15(8):837–43. doi:10.1023/B:- Million Women Study Collaboration (2007). Cancer
CACO.0000043434.21558.ea PMID:15456997 incidence and mortality in relation to body mass
Hoyo C, Cook MB, Kamangar F, Freedman ND, index in the Million Women Study: cohort study. BMJ,
Whiteman DC, Bernstein L, et al. (2012). Body mass 335(7630):1134. doi:10.1136/bmj.39367.495995.AE
index in relation to oesophageal and oesophagogastric PMID:17986716
junction adenocarcinomas: a pooled analysis from the Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen
International BEACON Consortium. Int J Epidemiol, M (2008). Body-mass index and incidence of cancer:
41(6):1706–18. doi:10.1093/ije/dys176 PMID:23148106 a systematic review and meta-analysis of prospec-
IARC (2002). Weight control and physical activity. Lyon, tive observational studies. Lancet, 371(9612):569–78.
France: IARC Press (IARC Handbooks of Cancer doi:10.1016/S0140-6736(08)60269-X PMID:18280327
Prevention, Vol. 6). Available from: http://publications. Samanic C, Chow WH, Gridley G, Jarvholm B,
iarc.fr/376. Fraumeni JF Jr (2006). Relation of body mass index
Lagergren J, Bergström R, Nyrén O (1999). Association to cancer risk in 362,552 Swedish men. Cancer Causes
between body mass and adenocarcinoma of the esoph- Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9
agus and gastric cardia. Ann Intern Med, 130(11):883– PMID:16841257
90. doi:10.7326/0003-4819-130-11-199906010-00003 Speliotes EK, Willer CJ, Berndt SI, Monda KL, Thorleifsson
PMID:10375336 G, Jackson AU, et al.; MAGIC; Procardis Consortium
Lahmann PH, Pandeya N, Webb PM, Green AC, (2010). Association analyses of 249,796 individ-
Whiteman DC; Australian Cancer Study (2012). Body uals reveal 18 new loci associated with body mass
mass index, long-term weight change, and esophageal index. Nat Genet, 42(11):937–48. doi:10.1038/ng.686
squamous cell carcinoma: is the inverse association PMID:20935630
modified by smoking status? Cancer, 118(7):1901–9. Steffen A, Huerta JM, Weiderpass E, Bueno-de-Mesquita
doi:10.1002/cncr.26455 PMID:21960146 HB, May AM, Siersema PD, et al. (2015). General
Lindblad M, Rodríguez LA, Lagergren J (2005). Body and abdominal obesity and risk of esophageal and
mass, tobacco and alcohol and risk of esophageal, gastric adenocarcinoma in the European Prospective
132
Investigation into Cancer and Nutrition. Int J Cancer,

137(3):646–57. doi:10.1002/ijc.29432 PMID:25598323
Steffen A, Schulze MB, Pischon T, Dietrich T, Molina
E, Chirlaque MD, et al. (2009). Anthropometry
and esophageal cancer risk in the European
Prospective Investigation into Cancer and Nutrition.
Cancer Epidemiol Biomarkers Prev, 18(7):2079–89.
PMID:19567501
Thrift AP, Shaheen NJ, Gammon MD, Bernstein L, Reid BJ,
Onstad L, et al. (2014). Obesity and risk of esophageal
adenocarcinoma and Barrett’s esophagus: a Mendelian
randomization study. J Natl Cancer Inst, 106(11):dju252.
doi:10.1093/jnci/dju252 PMID:25269698
Tran GD, Sun XD, Abnet CC, Fan JH, Dawsey SM, Dong
ZW, et al. (2005). Prospective study of risk factors
for esophageal and gastric cancers in the Linxian
General Population Trial cohort in China. Int J Cancer,
113(3):456–63. doi:10.1002/ijc.20616 PMID:15455378
Vaughan TL, Davis S, Kristal A, Thomas DB (1995).
Obesity, alcohol, and tobacco as risk factors for cancers
of the esophagus and gastric cardia: adenocarcinoma
versus squamous cell carcinoma. Cancer Epidemiol
Biomarkers Prev, 4(2):85–92. PMID:7742727
Whiteman DC, Sadeghi S, Pandeya N, Smithers BM,
Gotley DC, Bain CJ, et al.; Australian Cancer Study
(2008). Combined effects of obesity, acid reflux and
smoking on the risk of adenocarcinomas of the oesoph-
agus. Gut, 57(2):173–80. doi:10.1136/gut.2007.131375
PMID:17932103
Wu AH, Wan P, Bernstein L (2001). A multiethnic popula-
tion-based study of smoking, alcohol and body size and
risk of adenocarcinomas of the stomach and esophagus
(United States). Cancer Causes Control, 12(8):721–32.
doi:10.1023/A:1011290704728 PMID:11562112
133
2.2.3 Cancer of the stomach (a) Cohort studies

In 2012, gastric cancer, or cancer of the (i) Gastric cancer NOS
stomach, was the fifth most commonly diag- Since 2000, at least 20 individual cohort
nosed cancer worldwide, with heterogeneous studies (Table 2.2.3a) and six meta-analyses or
geographical distribution (Jemal et al., 2014). pooled analyses (Table 2.2.3c) of prospective
Gastric cancer can generally be classified studies have examined associations of base-
into two subsites: cancer of the gastric cardia, line BMI with gastric cancer incidence and/or
which arises from the area of the stomach mortality. Most of the individual prospective
adjoining the gastro-oesophageal junction, and studies showed no associations with gastric
non-cardia gastric cancer, which develops in the cancer incidence or mortality (Table 2.2.3a).
distal stomach and represents about 73% of all A few studies found inconsistent evidence of
gastric cancer cases globally (Colquhoun et al., either positive or negative associations (Calle
2015). Several risk factors for gastric cardia and et al., 2003; Samanic et al., 2006; Jee et al., 2008;
non-cardia cancer have been identified. For Persson et al., 2008; Camargo et al., 2014).
example, infection with Helicobacter pylori has Although three pooled analyses and one
been strongly associated with non-cardia gastric meta-analysis also showed no association
cancer, whereas diets rich in smoked foods, between high BMI and incidence of gastric
salted foods (especially fish), or pickled foods, as cancer (Lindkvist et al., 2013) or incidence and/
well as cigarette smoking, appear to increase the or mortality (Renehan et al., 2008; Whitlock
risk of both types of gastric cancer (Kamangar et al., 2009; Parr et al., 2010), others were sugges-
et al., 2006; IARC, 2012). tive of a positive association (Yang et al., 2009;
In 2001, the IARC Handbook on weight Chen et al., 2013; Lin et al., 2014). In the most
control and physical activity (IARC, 2002) recent meta-analysis of 12 prospective studies of
reviewed the studies of cancer of the gastric gastric cancer incidence and mortality combined
cardia together with studies of oesophageal and more than 41 791 gastric cancer cases, strong
adenocarcinoma, but did not provide a sepa- associations with overweight and obesity were
rate evaluation for stomach cancer (cardia or reported in men only, but there was no evidence
non-cardia). Since then, numerous individual of heterogeneity of results according to sex (Chen
and pooled cohort studies and meta-analyses, as et al., 2013). The same study did not show hetero-
well as several case–control studies of anthropo- geneity in results between Asian and non-Asian
metric measures and risk of stomach cancer have populations.
been published. Results from studies that exam- No associations of weight or BMI in early
ined this association for gastric cancer not other- adulthood, usually defined as age 18–21 years,
wise specified (NOS) and separately for gastric with gastric cancer incidence or mortality were
cardia and non-cardia cancers are summarized found in three studies (Fujino et al., 2007; Merry
here and in Tables 2.2.3a, 2.2.3b, and 2.2.3c. et al., 2007; Tanaka et al., 2007), or of BMI change
Studies that had fewer than 75 incident cases during adulthood in relation to incidence of
or that overlapped with a more recent study, as gastric cancer (Merry et al., 2007; Rapp et al.,
well as those that considered gastric cardia and 2008). No prospective studies of waist circum-
oesophageal cancers together, were excluded. ference and total gastric cancer were identified.
134
(ii) Cancer of the gastric cardia Chen et al., 2013; Lin et al., 2014). However, in
Most individual prospective studies of the the Linxian General Population Trial, a signifi-
association between baseline BMI (or weight) cant inverse association was reported with a rela-
and cardia gastric cancer incidence (or incidence tive risk of 0.68 for BMI ≥ 23 kg/m2 versus BMI
and mortality) showed a positive association (see < 20 kg/m2 (Tran et al., 2005), and a significant
Table 2.2.3a), except for four studies (Tran et al., inverse association was also reported in a Swedish
2005; Samanic et al., 2006; Corley et al., 2008; cohort study (Ptrend < 0.01) (Samanic et al., 2006).
Steffen et al., 2015). In the large meta-analysis Conversely, one individual study suggested a
by Chen et al., overweight was associated with a positive association of BMI and/or weight and
21% higher risk (based on six studies) and obesity risk of non-cardia gastric cancer (O’Doherty
was associated with an 82% higher risk (based et al., 2012).
on seven studies) compared with normal BMI No associations were reported in the only
(18.5–24.9 kg/m2) (Chen et al., 2013). These find- study of BMI in early adulthood and adult BMI
ings were similar to those reported in an earlier change in relation to incidence of non-cardia
meta-analysis of three prospective studies (Yang gastric cancer (Merry et al., 2007), or in the three
et al., 2009). studies that examined waist circumference and
Associations of BMI in early adulthood risk of non-cardia gastric cancer (MacInnis et al.,
and adult BMI change with incidence of cardia 2006; O’Doherty et al., 2012; Steffen et al., 2015).
gastric cancer were examined in only one study
(b) Case–control studies
of mortality (Merry et al., 2007). In that study,
BMI at age 20 years was not associated with See Table 2.2.3b.
risk, whereas increasing BMI from age 20 years There were a total of 11 independent reports
to baseline showed a positive association from case–control studies on the association
(Ptrend = 0.02). of BMI with risk of gastric cancer, in China,
Although one study showed no association Europe, Japan, the Republic of Korea, the USA,
between sagittal abdominal diameter and risk of and Venezuela. With the exception of one hospi-
gastric cardia cancer (Corley et al., 2008), in the tal-based study (Kim et al., 2015), in which BMI
NIH-AARP cohort a 2.2-fold higher risk for the was measured at the time of initial endoscopic
fourth versus the first quartile of waist circum- diagnosis, BMI was assessed through self-reports
ference was reported, with a significant trend of height and body weight, referring to either a
(O’Doherty et al., 2012). A similar positive trend recent period (mostly 1 year) before disease diag-
of waist circumference and gastric cardia cancer nosis or a period in the more distant past (e.g.
risk (incidence and mortality) was also found in at age 18 years or 20 years), or both. In addition
the EPIC study (Steffen et al., 2015). to standard adjustments for age and sex, studies
were reported with variable adjustments for
(iii) Non-cardia gastric cancer further confounding factors such as smoking,
Findings from cohort studies and meta-ana- alcohol consumption, family history of gastric
lyses of excess body weight at baseline in rela- cancer, dietary variables, or H. pylori infection.
tion to incidence of non-cardia gastric cancer With regard to gastric cardia cancer, three
are inconsistent. Neither BMI nor weight was out of four studies showed a positive associa-
associated with risk in most individual prospec- tion of BMI with risk. Three studies specifically
tive studies (see Table 2.2.3a). Similarly, several addressing non-cardia cancer showed no asso-
meta-analyses did not show an association ciation of recent BMI with risk, whereas two
between BMI and risk either (Yang et al., 2009; studies reported a positive association of risk
135
with BMI at age 20 years. With regard to overall

gastric cancer – without specification by subsite
– three studies showed an increase in risk with
increasing BMI, one showed a decrease in risk,
and two showed no significant association.
136
Table 2.2.3a Cohort studies of measures of body fatness and cancer of the stomach
Reference Total number of Organ site or Exposure categories Exposed Relative risk Covariates Comments
Cohort subjects cancer type cases (95% CI)
Location Sex (ICD code)
Follow-up period Incidence/
mortality
Stomach not otherwise specified
Calle et al. (2003) 404 576 Stomach BMI Age, education
Cancer Prevention Men ICD-9: 151.0–151.9 18.5–24.9 388 1.00 level, smoking,
Study II (CPS II) Mortality 25–29.9 455 1.01 (0.88–1.16) physical
USA 30–34.9 84 1.20 (0.94–1.52) activity, alcohol
1982–1998 ≥ 35 18 1.94 (1.21–3.13) consumption,
[Ptrend] [0.03] marital
495 477 BMI status, race,
Women 18.5–24.9 304 1.00 aspirin use,
Mortality 25–29.9 134 0.89 (0.72–1.09) consumption
30–34.9 57 1.30 (0.97–1.74) of fat and
≥ 35 13 1.08 (0.61–1.89) vegetables; for
[Ptrend] [0.46] women, also
adjusted for
HRT use
Samanic et al. 4 500 700 Stomach Obesity Age, calendar Obesity defined
(2004) Men ICD-9: 151 White men: year as discharge
United States Incidence Non-obese 4989 1.00 diagnosis of
1969–1996 Non-obese 2089 1.00
Obese 99 0.98 (0.79–1.20)

137
138

mortality
Batty et al. (2005) 18 403 Stomach BMI Age,
Whitehall study Men 18.5–24.9 100 1.00 employment
of London-based Mortality 25.0–29.9 81 1.05 (0.76–1.44) grade, physical
male government ≥ 30 9 1.23 (0.59–2.58) activity,
employees [Ptrend] [0.60] smoking,
United Kingdom marital status,
1967–2002 prevalent
disease, weight
loss in past year,
BP medication,
height, skinfold
thickness,
systolic
BP, plasma
cholesterol,
glucose
intolerance,
diabetes
Kuriyama et al. 12 485 Stomach BMI Age, smoking,
(2005) Men ICD-9: 151.0–151.9 18.5–24.9 243 1.00 alcohol
Population-based Incidence 25.0–27.4 50 1.01 (0.74–1.37) consumption,
cohort 27..5–29.9 14 0.96 (0.56–1.65) diet, type
Japan ≥ 30 7 1.13 (0.53–2.41) of health
1984–1992 [Ptrend] [0.91] insurance; for
15 054 BMI women, also
Women 18.5–24.9 79 1.00 adjusted for
Incidence 25.0–27.4 26 1.19 (0.76–1.86) menopausal
27.5–29.9 17 1.80 (1.06–3.05) status, parity,
≥ 30 4 0.79 (0.29–2.17) age at menarche,
[Ptrend] [0.25] age at first
pregnancy
mortality
Lindblad et al. 11 023 Stomach BMI Age, sex,
(2005) Men and women < 20 29 1.05 (0.69–1.58) calendar year,
Case–control study Incidence 20–24 217 1.00 smoking,
nested in General 25–29 254 1.09 (0.90–1.32) alcohol
Practitioner ≥ 30 98 1.21 (0.94–1.56) consumption,
Research Database [Ptrend] [0.21] reflux
United Kingdom
1994–2001
Rapp et al. (2005) 67 447 Stomach BMI Age, smoking
VHM&PP Men ICD-9: 151 18.5–24.9 58 1.00 status,
(population-based Incidence 25–29.9 75 1.04 (0.73–1.47) occupation
cohort) ≥ 30 13 0.72 (0.40–1.33)
Austria [Ptrend] [0.44]
1985–2001 BMI Age, smoking
18.5–24.9 56 1.00 status,
25–29.9 36 0.78 (0.51–1.20) occupation
30–34.9 20 1.28 (0.76–2.15)
≥ 35 6 1.34 (0.57–3.13)
[Ptrend] [0.48]
Samanic et al. 362 552 Stomach BMI Attained age,
(2006) Men ICD-7: 151 18.5–24.9 666 1.00 calendar year,
Swedish Incidence 25–29.9 531 0.87 (0.77–0.97) smoking

Construction ≥ 30 84 0.83 (0.66–1.05)
Sweden
1958–1999
139
140

mortality
Fujino et al. (2007) 46 465 Stomach BMI Age, study area
JACC cohort Men < 18.5 54 1.00 (0.75–1.32)
Japan Mortality 18.5–24 569 1.00
1988–1997 25–29 89 0.78 (0.62–0.97)
≥ 30 7 1.04 (0.49–2.20)
Weight (kg)
< 55 280 1.00
55–62 260 0.88 (0.74–1.04)
≥ 63 198 0.83 (0.69–1.01)
Weight (kg) at age 20 yr
< 55 339 1.00
55–60 210 1.04 (0.84–1.30)
≥ 61 157 1.17 (0.93–1.48)
46 465 Stomach BMI Age, study area
Women < 18.5 37 1.44 (1.01–2.05)
Mortality 18.5–24 227 1.00
25–29 66 0.98 (0.74–1.30)
≥ 30 11 1.52 (0.82–2.80)
Weight (kg)
< 47 156 1.00
47–54 84 0.79 (0.60–1.03)
≥ 55 118 1.01 (0.78–1.29)
< 47 167 1.00
47–52 72 0.97 (0.70–1.34)
≥ 53 95 1.25 (0.92–1.70)
Máchová et al. 17 218 Stomach BMI 222 total Age, smoking, Nested case–
(2007) Men ICD-10: C16 18.5–24.9 1.00 hypertension, control study,
National Cancer Incidence 25–29.9 1.05 (0.74–1.47) height reporting odds
Registry ≥ 30 0.92 (0.57–1.50) ratios
Czech Republic 20 932 BMI 156 total Age, smoking,
1987–2002 Women 18.5–24.9 1.00 hypertension,
Incidence 25–29.9 0.81 (0.51–1.27) height
≥ 30 0.97 (0.60–1.57)
mortality
Merry et al. (2007) 4774 Stomach, BMI at baseline Age, sex,
Netherlands Men and women unspecified < 20 6 0.92 (0.38–2.25) smoking,
Cohort Study Incidence location 20–24.9 93 1.00 education
The Netherlands ICD-O-3: 25–29.9 67 0.85 (0.61–1.19) level, history of
1986–1999 C16.6–16.9 ≥ 30 7 0.77 (0.35–1.68) gastric ulcer or
Histology: 8140– [Ptrend] [0.33] bleeding
8141, 8190–8231, BMI at age 20 yr
8260–8263, 8310,
< 20 26 0.60 (0.37–0.99)
8430, 8480–8490,
20–21.4 49 1.00
8560, 8570–8572
21.5–22.9 40 0.92 (0.59–1.44)
23.0–24.9 26 0.70 (0.42–1.18)
≥ 25 12 0.82 (0.42–1.60)
[Ptrend] [0.72]
< 0 16 0.85 (0.47–1.55)
0–3.9 82 1.00
4–7.9 45 0.85 (0.56–1.27)
≥ 8 10 0.86 (0.41–1.80)
[Ptrend] [0.70]
Reeves et al. (2007) 1 222 630 Stomach BMI Incidence: Age,
Million Women Women ICD-10: C16 < 22.5 117 1.26 (1.05–1.51) geographical
Study Incidence and 22.5–24.9 121 1.00 (0.84–1.20) region, SES,

United Kingdom mortality 25–27.4 111 1.04 (0.86–1.25) reproductive
1996–2005 27.5–29.9 76 1.10 (0.88–1.38) history,
≥ 30 96 1.04 (0.84–1.27) smoking
BMI Mortality: status, alcohol
< 22.5 92 1.47 (1.19–1.81) consumption,
22.5–24.9 82 1.00 (0.80–1.24) physical activity,
25–27.4 85 1.16 (0.93–1.43) menopausal
27.5–29.9 64 1.34 (1.05–1.71) status,
≥ 30 80 1.24 (0.99–1.55) time since
menopause,
HRT use
141
142

mortality
Tanaka et al. (2007) 13 211 Stomach BMI at baseline Age, smoking, Too few incident
Population cohort Men ICD-9: 151 < 20.3 29 1.00 alcohol cases in women
from Takayama Mortality ICD-10: C16 20.3–22.2 20 0.68 (0.34–1.33) consumption, (results not
Japan > 22.2 16 0.53 (0.24–1.20) education level, shown)
1992–2000 [Ptrend] [0.12] physical activity,
BMI at age 20 yr marital status
< 20.3 12 1.00
20.3–22.2 33 2.53 (1.18–5.43)
> 22.2 41 1.72 (0.79–3.73)
[Ptrend] [0.76]
Jee et al. (2008) 770 556 Stomach BMI Age, smoking
Cohort from Men < 20.0 1808 1.04 (0.97–1.13)
the National Incidence 20.0–22.9 5602 1.07 (1.01–1.13)
Health Insurance 23.0–24.9 3839 1.00
Corporation 25.0–29.9 3188 1.09 (1.02–1.16)
Republic of Korea ≥ 30.0 131 1.31 (1.05–1.64)
1992–2006 [Ptrend] [0.50]
423 273 Stomach BMI Age, smoking
Women < 20.0 524 0.86 (0.75–1.00)
Incidence 20.0–22.9 1314 0.90 (0.80–1.00)
23.0–24.9 1035 1.00
25.0–29.9 1132 0.94 (0.84–1.05)
≥ 30.0 111 0.84 (0.64–1.11)
[Ptrend] [0.25]
Rapp et al. (2008) 28 711 Stomach BMI change per year Age, smoking
VHM&PP Men ICD-10: C16 < −0.1 11 0.75 (0.36–1.54) status, blood
(population-based Incidence −0.1– < 0.1 25 1.00 glucose,
cohort) 0.1– < 0.3 20 1.18 (0.65–2.13) occupational
Austria ≥ 0.3 10 1.22 (0.58–2.59) group, baseline
1985–2002 [Ptrend] [0.49] BMI
mortality
Rapp et al. (2008) 36 938 BMI change per year Age, smoking
(cont.) Women < −0.1 19 1.73 (0.82–3.63) status, blood
Incidence −0.1– < 0.1 12 1.00 glucose,
0.1– < 0.3 19 1.73 (0.84–3.57) occupational
≥ 0.3 9 1.11 (0.46–2.65) group, baseline
[Ptrend] [0.73] BMI
Sjödahl et al. (2008) 73 133 Stomach, BMI Age, sex,
Nord-Trondelag Men and women adenocarcinoma < 18.5 3 0.7 (0.1–5.2) physical activity,
Health Study Incidence ICD-7: 151.0, 151.8, 18.5–24.9 104 1.0 occupation, salt
Norway 151.9 25–29.9 110 1.0 (0.7–1.4) intake, smoking,
1984–2002 ≥ 30 32 1.1 (0.7–1.8) alcohol
[Ptrend] [0.74] consumption
Whitlock et al. 894 576 Stomach BMI, per 5 kg/m2 Study, sex, age,
(2009) Men and women ICD-9: 151 For BMI 15–25 934 0.86 (0.70–1.05) smoking
Pooled analysis of Mortality For BMI 25–50 651 1.11 (0.94–1.32)
57 cohort studies For BMI 15–50 0.98 (0.90–1.07)
Europe and North
America
Follow-up varied by
cohort
Parr et al. (2010) 326 387 Stomach BMI NR Age, sex,
Pooled analysis of Men and women ICD-9: 151 12–< 18.5 1.19 (0.87–1.62) smoking

39 cohort studies Mortality ICD-10: C16 18.5–24.9 1.00
Asia, Australia, and 25–29.9 1.05 (0.88–1.25)
New Zealand ≥ 30 1.04 (0.67–1.63)
1961–1999, median [Ptrend] [0.66]
follow-up 4 yr
Chen et al. (2012) 142 214 Stomach BMI Age, area,
Population-based Men 15–23.5 757 0.74 (0.59–0.94) smoking,
cohort of men Mortality 23.5–35 198 0.96 (0.61–1.49) alcohol
China consumption,
1990–2006 education level
143
144

mortality
Lindkvist et al. 289 866 Stomach BMI, quintiles Smoking, age, Ranges of BMI
(2013) Men ICD-7: 151 Q1 157 1.00 study cohort, quintiles not
Metabolic Incidence Q2 134 0.79 (0.62–0.99) year of birth specified
Syndrome and Q3 154 0.84 (0.67–1.05)
Cancer Project Q4 197 1.02 (0.83–1.26)
(Me-Can) pooled Q5 186 1.00 (0.80–1.24)
analysis of [Ptrend] [0.26]
prospective cohorts 288 834 Stomach BMI, quintiles Smoking, age, Ranges of BMI
Austria, Norway, Women ICD-7: 151 Q1 59 1.00 study cohort, quintiles not
and Sweden Incidence Q2 65 0.92 (0.65–1.31) year of birth specified
1972–2006, follow- Q3 63 0.73 (0.51–1.05)
up varied by cohort Q4 104 1.01 (0.72–1.40)
Q5 91 0.85 (0.61–1.20)
[Ptrend] [0.68]
Bhaskaran et al. 5 243 978 Stomach BMI 3337 total Age, sex, Stronger
(2014) Incidence ICD-10: C16 per 5 kg/m2 increase 1.03 (0.98–1.09) diabetes, association in
Population-based [Ptrend] [0.16] smoking, non-smokers
cohort: Clinical alcohol
Practice Research consumption,
Datalink SES, calendar
United Kingdom year
1987–2012
Camargo et al. 483 700 Stomach BMI 1000 total Age, sex,
(2014) Men and women ICD-10: C16.0–16.9 18.5–24.9 1.00 education
NIH-AARP cohort Incidence 25–29.9 1.05 (0.90–1.22) level, cigarette
USA 30–34.9 1.40 (1.16–1.68) smoking
1995–2006 ≥ 35 1.57 (1.21–2.04)
Weight, tertiles
T1 1.00
T2 1.00 (0.86–1.17)
T3 1.18 (1.01–1.38)
mortality
Gastric cardia
Samanic et al. 4 500 700 Gastric cardia Obesity Age, calendar Obesity defined
(2004) Men ICD-9: 151.0 White men: year as discharge
USA 277; ICD-9: 278.0
1969–1996 Only 5 cases were
available among
Black men
Lindblad et al. 10 195 Gastric cardia BMI Age, sex,
United Kingdom
1994–2001
Tran et al. (2005) 29 584 Gastric cardia BMI 1089 total Age, sex
Linxian General Men and women < 20 1.00
Population Trial Incidence 20–21 0.98 (0.84–1.16)
China 22 0.96 (0.81–1.13)
1986–2001 ≥ 23 0.95 (0.80–1.13)

[Ptrend] [0.51]
Samanic et al. 362 552 Gastric cardia BMI Attained age,
(2006) Men ICD-7: 151.0 18.5–24.9 108 1.00 calendar year,
Construction ≥ 30 16 1.09 (0.64–1.85)
Worker Cohort [Ptrend] [0.40]
Sweden
1958–1999
145
146

mortality
Merry et al. (2007) 4774 Gastric cardia BMI at baseline Age, sex
Netherlands Men and women ICD-O-3: C16.0 < 20 2 0.67 (0.16–2.80)
Cohort Study Incidence Histology: 8140– 20–24.9 68 1.00
The Netherlands 8141, 8190–8231, 25–29.9 76 1.32 (0.94–1.85)
1986–1999 8260–8263, 8310, ≥ 30 17 2.73 (1.56–4.79)
8430, 8480–8490, [Ptrend] [0.002]
8560, 8570–8572 BMI at age 20 yr
< 20 21 0.66 (0.39–1.14)
20–21.4 40 1.00
21.5–22.9 39 1.02 (0.65–1.60)
23.0–24.9 22 0.75 (0.44–1.28)
≥ 25 16 1.47 (0.81–2.70)
[Ptrend] [0.17]
< 0 10 0.68 (0.34–1.35)
0–3.9 70 1.00
4–7.9 45 1.22 (0.82–1.82)
≥ 8 13 2.07 (1.08–3.97)
[Ptrend] [0.02]
Abnet et al. (2008) 480 475 Gastric cardia BMI Age, sex,
NIH-AARP cohort Men and women ICD-O-3: C16.0 < 18.5 1 0.70 (0.10–5.06) cigarette
USA Incidence Histology: 18.5–24.9 76 1.00 smoking,
1995–2003 “adenocarcinoma” 25–29.9 128 1.06 (0.79–1.41) alcohol
30–34.9 71 1.70 (1.22–2.36) consumption,
≥ 35 31 2.46 (1.60–3.80) education level,
physical activity
mortality
Corley et al. (2008) 3150 Gastric cardia BMI Age, sex, year of
Nested case– Men and women ICD-10: C16.0 < 18.5 0 – health check-up
control of Kaiser Incidence Histology: 18.5–24.9 43 1.00 BMI results also
Permanente 8140–8573 25–29.9 40 0.91 (0.55–1.53) adjusted for
Multiphasic Health ≥ 30 16 2.04 (0.99–4.21) ethnicity
Check-up cohort per 1 kg/m2 increase 1.04 (0.98–1.09)
USA Sagittal abdominal diameter (cm)
1964–1973 < 20 16 1.00
20–22.4 12 0.69 (0.29–1.60)
22.5–25 12 1.17 (0.49–2.84)
≥ 25 14 1.28 (0.38–4.25)
per 1 cm increase 1.03 (0.95–1.11)
O’Doherty et al. 218 854 Gastric cardia BMI Age, sex, total
(2012) Men and women ICD-10: C16.0 < 18.5 2 2.57 (0.62–10.65) energy intake,
NIH-AARP cohort Incidence 18.5–24.9 50 1.00 antacid use,
USA 25–29.9 79 1.15 (0.80–1.65) aspirin use,
1995–2006 30–34.9 45 2.16 (1.41–3.29) NSAID use,
≥ 35 15 3.67 (2.00–6.71) marital status,
[Ptrend] [< 0.01] diabetes,
Weight, quartiles (sex-specific) cigarette
smoking,
Q1 28 1.00
education
Q2 46 1.66 (1.03–2.67)

level, ethnicity,
Q3 44 1.53 (0.93–2.51)
alcohol
Q4 73 2.52 (1.55–4.11)
consumption,
[Ptrend] [< 0.01]
physical activity,
diet
147
148

mortality
O’Doherty et al. WC, quartiles (sex-specific)
(2012) Q1 30 1.00
(cont.) Q2 38 1.32 (0.82–2.14)
Q3 51 1.29 (0.82–2.04)
Q4 72 2.22 (1.43–3.47)
[Ptrend] [< 0.01]
Camargo et al. 483 700 Gastric cardia BMI 478 total Age, sex,
(2014) Men and women ICD-10: C16.0 18.5–24.9 1.00 education
USA 30–34.9 1.64 (1.26–2.14) smoking
1995–2006 ≥ 35 2.24 (1.58–3.17)
Weight, tertiles
T1 1.00
T2 1.20 (0.94–1.52)
T3 1.53 (1.21–1.92)
Steffen et al. (2015) 391 456 Gastric cardia BMI, quintiles Age, centre, Sex-specific
EPIC cohort Men and women ICD-10: C16.0 Q1 31 1.00 sex, education quintiles for
10 European Incidence/mortality Q2 37 1.09 (0.68–1.77) level, smoking, weight, BMI,
countries Q3 48 1.37 (0.87–2.17) alcohol and WC. Cut-
1992–2008 Q4 41 1.20 (0.74–1.94) consumption, off points not
Q5 36 1.17 (0.71–1.92) physical activity, provided, only
[Ptrend] [0.53] diet, height the median
Weight, quintiles values for each
Q1 33 1.00
Q2 37 1.14 (0.71–1.84)
Q3 43 1.29 (0.81–2.08)
Q4 38 1.11 (0.68–1.83)
Q5 42 1.26 (0.75–2.10)
[Ptrend] [0.48]
mortality
Steffen et al. (2015) WC, quintiles
(cont.) Q1 22 1.00
Q2 31 1.20 (0.69–2.09)
Q3 40 1.41 (0.83–2.40)
Q4 42 1.52 (0.89–2.58)
Q5 45 1.59 (0.93–2.73)
[Ptrend] [0.06]
Gastric non-cardia
Samanic et al. 4 500 700 Gastric non-cardia Obesity Age, calendar Obesity defined
(2004) Men ICD-9: 151.x White men: year as discharge
1969–1996 Non-obese 1958 1.00
Obese 94 0.99 (0.80–1.22)
Lindblad et al. 10 327 Gastric non-cardia BMI Age, sex,
United Kingdom

1994–2001
Tran et al. (2005) 29 584 Gastric non-cardia BMI 363 total Age, sex
Linxian General Men and women < 20 1.00
Population Trial Incidence 20–21 1.00 (0.76–1.32)
China 22 0.91 (0.68–1.20)
1986–2001 ≥ 23 0.68 (0.49–0.93)
[Ptrend] [0.017]
149
150

mortality
MacInnis et al. 41 295 Gastric non-cardia BMI 68 total Sex, country of
(2006) Men and women ICD-9: 151.1–151.9 < 25 1.0 birth, education
Melbourne Incidence/mortality ICD-10: C16.1–16.9 25–29 0.5 (0.3–1.0) level, physical
Collaborative ≥ 30 1.0 (0.5–1.8) activity
Cohort Study [Ptrend] [0.76]
Australia Weight (kg)
1990–2004 Men: Women:
< 75 < 62 1.0
75–83 62–70 0.6 (0.3–1.1)
≥ 84 ≥ 71 1.1 (0.6–1.9)
[Ptrend] [0.62]
WC (cm)
Men: Women:
< 94 < 80 1.0
94–101 80–87 0.8 (0.4–1.4)
≥ 102 ≥ 88 1.1 (0.6–2.0)
[Ptrend] [0.57]
Samanic et al. 362 552 Gastric non-cardia BMI Attained age,
(2006) Men ICD-7: 151.x 18.5–24.9 558 1.00 calendar year,
Construction ≥ 30 68 0.78 (0.61–1.01)
Sweden
1958–1999
Merry et al. (2007) 4774 Gastric non-cardia BMI at baseline Age, sex,
Netherlands Men and women ICD-10: C16.1–16.5 < 20 12 1.80 (0.96–3.39) current
Cohort Study Incidence Histology: 8140– 20–24.9 115 1.00 smoking,
The Netherlands 8141, 8190–8231, 25–29.9 99 0.97 (0.73–1.30) number of
1986–1999 8260–8263, 8310, ≥ 30 9 0.68 (0.34–1.35) cigarettes
8430, 8480–8490, [Ptrend] [0.13] smoked per
8560, 8570–8572 day, smoking
duration,
education level
mortality
Merry et al. (2007) BMI at age 20 yr
(cont.) < 20 53 1.40 (0.91–2.15)
20–21.4 40 1.00
21.5–22.9 49 1.24 (0.80–1.91)
23.0–24.9 36 1.12 (0.69–1.80)
≥ 25 20 1.60 (0.91–2.83)
[Ptrend] [0.93]
< 0 17 0.77 (0.44–1.36)
0–3.9 106 1.00
4–7.9 61 0.85 (0.60–1.21)
≥ 8 14 0.86 (0.46–1.59)
[Ptrend] [0.77]
Abnet et al. (2008) 480 475 Gastric non-cardia BMI Age, sex,
NIH-AARP cohort Men and women ICD-O-3: < 18.5 7 2.97 (1.38–6.39) cigarette
USA Incidence C16.1–16.9 18.5–24.9 107 1.00 smoking,
1995–2003 Histology: 25–29.9 123 0.80 (0.61–1.04) alcohol
“adenocarcinoma” 30–34.9 61 1.08 (0.78–1.50) consumption,
≥ 35 17 0.84 (0.50–1.42) education level,
physical activity
Persson et al. 44 453 Stomach, non- BMI Age, family Similar results in
(2008) Women cardia < 20 53 1.00 history of postmenopausal

Japan Public Health Incidence ICD-10: C16.2-16.7 20–24.9 225 0.82 (0.61–1.11) gastric cancer, women only
Center-based ≥ 25 90 0.74 (0.53–1.04) study area
Prospective Study [Ptrend] [0.10]
Japan Stomach, BMI Similar results in
1990–2004 non-cardia, < 20 12 1.00 postmenopausal
differentiated 20–24.9 56 0.93 (0.50–1.74) women only
cancer type ≥ 25 29 1.12 (0.57–2.21)
ICD-10: C16.2-16.7 [Ptrend] [0.59]
151
152

mortality
Persson et al. Stomach, BMI Similar results in
(2008) non-cardia, < 20 37 1.00 postmenopausal
(cont.) undifferentiated 20–24.9 153 0.79 (0.55–1.14) women only
cancer type ≥ 25 52 0.60 (0.39–0.91)
ICD-10: C16.2-16.7 [Ptrend] [0.01]
Sjödahl et al. (2008) 73 133 Gastric non-cardia BMI Age, sex,
Nord-Trondelag Men and women ICD-7: 151.0, 151.8, < 18.5 2 0.9 (0.1–6.7) physical activity,
Health Study Incidence 151.9 18.5–24.9 84 1.0 occupation, salt
Norway 25–29.9 92 1.1 (0.7–1.6) intake, smoking,
1984–2002 ≥ 30 29 1.2 (0.7–2.1) alcohol
[Ptrend] [0.42] consumption
O’Doherty et al. 218 854 Gastric non-cardia BMI Age, sex, total
(2012) Men and women ICD-10: C16.1–16.7 < 18.5 1 1.34 (0.18–9.79) energy intake,
NIH-AARP cohort Incidence 18.5–24.9 37 1.00 antacid use,
USA 25–29.9 60 1.32 (0.86–2.00) aspirin use,
1995–2006 30–34.9 23 1.46 (0.84–2.51) NSAID use,
≥ 35 4 0.99 (0.34–2.84) marital status,
[Ptrend] [0.38] diabetes,
Weight, quartiles (sex-specific) cigarette
Q1 20 1.00 smoking,
Q2 35 1.93 (1.10–3.38) education
Q3 32 1.73 (0.96–3.10) level, ethnicity,
Q4 38 1.93 (1.05–3.54) alcohol
[Ptrend] [0.07] consumption,
physical activity,
WC, quartiles (sex-specific) diet
Q1 21 1.00
Q2 26 1.27 (0.71–2.26)
Q3 40 1.41 (0.82–2.41)
Q4 38 1.46 (0.83–2.55)
[Ptrend] [0.19]
mortality
Camargo et al. 483 700 Gastric non-cardia BMI 522 total Age, sex,
(2014) Men and women ICD-10: C16.1–16.6 18.5–24.9 1.00 education
USA 30–34.9 1.38 (0.99–1.92) smoking
1995–2006 ≥ 35 1.05 (0.61–1.82)
Weight, tertiles
T1 1.00
T2 1.00 (0.76–1.32)
T3 1.02 (0.77–1.34)
Steffen et al. (2015) 391 456 Gastric non-cardia BMI, quintiles Sex, education Sex-specific
EPIC cohort Men and women ICD-10: C16.1–16.9 Q1 36 1.00 level, smoking, quintiles for
10 European Incidence/mortality Q2 36 0.77 (0.48–1.22) alcohol weight, BMI,
countries Q3 33 0.61 (0.38–0.99) consumption, and WC. Cut-
1992–2008 Q4 49 0.78 (0.50–1.22) physical activity, off points not
Q5 70 0.99 (0.64–1.54) diet, height provided, only
[Ptrend] [0.41] the median
Weight, quintiles values for each
Q1 50 1.00
Q2 35 0.68 (0.44–1.06)
Q3 36 0.67 (0.43–1.06)
Q4 57 1.02 (0.68–1.55)
Q5 46 0.84 (0.53–1.32)

[Ptrend] [0.94]
WC, quintiles
Q1 25 1.00
Q2 25 0.81 (0.46–1.42)
Q3 33 0.89 (0.52–1.52)
Q4 66 1.58 (0.97–2.57)
Q5 55 1.14 (0.68–1.91)
[Ptrend] [0.12]
BMI, body mass index (in kg/m 2); BP, blood pressure; CI, confidence interval; EPIC, European Prospective Investigation into Cancer and Nutrition; HRT, hormone replacement therapy;
ICD, International Classification of Diseases; JACC, Japan Collaborative Cohort Study for Evaluation of Cancer Risk; NIH-AARP, National Institutes of Health–AARP Diet and Health
Study; NSAID, non-steroidal anti-inflammatory drug; SES, socioeconomic status; VHM&PP, Vorarlberg Health Monitoring and Prevention Program; WC, waist circumference; yr, year
or years
153
154

Table 2.2.3b Case–control studies of measures of body fatness and cancer of the stomach
Reference Total Organ site Exposure categories Exposed Relative risk Adjustment for Comments
Study number of cases (95% CI) confounding
location cases
Period Source of
controls
Stomach
Hansson et al. 338 Stomach BMI at age 20 yr Age, height No differences were
(1994) Population Men: observed in the
Sweden ≤ 21.20 37 1.00 associations by age at
1989–1992 21.21–22.60 40 1.06 (0.63–1.86) interview (age groups:
22.62–24.20 45 1.09 (0.66–1.82) < 59 yr, 60–69 yr, and
≥ 24.21 84 2.16 (1.35–3.46) ≥ 70 yr)
continuous 1.12 (1.05–1.20) No associations were
Women: found between BMI
≤ 19.20 12 1.00 and GC 20 yr before the
19.21–20.80 18 1.39 (0.60–3.23) interview
20.81–23.30 40 3.06 (1.43–6.58)
≥ 23.21 28 2.14 (0.96–4.78)
continuous 1.11 (1.02–1.21)
Muñoz et al. 292 Stomach BMI Age, sex Similar results for
(2001) Population < 18.5 51 11.0 (4.8–27.0) self-reported weight at
Venezuela 18.5–25.0 200 1.0 current age. Increased
1991–1997 > 25.0 41 0.3 (0.2–0.4) risk in overweight
cases with self-reported
weight in childhood,
adolescence, and early
adulthood
Inoue et al. Women: Stomach Current BMI Age, year, season Postmenopausal women
(2002) 365 Upper third 72 total of interview, only. P values for trend
Japan Population < 21.08 1.00 family history were non-significant
1988–1998 21.08–23.56 1.69 (0.91–3.12) of GC, smoking among all subsites, both
> 23.56 1.07 (0.54–2.10) status, intake of for current BMI and for
Middle third 155 total raw vegetables and BMI at age 20 yr
< 21.08 1.00 fish
21.08–23.56 0.75 (0.49–1.16)
> 23.56 0.80 (0.52–1.22)
Lower third 127 total
< 21.08 1.00
21.08–23.56 1.02 (0.63–1.66)
> 23.56 1.16 (0.72–1.89)
location cases
Period Source of
controls
Inoue et al. BMI at age 20 yr
(2002) Upper third 72 total
(cont.) < 21.08 1.00
21.08–23.56 1.33 (0.69–2.55)
> 23.56 1.33 (0.69–2.58)
Middle third 155 total
< 21.08 1.00
21.08–23.56 1.83 (1.14–2.94)
> 23.56 1.81 (1.12–2.93)
Lower third 127 total
< 21.08 1.00
21.08–23.56 0.88 (0.52–1.50)
> 23.56 1.31 (0.81–2.12)
Chung et al. Men: 374 Stomach Current BMI Age Study in young
(2010) Women: Men: individuals (ages
Republic of 270 > 35 vs ≤ 35 374 total 1.94 (1.63–2.37) 18–45 yr)
Korea Hospital Women:
1990–2008 > 35 vs ≤ 35 270 total 1.65 (1.34–2.04)
Praud et al. Men: 612 Stomach BMI Age, sex, study,
(2014) Women: Men: year of interview,
Italy 387 < 25 vs ≥ 25 646 total 0.85 (0.79–0.90) education level,
1985–2007 Hospital tobacco smoking,

[Ptrend] [< 0.0001]
Women: family history,
< 25 vs ≥ 25 348 total 0.86 (0.79–0.93) total energy intake
[Ptrend] [0.0009]
Kim et al. Men: 663 Stomach BMI measured at endoscopy Age, smoking No significant
(2015) Women: Men: status, drinking associations were
Republic of 335 < 23 286 1.00 status, family observed when
Korea Hospital 23− < 25 193 1.25 (0.87–1.81) history of GC, stratifying by cardia and
2003–2013 ≥ 25− < 30 175 1.33 (0.92–1.92) Helicobacter pylori non-cardia GC
≥ 30 9 1.27 (0.42–3.86) infection, atrophic
[Ptrend] [0.43] gastritis, intestinal
metaplasia, serum
pepsinogen I/II
155
ratio
156

location cases
Period Source of
controls
Kim et al. Women:
(2015) < 23 182 1.00
(cont.) 23− < 25 73 0.92 (0.6–1.43)
≥ 25− < 30 69 1.11 (0.70–1.77)
≥ 30 11 0.86 (0.33–2.26)
[Ptrend] [0.904]
Song et al. 1492 Stomach BMI at age 18 yr Age, smoking
(2015) Population Men: status, alcohol
Republic of 21.75 1.00 drinking status,
Korea ≥ 25.3 1.13 (1.01–1.55) regular exercise,
2010–2014 Women: family history of
21.75 1.00 GC, past medical
≥ 25.3 1.25 (1.01–1.55) history
Gastric cardia
Vaughan et al. 165 Gastric cardia, BMI, percentiles Age, sex, BMI percentiles
(1995) Population adenocarcinoma 1–10% 13 0.8 (0.4–1.8) education level, (derived from in- person
USA (13 10–49% 52 1.0 race, cigarette interviews) based on
counties of 50–89% 74 1.3 (0.8–2.1) smoking, alcohol distribution of controls
Washington 90–100% 25 1.6 (0.8–3.0) consumption for each sex separately
State)
1993–1990
Chow et al. 365 Gastric cardia BMI (sex-specific) Geographical BMI up to 1 yr before
(1998) Population Men: Women: location, age, sex, diagnosis for cases and
USA < 23.12 < 21.95 54 1.0 race, cigarette date of interview for
1993–1995 23.12–25.08 21.95–24.12 51 0.9 (0.6–1.5) smoking, controls
25.09–27.31 24.13–27.43 70 1.4 (0.9–2.1) respondent status
≥ 27.32 ≥ 27.44 86 1.6 (1.1–2.6)
[Ptrend] [0.008]
location cases
Period Source of
controls
Lagergren 262 Gastric cardia BMI 20 yr before interview Age, sex, tobacco
et al. (1999) Population < 22 47 1.0 smoking, alcohol
Sweden 22–24.9 100 1.3 (0.8–1.9) consumption, SES,
1995–1997 25–30 91 2.2 (1.4–3.4) reflux symptoms,
> 30 24 4.3 (2.1–8.7) intake of fruits
[Ptrend] [< 0.001] and vegetables,
BMI at age 20 yr, quartiles (sex-specific) energy intake,
physical activity
Men: Women:
< 20.7 < 19.3 52 1.0
20.7–22.1 19.3–20.4 46 0.8 (0.5–1.3)
22.2–23.7 20.5–22.1 65 1.2 (0.8–1.9)
> 23.7 > 22.1 99 1.9 (1.3–2.9)
[Ptrend] [< 0.001]
Wu et al. 277 Gastric cardia BMI at age 40 yr, quartiles (sex-specific) Smoking, age, sex,
(2001) Population Men: Women: 247 total race, education
USA (proxy ≤ 22 ≤ 21 1.00 level
1992–1997 control) > 22– ≤ 25 > 21– ≤ 23 1.49 (1.0–2.1)
> 25– ≤ 27 > 23– ≤ 25 1.45 (0.9–2.3)
> 27 > 25 2.08 (1.4–3.2)
[Ptrend] [0.016]

≤ 20 ≤ 18 1.00
> 20– ≤ 22 > 18– ≤ 20 1.13 (0.8–1.7)
> 22– ≤ 24 > 20– ≤ 22 1.36 (0.9–2.0)
> 24 > 22 1.71 (1.2–2.6)
[Ptrend] [0.006]
Gastric non-cardia
Chow et al. 365 Gastric non- BMI up to 1 yr before diagnosis (sex-specific) Geographical BMI up to 1 yr before
(1998) Population cardia Men: Women: location, age, sex, diagnosis for cases and
USA < 23.12 < 21.95 105 1.0 race, cigarette date of interview for
1993–1995 23.12–25.08 21.95–24.12 77 0.9 (0.6–1.4) smoking, controls
25.09–27.31 24.13–27.43 91 1.2 (0.8–1.8) respondent status
≥ 27.32 ≥ 27.44 92 1.2 (0.8–1.8)
157
[Ptrend] [2.14]
158

location cases
Period Source of
controls
Wu et al. 443 Gastric non- BMI at age 40 yr, quartiles (sex-specific) Smoking, age, sex, Results did not change
(2001) Population cardia Men: Women: 352 total race, education when stratifying by
USA ≤ 22 ≤ 21 1.00 level Whites/non-Whites or
1992–1997 > 22– ≤ 25 > 21– ≤ 23 0.86 (0.6–1.2) by sex
> 25– ≤ 27 > 23– ≤ 25 1.00 (0.7–1.5)
> 27 > 25 1.10 (0.8–1.6)
[Ptrend] [0.57]
≤ 20 ≤ 18 1.00
> 20– ≤ 22 > 18– ≤ 20 1.21 (0.9–1.7)
> 22– ≤ 24 > 20– ≤ 22 1.39 (1.0–2.0)
> 24 > 22 1.43 (1.0–2.1)
[Ptrend] [0.03]
BMI, body mass index (in kg/m 2); CI, confidence interval; GC, gastric cancer; SES, socioeconomic status; yr, year or years
Table 2.2.3c Meta-analyses of measures of body fatness and cancer of the stomach
Reference Total number Organ Exposure categories Relative risk Adjustment for Comments
Period of studies site (95% CI) confounding
Total number
of cases
Renehan et al. Men: Stomach BMI Age (all studies) and other
(2008) 8 prospective per 5 kg/m2 increase 0.97 (0.88–1.06) factors (not in all studies)
1996–2007 studies
817 incident
cases
Women: Stomach BMI
5 prospective per 5 kg/m2 increase 1.04 (0.90–1.20)
studies
325 incident
cases
Yang et al. 12 prospective Stomach BMI NR No differences in risk by
(2009) studies Overweight and obese vs normal 1.22 (1.06–1.41) sex; normal, overweight,
1950–2009 9492 incident Obese vs normal 1.36 (1.21–1.54) and obese are defined
cases Overweight vs normal 1.21 (1.08–1.36) in most studies as BMI
3 prospective Cardia BMI NR of 18.5–25, 25–29.9, and
studies Overweight and obese vs normal 1.55 (1.31–1.84) ≥ 30, respectively
Obese vs normal 2.06 (1.63–2.61)
Overweight vs normal 1.40 (1.16–1.68)
4 prospective Non- BMI NR
studies cardia Overweight and obese vs normal 1.18 (0.96–1.45)
Obese vs normal 1.26 (0.89–1.78)
Overweight vs normal 1.16 (0.94–1.43)
Chen et al. 12 prospective Stomach BMI Stronger associations in

(2013) studies 18.5– < 25 1.00 men in both BMI groups
1994–2012 41 791 incident 25–29.9 1.01 (0.96–1.07)
cases ≥ 30 1.06 (0.99–1.12)
7 prospective Cardia BMI
studies 18.5– < 25 1.00
25–29.9 1.21 (1.03–1.42)
≥ 30 1.82 (1.32–2.49)
8 prospective Non- BMI
studies cardia 18.5– < 25 1.00
25–29.9 0.93 (0.82–1.05)
≥ 30 1.00 (0.87–1.15)
159
160

Table 2.2.3c (continued)
Reference Total number Organ Exposure categories Relative risk Adjustment for Comments
Period of studies site (95% CI) confounding
Total number
of cases
Lin et al. 13 prospective Stomach BMI Age and others (not Stronger association
(2014) studies and 3 18.5– < 25 1.00 specified) of obesity with risk in
NR case–controls 25–29.9 1.13 (1.03–1.24) men (5 studies) and in
NR ≥ 30 1.04 (0.96–1.12) non-Asian population (11
Cardia BMI Age and others (not studies)
18.5– < 25 1.00 specified)
25–29.9 1.61 (1.15–2.24)
≥ 30 1.22 (1.05–1.42)
Non- BMI Age and others (not
cardia 18.5– < 25 1.00 specified)
25–29.9 0.83 (0.68–1.01)
≥ 30 0.94 (0.81–1.10)
BMI, body mass index (in kg/m 2); CI, confidence interval; CRC, colorectal cancer; HRT, hormone replacement therapy; IBD, inflammatory bowel disease; NR, not reported
References Corley DA, Kubo A, Zhao W (2008). Abdominal obesity

and the risk of esophageal and gastric cardia carci-
nomas. Cancer Epidemiol Biomarkers Prev, 17(2):352–8.
Abnet CC, Freedman ND, Hollenbeck AR, Fraumeni JF Jr, doi:10.1158/1055-9965.EPI-07-0748 PMID:18268119
Leitzmann M, Schatzkin A (2008). A prospective study Fujino Y; Japan Collaborative Cohort Study for Evaluation
of BMI and risk of oesophageal and gastric adeno- of Cancer (2007). Anthropometry, development history
carcinoma. Eur J Cancer, 44(3):465–71. doi:10.1016/j. and mortality in the Japan Collaborative Cohort Study
ejca.2007.12.009 PMID:18221867 for Evaluation of Cancer (JACC). Asian Pac J Cancer
Batty GD, Shipley MJ, Jarrett RJ, Breeze E, Marmot MG, Prev, 8(Suppl):105–12. PMID:18260709
Smith GD (2005). Obesity and overweight in rela- Hansson LE, Baron J, Nyrén O, Bergström R, Wolk A,
tion to organ-specific cancer mortality in London Lindgren A, et al. (1994). Early-life risk indicators
(UK): findings from the original Whitehall study. of gastric cancer. A population-based case-control
Int J Obes, 29(10):1267–74. doi:10.1038/sj.ijo.0803020 study in Sweden. Int J Cancer, 57(1):32–7. doi:10.1002/
PMID:15997248 ijc.2910570107 PMID:8150538
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, IARC (2002). Weight control and physical activity. Lyon,
Leon DA, Smeeth L (2014). Body-mass index and risk France: IARC Press (IARC Handbooks of Cancer
of 22 specific cancers: a population-based cohort study Prevention, Vol. 6). Available from: http://publications.
of 5.24 million UK adults. Lancet, 384(9945):755–65. iarc.fr/376.
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 IARC (2012). Personal habits and indoor combustions.
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ IARC Monogr Eval Carcinog Risks Hum, 100E:1–
(2003). Overweight, obesity, and mortality from cancer 575. Available from: http://publications.iarc.fr/122
in a prospectively studied cohort of U.S. adults. N Engl PMID:23193840
J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423 Inoue M, Ito LS, Tajima K, Yamamura Y, Kodera Y,
PMID:12711737 Takezaki T, et al. (2002). Height, weight, menstrual and
Camargo MC, Freedman ND, Hollenbeck AR, Abnet reproductive factors and risk of gastric cancer among
CC, Rabkin CS (2014). Height, weight, and body mass Japanese postmenopausal women: analysis by subsite
index associations with gastric cancer subsites. Gastric and histologic subtype. Int J Cancer, 97(6):833–8.
Cancer, 17(3):463–8. doi:10.1007/s10120-013-0312-4 doi:10.1002/ijc.10149 PMID:11857364
PMID:24174008 Jee SH, Yun JE, Park EJ, Cho ER, Park IS, Sull JW, et al.
Chen Y, Liu L, Wang X, Wang J, Yan Z, Cheng J, et al. (2008). Body mass index and cancer risk in Korean men
(2013). Body mass index and risk of gastric cancer: and women. Int J Cancer, 123(8):1892–6. doi:10.1002/
a meta-analysis of a population with more than ten ijc.23719 PMID:18651571
million from 24 prospective studies. Cancer Epidemiol Jemal A, Vineis P, Bray F, Torre L, Forman D (2014). The
Biomarkers Prev, 22(8):1395–408. doi:10.1158/1055- cancer atlas. 2nd ed. Atlanta (GA), USA: American
9965.EPI-13-0042 PMID:23697611 Cancer Society.
Chen Z, Yang G, Offer A, Zhou M, Smith M, Peto R, Kamangar F, Dawsey SM, Blaser MJ, Perez-Perez GI,
et al. (2012). Body mass index and mortality in Pietinen P, Newschaffer CJ, et al. (2006). Opposing
China: a 15-year prospective study of 220 000 men. risks of gastric cardia and noncardia gastric adenocar-
Int J Epidemiol, 41(2):472–81. doi:10.1093/ije/dyr208 cinomas associated with Helicobacter pylori seroposi-
PMID:22296991 tivity. J Natl Cancer Inst, 98(20):1445–52. doi:10.1093/
Chow WH, Blot WJ, Vaughan TL, Risch HA, Gammon jnci/djj393 PMID:17047193
MD, Stanford JL, et al. (1998). Body mass index and Kim HJ, Kim N, Kim HY, Lee HS, Yoon H, Shin CM, et al.
risk of adenocarcinomas of the esophagus and gastric (2015). Relationship between body mass index and the
cardia. J Natl Cancer Inst, 90(2):150–5. doi:10.1093/ risk of early gastric cancer and dysplasia regardless of
jnci/90.2.150 PMID:9450576 Helicobacter pylori infection. Gastric Cancer, 18(4):762–
Chung HW, Noh SH, Lim JB (2010). Analysis of demo- 73. doi:10.1007/s10120-014-0429-0 PMID:25240409
graphic characteristics in 3242 young age gastric cancer Kuriyama S, Tsubono Y, Hozawa A, Shimazu T, Suzuki Y,
patients in Korea. World J Gastroenterol, 16(2):256–63. Koizumi Y, et al. (2005). Obesity and risk of cancer in
doi:10.3748/wjg.v16.i2.256 PMID:20066747 Japan. Int J Cancer, 113(1):148–57. doi:10.1002/ijc.20529
Colquhoun A, Arnold M, Ferlay J, Goodman KJ, Forman PMID:15386435
D, Soerjomataram I (2015). Global patterns of cardia Lagergren J, Bergström R, Nyrén O (1999). Association
and non-cardia gastric cancer incidence in 2012. between body mass and adenocarcinoma of the esoph-
Gut, 64(12):1881–8. doi:10.1136/gutjnl-2014-308915 agus and gastric cardia. Ann Intern Med, 130(11):883–
PMID:25748648 90. doi:10.7326/0003-4819-130-11-199906010-00003
PMID:10375336
161
Lin XJ, Wang CP, Liu XD, Yan KK, Li S, Bao HH, et al. Rapp K, Klenk J, Ulmer H, Concin H, Diem G, Oberaigner
(2014). Body mass index and risk of gastric cancer: W, et al. (2008). Weight change and cancer risk in a
a meta-analysis. Jpn J Clin Oncol, 44(9):783–91. cohort of more than 65,000 adults in Austria. Ann
doi:10.1093/jjco/hyu082 PMID:24951830 Oncol, 19(4):641–8. doi:10.1093/annonc/mdm549
Lindblad M, Rodríguez LA, Lagergren J (2005). Body PMID:18056917
mass, tobacco and alcohol and risk of esophageal, Rapp K, Schroeder J, Klenk J, Stoehr S, Ulmer H, Concin
gastric cardia, and gastric non-cardia adenocarcinoma H, et al. (2005). Obesity and incidence of cancer: a
among men and women in a nested case-control study. large cohort study of over 145,000 adults in Austria.
Cancer Causes Control, 16(3):285–94. doi:10.1007/ Br J Cancer, 93(9):1062–7. doi:10.1038/sj.bjc.6602819
s10552-004-3485-7 PMID:15947880 PMID:16234822
Lindkvist B, Almquist M, Bjørge T, Stocks T, Borena W, Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D;
Johansen D, et al. (2013). Prospective cohort study of Million Women Study Collaboration (2007). Cancer
metabolic risk factors and gastric adenocarcinoma incidence and mortality in relation to body mass
risk in the Metabolic Syndrome and Cancer Project index in the Million Women Study: cohort study.
(Me-Can). Cancer Causes Control, 24(1):107–16. BMJ, 335(7630):1134 doi:10.1136/bmj.39367.495995.AE
doi:10.1007/s10552-012-0096-6 PMID:23149498 PMID:17986716
Máchová L, Cízek L, Horáková D, Koutná J, Lorenc J, Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen
Janoutová G, et al. (2007). Association between obesity M (2008). Body-mass index and incidence of cancer:
and cancer incidence in the population of the District a systematic review and meta-analysis of prospec-
Sumperk, Czech Republic. Onkologie, 30(11):538–42. tive observational studies. Lancet, 371(9612):569–78.
doi:10.1159/000108284 PMID:17992023 doi:10.1016/S0140-6736(08)60269-X PMID:18280327
MacInnis RJ, English DR, Hopper JL, Giles GG (2006). Samanic C, Chow WH, Gridley G, Jarvholm B,
Body size and composition and the risk of gastric Fraumeni JF Jr (2006). Relation of body mass index
and oesophageal adenocarcinoma. Int J Cancer, to cancer risk in 362,552 Swedish men. Cancer Causes
118(10):2628–31. doi:10.1002/ijc.21638 PMID:16353151 Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9
Merry AH, Schouten LJ, Goldbohm RA, van den Brandt PA PMID:16841257
(2007). Body mass index, height and risk of adenocarci- Samanic C, Gridley G, Chow WH, Lubin J, Hoover
noma of the oesophagus and gastric cardia: a prospec- RN, Fraumeni JF Jr (2004). Obesity and cancer
tive cohort study. Gut, 56(11):1503–11. doi:10.1136/ risk among white and black United States veterans.
gut.2006.116665 PMID:17337464 Cancer Causes Control, 15(1):35–43. doi:10.1023/B:-
Muñoz N, Plummer M, Vivas J, Moreno V, De Sanjosé CACO.0000016573.79453.ba PMID:14970733
S, Lopez G, et al. (2001). A case-control study of Sjödahl K, Jia C, Vatten L, Nilsen T, Hveem K, Lagergren
gastric cancer in Venezuela. Int J Cancer, 93(3):417–23. J (2008). Body mass and physical activity and risk of
doi:10.1002/ijc.1333 PMID:11433408 gastric cancer in a population-based cohort study
O’Doherty MG, Freedman ND, Hollenbeck AR, Schatzkin in Norway. Cancer Epidemiol Biomarkers Prev,
A, Abnet CC (2012). A prospective cohort study of 17(1):135–40. doi:10.1158/1055-9965.EPI-07-0704
obesity and risk of oesophageal and gastric adeno- PMID:18187390
carcinoma in the NIH-AARP Diet and Health Study. Song M, Choi JY, Yang JJ, Sung H, Lee Y, Lee HW, et al.
Gut, 61(9):1261–8. doi:10.1136/gutjnl-2011-300551 (2015). Obesity at adolescence and gastric cancer risk.
PMID:22174193 Cancer Causes Control, 26(2):247–56. doi:10.1007/
Parr CL, Batty GD, Lam TH, Barzi F, Fang X, Ho SC, s10552-014-0506-z PMID:25471061
et al.; Asia-Pacific Cohort Studies Collaboration (2010). Steffen A, Huerta JM, Weiderpass E, Bueno-de-Mesquita
Body-mass index and cancer mortality in the Asia- HB, May AM, Siersema PD, et al. (2015). General
Pacific Cohort Studies Collaboration: pooled analyses and abdominal obesity and risk of esophageal and
of 424,519 participants. Lancet Oncol, 11(8):741–52. gastric adenocarcinoma in the European Prospective
doi:10.1016/S1470-2045(10)70141-8 PMID:20594911 Investigation into Cancer and Nutrition. Int J Cancer,
Persson C, Inoue M, Sasazuki S, Kurahashi N, Iwasaki M, 137(3):646–57. doi:10.1002/ijc.29432 PMID:25598323
Ye W, et al.; JPHC Study Group (2008). Female repro- Tanaka T, Nagata C, Oba S, Takatsuka N, Shimizu H
ductive factors and the risk of gastric cancer in a large- (2007). Prospective cohort study of body mass index
scale population-based cohort study in Japan (JPHC in adolescence and death from stomach cancer in
study). Eur J Cancer Prev, 17(4):345–53. doi:10.1097/ Japan. Cancer Sci, 98(11):1785–9. doi:10.1111/j.1349-
CEJ.0b013e3282f521e4 PMID:18562960 7006.2007.00583.x PMID:17725807
Praud D, Bertuccio P, Bosetti C, Turati F, Ferraroni M, Tran GD, Sun XD, Abnet CC, Fan JH, Dawsey SM, Dong
La Vecchia C (2014). Adherence to the Mediterranean ZW, et al. (2005). Prospective study of risk factors
diet and gastric cancer risk in Italy. Int J Cancer, for esophageal and gastric cancers in the Linxian
134(12):2935–41. doi:10.1002/ijc.28620 PMID:24259274
162
General Population Trial cohort in China. Int J Cancer,

113(3):456–63. doi:10.1002/ijc.20616 PMID:15455378
Vaughan TL, Davis S, Kristal A, Thomas DB (1995).
Obesity, alcohol, and tobacco as risk factors for cancers
of the esophagus and gastric cardia: adenocarcinoma
versus squamous cell carcinoma. Cancer Epidemiol
Biomarkers Prev, 4(2):85–92. PMID:7742727
Whitlock G, Lewington S, Sherliker P, Clarke R, Emberson
J, Halsey J, et al.; Prospective Studies Collaboration
(2009). Body-mass index and cause-specific mortality
in 900 000 adults: collaborative analyses of 57 prospec-
tive studies. Lancet, 373(9669):1083–96. doi:10.1016/
S0140-6736(09)60318-4 PMID:19299006
Wu AH, Wan P, Bernstein L (2001). A multiethnic popula-
tion-based study of smoking, alcohol and body size and
risk of adenocarcinomas of the stomach and esophagus
(United States). Cancer Causes Control, 12(8):721–32.
doi:10.1023/A:1011290704728 PMID:11562112
Yang P, Zhou Y, Chen B, Wan HW, Jia GQ, Bai HL, et al.
(2009). Overweight, obesity and gastric cancer risk:
results from a meta-analysis of cohort studies. Eur J
Cancer, 45(16):2867–73. doi:10.1016/j.ejca.2009.04.019
PMID:19427197
163
2.2.4 Cancer of the liver (hepatocellular in Sweden, a relative risk for BMI ≥ 30 kg/m2
carcinoma) versus BMI < 25 kg/m2 of 3.13 (95% CI, 2.04–4.79)
was reported (Samanic et al., 2006).
Hepatocellular carcinoma (HCC) is the most At least eight other studies have examined
frequent primary malignancy of the liver (> 80% the association between BMI and liver cancer
of primary liver cancers) and occurs predomi- (hepatocellular and intrahepatic bile duct
nantly in patients with underlying chronic liver combined, or NOS) incidence and/or mortality
disease or cirrhosis. Worldwide, liver cancer (Table 2.2.4a). One study of Japanese men and
is among the most common causes of cancer women showed no evidence of association with
death; the highest rates of liver cancer incidence increased incidence (Kuriyama et al., 2005) [the
and mortality occur in some areas in Asia and number of liver cancer cases in the highest cate-
sub-Saharan Africa, as a result of chronic hepa- gories of BMI was small or zero in both sexes,
titis infection (Jemal et al., 2014). and therefore power was limited to detect an
In 2001, the Working Group of the IARC association]. Conversely, in the large Korea
Handbook on weight control and physical activity National Health Insurance Corporation Study,
(IARC, 2002) concluded that the evidence of an risk of liver cancer increased significantly for
association between avoidance of weight gain BMI ≥ 30 kg/m2 in men (relative risk [RR], 1.63;
and liver cancer was inadequate. Since then, 95% CI, 1.27–2.10) and in women (RR, 1.39;
numerous individual cohort studies with at least 95% CI, 1.00–1.94), compared with the reference
100 cases (Table 2.2.4a), case–control studies category of BMI of 23.0–24.9 kg/m2. Significant
(Table 2.2.4b), and pooled and meta-analyses Ptrend values were found in both men and women
of cohort studies and case–control studies (Jee et al., 2008). Strong positive associations
(Table 2.2.4c) have been published examining were also observed in another prospective cohort
the association of anthropometric factors with study of BMI in relation to liver cancer incidence,
liver cancer incidence and/or mortality. Notably, a data linkage study in the United Kingdom
because chronic liver disease is among the most where a 5 kg/m2 increase in BMI was associated
common risk factors for cancer of the liver, with a 19% increase in risk (95% CI, 1.12–1.27)
results from cohort studies of anthropometric (Bhaskaran et al., 2014).
factors in relation to liver cancer incidence and/ In general, studies of BMI in relation to
or mortality in patients with liver disease have liver cancer mortality (Calle et al., 2003) or
also been included. liver cancer incidence and mortality combined
(a) Cohort studies (Borena et al., 2012) showed strong positive asso-
ciations. For example, in the Cancer Prevention
(i) Body weight and body mass index Study II in the USA, there was a strong positive
Six cohort studies of BMI or weight in relation association between liver cancer mortality in men
to risk of HCC specifically (Samanic et al., 2006; (RR, 4.52; 95% CI, 2.94–6.94 for BMI ≥ 35 kg/m2
Joshi et al., 2008; Ohishi et al., 2008; Borena et al., vs 18.5–24.9 kg/m2; Ptrend < 0.001), and to a lesser
2012; Loomba et al., 2013; Schlesinger et al., 2013) extent in women (RR, 1.68; 95% CI, 0.93–3.05 for
have been published (Table 2.2.4a). Of these BMI ≥ 35 kg/m2 vs 18.5–24.9 kg/m2; Ptrend < 0.04)
studies, four showed statistically significant posi- (Calle et al., 2003). In the Japan Collaborative
tive associations and/or trends (Samanic et al., Cohort Study, there was evidence of an asso-
2006; Ohishi et al., 2008; Borena et al., 2012; ciation between higher BMI and liver cancer
Schlesinger et al., 2013). In a large cohort of men
164
mortality when men with liver disease were such as alcohol consumption, cigarette smoking,
excluded (Li et al., 2013). or diabetes history. The largest meta-analysis, by
Associations of measures of body weight Chen et al. (2012), which included 26 prospective
and liver cancer have been examined in at least cohorts from Asia, Europe, and the USA, found
six cohort studies of patients with cirrhosis, a stronger risk of primary liver cancer in relation
hepatitis infections, or other liver conditions. to higher BMI in patients with liver cirrhosis or
Of these studies, four showed statistically HBV or HCV infection (n = 9 cohorts, summary
significant positive associations or trends RR, 1.73; 95% CI, 1.28–2.35) compared with the
between BMI and risk of HCC (N’Kontchou BMI-associated risk observed in the general
et al., 2006; Ioannou et al., 2007; Yu et al., population (n = 17 cohorts, summary RR, 1.36;
2008; Ohki et al., 2008). In the study with the 95% CI, 1.20–1.53) [the P value for difference
largest number of HCC cases, the relative risk for was 0.15]. In the recent meta-analysis of the
BMI ≥ 30 kg/m2 versus BMI < 18.5 kg/m2 was WCRF Continuous Update Project, a 5 kg/m2
3.10 (95% CI, 1.41–6.81) in Japanese men and increase in BMI was associated with a 43% (95%
women who were patients at a liver clinic (Ohki CI, 1.19–1.70) increase in liver cancer incidence
et al., 2008). Two studies of cirrhosis patients and a 13% (95% CI, 1.00–1.28) increase in liver
also showed statistically significant 2.5–2.8-fold cancer mortality based on 8 and 4 cohort studies,
higher risks of HCC for obese versus normal- respectively, and the association was stronger in
weight patients (N’Kontchou et al., 2006; Ioannou studies in Europe (summary RR, 1.59; 95% CI,
et al., 2007). The association was approximately 1.35–1.87) than in studies in Asia (summary RR,
of the same magnitude in a prospective study 1.18; 95% CI, 1.04–1.34) (WCRF/AICR, 2015).
in Taiwan, China, of carriers of hepatitis B (ii) Weight at different ages and weight
virus (HBV) (RR, 1.96; 95% CI, 0.72–5.38 for change
BMI ≥ 30 kg/m2 vs 18.5–24.9 kg/m2; Ptrend = 0.048)
[only 4 obese men developed HCC during Only a few cohort studies examined asso-
follow-up] (Yu et al., 2008). A Japanese study of ciations of BMI and/or weight at earlier ages
patients with hepatitis C virus (HCV) infection or change in BMI or weight with risk of liver
also found evidence of a borderline positive asso- cancer. In the EPIC study, BMI at age 20 years
ciation when BMI was modelled as a continuous was overall not associated with liver cancer
measure in women (RR per 1 kg/m2 increase in mortality (Schlesinger et al., 2013). However, in
BMI, 1.09; 95% CI, 0.99–1.19) but not in men that study there was a positive dose–response
(RR per 1 kg/m2 increase in BMI, 1.01; 95% CI, relationship between the average annual weight
0.93–1.09) (Arano et al., 2011). change from age 20 years to age at reporting and
There have been numerous meta-analyses increased risk; the relative risk for each kilo-
(Larsson & Wolk, 2007; Renehan et al., 2008; gram per year increase in weight of HCC was
Chen et al., 2012; Rui et al., 2012; Tanaka et al., 3.51 (95% CI, 1.93–6.41), after adjustment for
2012; Wang et al., 2012; WCRF/AICR, 2015; weight at age 20 years and other confounding
Table 2.2.4c) and a large pooled analysis of 57 factors. In a large Swedish occupational cohort,
cohorts (Whitlock et al., 2009) on BMI and 6-year BMI change during adulthood in rela-
(primary) liver cancer incidence or mortality. tion to liver cancer incidence was examined
Overall, these meta-analyses showed an increased (Samanic et al., 2006). Although the results were
risk of liver cancer in individuals with higher somewhat suggestive of an increasing risk with
BMI independently of sex, geographical region, increasing BMI gain, there were only 55 cases in
duration of follow-up, and potential confounders total [and therefore statistical power was limited
165
to detect associations]. Similarly, in the Japan Finally, using waist circumference as a

Collaborative Cohort Study, change in weight measure of adiposity, a large population-based
between age 20 years and baseline was not case–control study with 3649 cases all aged
associated with liver cancer mortality in men 68 years or older identified through the
or women, although some evidence for a trend Surveillance, Epidemiology, and End Results
could be observed in women (Li et al., 2013). (SEER) Program of the United States National
Cancer Institute, and with 195 953 population
(iii) Waist circumference
control subjects, showed a significantly increased
A positive association between waist circum- risk of HCC in men with waist circumference
ference and incidence of HCC was shown in the greater than 40 inches (101 cm) and in women
EPIC study, which was the only study to examine with waist circumference greater than 35 inches
this association (Schlesinger et al., 2013). In that (89 cm), compared with men or women with a
study, each increase of 5 cm in waist circumfer- smaller waist circumference (OR, 1.93; 95% CI,
ence was associated on average with a 25% (95% 1.71–2.18) (Welzel et al., 2011).
CI, 1.17–1.33) increase in risk in men and women Several of the above-mentioned studies
combined. considered HBV or HCV infection as a confounder
(b) Case–control studies or effect modifier for the association between
BMI and risk of liver cancer. In the study in
A total of five case–control studies have been Italy (Polesel et al., 2009), HCC risk was signifi-
published since 2001 on the association of BMI cantly increased in obese subjects without HBV
with HCC in Canada, China, France, Italy, and and HCV infection (OR, 3.5; 95% CI, 1.3–9.2;
the USA (Table 2.2.4b). compared with BMI < 30 kg/m2) but not in
In the USA, a study of 622 cases and 660 subjects with HBV or HCV infection. The study
population control subjects showed an increased by Hassan et al. (2015) in the USA showed a syner-
risk of HCC for men and women with early gistic interaction between obesity and hepatitis
adulthood (mid-20s to mid-40s) obesity (BMI virus infection, with highly increased risk in
> 30 kg/m2) compared with normal-weight indi- obese subjects with HBV or HCV infection.
viduals (men: OR, 2.3; 95% CI, 1.2–4.4; women:
OR, 3.6; 95% CI, 1.5–8.9) (Hassan et al., 2015),
but no association was found for recalled BMI in
the mid-50s. A hospital-based case–control study
in Italy also showed an increased risk for subjects
with elevated recalled BMI at about age 30 years,
but not for BMI 1 year before cancer diagnosis (or
equivalent time frame for the controls) (Polesel
et al., 2009), and a study in France also showed a
direct association of HCC risk with recalled past
obesity in patients with non-viral liver cirrhosis
(Archambeaud et al., 2015). In contrast, a popula-
tion-based case–control study in Canada showed
no association between risk of liver cancer and
self-reported BMI 1 year before diagnosis (Pan
et al., 2004).
166
Table 2.2.4a Cohort studies of measures of body fatness and cancer of the liver
Reference Total no. of Organ site Exposure categories Exposed Relative risk Covariates Comments
Cohort subjects or cancer cases (95% CI)
Location Sex type
Follow-up Incidence/ (ICD code)
period mortality
Hepatocellular carcinoma
N’Kontchou 771 HCC BMI 220 total Patients with alcohol-
et al. (2006) Men and women < 25 1.0 or hepatitis C-related
Cohort of Incidence 25–30 2.0 (1.4–2.7) cirrhosis
patients with ≥ 30 2.8 (2.0–4.0)
cirrhosis
France
1994–2004
Samanic et al. 362 552 HCC BMI Attained age, Based on fewer
(2006) Men 18.5–24.9 73 1.00 calendar year, than 30 incident
Swedish Incidence 25–29.9 90 1.45 (1.06–1.98) smoking cases, no significant
Construction ≥ 30 31 3.13 (2.04–4.79) associations for
Worker Cohort [Ptrend] [< 0.001] cholangiocarcinoma
Sweden or adenocarcinoma of
1958–1999 the liver were found.
No associations
between BMI change
and liver cancer
overall observed
(n = 469)
Ioannou et al. 2126 HCC BMI Age, HCV infection,
(2007) Men and women ICD-9: 18.5–24.9 15 1.00 HBsAg, HBV core

Cohort of Incidence 155.0 25–29 45 2.8 (1.4–5.4) antibody, type 2
cirrhosis ≥ 30 40 2.5 (1.3–4.9) diabetes mellitus,
patients in the platelet count
Veterans Affairs
facility
USA
1994–2005
167
168

Location Sex type
period mortality
Joshi et al. 548 530 HCC BMI 989 total Age, serum
(2008) Men ICD-10: < 18.5 1.00 glucose, alcohol
Korean male Mortality C22.0 18.5–24.9 1.38 (0.90–2.11) consumption,
civil service 25–29.9 0.98 (0.85–1.12) tobacco use, HBsAg
workers cohort ≥ 30 1.08 (0.67–1.72)
Republic of
Korea
1999–2004
Ohishi et al. 868 HCC BMI 10 yr before diagnosis Hepatitis
(2008) Men and women < 19.6 38 1.31 (0.51–3.34) infection, alcohol
Nested case– Incidence 19.6–21.2 33 1.24 (0.43–3.54) consumption,
control in the 21.3–22.9 36 1.00 smoking, coffee
Adult Health 23–25 49 2.51 (0.99–6.37) consumption,
Study (atomic > 25 54 4.57 (1.85–11.3) diabetes, radiation
bomb survivors) per 1 kg/m2 1.12 (1.03–1.22) dose to liver
Japan [Ptrend] [0.01]
1970–2002
Ohki et al. 1431 HCC BMI 340 total Age, sex,
(2008) Men and women < 18.5 1.00 heavy alcohol
Hospital-based Incidence 18.5–24.9 1.52 (0.93–2.47) consumption,
cohort of 25–29.9 1.86 (1.09–3.16) diabetes mellitus,
patients with ≥ 30 3.10 (1.41–6.81) serum albumin
chronic hepatitis concentration,
C total bilirubin
Tokyo, Japan concentration, ALT
1994–2006 levels, prothrombin
time activity,
platelet counts, AFP
concentration
Yu et al. (2008) 2903 HCC BMI Age, number
Cohort of male Men < 18.5 3 1.55 (0.49–4.93) of clinic visits,
government Incidence 18.5–24.9 77 1.00 smoking, alcohol
employees (HBV 25–29.9 50 1.48 (1.04–2.12) consumption,
carriers) ≥ 30 4 1.96 (0.72–5.38) diabetes
Taiwan, China [Ptrend] [0.048]
1989–2005
Location Sex type
period mortality
Loomba et al. 2260 HCC BMI 136 Age, serum HBV Alcohol
(2010) Men per 1 kg/m2 increase 1.00 (0.93–1.06) DNA level, smoking, consumption–BMI
HBV-positive Incidence serum ALT level, interaction
men part of the HBeAg status,
REVEAL-HBV cirrhosis at baseline
cohort visit
Taiwan, China
1991–2004
Arano et al. 146 HCC BMI 67 Age, alcohol
(2011) Men per 1 kg/m2 increase 1.01 (0.93–1.09) consumption, serum
Hospital-based Incidence biomarkers, platelet
cohort of count, diabetes
patients with 179 HCC BMI 55 Age, alcohol
hepatitis C Women per 1 kg/m2 increase 1.09 (0.99–1.19) consumption, serum
Japan Incidence biomarkers, platelet
1994–2009 count, diabetes
Borena et al. 578 700 HCC BMI, per unit SD 155 total 1.51 (1.29–1.77) Age, cohort, year of
(2012) Men and women birth, sex, smoking
Me-Can cohorts Incidence and status
Austria, Norway, mortality
and Sweden
1972–2006,

follow-up varied
by cohort
Loomba et al. 23 712 HCC BMI Only univariate A significant
(2013) Men and women < 30 284 1.00 model available interaction was
Population- Incidence ≥ 30 21 1.47 (0.85–2.30) reported between
based cohort of alcohol drinkers
residents (4 days per week for
(7 townships) at least 1 yr) and BMI
Taiwan, China ≥ 30 kg/m2, with a
1991–2004 7-fold increased risk
of HCC
169
170

Location Sex type
period mortality
Schlesinger et al. Men and women HCC BMI, tertiles (sex-specific) Age, sex, Associations were
(2013) Incidence ICD-10: Men: Women: study centre, lost when BMI
EPIC cohort C22.0 < 24.93 < 23.04 33 1.00 education level, analyses were further
10 European 24.93–27.8 23.04–26.64 49 1.31 (0.84–2.05) smoking, alcohol adjusted for WC
countries ≥ 27.81 ≥ 26.65 95 2.28 (1.50–3.45) consumption, height
1992–2010 [Ptrend] [< 0.0001] Analysis of weight
per 5 kg/m2 1.55 (1.31–1.83) change also adjusted
Weight, tertiles for weight at age No significant
20 yr associations were
T1 46 1.00
T2 50 1.19 (0.78–1.80) observed with weight
T3 81 2.04 (1.36–3.06) at age 20 yr, when
[Ptrend] [< 0.001] comparing extreme
per 5 kg 1.18 (1.11–1.25) tertiles (Ptrend = 0.95)
Weight change from age 20 yr to age at reporting, tertiles
T1 30 1.00
T2 32 1.30 (0.77–2.19)
T3 46 2.48 (1.49–4.13)
[Ptrend] [< 0.001]
per kg/yr 3.51 (1.93–6.41)
WC, tertiles
T1 27 1.00
T2 50 1.45 (0.90–2.34)
T3 100 2.60 (1.66–4.07)
[Ptrend] [< 0.0001]
per 5 cm 1.25 (1.17–1.33)
Liver NOS
Calle et al. 404 576 Liver BMI Age, education level,
(2003) Men 18.5–24.9 222 1.00 smoking, physical
Cancer Mortality 25–29.9 296 1.13 (0.94–1.34) activity, alcohol
Prevention 30–34.9 78 1.90 (1.46–2.47) consumption,
Study II (CPS II) ≥ 35 24 4.52 (2.94–6.94) marital status,
USA [Ptrend] [< 0.001] race, aspirin use,
1982–1998 consumption of fat
and vegetables
Location Sex type
period mortality
Calle et al. 495 477 Liver BMI For women, also
(2003) Women 18.5–24.9 200 1.00 adjusted for HRT
(cont.) Mortality 25–29.9 96 1.02 (0.80–1.31) use
30–34.9 37 1.40 (0.97–2.00)
≥ 35 12 1.68 (0.93–3.05)
[Ptrend] [0.04]
Samanic et al. 4 500 700 Liver Obesity Age, calendar year Obesity defined as
(2004) Men ICD-9: 155 White men: discharge diagnosis
United States Incidence Non-obese 3612 1.00 of obesity: ICD-8:
Veterans cohort Obese 322 1.44 (1.28–1.61) 277; ICD-9: 278.0
USA Black men: Significantly different
1969–1996 Non-obese 1168 1.00 risk in White men
Obese 38 0.68 (0.49–0.94) and Black men
(P < 0.001)
Kuriyama et al. 12 485 Liver BMI Age, smoking,
(2005) Men ICD-9: 18.5–24.9 55 1.00 alcohol
Population- Incidence 155.0–155.2 25.0–27.4 9 0.80 (0.40–1.63) consumption,
based cohort 27.5–29.9 5 1.14 (0.46–2.87) consumption of
in Miyagi ≥ 30 – – red meat, fruits
Prefecture [Ptrend] [0.92] and vegetables, and
Japan 15 054 Liver BMI bean paste, type of
1984–1992 Women ICD-9: 18.5–24.9 220 1.00 health insurance;

Incidence 155.0–155.2 25.0–27.4 7 1.30 (0.54–3.16) for women, also
27.5–29.9 4 0.91 (0.30–2.80) adjusted for
≥ 30 – – menopausal status,
[Ptrend] [0.94] parity, age at
menarche, age at
first pregnancy
Jee et al. (2008) 770 556 Liver BMI Age, smoking
Cohort from Men < 20.0 862 0.90 (0.81–1.00)
National Health Incidence 20.0–22.9 3260 0.97 (0.90–1.04)
Insurance 23.0–24.9 2463 1.00
Corporation 25.0–29.9 2062 1.04 (0.96–1.13)
Republic of ≥ 30.0 112 1.63 (1.27–2.10)
Korea [Ptrend] [0.0002]
171
1992–2006
172

Location Sex type
period mortality
Jee et al. (2008) 423 273 Liver BMI Age, smoking
(cont.) Women < 20.0 195 0.85 (0.67–1.06)
Incidence 20.0–22.9 505 0.76 (0.64–0.91)
23.0–24.9 411 1.00
25.0–29.9 587 1.14 (1.97–1.35)
≥ 30.0 63 1.39 (1.00–1.94)
[Ptrend] [< 0.0001]
Whitlock et al. 894 576 Liver BMI Study, sex, age,
(2009) Men and women ICD-9: 155 For BMI 15–25 201 1.37 (0.87–2.15) smoking
Pooled analysis Mortality For BMI 25–50 221 1.61 (1.26–2.05)
of 57 cohort For BMI 15–50 422 1.47 (1.26–1.71)
studies
Europe, Japan,
and USA
Follow-up varied
by cohort
Parr et al. (2010) 326 387 Liver BMI 774 Age, sex, smoking
Pooled analysis Men and women ICD-9: 155 12.0–18.4 1.13 (0.78–1.64) status
of 39 cohort Mortality ICD-10: 18.5–24.9 1.00 (0.89–1.13)
studies C22 25–29.99 1.06 (0.83–1.36)
Asia, Australia, ≥ 30 1.10 (0.63–1.91)
and New per 5 kg/m2 1.11 (0.94–1.31)
Zealand [Ptrend] [0.58]
1961–1999,
median follow-
up 4 yr
Borena et al. 578 700 Liver, BMI, quintiles (mean) Age, smoking
(2012) Incidence and primary Q1 (20.7) 36 1.00 status, cohort, year
Me-Can cohorts mortality cancer Q2 (23.0) 38 0.91 (0.55–1.51) of birth, sex
Austria, Norway, ICD-7: Q3 (24.7) 45 0.97 (0.59–1.57)
and Sweden 155.0 Q4 (26.8) 53 1.02 (0.63–1.64)
1972–2006, Q5 (31.3) 94 1.92 (1.23–2.96)
follow-up varied [Ptrend] [0.001]
by cohort
Location Sex type
period mortality
Li et al. (2013) 31 018 Liver BMI at baseline Age, smoking In stratified
JACC cohort Men ICD-10: < 18.5 32 1.42 (0.93–2.15) status, alcohol analyses, significant
Japan Mortality C22.0–22.9 18.5–20.9 82 1.09 (0.81–1.48) consumption, associations were
1988–2009 21–22.9 88 1.00 physical activity, observed in men
23–24.9 73 1.04 (0.76–1.42) intake of coffee and without liver disease
≥ 25 63 1.15 (0.83–1.60) fish, education level, (Ptrend = 0.03)
[Ptrend] [0.37] area of residence,
BMI at age 20 yr diabetes, gall
bladder disease,
< 18.5 14 0.74 (0.42–1.29)
blood transfusions,
18.5–20.9 91 0.89 (0.68–1.18)
history of liver
21–22.9 115 1.00
disease
23–24.9 75 0.92 (0.69–1.24)
≥ 25 43 0.91 (0.64–1.31)
[Ptrend] [0.54]
Weight change (kg), age 20 yr to baseline
≤−10.0 27 0.68 (0.43–1.08)
−9.9 to −5.0 76 1.08 (0.80–1.46)
−4.9 to 4.9 124 1.00
5.0 to 9.9 55 1.06 (0.77–1.47)
≥ 10.0 56 0.98 (0.70–1.37)
[Ptrend] [0.88]

173
174

Location Sex type
period mortality
Li et al. (2013) 41 455 Liver BMI at baseline Age, smoking
(cont.) Women ICD-10: < 18.5 8 0.74 (0.35–1.60) status, alcohol
Mortality C22.0–22.9 18.5–20.9 36 1.08 (0.69–1.68) consumption,
21–22.9 42 1.00 physical activity,
23–24.9 41 1.16 (0.75–1.79) intake of coffee and
≥ 25 62 1.42 (0.95–2.13) fish, education level,
[Ptrend] [0.10] area of residence,
BMI at age 20 yr diabetes, gall
bladder disease,
< 18.5 11 0.98 (0.58–1.64)
blood transfusions,
18.5–20.9 17 0.85 (0.58–1.25)
history of liver
21–22.9 28 1.00
disease
23–24.9 13 0.91 (0.60–1.38)
≥ 25 14 0.73 (0.45–1.18)
[Ptrend] [0.18]
Weight change (kg), age 20 yr to baseline
≤−10.0 10 0.68 (0.34–1.40)
−9.9 to −5.0 24 0.83 (0.51–1.35)
−4.9 to 4.9 62 1.00
5.0 to 9.9 46 1.31 (0.89–1.94)
≥ 10.0 47 1.41 (0.94–2.11)
[Ptrend] [0.08]
Bhaskaran et al. 5 243 978 Liver BMI 1859 total Age, sex, diabetes, Similar association in
(2014) Men and women ICD-10: per 5 kg/m2 1.19 (1.12–1.27) smoking, alcohol never-smokers only
Clinical Practice Incidence C22 [Ptrend] [< 0.0001] consumption,
Research socioeconomic
Datalink status, calendar year
United Kingdom
1987–2012
AFP, α-fetoprotein; ALT, alanine aminotransferase; BMI, body mass index (in kg/m 2); CI, confidence interval; EPIC, European Prospective Investigation into Cancer and Nutrition;
HBeAg, hepatitis B envelope antigen; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; ICD, International
Classification of Diseases; JACC, Japan Collaborative Cohort Study for Evaluation of Cancer Risk; Me-Can, Metabolic Syndrome and Cancer Project; NOS, not otherwise specified; SD,
standard deviation; WC, waist circumference; yr, year or years
Table 2.2.4b Case–control studies of measures of body fatness and hepatocellular carcinoma
Reference Total number of Exposure categories Exposed Relative risk Adjustment for confounding Comments
Study location cases cases (95% CI)
Period Source of controls
Pan et al. Men: 225 BMI 1 yr before diagnosis Age, geographical region, Self-reported BMI
(2004) Women: 84 Men: 225 total education level, smoking,
Canada Population < 25 1.00 physical activity, total
1994–1997 25– < 30 0.99 (0.72–1.38) calorie intake, total vegetable
≥ 30 1.30 (0.85–1.97) consumption, dietary fibre
[Ptrend] [0.31] intake, multivitamin intake;
Women: 85 total for women, also adjusted for
< 25 1.00 menopausal status, parity, age
25– < 30 0.61 (0.35–1.07) at menarche, age at end of first
≥ 30 0.94 (0.48–1.84) pregnancy
[Ptrend] [0.40]
Polesel et al. 185 BMI 1 yr before interview Centre, sex, age, education, BMI calculated from
(2009) Hospital < 25 71 1.00 drinking status, lifetime self-reported weights
Italy 25– < 30 76 1.0 (0.5–1.9) maximal alcohol consumption, and heights
1999–2003 ≥ 30 38 1.9 (0.9–3.9) smoking status, cigarettes
5 kg/m2 increase 1.1 (0.8–1.5) smoked per day, HBsAg and/or
BMI at age 30 yr anti-HCV positivity
< 25 109 1.00
≥ 25 69 1.0 (0.6–1.7)
5 kg/m2 increase 0.8 (0.6–1.2)
BMI increase from age 30 yr
< 1 73 1.00
1– < 4 53 1.2 (0.6–2.4)
≥ 4 52 1.6 (0.8–3.2)

Welzel et al. 3649 WC (≥ 40 inches [101 cm] 308 total 1.93 (1.71–2.18) Age, sex, race, geographical
(2011) Population in men, ≥ 35 inches location, Medicare/Medicaid
USA [89 cm] in women) dual enrolment
1993–2005
Archambeaud 200 Maximum BMI 125 total Sex, age, diabetes, smoking
et al. (2015) Hospital < 30 1.00 (past or present)
France ≥ 30 1.56 (1.02–2.37)
2007–2010
175
176

Reference Total number of Exposure categories Exposed Relative risk Adjustment for confounding Comments
Study location cases cases (95% CI)
Hassan et al. Men: 473 Normal weight (reference) Sex, age, ethnicity, education BMI calculated from
(2015) Women: 149 Overweight at different ages level, HCV, HBV, alcohol self-reported weights
USA Population Men: consumption, cigarette and heights at different
2004–2013 Mid-20s 124 1.5 (0.9–2.3) smoking, history of diabetes, ages; overweight and
Mid-30s 172 1.3 (0.9–2.1) physical activity, family obesity defined as BMI
Mid-40s 174 0.9 (0.6–1.4) history of cancer 24–29.9 and BMI ≥ 30,
Mid-50s 170 0.5 (0.3–0.9) respectively
Women:
Mid-20s 11 2.4 (0.9–3.0)
Mid-30s 19 1.2 (0.5–2.6)
Mid-40s 29 0.8 (0.4–1.6)
Mid-50s 35 0.9 (0.5–1.7)
Obesity at different ages
Men:
Mid-20s 33 1.8 (0.8–4.1)
Mid-30s 58 3.1 (1.6–6)
Mid-40s 101 2.2 (1.2–4)
Mid-50s 104 0.8 (0.4–1.4)
Women:
Mid-20s 13 5.2 (1.6–7.2)
Mid-30s 15 3.3 (1.3–8.6)
Mid-40s 26 2.1 (1.1–4.5)
Mid-50s 30 1.2 (0.5–2.5)
Obesity in early adulthood Men:
(mid-20s to mid-40s) 192 2.3 (1.2–4.4)
Women:
54 3.6 (1.5–8.9)
BMI, body mass index (in kg/m ); CI, confidence interval; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HCV, hepatitis C virus; WC, waist circumference
2
Table 2.2.4c Meta-analyses of measures of body fatness and cancer of the liver
Reference Total number of Organ site Exposure categories Relative risk Adjustment for Comments
studies (95% CI) confounding
Total number of
cases
Larsson & Wolk 11 cohort studies Liver BMI Age and other covariates The relative risk for
(2007) (7 on overweight Overweight vs normal 1.17 (1.02–1.34) depending on study obesity compared
with 5037 cases, and Obese vs normal 1.89 (1.51–2.36) (calendar year, sex, race, with normal BMI was
10 on obesity with diabetes, education, stronger in men than in
6042 cases) marital status, smoking, women
physical activity, diet,
family history of cancer,
occupational group,
aspirin use, estrogen
replacement therapy in
women)
Renehan et al. 4 cohort studies Liver BMI Men: NR Results are from
(2008) in men and 1 in per 5 kg/m2 increase 1.24 (0.95–1.62) random-effects
women Women: models. Substantial
2070 1.07 (0.55–2.08) heterogeneity was
observed (I2 = 83.1% in
men)
Chen et al. 26 prospective Liver BMI All: Age (all studies), and Significant
(2012) cohort studies (primary < 25 1.00 most of the studies heterogeneity in the
25 337 cancer) 25–29.9 1.48 (1.31–1.67) included alcohol overall analyses, and
≥ 30 1.83 (1.59–2.11) consumption, HBV in analyses in men;
BMI Men: and/or HCV infection, effects significantly
diabetes mellitus different in men vs

< 25 1.00
25–29.9 1.42 (1.22–1.65) women; associations
≥ 30 1.91 (1.51–2.41) independent of
BMI Women: geographical location,
< 25 1.00 alcohol consumption,
25–29.9 1.18 (1.08–1.30) diabetes, or HBV/HCV
≥ 30 1.55 (1.30–1.85) infections
Rui et al. (2012) 8 cohort studies in Liver BMI NR Associations remained
men and women 18.5–24.9 1.00 significant after
11 616 25–30 1.13 (1.05–1.21) excluding 3 studies on
≥ 30 2.09 (1.72–2.45) cirrhosis cohorts
177
178

Total number of
cases
Tanaka et al. 9 cohort studies Liver BMI Different adjustment Study restricted to
(2012) and 3 case–control per 1 kg/m2 increase 1.13 (1.07−1.20) factors depending on Japanese populations
studies study (hepatitis, alcohol Overweight/obese
NR consumption, diabetes, individuals showed
smoking) a 74% increased risk
compared with those
with normal weight
Wang et al. 21 prospective Liver BMI All: Age (all studies). Other Significant
(2012) cohort studies (11 (primary per 5 kg/m2 increase 1.39 (1.25–1.55) covariates, depending on heterogeneity among
in men and 5 in cancer) Men: study studies was observed.
women) 1.26 (1.11–1.44) Non-linear association
17 624 Women: was reported, with a
1.18 (1.08–1.29) steeper increase in risk
from BMI > 32 kg/m2
WCRF/AICR 12 studies Liver BMI Heterogeneity between
(2015) Men and women per 5 kg/m2 1.30 (1.16–1.46) studies; non-linear
Incidence and associations; similar
mortality risks in men and
14 311 women; associations
8 studies Liver BMI stronger for incidence
Men and women per 5 kg/m2 1.43 (1.19–1.70) than for mortality, and
Incidence for European vs Asian
11 530 studies
4 studies Liver BMI
Men and women per 5 kg/m2 1.13 (1.00–1.28)
Mortality
2543
8 studies Liver BMI
Men per 5 kg/m2 1.21 (1.02–1.44)
Incidence and
mortality
11 180
Total number of
cases
WCRF/AICR 4 studies Liver BMI
(2015) Women per 5 kg/m2 1.21 (1.10–1.33)
(cont.) Incidence and
mortality
2337
Meta-analysis of Liver BMI
European studies: per 5 kg/m2 1.59 (1.35–1.87)
4 studies
Men and women
Incidence and
mortality
588
Meta-analysis of Liver BMI
Asian studies: per 5 kg/m2 1.18 (1.04–1.34)
7 studies
Men and women
Incidence and
mortality
12 520
BMI, body mass index (in kg/m 2); CI, confidence interval; HBV, hepatitis B virus; HCV, hepatitis C virus; NR, not reported; WCRF/AICR, World Cancer Research Fund/American
Institute for Cancer Research

179
References Joshi S, Song YM, Kim TH, Cho SI (2008). Socio-economic

status and the risk of liver cancer mortality: a prospec-
tive study in Korean men. Public Health, 122(11):1144–
Arano T, Nakagawa H, Tateishi R, Ikeda H, Uchino K, 51. doi:10.1016/j.puhe.2008.04.003 PMID:18793784
Enooku K, et al. (2011). Serum level of adiponectin and Kuriyama S, Tsubono Y, Hozawa A, Shimazu T, Suzuki Y,
the risk of liver cancer development in chronic hepatitis Koizumi Y, et al. (2005). Obesity and risk of cancer in
C patients. Int J Cancer, 129(9):2226–35. doi:10.1002/ Japan. Int J Cancer, 113(1):148–57. doi:10.1002/ijc.20529
ijc.25861 PMID:21170963 PMID:15386435
Archambeaud I, Auble H, Nahon P, Planche L, Fallot G, Larsson SC, Wolk A (2007). Overweight, obesity and risk
Faroux R, et al. (2015). Risk factors for hepatocellular of liver cancer: a meta-analysis of cohort studies. Br J
carcinoma in Caucasian patients with non-viral Cancer, 97(7):1005–8. PMID:17700568
cirrhosis: the importance of prior obesity. Liver Int, Li Y, Yatsuya H, Yamagishi K, Wakai K, Tamakoshi A,
35(7):1872–6. doi:10.1111/liv.12767 PMID:25522809 Iso H (2013). Body mass index and weight change
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, during adulthood are associated with increased
Leon DA, Smeeth L (2014). Body-mass index and risk mortality from liver cancer: the JACC Study.
of 22 specific cancers: a population-based cohort study J Epidemiol, 23(3):219–26. doi:10.2188/jea.JE20120199
of 5.24 million UK adults. Lancet, 384(9945):755–65. PMID:23604063
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 Loomba R, Yang HI, Su J, Brenner D, Barrett-Connor E,
Borena W, Strohmaier S, Lukanova A, Bjørge T, Lindkvist Iloeje U, et al. (2013). Synergism between obesity and
B, Hallmans G, et al. (2012). Metabolic risk factors and alcohol in increasing the risk of hepatocellular carci-
primary liver cancer in a prospective study of 578,700 noma: a prospective cohort study. Am J Epidemiol,
adults. Int J Cancer, 131(1):193–200. doi:10.1002/ 177(4):333–42. doi:10.1093/aje/kws252 PMID:23355498
ijc.26338 PMID:21805476 Loomba R, Yang HI, Su J, Brenner D, Iloeje U, Chen CJ
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ (2010). Obesity and alcohol synergize to increase the
(2003). Overweight, obesity, and mortality from cancer risk of incident hepatocellular carcinoma in men.
in a prospectively studied cohort of U.S. adults. N Engl Clin Gastroenterol Hepatol, 8(10):891–8, 898.e1–2.
J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423 doi:10.1016/j.cgh.2010.06.027 PMID:20621202
PMID:12711737 N’Kontchou G, Paries J, Htar MT, Ganne-Carrie N,
Chen Y, Wang X, Wang J, Yan Z, Luo J (2012). Excess Costentin L, Grando-Lemaire V, et al. (2006). Risk
body weight and the risk of primary liver cancer: an factors for hepatocellular carcinoma in patients with
updated meta-analysis of prospective studies. Eur J alcoholic or viral C cirrhosis. Clin Gastroenterol
Cancer, 48(14):2137–45. doi:10.1016/j.ejca.2012.02.063 Hepatol, 4(8):1062–8. doi:10.1016/j.cgh.2006.05.013
PMID:22446023 PMID:16844421
Hassan MM, Abdel-Wahab R, Kaseb A, Shalaby A, Phan Ohishi W, Fujiwara S, Cologne JB, Suzuki G, Akahoshi
AT, El-Serag HB, et al. (2015). Obesity early in adulthood M, Nishi N, et al. (2008). Risk factors for hepato-
increases risk but does not affect outcomes of hepatocellular carcinoma in a Japanese population: a nested
cellular carcinoma. Gastroenterology, 149(1):119–29. case-control study. Cancer Epidemiol Biomarkers Prev,
doi:10.1053/j.gastro.2015.03.044 PMID:25836985 17(4):846–54. doi:10.1158/1055-9965.EPI-07-2806
IARC (2002). Weight control and physical activity. Lyon, PMID:18398026
France: IARC Press (IARC Handbooks of Cancer Ohki T, Tateishi R, Sato T, Masuzaki R, Imamura J,
Prevention, Vol. 6). Available from: http://publications. Goto T, et al. (2008). Obesity is an independent risk
iarc.fr/376. factor for hepatocellular carcinoma development
Ioannou GN, Splan MF, Weiss NS, McDonald GB, in chronic hepatitis C patients. Clin Gastroenterol
Beretta L, Lee SP (2007). Incidence and predictors of Hepatol, 6(4):459–64. doi:10.1016/j.cgh.2008.02.012
hepatocellular carcinoma in patients with cirrhosis. PMID:18387499
Clin Gastroenterol Hepatol, 5(8):938–45, 945.e1–4. Pan SY, Johnson KC, Ugnat AM, Wen SW, Mao Y;
doi:10.1016/j.cgh.2007.02.039 PMID:17509946 Canadian Cancer Registries Epidemiology Research
Jee SH, Yun JE, Park EJ, Cho ER, Park IS, Sull JW, et al. Group (2004). Association of obesity and cancer risk
(2008). Body mass index and cancer risk in Korean men in Canada. Am J Epidemiol, 159(3):259–68. doi:10.1093/
and women. Int J Cancer, 123(8):1892–6. doi:10.1002/ aje/kwh041 PMID:14742286
ijc.23719 PMID:18651571 Parr CL, Batty GD, Lam TH, Barzi F, Fang X, Ho SC, et al.;
Jemal A, Vineis P, Bray F, Torre L, Forman D, editors Asia-Pacific Cohort Studies Collaboration (2010).
(2014). The cancer atlas. 2nd ed. Atlanta (GA), USA: Body-mass index and cancer mortality in the Asia-
American Cancer Society. Available from: http://www. Pacific Cohort Studies Collaboration: pooled analyses
cancer.org/canceratlas. of 424,519 participants. Lancet Oncol, 11(8):741–52.
doi:10.1016/S1470-2045(10)70141-8 PMID:20594911
180
Polesel J, Zucchetto A, Montella M, Dal Maso L, Crispo A, Yu MW, Shih WL, Lin CL, Liu CJ, Jian JW, Tsai KS, et al.
La Vecchia C, et al. (2009). The impact of obesity and (2008). Body-mass index and progression of hepatitis
diabetes mellitus on the risk of hepatocellular carci- B: a population-based cohort study in men. J Clin
noma. Ann Oncol, 20(2):353–7. doi:10.1093/annonc/ Oncol, 26(34):5576–82. doi:10.1200/JCO.2008.16.1075
mdn565 PMID:18723550 PMID:18955457
Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen M
(2008). Body-mass index and incidence of cancer:
a systematic review and meta-analysis of prospec-
tive observational studies. Lancet, 371(9612):569–78.
doi:10.1016/S0140-6736(08)60269-X PMID:18280327
Rui R, Lou J, Zou L, Zhong R, Wang J, Xia D, et al. (2012).
Excess body mass index and risk of liver cancer:
a nonlinear dose-response meta-analysis of prospective
studies. PLoS One, 7(9):e44522. doi:10.1371/journal.
pone.0044522 PMID:23028553
Samanic C, Chow WH, Gridley G, Jarvholm B,
Fraumeni JF Jr (2006). Relation of body mass index
to cancer risk in 362,552 Swedish men. Cancer Causes
Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9
PMID:16841257
Samanic C, Gridley G, Chow WH, Lubin J, Hoover
RN, Fraumeni JF Jr (2004). Obesity and cancer
risk among white and black United States veterans.
Cancer Causes Control, 15(1):35–43. doi:10.1023/B:-
CACO.0000016573.79453.ba PMID:14970733
Schlesinger S, Aleksandrova K, Pischon T, Fedirko V,
Jenab M, Trepo E, et al. (2013). Abdominal obesity,
weight gain during adulthood and risk of liver and
biliary tract cancer in a European cohort. Int J Cancer,
132(3):645–57. doi:10.1002/ijc.27645 PMID:22618881
Tanaka K, Tsuji I, Tamakoshi A, Matsuo K, Ito H, Wakai
K, et al.; Research Group for the Development and
Evaluation of Cancer Prevention Strategies in Japan
(2012). Obesity and liver cancer risk: an evaluation
based on a systematic review of epidemiologic evidence
among the Japanese population. Jpn J Clin Oncol,
42(3):212–21. doi:10.1093/jjco/hyr198 PMID:22241822
Wang Y, Wang B, Shen F, Fan J, Cao H (2012). Body mass
index and risk of primary liver cancer: a meta-ana-
lysis of prospective studies. Oncologist, 17(11):1461–8.
doi:10.1634/theoncologist.2012-0066 PMID:22956536
Diet, nutrition, physical activity and liver cancer.
Available from: http://wcrf.org/sites/default/files/
Liver-Cancer-2015-Report.pdf.
Welzel TM, Graubard BI, Zeuzem S, El-Serag HB, Davila
JA, McGlynn KA (2011). Metabolic syndrome increases
the risk of primary liver cancer in the United States:
a population-based case-control study. Hepatology,
54(2):463–71. doi:10.1002/hep.24397 PMID:21538440
S0140-6736(09)60318-4 PMID:19299006
181
2.2.5 Cancer of the gall bladder 1.12–1.52; Ptrend < 0.0001) (Bhaskaran et al., 2014).
In a nationwide prospective study in the USA,
Cancer of the gall bladder cancer is uncom- there was evidence of a strong positive association
mon, and almost all gall bladder cancers are between being obese (BMI ≥ 30 kg/m2) and risk
adenocarcinomas. In 2001, the Working Group of gall bladder cancer mortality in both women
of the IARC Handbook on weight control and (RR, 2.13; 95% CI, 1.56–2.90; Ptrend < 0.001) and
physical activity (IARC, 2002) concluded that men (RR, 1.76; 95% CI, 1.06–2.95; Ptrend = 0.02)
the evidence of an association between avoid- (Calle et al., 2003).
ance of weight gain and gall bladder cancer In one study that assessed waist circumfer-
was inadequate. Since then, numberous indi- ence in relation to risk of gall bladder cancer,
vidual studies and meta-analyses of anthropo- each increase of 5 cm in waist circumference was
metric measures of body fatness and risk of gall associated with a 17% (95% CI, 1.06–1.30) higher
bladder cancer have been published. Results are risk in men and women combined (Schlesinger
presented here for cohort studies with at least 50 et al., 2013). [These results should be interpreted
cases (Table 2.2.5a) and for case–control studies with caution because only 76 cases of gall bladder
(Table 2.2.5b) and meta-analyses (Table 2.2.5c). cancer were identified during follow-up in 359 156
(a) Cohort studies men and women included in the analysis.]
The association between weight change and
There are at least 11 individual informative risk of gall bladder cancer was also examined in
prospective studies of the associations of BMI the EPIC cohort. Average annual weight change
or weight with gall bladder cancer incidence from age 20 years to the age at cohort enrolment
or mortality (Table 2.2.5a). No association was was not associated with risk of gall bladder cancer
observed in three of these studies (Samanic et al., (Schlesinger et al., 2013).
2006; Ishiguro et al., 2008; Hemminki et al.,
2011). Findings from the other eight prospective (b) Case–control studies
studies showed statistically significant positive Of the case–control studies that examined
association between BMI or weight and risk of the association between BMI and risk of gall
gall bladder cancer (Calle et al., 2003; Samanic bladder cancer (Table 2.2.5b), seven showed
et al., 2004; Engeland et al., 2005; Kuriyama no association (Strom et al., 1995; Serra et al.,
et al., 2005; Jee et al., 2008; Schlesinger et al., 2002; Máchová et al., 2007; Grainge et al., 2009;
2013; Bhaskaran et al., 2014; Borena et al., 2014), Nakadaira et al., 2009; Alvi et al., 2011; Cha,
and in several of those studies there was a 2015), whereas in three studies there was a statis-
dose–response relationship. In a large study tically significant positive association between
of nearly 1.2 million public servants in the adult BMI and risk of gall bladder cancer, which
Republic of Korea (Jee et al., 2008), the relative appeared to be dose-related (Zatonski et al., 1997;
risk of gall bladder cancer incidence for BMI Ahrens et al., 2007; Hsing et al., 2008).
≥ 30 kg/m2 versus 23.0–24.9 kg/m2 was 1.44 (95%
CI, 0.98–2.12) in women (Ptrend = 0.0007) and 1.65 (c) Pooled analyses and meta-analyses
(95% CI, 1.11–2.44) in men (Ptrend = 0.0003). A
There have been one pooled analysis
large cohort study in the United Kingdom that
(Whitlock et al., 2009; Table 2.2.5a) and several
included more than 5.2 million men and women
meta-analyses of cohort and case–control studies
also showed a statistically significant positive
(Larsson & Wolk, 2007; Renehan et al., 2008; Tan
association between BMI and risk of gall bladder
et al., 2015; WCRF/AICR, 2015; Table 2.2.5c) that
cancer (RR per 5 kg/m2 increase, 1.31; 95% CI,
182
examined the relationship between BMI and gall

bladder cancer incidence or mortality.
All meta-results were significantly positive.
In the largest and most recent meta-analysis
(Tan et al., 2015), which included 12 prospective
studies in Asia, Europe, and the USA, both over-
weight (RR, 1.15; 95% CI, 1.02–1.29) and obesity
(RR, 1.62; 95% CI, 1.45–1.81) were statistically
significantly positively associated with risk of gall
bladder cancer. Similarly, results from the 2015
WCRF Continuous Update Project on BMI and
risk of gall bladder cancer showed a statistically
significant 25% (95% CI, 1.15–1.37) higher risk
per 5 kg/m2 increase reported in a dose–response
analysis based on eight prospective studies
(WCRF/AICR, 2015). In the WCRF analysis,
associations were similar between cancer inci-
dence and mortality, between men and women,
and between studies in Asia and in Europe. In a
meta-analysis of eight case–control studies, both
overweight (RR, 1.16) and obesity (RR, 1.37) were
associated with statistically significant higher
risks of gall bladder cancer (Tan et al., 2015).
183
184

Table 2.2.5a Cohort studies of measures of body fatness and cancer of the gall bladder
Reference Total number Organ site Exposure Exposed cases Relative risk Covariates Comments
Cohort of subjects (ICD code) categories (95% CI)
Location Sex
mortality
Calle et al. (2003) 495 477 Gall BMI Age, education level,
Cancer Prevention Women bladder and 18.5–24.9 159 1.00 smoking, physical
Study II (CPS II) Mortality extrahepatic 25–29.9 86 1.12 (0.86–1.47) activity, alcohol
USA bile ducts 30–34 59 2.13 (1.56–2.90) consumption, marital
1982–1998 ICD-9: [Ptrend] [< 0.001] status, race, aspirin
404 576 156.0–156.9 BMI use, fat consumption,
Men 18.5–24.9 66 1.00 vegetable consumption;
Mortality 25–29.9 94 1.34 (0.97–1.84) for women, also adjusted
30–34 20 1.76 (1.06–2.94) for HRT use
[Ptrend] [0.02]
Samanic et al. 4 500 700 Gall Obesity Age, calendar year Obesity defined as
(2004) Men bladder and White men: discharge diagnosis
United States Incidence extrahepatic Non-obese 265 1.00 of obesity: ICD-8:
Veterans cohort bile ducts Obese 26 1.70 (1.13–2.57) 277; ICD-9: 278.0
USA ICD-9: 156 Black men:
1969–1996 Non-obese 45 1.00
Obese 2 0.93 (0.23–3.86)
Engeland et al. 1 037 892 Gall bladder BMI 1087 total Age, birth cohort, height
(2005) Women ICD-7: 155.1 < 18.5 1.02 (0.54–1.91)
Norwegian men Incidence 18.5–24.9 1.00
and women 25.0–29.9 1.27 (1.10–1.47)
Norway ≥ 30 1.88 (1.60–2.21)
1963–2002 [Ptrend] [< 0.001]
963 619 BMI 628 total
Men < 18.5 0.31 (0.04–2.24)
Incidence 18.5–24.9 1.00
25.0–29.9 1.00 (0.84–1.17)
≥ 30 1.38 (1.01–1.89)
[Ptrend] [0.2]
Location Sex
mortality
Kuriyama et al. 15 054 Gall BMI Age, smoking,
(2005) Women bladder and 18.5–24.9 12 1.00 alcohol consumption,
Japanese men and Incidence extrahepatic 25.0–27.4 3 0.83 (0.23–2.98) consumption of red meat,
women bile ducts 27.5–29.9 5 3.43 (1.19–9.94) fruits and vegetables,
Japan ICD-9: ≥ 30 4 4.45 (1.39–14.23) and bean paste, type of
1984–1992 156.0–156.9 [Ptrend] [0.004] health insurance; for
12 485 BMI women, also adjusted
Men 18.5–24.9 8 1.00 for menopausal status,
Incidence 25.0–27.4 1 0.46 (0.05–3.93) parity, age at menarche,
27.5–29.9 – – age at first pregnancy
≥ 30 – –
[Ptrend] [0.48]
Samanic et al. 362 552 Gall bladder BMI Attained age, calendar
(2006) Men ICD-7: 155.1 18.5–24.9 53 1.00 year, smoking
Swedish Incidence 25–29.9 45 0.93 (0.62–1.39)
Construction ≥ 30 11 1.40 (0.73–2.70)
Worker Cohort [Ptrend] [> 0.5]
Sweden
1958–1999
Ishiguro et al. 53 187 Gall bladder BMI Age, study area,
(2008) Women ICD-O-3: < 23 35 1.00 cholelithiasis, diabetes,
Japan Public Incidence C23.9, C24.0 23.0–24.9 9 0.47 (0.22–0.98) smoking, alcohol

Health Center 25.0–26.9 8 0.62 (0.29–1.34) consumption
Japan ≥ 27.0 11 0.94 (0.48–1.88)
1990–2004 [Ptrend] [0.50]
48 681 BMI
Men < 23 14 1.00
Incidence 23.0–24.9 6 0.74 (0.28–1.92)
25.0–26.9 6 1.26 (0.48–3.33)
≥ 27.0 4 1.39 (0.45–4.34)
[Ptrend] [0.52]
185
186

Location Sex
mortality
Ishiguro et al. 101 868 BMI
(2008) Men and < 23 49 1.00
(cont.) women 23.0–24.9 15 0.55 (0.31–0.98)
Incidence 25.0–26.9 14 0.80 (0.44–1.46)
≥ 27.0 15 1.06 (0.59–1.90)
[Ptrend] [0.82]
Jee et al. (2008) 443 273 Gall bladder BMI Age, smoking Excluded first 2 yr of
Cohort from Women (NOS) < 20.0 121 0.97 (0.78–1.21) follow-up
the National Incidence 20.0–22.9 302 1.12 (0.90–1.41) Update of study by
Health Insurance 23.0–24.9 262 1.00 Oh et al. (2005)
Corporation 25.0–29.9 341 1.27 (1.02–2.12)
Republic of Korea ≥ 30.0 36 1.44 (0.98–2.12)
1992–2006 [Ptrend] [0.0007]
770 556 BMI
Men < 20.0 246 0.80 (0.68–0.94)
Incidence 20.0–22.9 787 0.86 (0.77–0.96)
23.0–24.9 670 1.00
25.0–29.9 542 0.97 (0.86–1.10)
≥ 30.0 31 1.65 (1.11–2.44)
[Ptrend] [0.0003]
Whitlock et al. 894 576 Gall BMI, per 5 kg/m2 120 1.29 (0.90–1.85) Age, sex, smoking status,
(2009) Men and bladder and study
Pooled analysis of women extrahepatic
57 cohort studies Mortality bile ducts
Europe, Japan, ICD-9: 156
and USA
Follow-up varied
by cohort
Hemminki et al. 30 020 Gall bladder Obesity 19 SIR Age, sex, time period, Overlap with study
(2011) Men and ICD-7: 155.1 Observed vs 1.55 (0.93–2.43) region, SES by Wolk et al. (2001)
Swedish hospital women expected rates is unclear
patients Incidence Obesity defined as
Sweden hospital discharge
1964–2006 diagnosis
Location Sex
mortality
Schlesinger et al. 359 156 Gall bladder BMI, per 5 kg/m2 76 total 1.28 (0.99–1.65) Age, sex, study centre,
(2013) (191 856 for ICD-10: 0.70 (0.43–1.15)* education level, smoking,
EPIC cohort weight change) C23.9 alcohol consumption,
10 European Men and Baseline weight, 76 total 1.11 (1.00–1.22) height
countries women per 5 kg *additional adjustment
1992–2010 Incidence for WC
Weight change 37 total 1.76 (0.59–5.29) **additional adjustment
(kg/year) for BMI
WC, per 5 cm 76 total 1.17 (1.06–1.30)
1.33 (1.10–1.62)**
Bhaskaran et al. 5 243 978 Gall bladder BMI 303 total HR (99% CI) Age, sex, diabetes,
(2014) Men and ICD-10: C23 per 5 kg/m2 1.31 (1.12–1.52) smoking, alcohol
Clinical Practice women [Ptrend] [< 0.0001] consumption, SES,
Research Datalink Incidence calendar year
United Kingdom
1987–2012
Borena et al. 578 700 Gall bladder BMI, quintiles (mean) Smoking status, baseline
(2014) Men and ICD-7: 155.1 Q1 (20.7) 20 1.00 age, cohort, sex, year of
Metabolic women Q2 (23.0) 26 1.12 (0.58–2.19) birth
Syndrome and Incidence Q3 (24.7) 38 1.49 (0.80–2.76)
Cancer Project Q4 (26.8) 47 1.70 (0.93–3.09)
(Me-Can) cohort Q5 (31.3) 53 1.94 (1.08–3.51)

Austria, Norway, [Ptrend] [0.08]
and Sweden BMI
1972–2006 < 25 77 1.00
≥ 25 107 1.52 (1.12–2.10)
BMI, body mass index (in kg/m 2); CI, confidence interval; EPIC, European Prospective Investigation into Cancer and Nutrition; HR, hazard ratio; HRT, hormone replacement therapy;
SES, socioeconomic status; SIR, standardized incidence ratio; WC, waist circumference
187
188

Table 2.2.5b Case–control studies of measures of body fatness and cancer of the gall bladder
Reference Total number of Organ site Exposure categories Exposed cases Relative risk Adjustment for
Study location cases (95% CI) confounding
Period Total number of
controls
Source of controls
Strom et al. (1995) 84 BMI, most of adult life Age, sex, country
Bolivia and Mexico 126 < 24 33 1.0
1984–1988 Men and women 24–25.9 17 1.5 (0.5–4.6)
Hospital 26–28 12 2.2 (0.7–8.4)
> 28 3 1.6 (0.4–6.1)
BMI, maximum ever
< 24 12 1.0
24–25.9 15 1.6 (0.4–7.6)
26–28 22 1.3 (0.3–5.6)
> 28 19 2.6 (0.5–18.6)
Zatonski et al. (1997) Men: BMI, quartiles Men:
Australia, Canada, 44 Q1 9 1.0
The Netherlands, and 815 Q2 11 1.0 (0.3–3.0)
Poland Population Q3 13 0.7 (0.3–2.0)
1983–1988 Q4 11 1.0 (0.3–2.8)
[Ptrend] [0.74]
Women: BMI, quatiles Women:
152 Q1 30 1.0
700 Q2 37 1.7 (0.9–3.1)
Population Q3 22 1.5 (0.3–3.0)
Q4 56 2.1 (1.2–3.8)
[Ptrend] [0.02]
Serra et al. (2002) 114 BMI Age, sex
Chile 114 < 25 53 1.0
1992–1995 Hospital 25–29.9 42 0.8 (0.4–1.4)
≥ 30 19 0.9 (0.4–1.8)
Ahrens et al. (2007) 104 Gall bladder ICD-O: BMI 62 total Age, country, history of
Europe 1401 C23.9 ≤ 25 1.0 gallstones
1995–1997 (men only) 25– < 27 1.8 (0.4–7.2)
Population 27– < 30 11.0 (2.9–41.9)
≥ 30 13.3 (1.4–123)
controls
Source of controls
Máchová et al. (2007) 93 BMI Age, smoking, height,
Czech Republic 37 772 Men: 14 total hypertension
1987–2002 Population 18.5–24.9 1.00
25–30 1.01 (0.24–4.32)
≥ 30 0.76 (0.08–7.41)
Women: 79 total
18.5–24.9 1.00
25–30 1.07 (0.58–1.95)
≥ 30 0.73 (0.36–1.50)
Hsing et al. (2008) 627 Gall bladder, Usual adult BMI Age (continuous), sex
China 959 excluding < 18.5 17 0.62 (0.35–1.09) (male, female), and
1997–2001 Population extrahepatic bile 18.5–22.9 30 1.0 education level (none/
ducts and ampulla of 23.0–24.9 73 1.20 (0.85–1.68) primary, junior middle,
Vater ≥ 25 145 1.56 (1.17–2.10) senior, some college)
[Ptrend] [< 0.001]
Maximum adult BMI
< 18.5 6 1.24 (0.47–3.29)
18.5–22.9 74 1.00
23.0–24.9 83 1.35 (0.94–1.95)
≥ 25 185 1.48 (1.08–2.03)
[Ptrend] [0.02]
BMI change in adulthood

≤ 0.74 74 1.00
0.75–2.77 62 0.93 (0.62–1.39)
2.78–5.21 86 1.45 (0.98–2.14)
> 5.21 93 1.47 (1.00–2.16)
[Ptrend] [0.01]
Grainge et al. (2009) 184 Gall bladder, BMI Smoking, alcohol
United Kingdom 3007 excluding < 25 36 1.00 consumption, NSAID use
1987–2002 Population cholangiocarcinomas 25–29.9 31 1.03 (0.62–1.72)
and unspecified ≥ 30 19 1.51 (0.83–2.75)
biliary tract cancers
189
190

controls
Source of controls
Nakadaira et al. 41 BMI Age
(2009) 30 ≤ 24.9 13 1.0
Hungary Hospital 25–29.9 9 1.5 (0.4–5.0)
2003–2006 ≥ 30 19 0.8 (0.3–1.8)
Alvi et al. (2011) 60 BMI Sex, hypertension,
Pakistan 120 < 23 14 1.00 diabetes, smoking
1988–2007 (70% of cases were > 23 46 1.98 (0.62–6.28)
women)
Hospital
Shebl et al. (2011) 627 Gall bladder, WC (cm) Age, sex, BMI
China 959 excluding Low 83 1.00
1997–2001 Population extrahepatic bile High (men: ≥ 90; women: 111 0.98 (0.65–1.47)
ducts and ampulla of ≥ 80)
Vater
Cha (2015) 78 BMI Age, sex, hypertension,
Republic of Korea 78 < 23 18 1.00 diabetes mellitus, vascular
2008–2013 Population ≥ 23 23 0.74 (0.28–1.97) occlusive disease, alcohol
consumption, smoking,
polypoid lesions of gall
bladder, gallstone disease
BMI, body mass index (in kg/m 2); CI, confidence interval; NSAID, non-steroidal anti-inflammatory drug; WC, waist circumference
Table 2.2.5c Meta-analyses of measures of body fatness and cancer of the gall bladder
Reference Number and Population Exposure categories Exposed Relative risk Covariates, comments
Period type of studies Incidence/mortality cases (95% CI)
Larsson & Wolk 8 cohort studies Men and women Obese vs normal (definition 1.69 (1.48–1.92) See also Table 2.2.5b
(2007) Incidence and mortality varies by study)
1966–2007
8 cohort studies, Men and women Obese vs normal (definition - 1.66 (1.47–1.88)
4 case–control Incidence and mortality varies by study)
studies
Renehan et al. 2 cohort studies Women BMI, per 5 kg/m2 1111 total 1.59 (1.02–2.47) Also split up by
(2008) Incidence geographical region
1966–2007 4 cohort studies Men BMI, per 5 kg/m2 928 total 1.09 (0.99–1.21)
Incidence
Tan et al. (2015) 12 cohort studies Men and women Overweight 5101 total 1.15 (1.02–1.29) Normal BMI used as
Cohort studies: Incidence and mortality Obese 1.62 (1.45–1.81) reference
1964–2006
Case–control 8 case–control Overweight 1.16 (0.96–1.41)
studies: studies Obese 1.37 (1.10–1.71)
1984–2007 12 cohort BMI
studies, 8 case– Overall:
control studies 25–30 1.14 (1.04–1.25)
> 30 1.56 (1.41–1.73)
Men:
25–30 1.06 (0.94–1.20)
> 30 1.42 (1.21–1.66)
Women:
25–30 1.26 (1.13–1.40)
> 30 1.67 (1.38–2.02)

WCRF/AICR 8 cohort studies Men and women BMI, per 5 kg/m2 6004 total 1.25 (1.15–1.37)
(2015) Incidence and mortality
NR
BMI, body mass index (in kg/m 2); CI, confidence interval; ICD, International Classification of Diseases; NR, not reported; WCRF/AICR, World Cancer Research Fund/American
Institute for Cancer Research
191
References IARC (2002). Weight control and physical activity. Lyon,

France: IARC Press (IARC Handbooks of Cancer
Prevention, Vol. 6). Available from: http://publications.
Ahrens W, Timmer A, Vyberg M, Fletcher T, Guénel P, iarc.fr/376.
Merler E, et al. (2007). Risk factors for extrahepatic Ishiguro S, Inoue M, Kurahashi N, Iwasaki M, Sasazuki
biliary tract carcinoma in men: medical conditions S, Tsugane S (2008). Risk factors of biliary tract cancer
and lifestyle: results from a European multicentre in a large-scale population-based cohort study in Japan
case-control study. Eur J Gastroenterol Hepatol, (JPHC study); with special focus on cholelithiasis,
19(8):623–30. doi:10.1097/01.meg.0000243876.79325. body mass index, and their effect modification. Cancer
a1 PMID:17625430 Causes Control, 19(1):33–41. doi:10.1007/s10552-007-
Alvi AR, Siddiqui NA, Zafar H (2011). Risk factors of 9067-8 PMID:17906958
gallbladder cancer in Karachi – a case-control study. Jee SH, Yun JE, Park EJ, Cho ER, Park IS, Sull JW, et al.
World J Surg Oncol, 9(1):164. doi:10.1186/1477-7819-9- (2008). Body mass index and cancer risk in Korean men
164 PMID:22151791 and women. Int J Cancer, 123(8):1892–6. doi:10.1002/
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, ijc.23719 PMID:18651571
Leon DA, Smeeth L (2014). Body-mass index and risk Kuriyama S, Tsubono Y, Hozawa A, Shimazu T, Suzuki Y,
of 22 specific cancers: a population-based cohort study Koizumi Y, et al. (2005). Obesity and risk of cancer in
of 5.24 million UK adults. Lancet, 384(9945):755–65. Japan. Int J Cancer, 113(1):148–57. doi:10.1002/ijc.20529
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 PMID:15386435
Borena W, Edlinger M, Bjørge T, Häggström C, Lindkvist Larsson SC, Wolk A (2007). Obesity and the risk of
B, Nagel G, et al. (2014). A prospective study on gallbladder cancer: a meta-analysis. Br J Cancer,
metabolic risk factors and gallbladder cancer in the 96(9):1457–61. PMID:17375043
Metabolic Syndrome and Cancer (Me-Can) collabora- Máchová L, Cízek L, Horáková D, Koutná J, Lorenc J,
tive study. PLoS One, 9(2):e89368. doi:10.1371/journal. Janoutová G, et al. (2007). Association between obesity
pone.0089368 PMID:24586723 and cancer incidence in the population of the District
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ Sumperk, Czech Republic. Onkologie, 30(11):538–42.
(2003). Overweight, obesity, and mortality from cancer doi:10.1159/000108284 PMID:17992023
in a prospectively studied cohort of U.S. adults. N Engl Nakadaira H, Lang I, Szentirmay Z, Hitre E, Kaster M,
J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423 Yamamoto M (2009). A case-control study of gall-
PMID:12711737 bladder cancer in Hungary. Asian Pac J Cancer Prev,
Cha BH (2015). Epidemiological characteristics of gall- 10(5):833–6. PMID:20104974
bladder cancer in Jeju Island: a single-center, clin- Oh SW, Yoon YS, Shin SA (2005). Effects of excess
ically based, age-sex-matched, case-control study. weight on cancer incidences depending on cancer
Asian Pac J Cancer Prev, 16(18):8451–4. doi:10.7314/ sites and histologic findings among men: Korea
APJCP.2015.16.18.8451 PMID:26745100 National Health Insurance Corporation Study. J Clin
Engeland A, Tretli S, Austad G, Bjørge T (2005). Height Oncol, 23(21):4742–54. doi:10.1200/JCO.2005.11.726
and body mass index in relation to colorectal and PMID:16034050
gallbladder cancer in two million Norwegian men Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen M
and women. Cancer Causes Control, 16(8):987–96. (2008). Body-mass index and incidence of cancer:
doi:10.1007/s10552-005-3638-3 PMID:16132807 a systematic review and meta-analysis of prospec-
Grainge MJ, West J, Solaymani-Dodaran M, Aithal GP, tive observational studies. Lancet, 371(9612):569–78.
Card TR (2009). The antecedents of biliary cancer: a doi:10.1016/S0140-6736(08)60269-X PMID:18280327
primary care case-control study in the United Kingdom. Samanic C, Chow WH, Gridley G, Jarvholm B,
Br J Cancer, 100(1):178–80. doi:10.1038/sj.bjc.6604765 Fraumeni JF Jr (2006). Relation of body mass index
PMID:19018260 to cancer risk in 362,552 Swedish men. Cancer Causes
Hemminki K, Li X, Sundquist J, Sundquist K (2011). Obesity Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9
and familial obesity and risk of cancer. Eur J Cancer PMID:16841257
Prev, 20(5):438–43. doi:10.1097/CEJ.0b013e32834761c0 Samanic C, Gridley G, Chow WH, Lubin J, Hoover
PMID:21606843 RN, Fraumeni JF Jr (2004). Obesity and cancer
Hsing AW, Sakoda LC, Rashid A, Chen J, Shen MC, Han risk among white and black United States veterans.
TQ, et al. (2008). Body size and the risk of biliary tract Cancer Causes Control, 15(1):35–43. doi:10.1023/B:-
cancer: a population-based study in China. Br J Cancer, CACO.0000016573.79453.ba PMID:14970733
99(5):811–5. doi:10.1038/sj.bjc.6604616 PMID:18728671 Schlesinger S, Aleksandrova K, Pischon T, Fedirko V,
192

132(3):645–57. doi:10.1002/ijc.27645 PMID:22618881
Serra I, Yamamoto M, Calvo A, Cavada G, Báez S, Endoh
K, et al. (2002). Association of chili pepper consump-
tion, low socioeconomic status and longstanding
gallstones with gallbladder cancer in a Chilean popul-
ation. Int J Cancer, 102(4):407–11. doi:10.1002/ijc.10716
PMID:12402311
Shebl FM, Andreotti G, Meyer TE, Gao YT, Rashid A,
Yu K, et al. (2011). Metabolic syndrome and insulin
resistance in relation to biliary tract cancer and stone
risks: a population-based study in Shanghai, China.
Br J Cancer, 105(9):1424–9. doi:10.1038/bjc.2011.363
PMID:21915122
Strom BL, Soloway RD, Rios-Dalenz JL, Rodriguez-
Martinez HA, West SL, Kinman JL, et al. (1995). Risk
factors for gallbladder cancer. An international collab-
orative case-control study. Cancer, 76(10):1747–56.
doi:10.1002/1097-0142(19951115)76:10<1747::AID-CN-
CR2820761011>3.0.CO;2-L PMID:8625043
Tan W, Gao M, Liu N, Zhang G, Xu T, Cui W (2015).
Body mass index and risk of gallbladder cancer:
systematic review and meta-analysis of observa-
tional studies. Nutrients, 7(10):8321–34. doi:10.3390/
nu7105387 PMID:26426043
Diet, nutrition, physical activity and gallbladder
cancer. Available from: https://wcrf.org/sites/default/
files/Gallbladder-Cancer-2015-Report.pdf.
S0140-6736(09)60318-4 PMID:19299006
Wolk A, Gridley G, Svensson M, Nyrén O, McLaughlin
JK, Fraumeni JF, et al. (2001). A prospective study
of obesity and cancer risk (Sweden). Cancer Causes
Control, 12(1):13–21. doi:10.1023/A:1008995217664
PMID:11227921
Zatonski WA, Lowenfels AB, Boyle P, Maisonneuve P,
Bueno de Mesquita HB, Ghadirian P, et al. (1997).
Epidemiologic aspects of gallbladder cancer: a
case-control study of the SEARCH Program of the
International Agency for Research on Cancer. J Natl
Cancer Inst, 89(15):1132–8. doi:10.1093/jnci/89.15.1132
PMID:9262251
193
IARC HANDBOOKS FOR CANCER PREVENTION – 16
2.2.6 Cancers of the biliary tract examined in two prospective studies. In the Japan
Public Health Center Study, the relative risk esti-
Intrahepatic bile duct cancers occur in mates for the highest versus lowest categories of
the smaller bile duct branches within the liver BMI in men and in women were greater than 1
and comprise about 10% of bile duct cancers. but were not statistically significant (Ishiguro
Extrahepatic bile duct cancers occur outside of et al., 2008). In the EPIC study, in which cancers
the liver. Perihilar (also called hilar) extrahepatic of the extrahepatic bile ducts included cancers of
bile duct cancers occur where the left and right the gall bladder, associations of BMI and weight
hepatic ducts join and are the most common in men and women combined were not statisti-
type of bile duct cancer, accounting for about cally significant. [The median BMI in the highest
two thirds of all bile duct cancers. Nearly all bile tertile was 29.9 kg/m2 for men and 29.6 kg/m2
duct cancers are cholangiocarcinomas, of which for women, which includes people with a BMI
most are adenocarcinomas. This section reviews in the overweight and obese category.] Similarly,
studies of all subtypes of cancer of the biliary no association was found with average annual
tract. weight change from age 20 years, or with waist
(a) Cohort studies circumference (Schlesinger et al., 2013).
In a meta-analysis of gall bladder or biliary
See Table 2.2.6a (web only; available at: http:// tract cancer incidence or mortality that included
publications.iarc.fr/570). seven studies, BMI was statistically significantly
Only one prospective study (i.e. the EPIC positively associated with risk in men and women
cohort) assessed the association between body combined (RR for highest vs lowest category of
weight and intrahepatic bile duct cancer specif- BMI, 1.40; 95% CI, 1.15–1.65) (Park et al., 2014).
ically; relative risk estimates for all measure-
ments (BMI, weight, waist or hip circumference, (b) Case–control studies
waist-to-hip ratio, or weight change) as contin- See Table 2.2.6b (web only; available at: http://
uous measures were greater than 1, but none publications.iarc.fr/570).
of the associations were statistically significant The associations of BMI with cancers of the
(Schlesinger et al., 2013). biliary tract system (including gall bladder or
The association between BMI and extra- not) were examined in six population- or hospi-
hepatic bile duct cancer specifically (excluding tal-based case–control studies.
the gall bladder) was examined in the Japan For extrahepatic bile duct cancer, two
Public Health Center Study. In that study, BMI population-based case–control studies, one in
was positively associated with risk in men and Europe (Ahrens et al., 2007) and one in China
women combined, with a relative risk of 1.78 (Hsing et al., 2008), showed a statistically signif-
for BMI ≥ 27 kg/m2 compared with < 23 kg/m2 icant higher risk for BMI > 25 kg/m2 versus
(Ptrend = 0.03) (Ishiguro et al., 2008). 18.5–25 kg/m2, whereas a lower risk with high
The association between BMI and intra- or BMI was observed in one study in China (Kato
extrahepatic bile duct cancer was examined in the et al., 1989). No association was observed in a
Korea National Health Insurance Corporation study of cholangiocarcinoma (Grainge et al.,
Study, which included only men and found a 2009). No association was observed with waist
statistically significant positive dose–response circumference in the only study that examined
relationship (Ptrend = 0.005) (Oh et al., 2005). such association (Shebl et al., 2011).
The association between BMI and cancer of
the bile ducts and gall bladder combined was
194
For overall biliary tract cancer, a 2.5-fold Shebl FM, Andreotti G, Meyer TE, Gao YT, Rashid A,
increase in risk was observed with BMI ≥ 30 kg/m2 Yu K, et al. (2011). Metabolic syndrome and insulin
resistance in relation to biliary tract cancer and stone
at age 35 years, but not with BMI 1–5 years before risks: a population-based study in Shanghai, China.
study entry (Ahrens et al., 2007). Br J Cancer, 105(9):1424–9. doi:10.1038/bjc.2011.363
PMID:21915122
References
Ahrens W, Timmer A, Vyberg M, Fletcher T, Guénel P,
Merler E, et al. (2007). Risk factors for extrahepatic
biliary tract carcinoma in men: medical conditions
and lifestyle: results from a European multicentre
case-control study. Eur J Gastroenterol Hepatol,
19(8):623–30. doi:10.1097/01.meg.0000243876.79325.
a1 PMID:17625430
Grainge MJ, West J, Solaymani-Dodaran M, Aithal GP,
Card TR (2009). The antecedents of biliary cancer: a
primary care case-control study in the United Kingdom.
Br J Cancer, 100(1):178–80. doi:10.1038/sj.bjc.6604765
PMID:19018260
Hsing AW, Sakoda LC, Rashid A, Chen J, Shen MC, Han
TQ, et al. (2008). Body size and the risk of biliary tract
cancer: a population-based study in China. Br J Cancer,
99(5):811–5. doi:10.1038/sj.bjc.6604616 PMID:18728671
Ishiguro S, Inoue M, Kurahashi N, Iwasaki M, Sasazuki
S, Tsugane S (2008). Risk factors of biliary tract cancer
in a large-scale population-based cohort study in Japan
(JPHC study); with special focus on cholelithiasis,
body mass index, and their effect modification. Cancer
Causes Control, 19(1):33–41. doi:10.1007/s10552-007-
9067-8 PMID:17906958
Kato K, Akai S, Tominaga S, Kato I (1989). A case-con-
trol study of biliary tract cancer in Niigata
Prefecture, Japan. Jpn J Cancer Res, 80(10):932–8.
doi:10.1111/j.1349-7006.1989.tb01629.x PMID:2515177
Oh SW, Yoon YS, Shin SA (2005). Effects of excess
weight on cancer incidences depending on cancer
sites and histologic findings among men: Korea
National Health Insurance Corporation Study. J Clin
Oncol, 23(21):4742–54. doi:10.1200/JCO.2005.11.726
PMID:16034050
Park M, Song Y, Je Y, Lee JE (2014). Body mass index and
biliary tract disease: a systematic review and meta-ana-
lysis of prospective studies. Prev Med, 65:13–22.
doi:10.1016/j.ypmed.2014.03.027 PMID:24721739
Schlesinger S, Aleksandrova K, Pischon T, Fedirko V,
132(3):645–57. doi:10.1002/ijc.27645 PMID:22618881
195
2.2.7 Cancer of the pancreas meta-analyses or pooled analyses (Table 2.2.7c)

are summarized in this section.
Cancer of the pancreas is the seventh leading
cause of cancer death worldwide (Ferlay et al., (a) Cohort studies
2015). Even in developed countries, few individ-
Since 2000, more than 30 individual cohort
uals diagnosed with pancreatic cancer survive
studies including pooled analyses have reported
more than 5 years (Sirri et al., 2016). Pancreatic
on the associations of excess body fatness
cancer incidence and mortality rates have been
with pancreatic cancer incidence or mortality
increasing both in the USA (Kohler et al., 2015)
(Table 2.2.7a). In addition, seven meta-analyses
and in western Europe (Bosetti et al., 2013),
of cohort studies have been published since then
despite declines in cigarette smoking, an estab-
(Table 2.2.7c).
lished risk factor for pancreatic cancer. It has
BMI, usually ascertained at study enrolment
been suggested that these increases may be at
at or after middle age, was by far the most common
least partly attributable to increases in the preva-
measure of excess body weight examined in these
lence of obesity (Ma & Jemal, 2013). Notably, type
cohort studies. In a comprehensive meta-analysis
2 diabetes mellitus, which is caused by obesity,
by the WCRF Continuous Update Project that
is also an established risk factor for pancreatic
included 23 cohort studies of pancreatic cancer
cancer, and the incidence of diabetes is also
incidence and more than 9500 incident cases
increasing.
of pancreatic cancer, the summary relative risk
The great majority (> 85%) of pancreatic
for a continuous 5 kg/m2 increase in BMI was
tumours are ductal adenocarcinomas and derive
1.10 (95% CI, 1.07–1.14), with similar results in
from the exocrine component of the pancreas.
men and in women (WCRF/AICR, 2012). Other
Other pancreatic tumours are a more heter-
meta-analyses or pooled cohort studies, all with
ogeneous collection of different tumour
considerable overlap in study populations, have
types and include, among others, acinar cell
reported similar results per 5 kg/m2 increase in
carcinoma of the pancreas (about 5% of
BMI (Larsson et al., 2007; Renehan et al., 2008;
exocrine pancreatic cancers), cystadenocar-
Genkinger et al., 2011, 2015).
cinomas, adenosquamous carcinomas, pancre-
The largest study that presented categorical
atic mucinous cystic neoplasms, and pancreatic
BMI results was an analysis that included nearly
neuroendocrine (islet cell) tumours (1–2% of all
6000 pancreatic cancer deaths in White men
pancreatic cancers).
and women in the Cancer Prevention Study II
In 2001, the Working Group of the IARC
(Arnold et al., 2009). In that analysis, obesity (i.e.
Handbook on weight control and physical
BMI ≥ 30 kg/m2) was associated on average with
activity (IARC, 2002) concluded that the
a 40% higher risk of pancreatic cancer mortality
evidence of an association between avoid-
compared with normal BMI (18.5– < 25 kg/m2),
ance of weight gain and pancreatic cancer was
and results were similar in men and in women
inadequate. Because of the high case fatality
separately. [No associations were found in Black
of pancreatic cancer, results from studies of
men and women, but the sample size was very
pancreatic cancer incidence and mortality
small compared with the group of White men
can be considered comparable. Results from
and women.]
individual cohort studies with more than 100
Relatively few large studies of BMI and
cases of pancreatic cancer (Table 2.2.7a), from
pancreatic cancer have been conducted in popu-
case–control studies (Table 2.2.7b), and from
lations that were not predominantly of European
descent. Relative risks from the largest study in
196
African Americans, a pooled analysis of seven risk (RR per 5 kg/m2 increase, 1.05, 95% CI,
cohorts (Bethea et al., 2014), and from a study in 1.01–1.10).
the Republic of Korea with 1860 cases (Jee et al., Several other individual cohort studies
2008), the largest in an Asian population, appear examined associations of change in weight
consistent with those observed in meta-analyses (Samanic et al., 2006, Lin et al., 2007, Luo et al.,
of populations of Caucasians. However, BMI was 2008, Johansen et al., 2009) or change in BMI
not associated with risk of pancreatic cancer in a (Verhage et al., 2007) with risk of pancreatic
pooled analysis of the Asia Cohort Consortium cancer. None of these studies reported statisti-
(Lin et al., 2013b). cally significant associations, except for a study
Some evidence suggests that the association in Sweden that found higher risk in a small group
between BMI and pancreatic cancer may differ by of men with a weight increase of 15% or more in
smoking status. In the large NIH-AARP cohort, 6 years (Samanic et al., 2006) and another study
there was a statistically significant interaction that reported significant positive associations
between BMI and smoking status, with a positive in a small group of men with a BMI increase of
association between BMI and risk of pancreatic 8 kg/m2 or more since age 20 years (Verhage et
cancer in never-smokers and in former smokers al., 2007).
but not in current smokers (Stolzenberg-Solomon Several individual cohort studies have exam-
et al., 2013). Similarly, increased BMI was asso- ined associations of waist circumference with
ciated with higher risk of pancreatic cancer in risk of pancreatic cancer (Larsson et al., 2005;
never-smokers and in former smokers in other Berrington de González et al., 2006; Luo et al.,
studies, although these interactions were not 2008; Stolzenberg-Solomon et al., 2008). In the
statistically significant (Genkinger et al., 2011; WCRF meta-analysis, the relative risk per 10 cm
Aune et al., 2012). increase in waist circumference was 1.11 (95%
A limited number of individual cohort CI, 1.05–1.18) (WCRF/AICR, 2012). In addition,
studies have examined the association between waist circumference was examined in a large
BMI in early adulthood, usually defined as age pooled analysis of pancreatic cancer mortality
18–21 years, and pancreatic cancer incidence including data from 11 cohort studies (Genkinger
or mortality (Patel et al., 2005; Lin, et al., 2007; et al., 2015); a higher waist circumference was
Verhage et al., 2007; Stolzenberg-Solomon et al., associated with increased risk of pancreatic
2013), with mixed results. [These studies calcu- cancer mortality (RR per 10 cm increase, 1.07;
lated BMI in early adulthood based on weight in 95% CI, 1.00–1.14), and no differences in risk
young adulthood recalled by participants who were observed between men and women.
were middle-aged or older.]
The largest analysis of BMI in early adult- (b) Case–control studies
hood, as well as BMI change after early adult- A total of 15 independent case–control
hood in relation to pancreatic cancer mortality, studies, conducted in Canada, China, Europe,
is a pooled analysis including more than 3000 Japan, North Africa (Egypt), and the USA,
pancreatic cancer deaths from 14 cohorts reported on the association of BMI with cancer
(Genkinger et al., 2015). In that pooled analysis, of the pancreas (Table 2.2.7b). In all studies, the
an increase of 5 kg/m2 in BMI in early adulthood assessment of BMI was based on self-reported
was associated with a relative risk of 1.18 (95% CI, height and usual body weight or body weight
1.11–1.25), and BMI change after early adulthood during a relatively recent time frame before
was also significantly associated with increased cancer diagnosis. In a few studies, additional
self-reports were also obtained for body weight
197
up to 20 years before cancer diagnosis, or body A second study, also in the USA (Li et al., 2009),
weight at various pre-specified ages in the more reported a positive association of BMI with risk
distant past. In all but two studies (Pezzilli et al., of pancreatic cancer both in ever-smokers (OR
2005; Lo et al., 2007), the estimated association of per 5 kg/m2 increase, 1.75; 95% CI, 1.37–2.22) and
BMI with risk of pancreatic cancer was adjusted in never-smokers (OR, 1.46; 95% CI, 1.16–1.84).
for smoking, as well as for various other potential One case–control study in the USA (with 309
confounding factors. cases and 602 controls) specifically addressed the
For usual BMI before disease onset, 7 of the 14 association of BMI with pancreatic neuroendo-
studies reported statistically significant increases crine tumours, a rare pancreatic cancer tumour,
in risk, either overall or in sex-stratified analyses and observed an increased risk in individuals
(Silverman et al., 1998; Hanley et al., 2001; Eberle who were obese (BMI ≥ 30 kg/m2) compared
et al., 2005; Anderson et al., 2009; Li et al., 2009; with those with a lower BMI (OR, 1.65; 95% CI,
Halfdanarson et al., 2014; Zheng et al., 2016). 1.11–2.45) (Halfdanarson et al., 2014).
Of the remaining studies, the majority showed
odds ratios above 1.0. In studies presenting
sex-stratified analyses, positive associations with
BMI appeared to be somewhat stronger and
more often significant for men than for women
(Hanley et al., 2001; Silverman, 2001; Eberle et
al., 2005; Fryzek et al. 2005; Li et al., 2009).
The study by Fryzek et al. (2005) in the USA
showed inverse associations of current BMI
(at diagnosis) and cancer of the pancreas and
no association with BMI 5 years before inter-
view. However, analyses based on recalled BMI
20 years before interview showed a statistically
significant direct association with risk of pancre-
atic cancer, although in men only. In a similar
type of analysis, a case–control study in the
Czech Republic and Slovakia (Urayama et al.,
2011) also showed a statistically significant asso-
ciation of pancreatic cancer with recalled BMI
at age 20 years and at age 40 years, but not with
BMI 2 years before interview (OR, 0.98; 95% CI,
0.85–1.13).
In two studies, associations of BMI with risk
of pancreatic cancer were estimated separately
for never-smokers and ever-smokers.
In the USA, Fryzek et al. (2005) reported
a statistically significant and up to 3.3-fold
increase in risk of pancreatic cancer (95% CI,
1.2–9.2) only in never-smokers in the highest
category of BMI compared with those with low
BMI, and no relationship was found in smokers.
198
Table 2.2.7a Cohort studies of measures of body fatness and cancer of the pancreas
Reference Total Organ site or Exposure categories Exposed Relative risk Covariates Comments
Cohort number of cancer type cases (95% CI)
Follow-up Sex
period Incidence/
mortality
Friedman & van 450 cases, Pancreas BMI, per 1 kg/m2 increase 450 1.02 (1.00–1.04) Age, cigarette
den Eeden (1993) 2687 smoking, race
Nested case– controls Weight, per 5 kg 1.06 (1.01–1.11)
control study Men and
within Kaiser women
Permanente Incidence
USA
1964–1988
Gapstur et al. 20 475 Pancreas BMI Age
(2000) Men ICD-8: 157 < 24.129 10 1.00
Chicago Heart Mortality 24.129–26.292 21 1.76 (0.83–3.74)
Association 26.293–28.630 23 1.68 (0.80–3.53)
Detection Project ≥ 28.631 42 3.04 (1.52–6.08)
in Industry 15 183 BMI Age
Cohort Women < 20.978 9 1.00
USA Mortality 20.978–23.240 6 0.48 (0.17–1.26)
1967–1995 23.241–26.156 16 1.09 (0.47–2.51)
≥ 26.157 12 0.73 (0.30–1.80)
Isaksson et al. 21 884 Pancreas BMI Age, sex, smoking No associations were
(2002) Men and < 18.5 5 2.30 (0.93–5.71) observed for adult
Swedish Twin women 18.5–24.99 84 1.00 weight gain (in kg)

Registry Incidence 25–30 70 1.36 (0.99–1.88)
Sweden > 30 4 0.56 (0.20–1.52)
1969–1997
Samanic et al. 4 500 700 Pancreas Obesity Age, calendar year Obesity defined as
(2004) Men ICD-9: 157 White men: discharge diagnosis
United States Incidence Non-obese 5483 1.00 of obesity: ICD-8:
USA Black men:
1969–1996 Non-obese 1638 1.00
Obese 83 1.07 (0.86–1.34)
199
200

Follow-up Sex
period Incidence/
mortality
Batty et al. (2005) 18 403 Pancreas BMI Age, employment
Whitehall Study Men ICD-8/9: 157 18.5–24.9 75 1.00 grade, physical
United Kingdom Mortality ICD-10: C25 25.0–29.9 69 1.18 (0.83–1.68) activity, smoking,
1967–2002 ≥ 30 3 0.58 (0.18–1.91) marital status,
[Ptrend] [0.80] prevalent disease,
weight loss in
past year, BP
medication, height,
skinfold thickness,
systolic BP, plasma
cholesterol, glucose
intolerance, diabetes
Larsson et al. 83 053 Pancreas BMI Age, education level, In stratified analyses,
(2005) Men and ICD-9: 157, < 20 5 0.96 (0.38–2.46) physical activity, stronger associations
Swedish women excluding 157.4 20–24.9 50 1.00 smoking, alcohol with BMI in men
Mammography Incidence 25–29.9 54 1.25 (0.84–1.86) consumption, sex than in women
Cohort (SMC) ≥ 30 19 1.81 (1.04–3.15)
1987–2004 WC (cm), quartiles (sex-specific)
Cohort of Men: Women:
Swedish Men < 90 < 76 16 1.00
(COSM) 90–94 76–81 20 1.15 (0.59–2.25)
Sweden 95–101 82–89 34 1.59 (0.87–2.93)
1997–2004 ≥ 102 ≥ 90 36 1.72 (0.93–3.20)
[Ptrend] [0.05]
Follow-up Sex
period Incidence/
mortality
Patel et al. (2005) 145 627 Pancreas BMI at baseline Age, smoking, In stratified analyses,
Cancer Men and ICD-9: < 25 50 1.00 years since quitting association with
Prevention women 157.0–157.9 25–29.9 33 1.03 (0.76–1.38) smoking, education BMI at baseline was
Study II (CPS II) Incidence ICD-10: ≥ 30 22 2.08 (1.48–2.93) level, family history stronger in men than
Nutrition Cohort and C25.0–25.9 [Ptrend] [0.0001] of pancreatic cancer, in women
1992–1999 mortality BMI at age 18 yr gall bladder disease,
< 21 59 1.00 diabetes, height,
21–22.9 25 1.07 (0.77–1.49) energy intake,
≥ 23 17 1.33 (0.95–1.85) physical activity
[Ptrend] [0.11]
Adult weight change (kg)
< −2.27 4 1.74 (0.94–3.22)
−2.27 to 4.54 20 1.00
4.55–9.07 18 1.12 (0.70–1.79)
9.08–13.61 21 0.97 (0.60–1.58)
≥ 13.62 38 0.96 (0.61–1.52)
[Ptrend] [0.16]
Sinner et al. 28 002 Pancreas BMI Age, smoking status,
(2005) Women ICD-10: C25 < 25 84 1.00 multivitamin use
Iowa Women’s Incidence 25–29.9 72 0.94 (0.69–1.29)
Health Study ≥ 30 53 1.14 (0.81–1.62)

USA
1986–2001
Berrington de 438 405 Pancreas BMI Sex, smoking,
González et al. Men and < 20 9 0.67 (0.33–1.37) diabetes
(2006) women 20–22.9 48 1.00 Weight and WC
EPIC cohort Incidence 23–24.9 85 0.99 (0.69–1.41) estimates also
10 European 25–26.9 71 0.82 (0.56–1.19) adjusted for height
countries 27–29.9 43 0.76 (0.50–1.16)
1991–2004 30–34.9 50 1.16 (0.77–1.76)
≥ 35 13 1.19 (0.64–2.23)
per 5 kg/m2 1.09 (0.95–1.24)
[Ptrend] [0.24]
201
202

Follow-up Sex
period Incidence/
mortality
Berrington de Weight (kg), quartiles (sex-specific)
González et al. Men: Women:
(2006) < 73 < 58 66 1.00
(cont.) 73–79 58–63 65 0.90 (0.63–1.28)
80–87 64–71 85 1.02(0.73–1.44)
≥ 88 ≥ 72 103 1.14 (0.82–1.61)
per 5 kg 1.05 (0.99–1.10)
[Ptrend] [0.06]
WC (cm), quartiles (sex-specific)
Men: Women:
< 88 < 73 51 1.00
88–93 73–78 59 0.96 (0.65–1.41)
94–100 79–87 79 1.05(0.72–1.53)
≥ 101 ≥ 88 91 1.33 (0.93–1.92)
per 10 cm 1.24 (1.04–1.48)
[Ptrend] [0.03]
Samanic et al. 362 552 Pancreas BMI Attained age,
(2006) (107 815 ICD-7: 157 18.5–24.9 352 1.00 calendar year,
Swedish in weight 25–29.9 289 0.95 (0.82–1.12) smoking
Construction change ≥ 30 57 1.16 (0.87–1.53)
Worker Cohort analysis) [Ptrend] [> 0.5]
Sweden Men 6-yr weight change
1958–1999 Incidence −4% to +4.9% 86 1.00
5–9.9% 41 1.45 (1.00–2.11)
10–14.9% 13 1.53 (0.85–2.77)
≥ 15% 7 2.67 (1.22–5.84)
[Ptrend] [> 0.5]
Follow-up Sex
period Incidence/
mortality
Lin et al. (2007) 43 579 Pancreas BMI at baseline Age, smoking,
JACC cohort Men ICD-10: C25 < 20 46 1.12 (0.76–1.63) diabetes, gall bladder
Japan Mortality 20–22.4 71 1.00 disease
1988–2003 22.5–24.9 57 0.94 (0.66–1.34)
25–27.4 26 1.02 (0.65–1.62)
27.5–29.9 6 0.62 (0.23–1.70)
≥ 30 1 0.58 (0.08–4.16)
[Ptrend] [0.47]
BMI at age 20 yr
< 20 27 1.39 (0.86–2.24)
20–22.4 45 1.00
22.5–24.9 45 1.13 (0.75–1.71)
25–27.4 21 1.54 (0.92–2.58)
27.5–29.9 6 1.65 (0.70–3.86)
≥ 30 4 3.51 (1.26–9.78)
[Ptrend] [0.01]
Weight change (kg)
< −5 45 1.63 (1.05–2.53)
−5 to < 0 22 1.39 (0.82–2.33)
0 47 1.00
> 0–4.9 12 1.11 (0.58–2.12)

≥ 5 21 0.85 (0.49–1.47)
59 107 Pancreas BMI at baseline Age, smoking,
Women ICD-10: C25 < 20 33 1.15 (0.74–1.80) diabetes, gall bladder
Mortality 20–22.4 50 1.00 disease
22.5–24.9 62 1.33 (0.91–1.95)
25–27.4 30 1.21 (0.77–1.92)
27.5–29.9 16 1.57 (0.86–2.86)
≥ 30 4 1.04 (0.37–2.89)
[Ptrend] [0.28]
203
204

Follow-up Sex
period Incidence/
mortality
Lin et al. (2007) BMI at age 20 yr
(cont.) < 20 25 0.81 (0.50–1.31)
20–22.4 51 1.00
22.5–24.9 48 1.08 (0.73–1.61)
25–27.4 15 0.69 (0.39–1.23)
27.5–29.9 3 0.46 (0.14–1.48)
≥ 30 1 0.43 (0.06–3.15)
[Ptrend] [0.09]
Weight change (kg)
< −5 14 0.41 (0.22–0.74)
−5 to < 0 14 0.61 (0.34–1.11)
0 59 1.00
> 0–4.9 13 0.70 (0.38–1.28)
≥ 5 43 0.93 (0.60–1.45)
Luo et al. (2007) 47 499 Pancreas BMI Smoking, diabetes,
Japan Public Men ICD-10: C25 14–20.9 37 1.4 (0.8–2.5) physical activity,
Health Center Incidence 21–24.9 69 1.0 study area, age,
Prospective 25–40 22 0.7 (0.4–1.1) alcohol use, history
Study [Ptrend] [0.01] of cholelithiasis
Japan 52 161 BMI
1990–2003 Women 14–20.9 14 0.7 (0.4–1.3)
Incidence 21–24.9 49 1.0
25–40 33 1.1 (0.7–1.6)
[Ptrend] [0.3]
Máchová et al. 17 110 Pancreas BMI 114 total Age, smoking, Nested case–control
(2007) Men ICD-10: C25 18.5–24.9 1.00 hypertension, height study, reporting odds
National Cancer Incidence 25–29.9 1.24 (0.74–2.07) ratios
Registry, nested ≥ 30 1.81 (0.98–3.31)
case–control 20 856 BMI 80 total
study Women 18.5–24.9 1.00
Czech Republic Incidence 25–29.9 0.68 (0.37–1.26)
1987–2002 ≥ 30 0.95 (0.50–1.79)
Follow-up Sex
period Incidence/
mortality
Nöthlings et al. 77 255 Pancreas BMI Ethnicity, smoking,
(2007) Men ICD-10: < 25 110 1.00 family history of
Multiethnic Incidence C25.0–25.3, 25–29.9 89 0.99 (0.74–1.33) pancreatic cancer,
Cohort Study C25.7–25.9 ≥ 30 38 1.51 (1.02–2.26) diabetes, age, energy
USA [Ptrend] [0.085] intake, intake of
1993–2002 90 175 BMI red meat, intake
Women < 25 52 1 of processed meat,
Incidence 25–29.9 62 0.80 (0.59–1.09) physical activity
≥ 30 62 0.65 (0.43–0.99)
[Ptrend] 61 [0.031]
Verhage et al. 2366 Pancreas BMI at baseline Age, smoking, When restricting
(2007) Men ICD-10: C25 < 23 44 1.10 (0.72–1.69) diabetes, to microscopically
Netherlands Incidence 23–24.9 67 1.00 hypertension confirmed exocrine
Cohort Study on 25–26.9 50 0.93 (0.61–1.39) pancreatic cancer,
Diet and Cancer 27–29.9 39 1.17 (0.75–1.81) significant positive
The Netherlands ≥ 30 20 2.69 (1.47–4.92) associations were
1986–1999 [Ptrend] [0.141] found with increased
per 1 kg/m2 1.05 (0.99–1.12) BMI and weight at
BMI at age 20 yr baseline, and with
< 20 35 1.00 BMI change since
20–20.9 26 0.80 (0.46–1.40) age 20 yr

21–22.9 60 0.99 (0.62–1.59)
≥ 23 52 1.07 (0.67–1.73)
[Ptrend] [0.56]
per 1 kg/m2 1.03 (0.96–1.10)
BMI change since age 20 yr
< 0 14 0.99 (0.53–1.85)
0–3.9 84 1.00
4–7.9 60 1.34 (0.90–1.99)
≥ 8 15 2.21 (1.09–4.49)
[Ptrend] [0.052]
per 1 kg/m2 1.07 (0.99–1.15)
205
206

Follow-up Sex
period Incidence/
mortality
Verhage et al. Weight at baseline (kg)
(2007) < 65 74 1.00
(cont.) 65–69 47 1.16 (0.76–1.76)
70–74 46 1.13 (0.75–1.70)
75–79 21 0.92 (0.53–1.59)
≥ 80 36 1.55 (0.99–2.45)
[Ptrend] [0.18]
continuous per kg 1.01 (0.99–1.03)
2438 Pancreas BMI at baseline Age, smoking, When restricting
Women ICD-10: C25 < 23 46 1.02 (0.66–1.58) diabetes, to microscopically
Incidence 23–24.9 45 1.00 hypertension confirmed exocrine
25–26.9 55 1.69 (1.11–2.58) pancreatic cancer,
27–29.9 38 1.41 (0.89–2.25) significant positive
≥ 30 19 1.31 (0.74–2.31) associations were
[Ptrend] [0.052] found with increased
per 1 kg/m2 1.04 (1.00–1.08) weight at baseline
BMI at age 20 yr and with BMI change
< 20 65 1.00 since age 20 yr. A
20–20.9 27 0.93 (0.58–1.51) significant Ptrend was
21–22.9 42 0.69 (0.46–1.04) also observed with
≥ 23 52 0.97 (0.66–1.44) increased BMI at
[Ptrend] [0.535] baseline
per 1 kg/m2 1.02 (0.95–1.09)
BMI change since age 20 yr
< 0 15 0.67 (0.37–1.21)
0–3.9 76 1.00
4–7.9 63 1.08 (0.75–1.55)
≥ 8 31 1.72 (1.11–2.67)
[Ptrend] 185 [0.004]
per 1 kg/m2 1.05 (1.01–1.10)
Follow-up Sex
period Incidence/
mortality
Verhage et al. Weight at baseline (kg)
(2007) < 65 59 1.00
(cont.) 65–69 42 1.23 (0.81–1.88)
70–74 39 1.30 (0.84–1.99)
75–79 31 1.58 (0.99–2.52)
≥ 80 39 1.64 (1.07–2.52)
[Ptrend] [0.010]
continuous per kg 1.02 (1.01–1.03)
Jee et al. (2008) 770 556 Pancreas BMI Age, smoking
National Health Men < 20.0 199 0.87 (0.71–1.08)
Insurance Incidence 20.0–22.9 678 1.01 (0.87–1.16)
Corporation 23.0–24.9 524 1.00
Republic of 25.0–29.9 442 1.06 (0.90–1.24)
Korea ≥ 30.0 17 1.34 (0.75–2.38)
1992–2006 [Ptrend] [0.1139]
423 273 BMI Age, smoking
Women < 20.0 80 0.88 (0.62–1.24)
Incidence 20.0–22.9 246 1.09 (0.84–1.40)
23.0–24.9 178 1.00
25.0–29.9 253 1.35 (1.05–1.74)
≥ 30.0 34 1.80 (1.14–2.86)

[Ptrend] [0.0014]
Luo et al. (2008) 138 503 Pancreas BMI Age, treatment Study of
Women’s Health Women < 22.0 25 0.8 (0.5–1.2) assignments, postmenopausal
Initiative Incidence 22.0–24.9 62 1.0 cigarette smoking, women
USA 25.0–29.9 84 0.9 (0.6–1.2) diabetes
1993–2005 30.0–34.9 56 1.1 (0.7–1.5)
≥ 35.0 24 0.8 (0.5–1.3)
[Ptrend] [0.9]
207
208

Follow-up Sex
period Incidence/
mortality
Luo et al. (2008) WC (cm), quintiles (range, median)
(cont.) 35.0–74.5, 70.5 41 1.0
74.6–81.0, 78.0 50 1.1 (0.7–1.7)
81.1–88.0, 85.0 46 1.0 (0.7–1.6)
88.1–97.4, 92.4 63 1.4 (0.9–2.0)
97.5–194.2, 105.0 51 1.1 (0.7–1.6)
[Ptrend] [0.6]
per 10 cm 1.05 (0.95–1.15)
Type of weight change:
Stable weight 85 1.0
Steady gain in weight 77 0.9 (0.6–1.2)
Lost weight and kept it off 5 0.6 (0.3–1.5)
Weight up and down 83 0.9 (0.7–1.2)
(> 10 lb)
Stolzenberg- 293 562 Pancreatic BMI Age, smoking,
Solomon et al. Men adenocarcinoma 18.5– < 25.0 110 1.00 race, energy intake,
(2008) Incidence ICD-10: 25.0–29.9 227 1.22 (0.97–1.54) energy-adjusted total
NIH-AARP C25.0–25.9 30.0–34.9 66 1.09 (0.80–1.48) fat intake, diabetes;
cohort Excludes ≥ 35.0 26 1.61 (1.05–2.49) for WC, also
USA endocrine [Ptrend] [0.07] adjusted for BMI
1995–2000 tumours WC (cm)
< 88.9 40 1.00
88.9–93.3 35 1.00 (0.62–1.61)
93.3–98.4 39 0.81 (0.49–1.32)
98.4–106 46 0.96 (0.58–1.58)
≥ 106 52 0.95 (0.54–1.67)
[Ptrend] [0.91]
Follow-up Sex
period Incidence/
mortality
Stolzenberg- 201 473 BMI Age, smoking,
Solomon et al. Women 18.5– < 25.0 84 1.00 race, energy intake,
(2008) Incidence 25.0–29.9 84 1.33 (0.98–1.81) energy-adjusted total
(cont.) 30.0–34.9 38 1.40 (0.95–2.07) fat intake, diabetes;
35.0 19 1.29 (0.78–2.16) for WC, also
[Ptrend] [0.09] adjusted for BMI
WC (cm)
< 74.9 14 1.00
74.9–83.2 24 1.74 (0.89–3.41)
83.2–92.1 28 1.88 (0.92–3.85)
≥ 92.1 34 2.53 (1.13–5.65)
[Ptrend] [0.04]
Arnold et al. 48 525 Pancreas BMI Age, diabetes,
(2009) Black men ICD-9: 157 < 18.5 2 0.44 (0.11–1.77) family history of
Cancer and women 18.5–24.9 122 1.00 pancreatic cancer,
Prevention Study Mortality 25–29.9 136 0.89 (0.70–1.40) cholecystectomy,
II (CPS II) ≥ 30 80 1.06 (0.80–1.42) smoking status;
USA 17 602 BMI analysis for men and
1984–2004 Black men < 18.5 0 – women also adjusted
Mortality 18.5–24.9 45 1.00 for sex
25–29.9 65 1.02 (0.69–1.49)

≥ 30 33 1.66 (1.05–2.63)
30 923 BMI
Black < 18.5 2 0.60 (0.15, 2.44)
women 18.5–24.9 77 1.00
Mortality 25–29.9 71 0.82 (0.59–1.14)
≥ 30 47 0.82 (0.56–1.18)
1 011 864 BMI
White men < 18.5 86 0.93 (0.75–1.16)
and women 18.5–24.9 2644 1.00
Mortality 25–29.9 2351 1.15 (1.08–1.22)
≥ 30 690 1.40 (1.28–1.52)
209
210

Follow-up Sex
period Incidence/
mortality
Arnold et al. 444 351 BMI
(2009) White men < 18.5 19 0.83 (0.53–1.31)
(cont.) Mortality 18.5–24.9 1080 1.00
25–29.9 1479 1.11 (1.02–1.20)
≥ 30 336 1.42 (1.25–1.60)
567 513 BMI
White < 18.5 67 0.97 (0.76–1.24)
women 18.5–24.9 1564 1.00
Mortality 25–29.9 872 1.20 (1.10–1.30)
≥ 30 354 1.37 (1.22–1.54)
Johansen et al. 33 325 Pancreas BMI Age, sex, smoking,
(2009) Men and ICD-7: 157 < 20 10 0.84 (0.44–1.61) alcohol consumption
Malmö women ICD-10: C25 20–24.9 101 1.00
Preventive Incidence 25–29.9 54 0.83 (0.60–1.16)
Project ≥ 30 18 1.38 (0.83–2.28)
Sweden continuous 1.04 (0.995–1.08)
1974–2004 Weight gain > 10 kg
No 118 1.00
Yes 52 1.07 (0.77–1.48)
Missing 13 0.65 (0.34–1.27)
Meinhold et al. 27 035 Pancreas BMI, quartiles Age, smoking,
(2009) Men ICD-9: 157, Q1 117 1.00 energy intake,
ATBC subcohort Incidence excluding 157.4 Q2 139 0.97 (0.76–1.24) diabetes mellitus
of non-diabetics Q3 41 1.03 (0.72–1.47) (self-reported)
Finland Q4 8 1.42 (0.69–2.93)
1985–2004 continuous 1.01 (0.94–1.08)
[Ptrend] [0.80]
Stevens et al. 1.29 million Pancreas BMI RR (floating SE) Age, region, SES,
(2009) Women ICD-10: C25 < 22.5 246 1.02 (0.07) smoking, height
Million Women Incidence 22–24.9 311 1.00 (0.06)
Study 25–27.4 260 0.99 (0.06)
USA 27.5–29.9 188 1.17 (0.09)
1996–2006 30–32.4 119 1.27 (0.12)
≥ 32.5 152 1.42 (0.12)
Follow-up Sex
period Incidence/
mortality
Stevens et al. 1.29 million BMI RR (floating SE) Age, region, SES,
(2009) Women < 22.5 334 1.08 (0.06) smoking, height
(cont.) Mortality 22–24.9 400 1.00 (0.05)
25–27.4 347 1.03 (0.05)
27.5–29.9 227 1.09 (0.07)
30–32.4 139 1.14 (0.10)
≥ 32.5 188 1.36 (0.10)
Whitlock et al. 894 576 Pancreas BMI, per 5 kg/m2 Study, sex, age,
(2009) Men and ICD-9: 157 For BMI 15–25 470 0.87 (0.65–1.17) baseline smoking
Pooled analysis women For BMI 25–50 520 1.04 (0.86–1.25)
of 57 cohort Mortality For BMI 15–50 1.07 (0.97–1.19)
studies
Europe, Japan,
and USA
Follow-up varied
by cohort
Arslan et al. 2170 (men: Pancreas BMI Cohort, age, sex, Non-significant
(2010) 1059; < 18.5 19 0.84 (0.44–1.59) anthropometry positive associations
Pancreatic women: ≥ 18.5– < 25.0 759 1.00 source, smoking, were observed with
Cancer Cohort 1111) ≥ 25– < 30 868 1.15 (1.00–1.33) diabetes history WC (Ptrend = 0.09)
Consortium Incidence ≥ 30– < 35 325 1.13 (0.93–1.37)

(PanScan) pooled ≥ 35 124 1.26 (0.93–1.71)
analysis, nested [Ptrend] [0.047]
case–control
Follow-up varies
by cohort
Jiao et al. (2010) 943 759 Pancreatic BMI Age, sex, cohort,
Pooled analysis Men and adenocarcinoma 16.5–18.4 17 0.89 (0.55–1.44) smoking
of 7 cohort women ICD-10: C25 18.5–24.9 855 1.00
studies Incidence excluding C25.4 25–29.9 1109 1.13 (1.03–1.23)
Follow-up varies ICD-8/9: 157 30–34.9 381 1.19 (1.05–1.35)
by cohort excluding 157.4 ≥ 35 92 1.19 (0.96–1.48)
[Ptrend] [0.001]
per 5 kg/m2 1.08 (1.03–1.14)
211
212

Follow-up Sex
period Incidence/
mortality
Jiao et al. (2010) 458 070 BMI
(cont.) Men 16.5–18.4 7 0.88 (0.42–1.86)
Incidence 18.5–24.9 465 1.00
25–29.9 793 1.11 (0.99–1.25)
30–34.9 240 1.11 (0.95–1.30)
≥ 35 43 1.34 (0.98–1.84)
[Ptrend] [0.03]
per 5 kg/m2 1.05 (0.98–1.12)
485 689 BMI
Women 16.5–18.4 10 0.91 (0.48–1.70)
Incidence 18.5–24.9 390 1.00
25–29.9 316 1.15 (0.99–1.34)
30–34.9 141 1.34 (1.11–1.64)
≥ 35 49 1.09 (0.81–1.47)
[Ptrend] [0.01]
per 5 kg/m2 1.12 (1.05–1.19)
Parr et al. (2010) 326 387 Pancreas BMI Age, sex, smoking
Pooled analysis Men and ICD-9: 157 < 18.5 11 0.71 (0.38–1.31)
of 39 cohort women ICD-10: C25 18.5–24.9 114 1.00 (0.86–1.16)
studies Mortality 25–29.9 65 0.93 (0.75–1.15)
Asia, Australia, ≥ 30 90 0.75 (0.48–1.18)
and New Zealand per 5 kg/m2 21 0.93 (0.78–1.11)
1961–1999, [Ptrend] [0.24]
median follow-
up 4 yr
Genkinger et al. Women: Pancreas BMI at baseline All: Smoking, No statistically
(2011) 531 755 < 21 196 1.16 (0.96–1.40) diabetes, alcohol significant
Pooling project Men: 314 585 21–22.9 290 1.00 consumption, energy interaction by sex
of prospective Incidence 23–24.9 457 1.07 (0.92–1.25) intake, age, baseline was found for BMI
studies of diet and 25–29.9 847 1.18 (1.03–1.36) year at baseline, BMI in
and cancer (14 mortality ≥ 30 345 1.47 (1.23–1.75) early adulthood, or
cohort studies) [Ptrend] [< 0.001] BMI change
per 5 kg/m2 1.14 (1.07–1.21)
Follow-up Sex
period Incidence/
mortality
Genkinger et al. Women: Pancreas BMI at baseline Women:
(2011) 531 755 < 21 148 1.15 (0.92–1.44)
(cont.) Men: 314 585 21–22.9 177 1.00
Incidence 23–24.9 221 1.08 (0.88–1.32)
and 25–29.9 378 1.29 (1.04–1.61)
mortality ≥ 30 192 1.46 (1.17–1.80)
[Ptrend] [0.002]
per 5 kg/m2 1.13 (1.06–1.21)
BMI at baseline Men:
< 21 48 1.19 (0.85–1.68)
21–22.9 113 1.00
23–24.9 236 1.07 (0.85–1.34)
25–29.9 469 1.09 (0.88–1.34)
≥ 30 153 1.50 (1.07–2.11)
[Ptrend] [0.06]
per 5 kg/m2 1.14 (1.01–1.29)
BMI in early adulthood All:
< 18.5 163 0.95 (0.79–1.15)
18.5–20.9 519 0.99 (0.87–1.13)
21–22.9 426 1.00
23–24.9 276 1.09 (0.92–1.29)

≥ 25 214 1.21 (1.01–1.45)
[Ptrend] [0.03]
per 5 kg/m2 1.20 (1.10–1.30)
BMI in early adulthood Women:
< 18.5 121 0.92 (0.70–1.21)
18.5–20.9 351 0.96 (0.81–1.14)
21–22.9 239 1.00
23–24.9 113 0.98 (0.78–1.24)
≥ 25 94 1.16 (0.90–1.50)
[Ptrend] [0.18]
per 5 kg/m2 1.14 (1.02–1.28)
213
214

Follow-up Sex
period Incidence/
mortality
Genkinger et al. Women: Pancreas BMI in early adulthood Men:
(2011) 531 755 < 18.5 42 1.02 (0.72–1.45)
(cont.) Men: 314 585 18.5–20.9 168 1.03 (0.78–1.35)
Incidence 21–22.9 187 1.00
and 23–24.9 163 1.19 (0.87–1.62)
mortality ≥ 25 120 1.21 (0.88–1.68)
[Ptrend] [0.06]
per 5 kg/m2 1.27 (1.12–1.44)
BMI change All:
< −2 79 1.44 (1.13–1.85)
−2 to +2 391 1
2–5 493 0.98 (0.85–1.12)
5–10 491 1.13 (0.98–1.30)
> 10 144 1.40 (1.13–1.72)
[Ptrend] [0.04]
Klein et al. (2013) 3349 Pancreas BMI NR Sex, age, study
Pancreatic Men and < 18.5 0.91 (0.54–1.53)
Cancer Cohort women 18.5–24.9 1.00
Consortium Incidence 25–30 1.08 (0.96–1.22)
(PanScan) > 30 1.26 (1.09–1.45)
Lin et al. (2013b) 799 542 Pancreas BMI All: Age, sex, cohort, No associations
Pooled analysis Men and < 18.5 116 1.04 (0.84–1.30) smoking, type 2 were observed when
of 16 cohort women 18.5–19.9 130 0.82 (0.67–1.00) diabetes results were stratified
studies from Mortality 20–22.4 432 0.91 (0.80–1.05) by Asian region (i.e.
Asia Cohort 22.5–24.9 454 1.00 East Asia vs South
Consortium 25–27.4 232 0.95 (0.80–1.11) Asia)
Follow-up varies 27.5–29.9 89 1.01 (0.80–1.29)
by cohort ≥ 30 36 0.96 (0.67–1.37)
Follow-up Sex
period Incidence/
mortality
Lin et al. (2013b) BMI Women:
(cont.) < 18.5 53 0.89 (0.64–1.24)
18.5–19.9 59 0.85 (0.63–1.15)
20–22.4 174 0.78 (0.63–0.96)
22.5–24.9 213 1.00
25–27.4 129 1.01 (0.81–1.27)
27.5–29.9 52 1.02 (0.74–1.39)
≥ 30 28 1.09 (0.72–1.65
BMI Men:
< 18.5 63 1.20 (0.90–1.61)
18.5–19.9 71 0.80 (0.61–1.05)
20–22.4 258 1.03 (0.86–1.24)
22.5–24.9 241 1.00
25–27.4 103 0.87 (0.69–1.10)
27.5–29.9 37 0.99 (0.69–1.42)
≥ 30 8 0.64 (0.30–1.35)
Stolzenberg- 501 698 Pancreatic BMI at age 18 yr Smoking, total fat
Solomon et al. Men and adenocarcinoma < 18.5 188 1.08 (0.92–1.27) consumption, energy
(2013) women ICD-10: 18.5–22.4 652 1.00 intake, sex
NIH-AARP Incidence C25.0–25.9 22.5–24.9 216 1.07 (0.92–1.25)
cohort 25–27.4 91 1.11 (0.89–1.39)

USA ≥ 27.5 59 1.56 (1.19–2.03)
1995–2006 [Ptrend] [0.005]
BMI at age 35 yr
< 18.5 34 1.04 (0.73–1.48)
18.5–22.4 405 1.00
22.5–24.9 350 1.08 (0.94–1.25)
25–29.9 346 1.22 (1.05–1.41)
≥ 30 71 1.37 (1.06–1.79)
[Ptrend] [0.001]
215
216

Follow-up Sex
period Incidence/
mortality
Stolzenberg- BMI at age 50 yr
Solomon et al. < 18.5 27 1.26 (0.85–1.85)
(2013) 18.5–24.9 532 1.00
(cont.) 25–29.9 499 1.13 (1.00–1.29)
≥ 30 148 1.22 (1.02–1.47)
[Ptrend] [0.01]
BMI at age > 50 yr
< 18.5 251.18 (0.79–1.75)
18.5–24.9 6891.00
25–29.9 9341.09 (0.98–1.20)
30–34.9 3401.14 (1.00–1.30)
≥ 35 1341.29 (1.07–1.55)
[Ptrend] [0.01]
Bhaskaran et al. 5 243 978 Pancreas BMI, per 5 kg/m2 3851 total 1.05 (1.00–1.10) Age, diabetes, A 11% significant
(2014) Men and ICD-10: C25 smoking, alcohol risk was observed
Clinical Practice women consumption, SES, when restricting to
Research Incidence calendar year, sex non-smokers only
Datalink
United Kingdom
1987–2012
Bethea et al. 239 597 Pancreas BMI Age, smoking,
(2014) Men and ICD-10: C25 18.5–24.9 187 1.00 education
Pooled study women ICD-9: 157 25–29.9 270 1.08 (0.90–1.31) level, marital
of African Mortality 30–34.9 128 1.25 (0.99–1.57) status, alcohol
Americans (7 ≥ 35 60 1.31 (0.97–1.77) consumption,
cohorts) [Ptrend] [0.03] physical activity;
USA NR BMI analysis for men and
Follow-up times Men 18.5–24.9 68 1.00 women also adjusted
differ across Mortality 25–29.9 123 1.15 (0.85–1.55) for sex
cohorts (at least 30–34.9 45 1.36 (0.93–2.00)
5 yr) ≥ 35 10 1.14 (0.58–2.24)
[Ptrend] [0.20]
Follow-up Sex
period Incidence/
mortality
Bethea et al. NR BMI
(2014) Women 18.5–24.9 119 1.00
(cont.) Mortality 25–29.9 147 1.03 (0.80–1.31)
30–34.9 83 1.16 (0.87–1.55)
≥ 35 50 1.34 (0.95–1.89)
[Ptrend] [0.08]
Untawale et al. 51 251 Pancreas BMI Age, sex, enrolment
(2014) Men and < 18.5 23 1.89 (1.15–3.09) year, dialect,
Singapore women 18.5–21.4 55 1.34 (0.92–1.96) education level,
Chinese Health Incidence 21.5–24.4 53 1.00 diabetes, smoking
Study 24.5–27.4 47 1.46 (0.99–2.17) history, alcohol
China ≥ 27.5 16 1.02 (0.58–1.79) consumption, diet,
1993–2011 [Ptrend] [0.08] physical activity,
sleep duration,
energy intake
Genkinger et al. 1 564 218 Pancreas BMI at baseline Age, race, education The positive
(2015) for BMI at ICD-9: 157 15–18.4 51 1.10 (0.83–1.47) level, marital association of WC
National Cancer baseline ICD-10: C25 18.5–21 296 1.01 (0.87–1.16) status, alcohol with increased risk
Institute BMI 1 096 492 for 21–22.9 574 1.00 consumption, of pancreatic cancer
and Mortality BMI in early 23–24.9 908 1.12 (1.01–1.24) physical activity, mortality remained
Cohort adulthood 25–27.4 1134 1.14 (1.03–1.26) smoking status significant when

Consortium 647 478 for 27.5–29.9 653 1.14 (1.01–1.27) additionally adjusting
(pooled analysis WC 30–34.9 617 1.27 (1.13–1.43) for BMI
of 20 cohort Men and 35– < 60 212 1.34 (1.14–1.57) No differences
studies) women continuous 1.09 (1.05–1.12) between men
Follow-up varies Mortality and women in
by cohort associations with
BMI at baseline and
in early adulthood, or
with WC
Stronger positive
associations of
pancreatic cancer risk
with BMI change in
women than in men
217
218

Follow-up Sex
period Incidence/
mortality
Genkinger et al. BMI in early adulthood
(2015) 15–18.4 376 1.01 (0.89–1.14)
(cont.) 18.5–21 1036 0.98 (0.89–1.08)
21–22.9 814 1.00
23–24.9 510 1.13 (1.01–1.26)
25–27.4 331 1.36 (1.20–1.55)
27.5–29.9 93 1.48 (1.20–1.84)
30–39.9 61 1.43 (1.11–1.85)
per 5 kg/m2 1.18 (1.11–1.25)
BMI change
< −2.5 117 1.24 (1.01–1.53)
−2.5 to 0 269 1.12 (0.97–1.29)
0–2.4 658 1.00
2.5–4.9 828 1.07 (0.97–1.19)
5–7.4 640 1.11 (0.99–1.24)
7.5–9.9 357 1.11 (0.98–1.27)
≥ 10 354 1.28 (1.12–1.47)
per 5 kg/m2 1.05 (1.01–1.10)
WC (cm), quartiles (sex-specific)
Men: Women:
< 90 < 70 385 1.00
90–99 70–79 660 1.11 (0.98–1.27)
110–109 80–89 531 1.26 (1.10–1.45)
≥ 110 ≥ 90 371 1.31 (1.12–1.54)
per 10 cm 1.09 (1.04–1.13)
[Ptrend] [< 0.0001]
Meyer et al. 35 703 Pancreas BMI 127 total Sex, age, survey,
(2015) Men and ICD-8: 157 < 25 1.00 alcohol consumption,
Swiss cohort women ICD-10: C25 25–29.9 1.20 (0.81–1.78) physical activity,
study Mortality ≥ 30 1.60 (0.93–2.75) civil status, years
Switzerland of education,
1977–2008 nationality, diet
ATBC, Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study; BMI, body mass index (in kg/m 2); BP, blood pressure; CI, confidence interval; EPIC, European Prospective
Investigation into Cancer and Nutrition; ICD, International Classification of Diseases; JACC, Japan Collaborative Cohort Study for Evaluation of Cancer Risk; NIH-AARP, National
Institutes of Health–AARP Diet and Health Study; NR, not reported; SE, standard error; SES, socioeconomic status; WC, waist circumference; yr, year or years
Table 2.2.7b Case–control studies of measures of body fatness and cancer of the pancreas
Bueno de Mesquita Men: 89 BMI 2 yr before diagnosis Men: 10-yr age group, response
et al. (1990) Women: 79 < 23 20 1.00 status, total smoking
The Netherlands Population > 27.9 20 0.88 (0.40–1.90)
1984–1988 [Ptrend] [> 0.50]
BMI 2 yr before diagnosis Women:
< 21.6 15 1.00
> 28.7 12 1.10 (0.46–2.80)
[Ptrend] [> 0.90]
Ghadirian et al. 179 BMI Age, sex, response status,
(1991) Population < 21.1 42 1.00 cigarette smoking
Canada > 26.5 40 0.88 (0.42–1.80)
1984–1988
Ji et al. (1996) Men: 255 BMI Men: Age, income, smoking,
China Women: 183 < 19.4 72 1.0 physical activity, response
1990–1993 Population > 22.5 59 1.40 (0.91–2.10) status, diabetes, vitamin C,
[Ptrend] [0.14] total energy
BMI Women: In women only: green tea
< 19.4 43 1.00 drinking
> 23.2 54 1.50 (0.85–2.50)
[Ptrend] [0.57]
Hanley et al. (2001) 312 BMI 2 yr before interview Men: Age, province, percentage Men who reported
Canada (7 Population < 23.7 31 1.0 weight change, energy a 2.9% or greater
Canadian 23.7– < 25.8 44 1.79 (1.01–3.19) intake, composite index of decrease in weight
provinces) 25.8– < 28.3 40 1.36 (0.74–2.49) physical activity from their maximum

1994–1997 ≥ 28.3 57 1.90 (1.08–3.35) lifetime weight were at
[Ptrend] [0.03] significantly reduced
risk of pancreatic
cancer
BMI 2 yr before interview Women: Age, province, energy Women who reported
< 22.1 32 1.0 intake, age at first a 12.5% or greater
22.1– < 24.5 22 0.64 (0.35–1.18) menstruation, cigarette decrease in weight
24.5– < 27.4 34 0.78 (0.44–1.40) smoking from their maximum
≥ 27.4 51 1.21 (0.70–2.06) lifetime weight were at
[Ptrend] [0.39] significantly reduced
risk of pancreatic
cancer
219
220

Silverman (2001) Men: 218 BMI Men: Age at diagnosis/interview, An interaction was
USA (Atlanta, Women: 213 17.35–23.13 51 1.0 race, area, diabetes observed between BMI
Detroit, New Population 23.17–25.07 39 0.8 (0.5–1.3) mellitus, gall bladder and total energy intake
Jersey) 25.09–27.18 55 1.1 (0.7–1.7) disease, cigarette smoking, in relation to pancreatic
1986–1989 ≥ 27.2 73 1.5 (1.0–2.3) alcohol consumption, cancer risk; those with
[Ptrend] [0.019] income (men), marital high BMI and high
status (women), energy energy intake were at
BMI Women:
intake from food 60% increased risk.
20.49–27.54 40 1.0
27.56–30.25 54 1.4 (0.9–2.3)
30.30–34.21 57 1.5 (0.9–2.4)
≥ 34.43 62 1.5 (0.9–2.5)
[Ptrend] [0.129]
Eberle et al. (2005) Men: 291 Adult BMI Men: Age, cigarette smoking
USA Women: 241 < 23.1 48 1.0 only for usual BMI in men
1995–1999 Population 23.1– < 25.1 70 1.6 (1.04–2.5)
25.1– < 27.1 75 1.6 (1.1–2.5)
≥ 27.1 95 2.1 (1.4–3.2)
[Ptrend] [0.0007]
Adult BMI Women:
< 21.5 67 1.0
21.5– < 23.4 51 0.72 (0.47–1.1)
23.4– < 25.8 62 0.86 (0.58–1.3)
≥ 25.8 61 0.91 (0.61–1.4)
[Ptrend] [NS]
BMI at age 25 yr Men:
< 20.9 44 1.0
20.9– < 22.8 76 1.7 (1.1–2.6)
22.8– < 24.7 79 1.8 (1.2–2.8)
≥ 24.7 91 2.0 (1.4–3.1)
[Ptrend] [0.001]
BMI at age 25 yr Women:
< 19.7 54 1.0
19.7– < 21.0 50 0.88 (0.57–1.4)
21.0– < 22.5 64 1.2 (0.77–1.7)
≥ 22.5 72 1.3 (0.84–1.9)
[Ptrend] [0.13]
Fryzek et al. (2005) Men: 119 Current BMI, quartiles Age, sex, race, county
USA (South- Women: 112 Q1: ≤ 24.4 33 1.0 group, smoking, relative
eastern Michigan) Population Q2: 24.5–27.3 59 0.4 (0.3–0.7) with pancreatic cancer,
1996–1999 Q3: 27.4–31.5 22 0.2 (0.1–0.3) income, medical history of
Q4: 31.5–67.8 17 0.1 (0.0–0.2) diabetes
[Ptrend] [< 0.0001]
BMI 5 yr before interview, quartiles
Q1: ≤ 24.1 46 1.0
Q2: 24.2–26.5 56 1.1 (0.6–1.8)
Q3: 26.6–30.3 68 1.3 (0.8–2.2)
Q4: 30.4–68.5 61 1.0 (0.6–1.8)
[Ptrend] [0.77]
BMI 20 yr before interview, quartiles
All:
Q1: 0.0–22.2 43 1.0
Q2: 22.3–24.4 48 1.1 (0.6–1.9)
Q3: 24.5–27.4 71 1.6 (0.9–2.6)
Q4: 27.5–43.0 69 1.4 (0.8–2.5)
[Ptrend] [0.15]
Men:
Q1: 0.0–22.2 8 1.0
Q2: 22.3–24.4 25 1.6 (0.6–4.1)
Q3: 24.5–27.4 43 2.6 (1.0–6.4)
Q4: 27.5–43.0 43 2.4 (1.0–6.2)

[Ptrend] [0.048]
Women:
Q1: 0.0–22.2 35 1.0
Q2: 22.3–24.4 23 1.2 (0.6–2.5)
Q3: 24.5–27.4 28 1.5 (0.7–3.0)
Q4: 27.5–43.0 26 1.4 (0.7–3.0)
[Ptrend] [0.37]
221
222

Fryzek et al. (2005) BMI, ever-smokers
(cont.) ≤ 22.2 34 1.0
22.3–24.4 32 1.0 (0.5–1.8)
24.5–27.4 52 1.7 (0.9–3.1)
27.5–43.0 36 0.9 (0.5–1.8)
[Ptrend] [0.94]
BMI, never-smokers
≤ 22.2 9 1.0
22.3–24.4 16 1.6 (0.6–0.46)
24.5–27.4 19 1.5 (0.5–4.0)
27.5–43.0 33 3.3 (1.2–9.2)
[Ptrend] [0.014]
Pezzilli et al. (2005) 400 BMI before diagnosis Matched for sex, age
Italy Hospital < 23 110 1.01 (0.72–1.41) (± 5 yr), social class,
23–29.9 246 1.00 geographical region
≥ 30 44 0.96 (0.60–1.53)
Lo et al. (2007) 194 BMI 1 yr before Age, sex, residence
Egypt Hospital < 27 99 1.0
2001–2004 27–31 59 1.4 (0.9–2.2)
≥ 32 28 1.5 (0.8–2.9)
Anderson et al. 422 BMI 1 yr before Age, education level,
(2009) Population < 25 148 1.00 smoking status, family
Canada (Ontario) 25–29.9 183 1.77 (1.19–2.62) history of pancreatic
2003–2007 ≥ 30 83 3.51 (1.92–6.39) cancer, weekly fruit
servings, alcohol
consumption, caffeinated
beverages, allergies
Li et al. (2009) 841 (men: 496; Mean lifetime BMI, All: Age, race, sex, smoking, Associations were
USA (Texas) women: 282) per 5 kg/m2 increase 841 1.55 (1.32–1.84) alcohol consumption, somewhat stronger in
2004–2008 Population (proxy Men: history of diabetes, family ever-smokers than in
controls) 496 1.80 (1.45–2.23) history of cancer never-smokers (1.75 vs
Women: 1.46)
345 1.32 (1.02–1.70) When stratifying
BMI by age ranges,
the greatest risk of
pancreatic cancer was
found at the ages of
onset of overweight
and/or obesity between
14–19 yr and 20–29 yr
Urayama et al. 574 BMI at age 20 yr Centre, age at interview,
(2011) Population 18.5–21.1 101 1.00 sex, diabetes mellitus,
Czech Republic and 21.2–22.8 113 1.15 (0.79–1.69) chronic pancreatitis,
Slovakia 22.9–24.5 161 1.81 (1.24–2.63) smoking, alcohol
2004–2009 > 24.5 164 1.79 (1.23–2.61) consumption
per 5 kg/m2 1.45 (1.15–1.84)
BMI at age 40 yr
18.5–23.0 106 1.00
23.1–24.8 114 1.04 (0.72–1.52)
24.9–27.3 154 1.40 (0.97–2.03)
> 27.3 173 1.57 (1.09–2.27)

per 5 kg/m2 1.24 (1.04–1.47)
BMI 2 yr before interview
18.5–24.3 131 1.00
24.4–27.1 151 1.07 (0.75–1.52)
27.2–30.4 153 1.04 (0.73–1.47)
> 30.4 130 0.91 (0.63–1.30)
per 5 kg/m2 0.98 (0.85–1.13)
Lin et al. (2013a) 360 (men: 145; BMI in the yr before study entry Age, sex, history of
Japan women: 215) < 25 278 1.00 diabetes, cigarette
2010–2012 Hospital 25.0–29.9 64 0.96 (0.65–1.43) smoking
≥ 30 16 1.21 (0.53–2.77)
223
224

Zheng et al. (2016) 323 Current BMI Age, sex, race, residential
China Population (family < 24.0 197 1.00 areas, smoking, tea
2011–2013 members of other ≥ 24.0 126 1.77 (1.22–2.57) drinking, mental
inpatients) pressure, family history
of pancreatic cancer,
diabetes, gallstone, intake
of pickles and vegetables
Pancreatic neuroendocrine tumours
Halfdanarson et al. 309 Current BMI
(2014) Hospital < 30 141 1.00
USA (Mayo Clinic ≥ 30 61 1.65 (1.11–2.45)
Rochester
2004–2011
BMI, body mass index (in kg/m 2); CI, confidence interval; NS, not significant; yr, year or years
Table 2.2.7c Meta-analyses of measures of body fatness and cancer of the pancreas
Reference Total number of Exposure categories Relative risk Adjustment for Comments
Total number of cases
Michaud et al. (2001) 2 cohort studies BMI Height, BMI at baseline,
350 < 23 1.00 age, smoking, history
23–24.9 1.09 (0.79–1.49) of diabetes mellitus,
25.0–26.9 1.29 (0.92–1.80) cholecystectomy
27.0–39.9 1.30 (0.91–1.87)
≥ 30 1.72 (1.19–2.48)
[Ptrend] [0.003]
Berrington de Gonzalez et 6 case–control studies BMI, per 1 kg/m2 1.02 (1.01–1.03) Age (all), smoking and No differences were
al. (2003) 8 cohort studies increase diabetes (not all studies) observed between men and
6391 women or when stratifying
by study design (cohort vs
case–control)
Larsson et al. (2007) 21 prospective studies BMI, per 5 kg/m2 All: All studies adjusted for
(13 in men and 10 in increase 1.12 (1.06–1.17) age, cigarette smoking;
women) Men: 13 studies also adjusted
8062 1.16 (1.05–1.28) for diabetes
Women:
1.10 (1.02–1.19)
Renehan et al. (2008) 12 prospective studies BMI, per 5 kg/m2 Men: Method of BMI When stratifying by region,
All studies: increase 1.07 (0.93–1.23) determination, extent the highest risk ratios were
Men: 2390 Women: of cancer site-specific reported in North America
Women: 2053 1.12 (1.03–1.23) risk factor adjustment, (n = 2 studies)
Studies with both Men: geographical region
sexes: 1.07 (0.83–1.39)

Men: 839 Women:
Women: 778 1.12 (0.95–1.33)
Guh et al. (2009) 10 prospective studies BMI Men:
(4 in men and 6 in Normal 1.00
women) Overweight 1.28 (0.94–1.75)
NR Obesity 2.29 (1.65–3.19)
BMI Women:
Normal 1.00
Overweight 1.24 (0.98–1.56)
Obesity 1.60 (1.17–2.20)
225
226

Reference Total number of Exposure categories Relative risk Adjustment for Comments
Total number of cases
Aune et al. (2012) 23 prospective studies BMI, per 5 kg/m2 All (23 studies): Non-linear association
9504 increase 1.10 (1.07–1.14) between BMI and pancreatic
Men (14 studies): cancer risk, with the most
1.13 (1.04–1.22) pronounced increase in risk
Women (15 studies): in those with BMI > 35
1.10 (1.04–1.16)
Never-smoker (5
studies):
1.11 (1.04–1.17)
Ever-smoker (4 studies):
1.03 (0.95–1.10)
WCRF/AICR (2012) 23 cohort studies BMI, per 5 kg/m2 Incidence: NR No differences were
9504 increase 1.10 (1.07–1.14) observed between men and
BMI, per 5 kg/m2 Mortality: women. Some evidence for
increase 1.10 (1.02–1.19) a non-linear dose–response
5 cohort studies WC, per 10 cm increase 1.11 (1.05–1.18) NR with an increase in risk from
949 BMI ≥ 25
4 cohort studies BMI at age 20 yr, per 1.12 (0.97–1.29) NR
900 5 kg/m2 increase
BMI, body mass index (in kg/m 2); CI, confidence interval; NR, not reported; WC, waist circumference; WCRF/AICR, World Cancer Research Fund/American Institute for Cancer
Research; yr, year or years
References Bosetti C, Bertuccio P, Malvezzi M, Levi F, Chatenoud

L, Negri E, et al. (2013). Cancer mortality in Europe,
2005-2009, and an overview of trends since 1980. Ann
Anderson LN, Cotterchio M, Gallinger S (2009). Lifestyle, Oncol, 24(10):2657–71. doi:10.1093/annonc/mdt301
dietary, and medical history factors associated with PMID:23921790
pancreatic cancer risk in Ontario, Canada. Cancer Bueno de Mesquita HB, Moerman CJ, Runia S, Maisonneuve
Causes Control, 20(6):825–34. doi:10.1007/s10552-009- P (1990). Are energy and energy-providing nutrients
9303-5 PMID:19194662 related to exocrine carcinoma of the pancreas? Int
Arnold LD, Patel AV, Yan Y, Jacobs EJ, Thun MJ, Calle J Cancer, 46(3):435–44. doi:10.1002/ijc.2910460319
EE, et al. (2009). Are racial disparities in pancreatic PMID:2394510
cancer explained by smoking and overweight/obesity? Eberle CA, Bracci PM, Holly EA (2005). Anthropometric
Cancer Epidemiol Biomarkers Prev, 18(9):2397–405. factors and pancreatic cancer in a population-based
doi:10.1158/1055-9965.EPI-09-0080 PMID:19723915 case-control study in the San Francisco Bay area.
Arslan AA, Helzlsouer KJ, Kooperberg C, Shu XO, Cancer Causes Control, 16(10):1235–44. doi:10.1007/
Steplowski E, Bueno-de-Mesquita HB, et al.; s10552-005-0354-y PMID:16215874
Pancreatic Cancer Cohort Consortium (PanScan) Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C,
(2010). Anthropometric measures, body mass index, Rebelo M, et al. (2015). Cancer incidence and mortality
and pancreatic cancer: a pooled analysis from the worldwide: sources, methods and major patterns in
Pancreatic Cancer Cohort Consortium (PanScan). GLOBOCAN 2012. Int J Cancer, 136(5):E359–86.
Arch Intern Med, 170(9):791–802. doi:10.1001/ doi:10.1002/ijc.29210 PMID:25220842
archinternmed.2010.63 PMID:20458087 Friedman GD, van den Eeden SK (1993). Risk factors for
Aune D, Greenwood DC, Chan DS, Vieira R, Vieira AR, pancreatic cancer: an exploratory study. Int J Epidemiol,
Navarro Rosenblatt DA, et al. (2012). Body mass index, 22(1):30–7. doi:10.1093/ije/22.1.30 PMID:8449644
abdominal fatness and pancreatic cancer risk: a system- Fryzek JP, Schenk M, Kinnard M, Greenson JK, Garabrant
atic review and non-linear dose-response meta-ana- DH (2005). The association of body mass index
lysis of prospective studies. Ann Oncol, 23(4):843–52. and pancreatic cancer in residents of southeastern
doi:10.1093/annonc/mdr398 PMID:21890910 Michigan, 1996–1999. Am J Epidemiol, 162(3):222–8.
Batty GD, Shipley MJ, Jarrett RJ, Breeze E, Marmot MG, doi:10.1093/aje/kwi183 PMID:15987732
Smith GD (2005). Obesity and overweight in relation Gapstur SM, Gann PH, Lowe W, Liu K, Colangelo L,
to organ-specific cancer mortality in London (UK): Dyer A (2000). Abnormal glucose metabolism and
findings from the original Whitehall study. Int J Obes pancreatic cancer mortality. JAMA, 283(19):2552–8.
(Lond), 29(10):1267–74. doi:10.1038/sj.ijo.0803020 doi:10.1001/jama.283.19.2552 PMID:10815119
PMID:15997248 Genkinger JM, Kitahara CM, Bernstein L, Berrington
Berrington de González A, Spencer EA, Bueno-de- de Gonzalez A, Brotzman M, Elena JW, et al. (2015).
Mesquita HB, Roddam A, Stolzenberg-Solomon Central adiposity, obesity during early adulthood, and
R, Halkjaer J, et al. (2006). Anthropometry, phys- pancreatic cancer mortality in a pooled analysis of
ical activity, and the risk of pancreatic cancer in the cohort studies. Ann Oncol, 26(11):2257–66. doi:10.1093/
European Prospective Investigation into Cancer annonc/mdv355 PMID:26347100
and Nutrition. Cancer Epidemiol Biomarkers Prev, Genkinger JM, Spiegelman D, Anderson KE, Bernstein L,
15(5):879–85. doi:10.1158/1055-9965.EPI-05-0800 van den Brandt PA, Calle EE, et al. (2011). A pooled
PMID:16702364 analysis of 14 cohort studies of anthropometric factors
Berrington de Gonzalez A, Sweetland S, Spencer E (2003). and pancreatic cancer risk. Int J Cancer, 129(7):1708–
A meta-analysis of obesity and the risk of pancre- 17. doi:10.1002/ijc.25794 PMID:21105029
atic cancer. Br J Cancer, 89(3):519–23. doi:10.1038/ Ghadirian P, Simard A, Baillargeon J, Maisonneuve P,
sj.bjc.6601140 PMID:12888824 Boyle P (1991). Nutritional factors and pancreatic cancer
Bethea TN, Kitahara CM, Sonderman J, Patel AV, Harvey in the francophone community in Montréal, Canada.
C, Knutsen SF, et al. (2014). A pooled analysis of body Int J Cancer, 47(1):1–6. doi:10.1002/ijc.2910470102
mass index and pancreatic cancer mortality in African PMID:1845960
Americans. Cancer Epidemiol Biomarkers Prev, Guh DP, Zhang W, Bansback N, Amarsi Z, Birmingham
23(10):2119–25. doi:10.1158/1055-9965.EPI-14-0422 CL, Anis AH (2009). The incidence of co-morbidi-
PMID:25017247 ties related to obesity and overweight: a systematic
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, review and meta-analysis. BMC Public Health, 9(1):88.
Leon DA, Smeeth L (2014). Body-mass index and risk doi:10.1186/1471-2458-9-88 PMID:19320986
of 22 specific cancers: a population-based cohort study Halfdanarson TR, Bamlet WR, McWilliams RR, Hobday
of 5.24 million UK adults. Lancet, 384(9945):755–65. TJ, Burch PA, Rabe KG, et al. (2014). Risk factors for
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 pancreatic neuroendocrine tumors: a clinic-based
227
case-control study. Pancreas, 43(8):1219–22. Larsson SC, Permert J, Håkansson N, Näslund I, Bergkvist
doi:10.1097/MPA.0000000000000234 PMID:25291526 L, Wolk A (2005). Overall obesity, abdominal adiposity,
Hanley AJ, Johnson KC, Villeneuve PJ, Mao Y; Canadian diabetes and cigarette smoking in relation to the risk
Cancer Registries Epidemiology Research Group of pancreatic cancer in two Swedish population-based
(2001). Physical activity, anthropometric factors and cohorts. Br J Cancer, 93(11):1310–5. doi:10.1038/
risk of pancreatic cancer: results from the Canadian sj.bjc.6602868 PMID:16288300
enhanced cancer surveillance system. Int J Cancer, Li D, Morris JS, Liu J, Hassan MM, Day RS, Bondy ML,
94(1):140–7. doi:10.1002/ijc.1446 PMID:11668489 et al. (2009). Body mass index and risk, age of onset,
IARC (2002). Weight control and physical activity. Lyon, and survival in patients with pancreatic cancer.
France: IARC Press (IARC Handbooks of Cancer JAMA, 301(24):2553–62. doi:10.1001/jama.2009.886
Prevention, Vol. 6). Available from: http://publications. PMID:19549972
iarc.fr/376. Lin Y, Fu R, Grant E, Chen Y, Lee JE, Gupta PC, et al.
Isaksson B, Jonsson F, Pedersen NL, Larsson J, Feychting (2013b). Association of body mass index and risk of
M, Permert J (2002). Lifestyle factors and pancre- death from pancreas cancer in Asians: findings from
atic cancer risk: a cohort study from the Swedish the Asia Cohort Consortium. Eur J Cancer Prev,
Twin Registry. Int J Cancer, 98(3):480–2. doi:10.1002/ 22(3):244–50.
ijc.10256 PMID:11920604 Lin Y, Kikuchi S, Tamakoshi A, Yagyu K, Obata Y,
Jee SH, Yun JE, Park EJ, Cho ER, Park IS, Sull JW, et al. Inaba Y, et al.; JACC Study Group (2007). Obesity,
(2008). Body mass index and cancer risk in Korean men physical activity and the risk of pancreatic cancer in
and women. Int J Cancer, 123(8):1892–6. doi:10.1002/ a large Japanese cohort. Int J Cancer, 120(12):2665–71.
ijc.23719 PMID:18651571 doi:10.1002/ijc.22614 PMID:17304505
Ji BT, Hatch MC, Chow WH, McLaughlin JK, Dai Lin Y, Ueda J, Yagyu K, Ishii H, Ueno M, Egawa N, et al.
Q, Howe GR, et al. (1996). Anthropometric and (2013a). Association between variations in the fat mass
reproductive factors and the risk of pancreatic and obesity-associated gene and pancreatic cancer risk:
cancer: a case-control study in Shanghai, China. a case-control study in Japan. BMC Cancer, 13(1):337.
Int J Cancer, 66(4):432–7. doi:10.1002/(SICI)1097- doi:10.1186/1471-2407-13-337 PMID:23835106
0215(19960516)66:4<432::AID-IJC4>3.0.CO;2-X Lo AC, Soliman AS, El-Ghawalby N, Abdel-Wahab M,
PMID:8635856 Fathy O, Khaled HM, et al. (2007). Lifestyle, occupa-
Jiao L, Berrington de Gonzalez A, Hartge P, Pfeiffer RM, tional, and reproductive factors in relation to pancre-
Park Y, Freedman DM, et al. (2010). Body mass index, atic cancer risk. Pancreas, 35(2):120–9. doi:10.1097/
effect modifiers, and risk of pancreatic cancer: a pooled mpa.0b013e318053e7d3 PMID:17632317
study of seven prospective cohorts. Cancer Causes Luo J, Iwasaki M, Inoue M, Sasazuki S, Otani T, Ye W,
Control, 21(8):1305–14. doi:10.1007/s10552-010-9558-x et al.; JPHC Study Group (2007). Body mass index,
PMID:20383573 physical activity and the risk of pancreatic cancer in
Johansen D, Borgström A, Lindkvist B, Manjer J (2009). relation to smoking status and history of diabetes: a
Different markers of alcohol consumption, smoking large-scale population-based cohort study in Japan –
and body mass index in relation to risk of pancre- the JPHC study. Cancer Causes Control, 18(6):603–12.
atic cancer. A prospective cohort study within the doi:10.1007/s10552-007-9002-z PMID:17401636
Malmö Preventive Project. Pancreatology, 9(5):677–86. Luo J, Margolis KL, Adami HO, LaCroix A, Ye W; Women’s
doi:10.1159/000212088 PMID:19684432 Health Initiative Investigators (2008). Obesity and risk
Klein AP, Lindström S, Mendelsohn JB, Steplowski E, of pancreatic cancer among postmenopausal women:
Arslan AA, Bueno-de-Mesquita HB, et al. (2013). An the Women’s Health Initiative (United States). Br
absolute risk model to identify individuals at elevated J Cancer, 99(3):527–31. doi:10.1038/sj.bjc.6604487
risk for pancreatic cancer in the general popul- PMID:18628761
ation. PLoS One, 8(9):e72311. doi:10.1371/journal. Ma J, Jemal A (2013). The rise and fall of cancer mortality
pone.0072311 PMID:24058443 in the USA: why does pancreatic cancer not follow the
Kohler BA, Sherman RL, Howlader N, Jemal A, Ryerson trend? Future Oncol, 9(7):917–9. doi:10.2217/fon.13.76
AB, Henry KA, et al. (2015). Annual Report to the PMID:23837751
Nation on the Status of Cancer, 1975-2011, featuring Máchová L, Cízek L, Horáková D, Koutná J, Lorenc J,
incidence of breast cancer subtypes by race/ethnicity, Janoutová G, et al. (2007). Association between obesity
poverty, and state. J Natl Cancer Inst, 107(6):djv048. and cancer incidence in the population of the District
doi:10.1093/jnci/djv048 PMID:25825511 Sumperk, Czech Republic. Onkologie, 30(11):538–42.
Larsson SC, Orsini N, Wolk A (2007). Body mass index doi:10.1159/000108284 PMID:17992023
and pancreatic cancer risk: a meta-analysis of prospec- Meinhold CL, Berrington de Gonzalez A, Albanes D,
tive studies. Int J Cancer, 120(9):1993–8. doi:10.1002/ Weinstein SJ, Taylor PR, Virtamo J, et al. (2009).
ijc.22535 PMID:17266034 Predictors of fasting serum insulin and glucose and
228
the risk of pancreatic cancer in smokers. Cancer Causes Silverman DT, Swanson CA, Gridley G, Wacholder S,
Control, 20(5):681–90. doi:10.1007/s10552-008-9281-z Greenberg RS, Brown LM, et al. (1998). Dietary and
PMID:19083105 nutritional factors and pancreatic cancer: a case-con-
Meyer J, Rohrmann S, Bopp M, Faeh D; Swiss National trol study based on direct interviews. J Natl Cancer
Cohort Study Group (2015). Impact of smoking and Inst, 90(22):1710–9. doi:10.1093/jnci/90.22.1710
excess body weight on overall and site-specific cancer PMID:9827525
mortality risk. Cancer Epidemiol Biomarkers Prev, Sinner PJ, Schmitz KH, Anderson KE, Folsom AR (2005).
24(10):1516–22. doi:10.1158/1055-9965.EPI-15-0415 Lack of association of physical activity and obesity
PMID:26215293 with incident pancreatic cancer in elderly women.
Michaud DS, Giovannucci E, Willett WC, Colditz GA, Cancer Epidemiol Biomarkers Prev, 14(6):1571–3.
Stampfer MJ, Fuchs CS (2001). Physical activity, doi:10.1158/1055-9965.EPI-05-0036 PMID:15941975
obesity, height, and the risk of pancreatic cancer. Sirri E, Castro FA, Kieschke J, Jansen L, Emrich K,
JAMA, 286(8):921–9. doi:10.1001/jama.286.8.921 Gondos A, et al. (2016). Recent trends in survival of
PMID:11509056 patients with pancreatic cancer in Germany and the
Nöthlings U, Wilkens LR, Murphy SP, Hankin JH, United States. Pancreas, 45(6):908–14. doi:10.1097/
Henderson BE, Kolonel LN (2007). Body mass index MPA.0000000000000588 PMID:26745860
and physical activity as risk factors for pancreatic Stevens RJ, Roddam AW, Spencer EA, Pirie KL, Reeves
cancer: the Multiethnic Cohort Study. Cancer Causes GK, Green J, et al.; Million Women Study Collaborators
Control, 18(2):165–75. doi:10.1007/s10552-006-0100-0 (2009). Factors associated with incident and fatal
PMID:17219012 pancreatic cancer in a cohort of middle-aged women.
Parr CL, Batty GD, Lam TH, Barzi F, Fang X, Ho SC Int J Cancer, 124(10):2400–5. doi:10.1002/ijc.24196
et al.; Asia-Pacific Cohort Studies Collaboration (2010). PMID:19165860
Body-mass index and cancer mortality in the Asia- Stolzenberg-Solomon RZ, Adams K, Leitzmann M,
Pacific Cohort Studies Collaboration: pooled analyses Schairer C, Michaud DS, Hollenbeck A, et al. (2008).
of 424,519 participants. Lancet Oncol, 11(8):741–52. Adiposity, physical activity, and pancreatic cancer
doi:10.1016/S1470-2045(10)70141-8 PMID:20594911 in the National Institutes of Health-AARP Diet
Patel AV, Rodriguez C, Bernstein L, Chao A, Thun MJ, and Health Cohort. Am J Epidemiol, 167(5):586–97.
Calle EE (2005). Obesity, recreational physical activity, doi:10.1093/aje/kwm361 PMID:18270373
and risk of pancreatic cancer in a large U.S. cohort. Stolzenberg-Solomon RZ, Schairer C, Moore S, Hollenbeck
Cancer Epidemiol Biomarkers Prev, 14(2):459–66. A, Silverman DT (2013). Lifetime adiposity and risk
doi:10.1158/1055-9965.EPI-04-0583 PMID:15734973 of pancreatic cancer in the NIH-AARP Diet and
Pezzilli R, Morselli-Labate AM, Migliori M, Manca M, Health Study cohort. Am J Clin Nutr, 98(4):1057–65.
Bastagli L, Gullo L (2005). Obesity and the risk of pancre- doi:10.3945/ajcn.113.058123 PMID:23985810
atic cancer: an Italian multicenter study. Pancreas, Untawale S, Odegaard AO, Koh WP, Jin AZ, Yuan JM,
31(3):221–4. doi:10.1097/01.mpa.0000175163.42327.91 Anderson KE (2014). Body mass index and risk of
PMID:16163052 pancreatic cancer in a Chinese population. PLoS
Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen ONE, 9(1):e85149. doi:10.1371/journal.pone.0085149
M (2008). Body-mass index and incidence of cancer: PMID:24454807
a systematic review and meta-analysis of prospec- Urayama KY, Holcatova I, Janout V, Foretova L, Fabianova
tive observational studies. Lancet, 371(9612):569–78. E, Adamcakova Z, et al. (2011). Body mass index and
doi:10.1016/S0140-6736(08)60269-X PMID:18280327 body size in early adulthood and risk of pancreatic
Samanic C, Chow WH, Gridley G, Jarvholm B, cancer in a central European multicenter case-con-
Fraumeni JF Jr (2006). Relation of body mass index trol study. Int J Cancer, 129(12):2875–84. doi:10.1002/
to cancer risk in 362,552 Swedish men. Cancer Causes ijc.25959 PMID:21520034
Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9 Verhage BA, Schouten LJ, Goldbohm RA, van den Brandt
PMID:16841257 PA (2007). Anthropometry and pancreatic cancer
Samanic C, Gridley G, Chow WH, Lubin J, Hoover risk: an illustration of the importance of microscopic
RN, Fraumeni JF Jr (2004). Obesity and cancer verification. Cancer Epidemiol Biomarkers Prev,
risk among white and black United States veterans. 16(7):1449–54. doi:10.1158/1055-9965.EPI-07-0201
Cancer Causes Control, 15(1):35–43. doi:10.1023/B:- PMID:17585056
CACO.0000016573.79453.ba PMID:14970733 WCRF/AICR (2012). Continuous Update Project Report.
Silverman DT (2001). Risk factors for pancreatic cancer: a Diet, nutrition, physical activity and pancreatic cancer.
case-control study based on direct interviews. Teratog Available from: https://www.wcrf.org/sites/default/
Carcinog Mutagen, 21(1):7–25. doi:10.1002/1520- files/Pancreatic-cancer-report.pdf.
6 8 6 6 ( 2 0 01) 2 1 : 1< 7: : A I D -T C M 3 >3 . 0 . C O ; 2 - A
PMID:11135318
229

S0140-6736(09)60318-4 PMID:19299006
Zheng Z, Zheng R, He Y, Sun X, Wang N, Chen T, et al.
(2016). Risk factors for pancreatic cancer in China: a
multicenter case-control study. J Epidemiol, 26(2):64–
70. doi:10.2188/jea.JE20140148 PMID:26441209
230
2.2.8 Cancer of the lung Few investigators have explored weight across
the life-course as related to lung cancer risk. In
The lung is the leading cancer site for deaths, general, BMI at cohort baseline (recruitment into
accounting for about 19% of all deaths from the cohort) seems to be more strongly (inversely)
cancer. Most (80–90%) cases of lung cancer can associated with lung cancer risk than is BMI
be attributed to long-term smoking. Because earlier in life (Olson et al., 2002; Fujino et al.,
of the large influence of tobacco smoking, any 2007; Kabat et al., 2008; Lam et al., 2013).
errors in estimating tobacco exposure could Several cohorts have included measurements
lead to errors in attribution of risk to any other of waist and hip circumferences (Olson et al.,
factor known to be associated with tobacco 2002; Kabat et al., 2008; Bethea et al., 2013). In
use, including adiposity, resulting in residual general, waist circumference and waist-to-hip
confounding, even after statistical adjustment ratio were less associated with lung cancer risk
for tobacco exposure, as measured. than was BMI.
Handbook on weight control and physical (b) Case–control studies
activity (IARC, 2002) concluded that the
There were a total of 11 independent reports
evidence of an association between avoidance
from case–control studies on the association of
of weight gain and lung cancer was inadequate.
BMI with lung cancer, conducted in Europe,
The 2007 WCRF review concluded that there
Japan, and the USA (Table 2.2.8b, web only;
was “limited evidence suggesting that low body
available at: http://publications.iarc.fr/570).
fatness (underweight) is a cause of lung cancer”
The studies were highly variable in size, some
(WCRF/AICR, 2007).
including fewer than 200 lung cancer cases,
(a) Cohort studies whereas others included about 1000 (El-Zein
et al., 2013), more than 2000 (Brennan et al.,
The evidence from cohort studies published 2009; ICARE study, France, Tarnaud et al., 2012),
since 2000 includes 18 reports (excluding anal- and more than 3000 (NECSS study, Canada, Pan
yses that were later updated and analyses based et al., 2004; Kabat & Wynder, 1992). In all studies
on fewer than 100 incident cases) and is summa- except those of Kubík et al. (2004) and Kanashiki
rized in Table 2.2.8a (web only; available at: et al. (2005), BMI was assessed on the basis of
http://publications.iarc.fr/570). self-reported height and body weight referring to
In general, studies consistently showed an a recent period (mostly 1 year or 2 years) before
inverse association between BMI and risk of lung disease diagnosis. Several studies collected
cancer. The inverse association is linear across recalled body weight in the more distant past, for
categories of BMI, with about 20–30% lower risk example at age 20–30 years (Goodman & Wilkens,
for those with BMI ≥ 30 kg/m2. The association 1993; Tarleton et al., 2012; Tarnaud et al., 2012;
is generally stronger for current smokers than El-Zein et al., 2013). In addition to various other
for never-smokers (Samanic et al., 2006; Kabat adjustments for potential confounding factors,
et al., 2008; Koh et al., 2010; Smith et al., 2012; all studies except one (Heck et al., 2009) adjusted
Bhaskaran et al., 2014). A meta-analysis of 29 for smoking, although the degree of the adjust-
cohort studies found consistency of the associa- ment varied from smoking status only (current,
tion by sex and region of the world, with a relative former, never) to lifetime cumulative exposure
risk estimate for obesity (compared with normal to tobacco smoke. The large studies by Kabat &
weight) of 0.78 (95% CI, 0.74–0.83) (Duan et al., Wynder (1992) in the USA, Pan et al. (2004) in
2015). Canada, Kanashiki et al. (2005) in Japan, and
231
Tarnaud et al. (2012) in France also provided adulthood showed no significant association
estimates within separate strata of current (Goodman & Wilkens, 1993; Tarleton et al., 2012;
smokers, former smokers, and never-smokers. El-Zein et al., 2013) with lung cancer risk or a
Furthermore, one study in the USA, by Rauscher weaker (inverse) association than that reported
et al. (2000), provided odds ratio estimates only for BMI shortly before diagnosis (Tarnaud et al.,
for former smokers and never-smokers (244 and 2012). In all four studies, cases tended to gain less
188 case–control pairs, respectively). weight during adult life than did controls. In one
Among the studies for which the reference study that analysed lung cancer risk according
time frames for BMI assessment were within 5 to weight gained since early adulthood (Tarleton
years before lung cancer diagnosis, all studies et al., 2012), weight gain was significantly
except that of Rauscher et al. (2000), which inversely related to lung cancer risk, and more
included only former smokers and never- so in current smokers than in never-smokers or
smokers, showed inverse associations of BMI former smokers.
with lung cancer risk. Several studies showed
an increased risk of lung cancer particularly (c) Mendelian randomization studies
in individuals with low BMI, compared with Two studies have applied Mendelian random-
individuals with BMI in the normal mid-range ization in the context of lung cancer (Table 2.2.8c,
or higher (Tarnaud et al., 2012: OR, 2.7; 95% web only; available at: http://publications.iarc.
CI, 1.2–6.2 for BMI < 18.5 vs 18.5– < 25 kg/m2 fr/570). Brennan et al. (2009) used the FTO
as reference category; El-Zein et al., 2013: OR, rs9939609 SNP, which is robustly associated with
2.30; 95% CI, 1.30–4.10 for BMI < 18.5 vs 18.5– BMI (Frayling et al., 2007; Scuteri et al., 2007;
< 25 kg/m2 as reference category; and Kanashiki Peeters et al., 2008), as an instrument for BMI.
et al., 2005: OR, 2.0; 95% CI, 1.2–3.4 for BMI Mendelian randomization analyses showed that
categories < 22.9 vs 22.9– < 25 kg/m2 as reference each 1 kg/m2 increase in BMI was associated
category). However, other studies showed a more with a reduced risk of lung cancer (OR, 0.85; 95%
linear inverse relationship between BMI and CI, 0.72–0.99; P = 0.04), including adenocarci-
relative risk over a wider range of BMI values, noma (OR, 0.51; 95% CI, 0.33–0.82; P = 0.004)
from < 18.5 kg/m2 to > 30 kg/m2. and squamous cell carcinoma (OR, 0.72; 95% CI,
In several larger studies that stratified the 0.57–0.90; P = 0.01). An inverse association was
analysis by current smokers, former smokers, and observed in never-smokers (OR, 0.57; 95% CI,
never–smokers, an increased risk in underweight 0.35–0.94; P = 0.03) but not in former smokers
individuals, and more generally an inverse rela- or current smokers.
tionship between BMI and lung cancer risk, was Gao et al. (2016) used genetic risk scores
observed only in current smokers and former comprising 15 SNPs for childhood BMI and 77
smokers (Kabat & Wynder, 1992; Pan et al., SNPs for adult BMI in Mendelian randomiza-
2004; Kanashiki et al., 2005; Tarleton et al., 2012; tion analyses to assess association between these
Tarnaud et al., 2012; El-Zein et al., 2013), whereas measures of adiposity and all lung cancer and lung
in never-smokers there was no significant asso- cancer subtypes. Each 1 kg/m2 increase in adult
ciation. The study of Rauscher et al. (2000), BMI was associated with a 5% increased risk of
which included only former smokers and never- all lung cancer (95% CI, 1.02–1.09; P = 2.9 × 10−3)
smokers, showed an increase in lung cancer risk and a 10% increased risk of squamous cell
with increasing BMI. carcinoma (95% CI, 1.04–1.16; P = 6.6 × 10−4)
In studies that collected information (assuming that a standard deviation was equiva-
about weight at ages 20–30 years, BMI in early lent to 4.5 kg/m2). There was no association with
232
childhood BMI. There was minimal evidence for Gao C, Patel CJ, Michailidou K, Peters U, Gong J,
a positive directional pleiotropy from Mendelian Schildkraut J, et al.; Colorectal Transdisciplinary Study
(CORECT); Discovery, Biology and Risk of Inherited
randomization Egger regression, and results Variants in Breast Cancer (DRIVE); Elucidating Loci
were null, suggesting that the positive association Involved in Prostate Cancer Susceptibility (ELLIPSE);
between adult BMI and both all lung cancer and Follow-up of Ovarian Cancer Genetic Association and
Interaction Studies (FOCI); and Transdisciplinary
squamous cell lung cancer may be overestimated. Research in Cancer of the Lung (TRICL) (2016).
[The Working Group noted that interpretation of Mendelian randomization study of adiposity-related
this finding is limited because individual-level traits and risk of breast, ovarian, prostate, lung and
data were not available on smoking status, which colorectal cancer. Int J Epidemiol, 45(3):896–908.
doi:10.1093/ije/dyw129 PMID:27427428
may be an important effect modifier. In addition, Goodman MT, Wilkens LR (1993). Relation of body size
there is a potential violation of the Mendelian and the risk of lung cancer. Nutr Cancer, 20(2):179–86.
randomization assumptions in this analysis.] doi:10.1080/01635589309514284 PMID:8233983
Heck JE, Andrew AS, Onega T, Rigas JR, Jackson BP,
Karagas MR, et al. (2009). Lung cancer in a U.S. popul-
ation with low to moderate arsenic exposure. Environ
References Health Perspect, 117(11):1718–23. PMID:20049123
Bethea TN, Rosenberg L, Charlot M, O’Connor GT,
Adams-Campbell LL, Palmer JR (2013). Obesity in
iarc.fr/376.
relation to lung cancer incidence in African American
Kabat GC, Kim M, Hunt JR, Chlebowski RT, Rohan TE
women. Cancer Causes Control, 24(9):1695–703.
(2008). Body mass index and waist circumference in
doi:10.1007/s10552-013-0245-6 PMID:23744044
relation to lung cancer risk in the Women’s Health
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I,
Initiative. Am J Epidemiol, 168(2):158–69. doi:10.1093/
Leon DA, Smeeth L (2014). Body-mass index and risk
aje/kwn109 PMID:18483121
of 22 specific cancers: a population-based cohort study
Kabat GC, Wynder EL (1992). Body mass index and lung
of 5.24 million UK adults. Lancet, 384(9945):755–65.
cancer risk. Am J Epidemiol, 135(7):769–74. doi:10.1093/
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328
oxfordjournals.aje.a116363 PMID:1595676
Brennan P, McKay J, Moore L, Zaridze D, Mukeria A,
Kanashiki M, Sairenchi T, Saito Y, Ishikawa H, Satoh H,
Szeszenia-Dabrowska N, et al. (2009). Obesity and
Sekizawa K (2005). Body mass index and lung cancer:
cancer: Mendelian randomization approach utilizing
a case-control study of subjects participating in a mass-
the FTO genotype. Int J Epidemiol, 38(4):971–5.
screening program. Chest, 128(3):1490–6. doi:10.1378/
doi:10.1093/ije/dyp162 PMID:19542184
chest.128.3.1490 PMID:16162748
Duan P, Hu C, Quan C, Yi X, Zhou W, Yuan M, et al.
Koh W-P, Yuan JM, Wang R, Lee HP, Yu MC (2010).
(2015). Body mass index and risk of lung cancer:
Body mass index and smoking-related lung cancer
systematic review and dose-response meta-analysis. Sci
risk in the Singapore Chinese Health Study. Br J
Rep, 5(1):16938. doi:10.1038/srep16938 PMID:26582414
Cancer, 102(3):610–4. doi:10.1038/sj.bjc.6605496
El-Zein M, Parent ME, Nicolau B, Koushik A, Siemiatycki
PMID:20010947
J, Rousseau MC (2013). Body mass index, lifetime
Kubík AK, Zatloukal P, Tomásek L, Pauk N, Havel L,
smoking intensity and lung cancer risk. Int J Cancer,
Krepela E, et al. (2004). Dietary habits and lung cancer
133(7):1721–31. doi:10.1002/ijc.28185 PMID:23553144
risk among non-smoking women. Eur J Cancer Prev,
Frayling TM, Timpson NJ, Weedon MN, Zeggini E,
13(6):471–80. doi:10.1097/00008469-200412000-00002
Freathy RM, Lindgren CM, et al. (2007). A common
PMID:15548939
variant in the FTO gene is associated with body mass
Lam TK, Moore SC, Brinton LA, Smith L, Hollenbeck
index and predisposes to childhood and adult obesity.
AR, Gierach GL, et al. (2013). Anthropometric meas-
Science, 316(5826):889–94. doi:10.1126/science.1141634
ures and physical activity and the risk of lung cancer
PMID:17434869
in never-smokers: a prospective cohort study. PLoS
Fujino Y; Japan Collaborative Cohort Study for Evaluation
One, 8(8):e70672. doi:10.1371/journal.pone.0070672
of Cancer (2007). Anthropometry, development history
PMID:23940620
and mortality in the Japan Collaborative Cohort Study
Olson JE, Yang P, Schmitz K, Vierkant RA, Cerhan JR,
for Evaluation of Cancer (JACC). Asian Pac J Cancer
Sellers TA (2002). Differential association of body mass
Prev, 8(Suppl):105–12. PMID:18260709
index and fat distribution with three major histologic
types of lung cancer: evidence from a cohort of older
233
women. Am J Epidemiol, 156(7):606–15. doi:10.1093/

aje/kwf084 PMID:12244029
Pan SY, Johnson KC, Ugnat AM, Wen SW, Mao Y;
Canadian Cancer Registries Epidemiology Research
Group (2004). Association of obesity and cancer risk
in Canada. Am J Epidemiol, 159(3):259–68. doi:10.1093/
aje/kwh041 PMID:14742286
Peeters A, Beckers S, Verrijken A, Roevens P, Peeters P,
Van Gaal L, et al. (2008). Variants in the FTO gene are
associated with common obesity in the Belgian popul-
ation. Mol Genet Metab, 93(4):481–4. doi:10.1016/j.
ymgme.2007.10.011 PMID:18055244
Rauscher GH, Mayne ST, Janerich DT (2000). Relation
between body mass index and lung cancer risk in
men and women never and former smokers. Am J
Epidemiol, 152(6):506–13. doi:10.1093/aje/152.6.506
PMID:10997540
Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9
PMID:16841257
Scuteri A, Sanna S, Chen W-M, Uda M, Albai G, Strait J, et al.
(2007). Genome-wide association scan shows genetic
variants in the FTO gene are associated with obesi-
ty-related traits. PLoS Genet, 3(7):e115. doi:10.1371/
journal.pgen.0030115 PMID:17658951
Smith L, Brinton LA, Spitz MR, Lam TK, Park Y,
Hollenbeck AR, et al. (2012). Body mass index and
risk of lung cancer among never, former, and current
smokers. J Natl Cancer Inst, 104(10):778–89. doi:10.1093/
jnci/djs179 PMID:22457475
Tarleton HP, Park SL, Zhu WM, Lee YC, Hashibe M,
Morgenstern H, et al. (2012). Body mass index change
in adulthood and lung and upper aerodigestive tract
cancers. Int J Cancer, 131(6):1407–16. doi:10.1002/
ijc.27383 PMID:22131048
Tarnaud C, Guida F, Papadopoulos A, Cénée S, Cyr D,
Schmaus A, et al. (2012). Body mass index and lung
cancer risk: results from the ICARE study, a large,
population-based case-control study. Cancer Causes
Control, 23(7):1113–26. doi:10.1007/s10552-012-9980-3
PMID:22610667
234
2.2.9 Cancer of the breast in women For premenopausal breast cancer, the risk
diminishes with increasing BMI on an approx-
In women, cancer of the breast constitutes imately linear scale, and for postmenopausal
about 25% of all incident cancers and about 15% breast cancer the risk increases on an approxi-
of all cancer deaths worldwide. There are several mately linear scale. Two large meta-analyses esti-
established risk factors for breast cancer, including mated a 7–8% decrease in premenopausal breast
age at menarche, age at menopause, age at first cancer risk and a 12–13% increase in postmeno-
birth, parity, breastfeeding, alcohol consump- pausal breast cancer risk per 5 kg/m2 (Renehan
tion, physical activity, and use of exogenous et al., 2008; WCRF/AICR, 2010).
estrogens. Breast cancer diagnosed before meno- Among those studies that have assessed the
pause differs from breast cancer diagnosed after association between BMI and breast cancer
menopause in both risk factors and clinical char- risk by estrogen receptor (ER) status (for post-
acteristics. There are several molecular subtypes menopausal and premenopausal breast cancer
of breast cancer; the most important aspect is the combined), the association was most robust for
presence or absence of estrogen receptors in the women with ER-positive tumours (MacInnis
tumour, because this substantially affects treat- et al., 2004; Suzuki et al., 2006; Vrieling et al.,
ment options and prognosis. 2010; Canchola et al., 2012; Bandera et al., 2015;
In 2001, the Working Group of the IARC Neuhouser et al., 2015).
Handbook on weight control and physical Among postmenopausal women, the major-
activity (IARC, 2002) concluded that there was ity of studies that have assessed the interaction
sufficient evidence for a cancer-preventive effect between obesity and use of HRT have found the
of avoidance of weight gain for postmenopausal association between BMI and breast cancer risk
breast cancer. to be apparent only among non-users of HRT
(a) Cohort studies (Feigelson et al., 2004; Lahmann et al., 2004;
Eliassen et al., 2006; Mellemkjaer et al., 2006;
The evidence published since 2000 includes Ahn et al., 2007; White et al., 2012). Similar
about 30 publications from cohort studies conclusions were reported by several meta-ana-
(excluding analyses that were later updated and lyses and systematic literature reviews (WCRF/
analyses based on fewer than 100 incident cases). AICR, 2010).
These findings are displayed for BMI at base-
line in Table 2.2.9a for postmenopausal women (ii) BMI or weight at earlier time points and
and Table 2.2.9b (web only; available at: http:// weight change
publications.iarc.fr/570) for premenopausal Several investigators have assessed the asso-
women, with comments on findings according ciation of BMI or weight at earlier time points
to other measures of body fatness, such as weight and weight change with subsequent breast cancer
changes over the life-course. risk.
For postmenopausal breast cancer, BMI in
(i) BMI
middle adulthood (ages 35–50 years) is associ-
In general, the findings are quite consistent ated with a risk similar to that with baseline BMI
across the studies, showing an inverse association (Ahn et al., 2007), but BMI in early adulthood
between baseline BMI and premenopausal breast (generally reported at age 18 years) is either not
cancer risk and a positive association between associated or modestly inversely associated with
baseline BMI and postmenopausal breast cancer postmenopausal breast cancer risk (Sweeney
risk.
235
et al., 2004; Ahn et al., 2007; Canchola et al., association (Table 2.2.9d; web only; available at:
2012; Bandera et al., 2015). http://publications.iarc.fr/570).
Weight gain since age 18 years has been Studies that assessed weight gave similar
shown to be associated with postmenopausal results to those with BMI for both postmeno-
breast cancer risk (Sweeney et al., 2004; Eliassen pausal women (Table 2.2.9e; web only; available
et al., 2006). Also, weight gain after age 50 years at: http://publications.iarc.fr/570) and premeno-
is positively associated with postmenopausal pausal women (Table 2.2.9f; web only; available
breast cancer risk (Eng et al., 2005). at: http://publications.iarc.fr/570).
Weight loss in adulthood has been exam- Comparable associations were observed for
ined in six studies (Eliassen et al., 2006; Ahn tumours that are both ER-positive and proges-
et al., 2007; Teras et al., 2011; Emaus et al., 2014; terone receptor (PR)-positive, especially for
Neuhouser et al., 2015; Rosner et al., 2015). Across postmenopausal women; see Table 2.2.9g for
these studies, there is not consistent evidence postmenopausal women and Table 2.2.9h (web
that weight loss from about age 50 years to the only; available at: http://publications.iarc.fr/570)
baseline of entry into the cohort affects post- for premenopausal women.
menopausal breast cancer risk. A meta-analysis based on 35 case–control
studies involving 71 216 subjects showed an
(iii) Waist circumference
increased risk of postmenopausal breast cancer
Seven cohort studies have included measure- (OR, 1.15; 95% CI, 1.07–1.24) but not of premen-
ments of waist circumference (Lahmann et al., opausal breast cancer, for which the estimates
2004; Sweeney et al., 2004; Krebs et al., 2006; were suggestive of an inverse association with
Ahn et al., 2007; Canchola et al., 2012; Fourkala higher BMI (overweight and obese subjects) (OR,
et al., 2014; Kabat et al., 2015). Waist circumfer- 0.93; 95% CI, 0.86–1.02) (Cheraghi et al., 2012).
ence (either as measured or as indicated by skirt
size) or waist-to-hip ratio was generally positively (ii) BMI and ethnicity
associated with postmenopausal breast cancer More than 20 studies were carried out
risk, and the strengths of those associations are in Caucasian women in North America and
approximately equivalent to those reported for western Europe (Wenten et al., 2002; Magnusson
BMI. et al., 2005; Tsakountakis et al., 2005; Verla-Tebit
& Chang-Claude, 2005; Dinger et al., 2006;
(b) Case–control studies Rosenberg et al., 2006; Kruk, 2007; Slattery et al.,
For the current evaluation, data from more 2007; Justenhoven et al., 2008; Berstad et al., 2010;
than 400 case–control studies published after Healy et al., 2010; Barnes et al., 2011; Cerne et al.,
2000 were reviewed. Only studies with more 2011; John et al., 2011; Rosato et al., 2011; Attner
than 100 cases are summarized. et al., 2012; Bandera et al., 2013a; Robinson et al.,
2014; John et al., 2015a, b; Sanderson et al., 2015),
(i) BMI
16 studies in women in East Asia (Hirose et al.,
In postmenopausal women, case–control 2001, 2003; Shu et al., 2001; Yoo et al., 2001;
studies yielded consistent results, with Adegoke et al., 2004; Chow et al., 2005; Nichols
increased risk of breast cancer with higher BMI et al., 2005; Tian et al., 2007; Wu et al., 2006;
(Table 2.2.9c). Gao et al., 2009; Shin et al., 2009; Shi et al., 2010;
In premenopausal women, the results are Bao et al., 2011; Kawai et al., 2013; Noh et al.,
less consistent despite the substantial number 2013; Sangrajrang et al., 2013; Minatoya et al.,
of studies; they mostly indicate an inverse 2014), 12 studies in Hispanic or Latina women
236
(de Vasconcelos et al., 2001; Wenten et al., 2002; both categories. Such lower BMI categories were
Ibarluzea et al., 2004; Ziv et al., 2006; Garmendia not specifically examined in most studies in
et al., 2007; Slattery et al., 2007; Justenhoven et al., South Asian women.
2008; John et al., 2011, 2015a, b; Ronco et al., (iii) Waist circumference
2012; Amadou et al., 2014), 8 studies in women in
South Asia (Gilani & Kamal, 2004; Mathew et al., As for BMI, results from case–control studies
2008; Montazeri et al., 2008; Dey et al., 2009; using waist circumference as an indicator of body
Dogan et al., 2011; Lodha et al., 2011; Ghiasvand fatness yielded consistent results in postmeno-
et al., 2012; Singh & Jangra, 2013), and 4 studies pausal women, with mostly positive associations
in Arab women (Alothaimeen et al., 2004; Dogan (Table 2.2.9k).
et al., 2011; Msolly et al., 2011; Elkum et al., 2014). In premenopausal women, the results of
Except for Asian populations, there are the 11 available studies were not consistent
not clear differences in risk estimates between (Table 2.2.9l; web only; available at: http://
ethnic groups for either postmenopausal women publications.iarc.fr/570); two studies (Bandera
(Table 2.2.9i; web only; available at: http:// et al., 2013b; Robinson et al., 2014) showed signif-
publications.iarc.fr/570) or premenopausal icant positive associations, whereas two studies
women (Table 2.2.9j; web only; available at: showed an inverse association (John et al., 2011 in
http://publications.iarc.fr/570). ER-positive, PR-positive tumours only; Amadou
The incidence of breast cancer in Hispanic et al., 2014). Interestingly, the significant positive
Whites is lower than that in non-Hispanic associations were observed in women of African
Whites. In the case–control studies that have ancestry.
evaluated the associations of BMI (or other Evidence is scarce about waist circumference
anthropometric measures) or weight change and risk of breast cancer by hormone receptor
with breast cancer risk and compared Hispanic status. The three studies in postmenopausal
Whites with non-Hispanic Whites (Wenten et al., women (John et al., 2011, 2013; Bandera et al.,
2002; Slattery et al., 2007; John et al., 2015b), the 2013b; Table 2.2.9k) provided conflicting results.
positive association observed in postmenopausal (iv) Change in BMI or weight
women was generally stronger in non-Hispanic Changes in BMI or weight were mostly
Whites than in Hispanic Whites. studied as an increase from the value at age 18,
Most studies in Asian women observed 21, 25, or 30 years to the value at the reference
an increased risk of breast cancer with higher date or 1 year before the reference date.
BMI, especially for postmenopausal women In postmenopausal women (Table 2.2.9m),
(Table 2.2.9i; web only; available at: http:// 12 of the 20 studies found a positive association
publications.iarc.fr/570) and/or tumours that between weight gain and risk of breast cancer
were hormone receptor-positive (ER-positive (Li et al., 2000; Trentham-Dietz et al., 2000; Shu
and/or PR-positive). However, the associations et al., 2001; Friedenreich et al., 2002; Carpenter
between BMI and breast cancer risk in post- et al., 2003, Eng et al., 2005; Han et al., 2006; Wu
menopausal women are observed at lower BMI et al., 2006; Shin et al., 2009), in three studies in
levels in Asian populations than in Caucasian non-Hispanic White women only (Wenten et al.,
populations. Some studies in East Asian women 2002; Slattery et al., 2007; John et al., 2013). One of
(Bao et al., 2011; Kawai et al., 2013) used BMI the two studies of BMI gain also found a positive
< 21 kg/m2 or BMI < 18.5 kg/m2 as a reference association (Hirose et al., 2001). The remaining
and categories of lower BMI for overweight and studies found no significant association.
obesity, and observed a positive association in
237
In the two studies that assessed weight gain decrease in risk (95% CI, 6–16%; P = 2.0 × 10−5) in
specifically after menopause (weight gain after ER-negative tumours (assuming that a standard
age 50 years or in the past 10 years) (Shu et al., deviation [SD] was equivalent to 4.5 kg/m2; Locke
2001; Eng et al., 2005), the association was still et al., 2015). Childhood BMI was inversely asso-
significant but was slightly weaker than that with ciated with all (OR per SD increase, 0.71; 95% CI,
weight change since early adulthood. 0.60–0.80; P = 6.5 × 10−5) and ER-negative breast
When premenopausal women were consid- cancer risk (OR per SD increase, 0.69; 95% CI,
ered (Table 2.2.9n; web only; available at: http:// 0.53–0.98; P = 5.8 × 10−3), where each SD increase
publications.iarc.fr/570), BMI change was was equivalent to 0.073 kg/m2 (Felix et al., 2016).
consistently not associated with risk of breast [There was minimal evidence for positive direc-
cancer in all four available studies (Hirose tional pleiotropy in the associations with child-
et al., 2001; Verla-Tebit & Chang-Claude, 2005; hood BMI, suggesting that estimates may be
Kawai et al., 2014; Robinson et al., 2014). Of 16 underestimated.]
studies, 10 confirmed no association between [Although the inverse association observed
body weight gain and breast cancer risk (Shu between adult BMI and breast cancer risk in this
et al., 2001; Friedenreich et al., 2002; Wenten study is inconsistent with the positive associ-
et al., 2002; Slattery et al., 2007; Wu et al., 2006; ations observed for postmenopausal women in
Berstad et al., 2010; Bandera et al., 2013a; Troisi observational studies, Mendelian randomization
et al., 2013; Robinson et al., 2014; Sanderson et al., analyses represent a lifelong predisposition to
2015). The remaining studies were inconsistent; increased BMI (especially because there is a high
two found an increased risk with increasing correlation between the otherwise independent
body weight gain (Shin et al., 2009; Cribb et al., childhood and adult BMI genetic risk scores).
2011), and three found a protective effect of body The results may suggest that the positive associ-
weight gain in at least one measure of exposure ation between adult BMI and breast cancer risk
(Verla-Tebit & Chang-Claude, 2005; John et al., may be driven by adult weight gain, as a result
2011; Sangaramoorthy et al., 2011). of environmental factors not captured by genetic
risk scores.]
(v) Weight loss
When assessing weight change during adult-
hood, several studies also assessed the impact of
weight loss on breast cancer risk (Trentham-Dietz
et al., 2000; de Vasconcelos et al., 2001; Eliassen
et al., 2006). The results were inconsistent, prob-
ably because of heterogeneity of ethnicity and
current BMI between studies.
(c) Mendelian randomization studies

One Mendelian randomization study has been
conducted to assess the association of childhood
and adult BMI with all and ER-negative breast
cancer risk (Gao et al., 2016; Table 2.2.9o). In this
study, each unit increase in adult BMI was asso-
ciated with a 9% decrease in risk (95% CI, 6–12%;
P = 2.5 × 10−7) in all breast cancers, and an 11%
238
Table 2.2.9a Cohort studies of body mass index and cancer of the breast in postmenopausal women
Reference Total number of Exposure Exposed cases Relative risk Covariates Comments
Cohort subjects categories (95% CI)
Location Incidence/
Follow-up period mortality
Feigelson et al. 62 756 BMI Non-HRT users Age, race, age at Positive association also with
(2004) Incidence < 22 187 1.00 menarche, age at adult weight gain
CPS2 cohort 22–24.9 304 1.06 (0.88–1.27) menopause, parity,
USA 25–26.9 182 1.11 (0.91–1.36) OC use, family history
1992–2001 27–29.9 233 1.41 (1.16–1.71) of BC in first-degree
30–34.9 204 1.74 (1.42–2.13) relative, benign breast
≥ 35 72 1.61 (1.22–2.12) disease, mammography,
[Ptrend] [< 0.0001] height, education level,
BMI Current HRT users physical activity, alcohol No association with adult
< 22 223 1.0 consumption weight gain
22–24.9 253 0.89 (0.74–1.06)
25–26.9 102 0.74 (0.59–0.94)
27–29.9 101 0.86 (0.68–1.09)
30–34.9 51 0.72 (0.53–0.98)
≥ 35 22 1.09 (0.70–1.69)
[Ptrend] [0.12]
Lahmann et al. 103 334 BMI, quintiles Non-HRT users Age, centre, education WC and WHR both showed
(2004) Incidence Q1 98 1.00 level, smoking, alcohol no association
EPIC cohort Q2 127 1.02 (0.78–1.33) consumption, parity, age
Europe Q3 206 1.35 (1.06–1.73) at first pregnancy, age at
1992–2002 Q4 241 1.38 (1.08–1.76) menarche
Q5 239 1.36 (1.06–1.75)
[Ptrend] [0.002]

BMI, quintiles HRT users
Q1 122 1.0
Q2 116 0.90 (0.69–1.17)
Q3 113 0.91 (0.70–1.19)
Q4 92 0.85 (0.64–1.13)
Q5 51 0.71 (0.50–1.10)
[Ptrend] [0.07]
MacInnis et al. 13 598 BMI, quartiles 357 total Age, education level, Association limited to ER+
(2004) Incidence Q1 1.0 country of birth, HRT cases
Population-based Q2 1.2 (0.9–1.5) use
cohort Q3 1.4 (1.0–1.9)
Australia Q4 –
1990–2003 [Ptrend] [0.02]
239
240

Location Incidence/
Sweeney et al. 36 658 BMI 55–64 yr Age, education level, Associations with WHR and
(2004) Incidence < 23.5 101 1.00 age at first birth, age at weight change since age 18 yr
Iowa women’s 23.5–26 78 0.86 (0.64–1.16) menarche, family history similar to those for BMI
cohort 26–29.5 119 1.26 (0.96–1.64) of BC, height
USA ≥ 29.5 130 1.34 (1.03–1.75)
1986–2001 [Ptrend] [0.004]
BMI 65–74 yr
< 23.5 274 1.00
23.5–26 306 1.21 (1.03–1.42)
26–29.5 335 1.26 (1.08–1.49)
≥ 29.5 382 1.48 (1.26–1.73)
[Ptrend] [< 0.0001]
BMI 75–84 yr
< 23.5 112 1.00
23.5–26 129 1.19 (0.92–1.53)
26–29.5 167 1.45 (1.14–1.85)
≥ 29.5 153 1.44 (1.12–1.84)
[Ptrend] [0.001]
Kuriyama et al. 15 054 BMI Age, smoking, alcohol
(2005) Incidence < 18.5–24.9 73 1.00 consumption, diet, age
Population-based 25–27.4 23 1.20 (0.75–1.93) at menopause, age at
cohort 27.5–29.9 12 1.55 (0.84–2.87) menarche, age at first
Japan ≥ 30 7 1.90 (0.87–4.15) pregnancy
1984–1992 [Ptrend] [0.04]
Rapp et al. (2005) 78 484 BMI NR Age, smoking,
Population-based Incidence 18.5–24.9 1.00 occupation
cohort 30–34.9 1.48 (1.12–1.95)
Austria ≥ 35 1.29 (0.79–2.11)
1985–2002 [Ptrend] [0.02]
Chang et al. (2006) 38 660 BMI Age, study centre, race,
USA Incidence < 22.4 139 1.00 family history of BC
PLCO cohort 22.5–24.9 177 1.20 (0.96–1.51) in first-degree relative,
1993–2003 25–27.4 168 1.24 (0.99–1.56) age at menarche, age at
27.5–29.9 114 1.42 (1.11–1.83) menopause, HRT use,
≥ 30 166 1.35 (1.06–1.70) education level
[Ptrend] [0.014]
Location Incidence/
Eliassen et al. 87 143 Weight change (kg), age 18 yr to baseline Age, age at menarche, Weight change since
(2006) Incidence loss ≥ 10 22 1.05 (0.64–1.70) parity, age at first birth, menopause associated more
NHS1 and NHS2 loss 5–9.9 35 1.14 (0.76–1.70) height, weight at age weakly. Association was
USA loss 2–4.9 33 0.77 (0.51–1.15) 18 yr, first-degree family much weaker among users
stable 85 1.00 history of BC, benign of HRT
gain 2–4.9 108 1.02 (0.77–1.36) breast disease, alcohol
gain 5–9.9 204 1.08 (0.83–1.39) consumption, use of
gain 10–19.9 435 1.34 (1.06–1.69) HRT, age at menopause
gain 20–24.9 159 1.55 (1.18–2.02)
gain ≥ 25 313 1.98 (1.55–2.53)
[Ptrend] [< 0.001]
Krebs et al. (2006) 7523 BMI, quartiles 350 total Age, HRT use, bone WC and WHR both showed
Cohort of older Incidence Q1 1.00 density, family history no association
women for Q2 0.82 (0.58–1.15) of BC, exercise,
osteoporosis Q3 1.01 (0.72–1.41) education level, parity,
USA Q4 1.29 (0.92–1.81) age at menarche, age at
1986 [Ptrend] [0.06] menopause, smoking
Average follow-up,
11.3 yr
Lukanova et al. 35 362 BMI Age, tobacco use
(2006) Incidence 18.5–24.9 213 1.00
Population-based 25–29.9 140 0.92 (0.74–1.14)
cohort ≥ 30 69 1.09 (0.83–1.43)

1994–2004
Mellemkjaer et al. 11 992 BMI Parity, age at first birth, WC and WHR both showed
(2006) Incidence < 18.5 7 1.23 (0.58–2.63) education level, benign no association
Population-based 18.5–24.9 237 1.00 breast disease, alcohol
cohort 25–29.9 130 0.88 (0.71–1.09) consumption
Denmark ≥ 30 42 0.94 (0.67–1.31)
1993–2002 [Ptrend] [0.74]
11 796 BMI Parity, age at first birth, WC and WHR both showed
Incidence < 18.5 1 – education level, benign no association
18.5–24.9 96 1.00 breast disease, alcohol
25–29.9 85 1.34 (1.00–1.80) consumption
≥ 30 35 1.17 (0.79–1.73)
241
[Ptrend] [0.28]
242

Location Incidence/
Silvera et al. (2006) 40 318 BMI 662 total Age, alcohol
Canadian Incidence < 25 1.00 consumption, smoking,
mammography 25–29.9 1.12 (0.91–1.38) HRT use, age at
screening cohort ≥ 30 1.26 (0.95–1.67) menarche, age at first
Canada [Ptrend] [0.08] birth, family history of
1980–2000 BC
Suzuki et al. 51 823 BMI ER+PR+: Age, family history of
(2006) Incidence < 18.5 11 1.03 (0.55–1.95) BC, age at menarche,
Swedish 18.5–24.9 345 1.00 parity, age at first birth,
mammography 25–29.9 249 1.23 (1.05–1.46) education level, OC use,
cohort ≥ 30 111 1.67 (1.34–2.07) HRT use, diet, alcohol
Sweden [Ptrend] [< 0.0001] consumption
1987–2003 BMI ER−PR−:
< 18.5 2 0.80 (0.20–3.27)
18.5–24.9 83 1.00
25–29.9 52 0.96 (0.67–1.38)
≥ 30 6 0.52 (0.26–1.04)
[Ptrend] [0.017]
Ahn et al. (2007) 99 039 BMI Non-HRT users: Age, age at first Associations with BMI at
NIH-AARP Incidence 15–18.4 6 0.64 (0.28–1.45) pregnancy, age at age 50 yr similar to BMI at
USA 18.5–22.4 134 1.00 menopause, age at first baseline. Association null
1995–2000 22.5–24.9 179 1.19 (0.95–1.49) birth, parity, smoking, at age 35 yr, inverse at age
25.0–27.4 197 1.35 (1.08–1.68) education level, race, 18 yr. Both WC and WHR
27.5–29.9 136 1.52 (1.29–1.94) family history of BC, positively associated with
30–34.9 175 1.55 (1.22–1.96) alcohol consumption, risk
35–39.9 77 1.89 (1.40–2.55) diet, physical activity,
≥ 40 44 2.08 (1.44–2.99) oophorectomy
[Ptrend] [< 0.001]
Location Incidence/
Ahn et al. (2007) 99 039 BMI HRT users: Age, age at first
(cont.) Incidence 15–18.4 11 0.79 (0.43–1.44) pregnancy, age at
18.5–22.4 280 1.00 menopause, age at first
22.5–24.9 313 1.13 (0.96–1.33) birth, parity, smoking,
25.0–27.4 257 1.19 (1.00–1.42) education level, race,
27.5–29.9 117 1.04 (0.83–1.30) family history of BC,
30–34.9 129 1.14 (0.91–1.42) alcohol consumption,
35–39.9 40 1.13 (0.80–1.61) diet, physical activity,
≥ 40 15 1.10 (0.64–1.88) oophorectomy
[Ptrend] [0.22]
Ericson et al. 11 699 BMI Age
(2007) Incidence < 25 183 1.00
Malmö cohort 25–29.9 147 1.20 (0.96–1.49)
Sweden ≥ 30 62 1.19 (0.89–1.59)
1991–2003 [Ptrend] [0.41]
Lundqvist et al. 14 058 older twins BMI Smoking, physical
(2007) (mean age at < 18.5 12 0.9 (0.5–1.5) activity, education level,
Twin cohort baseline, 56 yr) 18.5–24.9 411 1.0 diabetes
studies Incidence 25–29.9 274 1.2 (1.0–1.4)
Sweden and ≥ 30 59 1.3 (1.0–1.7)
Finland [Ptrend] [< 0.007]
1961–2004
Reeves et al. (2007) 1.2 million BMI Age, region, SES,

Population-based Incidence < 22.5 879 0.85 (0.80–0.91) reproductive history,
cohort 22.5–24.9 1336 1.00 smoking, alcohol
1996–2001 27.5–29.9 878 1.21 (1.13–1.29) activity, HRT use
≥ 30 1274 1.29 (1.22–1.36)
per 10 kg/m2 1.40 (1.21–1.49)
Reinier et al. 32 607 BMI 572 total Age, family history of
(2007) Incidence < 22.0 1.0 BC, age at first birth,
Mammography 22–24.9 1.2 (0.9–1.6) breast density
screening cohort 25.0–27.4 1.4 (1.0–1.8)
in Vermont 27.5–29.9 1.6 (1.1–2.1)
USA ≥ 30 1.9 (1.4–2.5)
1996–2002
243
244

Location Incidence/
Song et al. (2008) 107 481 BMI Age, smoking, alcohol
Korean medial Incidence < 18.5 11 0.54 (0.17–1.73) consumption, exercise
insurance cohort 18.5–20.9 59 0.87 (0.54–1.41)
Republic of Korea 21.0–22.9 132 1.00
1994–2003 23.0–24.9 186 1.27 (0.90–1.80)
25.0–26.7 159 1.52 (1.07–2.15)
27.0–29.9 130 1.97 (1.37–2.83)
≥ 30 36 1.64 (0.91–2.97)
per 1 kg/m2 1.08 (1.04–1.12)
Andreotti et al. 28 319 BMI Age, race, smoking,
(2010) Incidence < 18.5 5 – vegetable intake,
Agricultural 18.5–24.9 186 1.00 exercise, family history
workers 25–29.9 156 1.22 (0.93–1.60) of cancer
USA 30–34.9 93 1.62 (1.17–2.24)
1993–2005 ≥ 35 24 1.07 (0.61–1.87)
[Ptrend] [0.02]
Parr et al. (2010) 130 946 BMI 324 total Age, sex, tobacco use
39 cohorts Mortality < 12–18.4 0.71 (0.22–2.24)
Asia, Australia, 18.5–24.9 1.00
and New Zealand 25–29.9 1.13 (0.85–1.50)
1961–NR ≥ 30 1.63 (1.13–2.35)
[Ptrend] [0.03]
Canchola et al. 52 642 BMI ER+PR+: Age, race, parity, age No association with BMI
(2012) Incidence < 25 740 1.00 at menarche, age at at age 18 yr. WC positively
California 25–29.9 413 1.13 (1.00–1.28) first birth, family associated with risk
Teachers Study ≥ 30 218 1.20 (1.03–1.40) history of BC, alcohol
USA [Ptrend] [0.01] consumption, HRT use
1995–2008 BMI ER−PR−: No association with BMI at
< 25 156 1.00 age 18 yr. WC not associated
25–29.9 91 1.13 (0.87–1.47) with risk
≥ 30 33 0.77 (0.53–1.12)
[Ptrend] [0.36]
Location Incidence/
White et al. (2012) 35 495 BMI Never HRT users: Age, family history of Analyses available by race/
Population-based Incidence < 20 63 0.90 (0.69–1.18) BC, age at first birth, ethnicity: non-Hispanic
Multiethnic 20–24.9 316 1.00 age at menarche, parity, White, Latina, Japanese,
Cohort 25–29.9 396 1.35 (1.17–1.57) smoking, physical Native Hawaiian, African
USA ≥ 30 329 1.60 (1.36–1.87) activity, alcohol American
1993–2004 [Ptrend] [< 0.0001] consumption, height
28 200 BMI Current HRT users: Age, family history of
Incidence < 20 132 1.02 (0.84–1.23) BC, age at first birth,
20–24.9 610 1.00 age at menarche, parity,
25–29.9 376 1.04 (0.91–1.18) smoking, physical
≥ 30 190 1.14 (0.97–1.35) activity, alcohol
[Ptrend] [0.18] consumption, height
Fourkala et al. 1.2 million BMI 1090 Age, age at menarche,
(2014) Incidence per 1 kg/m2 1.06 (1.01–1.12) age at menopause
Ovarian cancer 1.2 million Skirt size 1090 Skirt size remained
screening cohort Incidence per 1 unit 1.05 (1.01–1.08) significant after adjustment
United Kingdom for BMI
2001–2012
Gaudet et al. 28 965 BMI Age, family history Similar association with
(2014) Incidence < 25 441 1.00 of BC, education WC , but in multivariate
CPS2 cohort 25–29.9 401 1.34 (1.17–1.54) level, height, age at adjustment, the BMI
USA ≥ 30 246 1.60 (1.36–1.89) menopause, tobacco use, association persisted but
1997–2006 per 1 kg/m2 1.04 (1.02–1.06) diabetes, race, age at first the WC association did not.

birth, physical activity, Cases overlap with Feigelson
alcohol consumption, et al. (2004)
OC use, HRT use
Bandera et al. 15 234 BMI ER+: Age, education level, Inverse association with BMI
(2015) Incidence < 25 254 1.00 study, family history of in young adulthood and risk.
Pooled data on 25–29.9 469 1.10 (0.93–1.30) BC, age at menarche, WHR positively associated
African American 30–34.9 361 1.21 (1.01–1.45) parity, breastfeeding, age with risk
women in 4 ≥ 35 329 1.32 (1.09–1.60) at first birth, HRT use,
cohorts [Ptrend] [0.002] OC use
USA
1995–2013
245
246

Location Incidence/
Bandera et al. BMI ER−: Age, education level, Inverse association with BMI
(2015) < 25 130 1.00 study, family history of in young adulthood and risk.
(cont.) 25–29.9 200 0.87 (0.69–1.11) BC, age at menarche, WHR positively associated
30–34.9 156 0.90 (0.70–1.17) parity, breastfeeding, age with risk
≥ 35 126 0.82 (0.63–1.08) at first birth, HRT use,
[Ptrend] [0.23] OC use
Kabat et al. (2015) 143 901 BMI, quintiles 7039 total Age, alcohol WC, WHR not associated
Women’s Health Incidence Q1 1.00 consumption, smoking, any more strongly than BMI
Initiative cohort Q2 1.09 (1.01–1.18) physical activity, age at
USA Q3 1.12 (1.04–1.21) menarche, age at first
1992–2013 Q4 1.23 (1.14–1.33) birth, parity, HRT use,
Q5 1.41 (1.31–1.53) family history of BC,
[Ptrend] [< 0.0001] ethnicity, education level
Dartois et al. 67 634 BMI Age, family history of Earlier study by Tehard &
(2016) Incidence < 18.5 84 – BC, education level, Clavel-Chapelon (2006)
E3N cohort 18.5–24.9 2310 1.00 height, age at menarche, showed similar association
France 25–29.9 610 1.19 (1.10–1.30) age at menopause, between WC and risk, but no
1990–2008 ≥ 30 134 1.25 (1.07–1.46) tobacco use, parity, associations with WHR
physical activity, alcohol
consumption, OC use,
HRT use
BC, breast cancer; BMI, body mass index (in kg/m 2); CI, confidence interval; CPS, Cancer Prevention Study; EPIC, European Prospective Investigation into Cancer and Nutrition; HRT,
hormone replacement therapy; NHS, Nurses’ Health Study; NIH-AARP, National Institutes of Health–AARP Diet and Health Study; NR, not reported; OC, oral contraceptive; PLCO,
Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial; SES, socioeconomic status; WC, waist circumference; WHR, waist-to-hip ratio; yr, year or years
Table 2.2.9c Case–control studies of body mass index and cancer of the breast in postmenopausal women
Reference Study populationa Exposure categories Exposed Relative risk Covariates

Study location Total number of cases cases (95% CI) Comments
Period Total number of controls
Source of controls
Li et al. (2000) 479 BMI at age 50–64 yr Age, family history of BC, parity
USA 435 ≤ 21.5 111 1.00
1988–1990 Population; Caucasian women 21.6–24.1 126 1.2 (0.9–1.8)
24.2–27.5 120 1.1 (0.8–1.6)
≥ 27.6 122 1.5 (1.1–2.3)
Trentham-Dietz et al. Postmenopausal women aged BMI Logistic conditional models on age and
(2000) 50–79 yr 11.62–21.94 841 1.0 state. Parity, age at FFTP, family history
USA 5031 21.95–24.02 920 1.0 (0.9–1.2) of BC, recent alcohol consumption,
January 1992–December 5255 24.03–26.44 971 1.1 (1.0–1.3) education level, age at menopause
1994 Population; matched by age 26.45–29.44 1013 1.2 (1.1–1.4)
29.45–54.87 1286 1.6 (1.4–1.9)
[Ptrend] [< 0.001]
de Vasconcelos et al. 177 Current BMI Age, parity, family history of BC,
(2001) 377 < 24.55 38 1.00 education level
Brazil Hospital/population; visitors 24.55–27.64 29 0.61 (0.33–1.14)
May 1995–February 1996 at hospital; 27 relatives of BC 27.65–30.79 35 0.84 (0.46–1.53)
patients ≥ 30.80 29 0.61 (0.33–1.14)
[Ptrend] [0.24]
Shu et al. (2001) 1459 aged 25–64 yr enrolled from BMI at diagnosis Age, education level, family history
China Shanghai Cancer Registry < 20.70 63 1.0 of BC, ever had fibroadenoma, age
August 1996–March 1998 1556 20.70–22.79 95 1.4 (0.9–2.1) at menarche, age at first live birth,
Population; randomly selected 22.80–25.09 134 1.5 (1.0–2.3) exercise, age at menopause
from female residents of 25.10–27.90 125 1.7 (1.1–2.6)

Shanghai (Shanghai Resident ≥ 28.0 83 2.0 (1.2–3.2)
Registry), matched to cases by [Ptrend] [0.003]
age, 5-yr interval
Yoo et al. (2001) 1154 aged ≥ 25 yr, with no BMI Age at interview, occupation, family
Japan previous history of cancer per 1 kg/m2 1.07 (1.04–1.10) history of BC, age at menarche, age at
1988–1992 21 714 menopause, age at FFTP, number of
Hospital FTPs, months of breastfeeding, alcohol
consumption, cigarette smoking,
weight, height
247
248


Source of controls
Friedenreich et al. (2002) 1233 BMI Current age, total energy intake, total
Canada, Alberta 1241 < 24.1 206 1.00 lifetime physical activity, education
1995–1997 Population; frequency-matched ≥ 24.1– < 27.3 179 0.93 (0.69–1.24) level, ever use of HRT, ever diagnosed
to cases by age, 5-yr interval, and ≥ 27.3– < 31.3 187 0.94 (0.70–1.26) with benign breast disease, first-degree
place of residence (urban/rural) ≥ 31.3 199 0.99 (0.74–1.32) family history of BC, ever alcohol
[Ptrend] [0.55] consumption, current smoking
Adebamowo et al. (2003) 234 BMI Age, age at menarche, regularity of
Nigeria, urban 273 ≥ 30 vs < 30 31 1.82 (0.78–4.31) periods; only natural menopause
1998–2000 Population
Carpenter et al. (2003) 1883 Caucasian (including BMI, 1 yr before diagnosis Age at FFTP, age at menarche, age
Canada, USA, and western Hispanic), born in Canada, USA, < 21.7 366 1.00 at menopause, family history of BC,
Europe or western Europe, diagnosed at 21.7–23.6 379 1.10 (0.88–1.37) interviewer, average MET hours per
Group I: March 1987– age ≥ 55 yr 23.7–27.0 497 1.18 (0.95–1.46) week of lifetime exercise activity
December 1989 1628 ≥ 27.1 641 1.34 (1.09–1.66)
Group II: January 1992– Population; matched to cases by [Ptrend] [0.005]
December 1992 neighbourhood
Group III: September
1995–April 1996
Li et al. (2003) 975 BMI at age 65–79 yr Age, income
USA 1007 < 23.32 209 1.00
1997–1999 Population 23.33–26.20 240 1.3 (1.0–1.7)
26.21–30.11 245 1.4 (1.1–1.9)
≥ 30.12 245 1.4 (1.0–1.8)
Pan et al. (2004) 1449 postmenopausal BMI 1449
Canada 2492 < 25 1.00
1994–1997 Population 25–30 1.17 (1.00–1.39)
≥ 30 1.66 (1.33–2.06)
[Ptrend] [< 0.0001]
Chow et al. (2005) Chinese women aged 24–85 yr BMI at diagnosis
Hong Kong Special 198 < 19 10 1.00
Administrative Region 353 19–23 38 1.78 (0.79–4.04)
1995–2000 Hospital; followed up for benign 23–27 42 1.73 (1.04–2.86)
breast disease; no BC 27–31 20 2.06 (1.08–3.93)
> 31 10 3.82 (1.03–14.27)
[Ptrend] [< 0.001]

Source of controls
Zhu et al. (2005) African American, aged 20–64 yr BMI at diagnosis Family history of BC, history of benign
USA 304, without previous cancer < 25 45 1.00 breast disease, alcohol consumption,
1995–1998 history, interviewed 1–3 yr after 25– < 30 55 1.50 (0.70–3.21) smoking, menstrual status, age at
diagnosis ≥ 30 61 2.32 (1.04–5.19) menarche, menstrual cycle length,
305 [Ptrend] [0.039] parity, age at first birth, miscarriages,
Population; no history of BC, history of radiotherapy, use of estrogen
matched to cases by age (5-yr other than for birth control, history of
intervals) and county; women losing weight, history of taking iron
were offered money to participate pills, age at first sexual intercourse,
daily energy intake, physical activity,
use of electric bedding devices, history
of infertility, demographic variables
Okobia et al. (2006) 250 BMI, mean (± SD) Age
Nigeria 250 Cases, 24.74 (± 6.89) 108 0.76 (0.44–1.32)
September 2002–April Hospital; patients recruited Controls, 25.03 (± 5.33)
2004 from the same hospitals as cases,
treated for non-malignant and
non-hormonal surgical disorders
Wu et al. (2006) Asian American (Chinese, BMI, recent Age, ethnicity, duration of residence
USA Japanese and Filipino) women ≤ 20.43 139 1.00 in the USA, education level, age at
1995–2001 aged 25–74 yr > 20.43–22.32 138 0.94 (0.65–1.36) menarche, number of live births, age at
1277 > 22.32–24.60 187 1.13 (0.79–1.62) menopause, intake of tea and soy during
1160 > 24.60 241 1.35 (0.95–1.93) adolescence and adult life, years of

Population; neighbourhood [Ptrend] [0.045] physical activity, height
controls, frequency-matched by
ethnicity and 5-yr age groups
Garmendia et al. (2007) 170 diagnosed within 2 mo before BMI Crude OR; controls matched to cases by
Chile, Santiago recruitment, aged 33–86 yr ≥ 30 122 0.66 (0.39–1.14) 5-yr age interval and place of residence
2005 170
Population; mammography
service of the same hospitals
Kruk (2007) 858 Current BMI Age, recreational activity, breastfeeding,
Poland 1085 < 22.5 78 1.00 stress, passive smoking
2003–2007 Hospital; controls frequency- 22.6– < 25.0 127 1.85 (0.98–2.84) Pinteraction = 0.002
matched by 5-yr age group and 25.0– < 30.0 221 2.13 (1.45–3.13)
place of residence (urban/rural) ≥ 30.0 122 2.62 (1.66–4.11)
249
[Ptrend] [< 0.0001]
250


Source of controls
Tian et al. (2007) 244 aged 22–87 yr BMI Age at enrolment, fasting status, levels
Taiwan, China 244 ≤ 24.45 54 1.00 of adiponectin
January 2004–December Hospital; recruited from health > 24.45 49 2.94 (1.53–5.68)
2005 examination clinics at the same
hospital and time, no history of
cancer, matched by menopausal
status, date of enrolment, and
duration of fasting
Mathew et al. (2008) 1866 BMI Age, centre, religion, marital status,
India 1873 < 25 559 1.00 education level, SES, residence status,
2002–2005 Accompanying persons to cancer 25–29.9 297 1.29 (1.00–1.66) parity, age at first birth, duration of
cases; matched by age ± 5 yr and ≥ 30 76 1.00 (0.64–1.54) breastfeeding, physical activity
residence type (urban/rural)
Montazeri et al. (2008) 116 in situ and invasive cancers BMI Age, age at menopause, family history of
Islamic Republic of Iran 116 18.5–24.9 23 1.00 BC, parity
1996–2000 Hospital; women presenting for 25–29.9 51 2.53 (1.20–5.35)
clinical breast examination ≥ 30 42 3.21 (1.15–8.47)
Nemesure et al. (2009) Women of African descent aged BMI at age ≥ 50 yr Age, HRT use, parity, family history of
Barbados ≥ 21 yr < 25 51 1.00 BC, history of benign breast disease,
July 2002–March 2006 222 25–30 42 0.67 (0.36–1.24) age at first pregnancy, age at menarche,
454 ≥ 30 49 0.70 (0.38–1.28) physical activity, other body size
Population; Barbados Statistical variable
Services; frequency-matched by
5-yr age group
Shin et al. (2009) 3452 aged 20–64 yr (phase 1), Current BMI
China 20–70 yr (phase 2) ≤ 20.9 192 1.0
1996–1998 (phase 1), 3474 21–22.9 285 1.3 (1.0–1.7)
2002–2005 (phase 2) Population; controls frequency- 23–24.9 348 1.5 (1.2–1.9)
matched to cases by age ≥ 25 543 1.8 (1.4–2.2)
[Ptrend] [< 0.001]

Source of controls
Berstad et al. (2010) 4575 BMI, 5 yr before reference date Age, race, education level, study site,
USA 4682 < 25 918 1.00 first-degree family history of BC, parity,
1994–1998 Caucasian: 25–29 579 0.98 (0.84–1.14) age at menopause, HRT use, BMI at age
2953 30–34 254 1.02 (0.82–1.26) 18 yr
3021 ≥ 35 149 1.09 (0.83–1.43)
African American: [Ptrend] [0.67]
1622
1661
Population
Healy et al. (2010) 200 BMI, quartiles
Ireland 519 (age-matched) Q4 vs Q1 2.2 (1.3–3.7) P = 0.002
NR healthy women > 30 vs 20–25 2.04 (1.3–3.3) P = 0.004
Ogundiran et al. (2010) 1233 BMI Age at diagnosis or interview, ethnicity,
Nigeria 1101 < 21 100 1.00 education level, age at menarche,
1998–2009 Population; community register 21–23.9 115 1.04 (0.63–1.71) number of live births, age at first live
of Ibadan 24–27.9 139 0.88 (0.55–1.41) birth, duration of breastfeeding, age
≥ 28 151 0.76 (0.48–1.21) at menopause, family history of BC,
[Ptrend] [0.15] benign breast disease, OC use, alcohol
consumption, height
Pinteraction = 0.85
Barnes et al. (2011) 3074 BMI at age 50–74 yr Family history of BC, benign breast
Germany 6386 ≤ 22.4 1354 1.00 disease, age at menarche, OC use,
2001 (Hamburg); 2002 Population; frequency-matched 22.5–24.9 993 1.06 (0.95–1.17) breastfeeding, parity, cause of

(Rhein-Neckar-Karlsruhe) by year of birth and study region 25–29.9 622 1.04 (0.92–1.18) menopause, age at menopause, alcohol
to 2005 ≥ 30 105 0.93 (0.73–1.19) consumption, HRT use, recent physical
activity, occupational status, year of
birth, study region, lifetime number of
mammograms
Cerne et al. (2011) Caucasian women BMI
Slovenia 784, aged 50–69 yr at diagnosis < 25 267 1.00
January 2006–December 709 25–30 327 1.34 (1.04–1.73)
2008 Hospital; no history of BC ≥ 30 190 1.89 (1.36–2.63)
251
252


Source of controls
Cribb et al. (2011) 207 BMI
Canada, Prince Edward 621 > 25 vs ≤ 25 61% 1.71 (1.08–2.70)
Island Population; women presenting
1999–2002 for routine mammography
screening; matched by age,
menopausal status, and family
history of BC
Rosato et al. (2011) 3869, postmenopausal BMI Age, study centre, study period,
Pooled analysis of 2 4082 < 30 3292 1.00 education level, alcohol consumption,
studies in Italy and Hospital; admitted for acute, ≥ 30 578 1.26 (1.11–1.44) age at menarche, age at first birth, age
Switzerland non-neoplastic diseases, not at menopause, HRT use, family history
1983–1994 (1st study), related to gynaecological or of BC
1991–2007 (2nd study) hormonal conditions, matched
by age and study centre
Attner et al. (2012) 2613 Obesity 2.1% 0.79 (0.52–1.19) 90–1461 days (4 yr) before diagnosis
Sweden, County of Scania 19 898 Obesity defined as comorbidity
2005–2007 Registry: Population Registry of diagnosis of obesity (ICD-10: E66)
Scania
Ghiasvand et al. (2012) 493 women aged ≥ 50 yr enrolled BMI Age, parity, age at menarche, education
Islamic Republic of Iran within 6 mo after diagnosis < 18.5 4 0.60 (0.17–2.11) level, occupation, height, family history
September 2005– 493 18.5–24.9 129 1.00 of BC
December 2008 (cases), Hospital; frequency-matched 25–29.9 208 1.39 (1.02–1.94)
May–August 2009 to cases by 5-yr age groups and ≥ 30 141 1.61 (1.18–2.30)
(controls) province of residence; no history [Ptrend] [0.01]
of BC
Ronco et al. (2012) 367 BMI 165 Age, residence, first-degree family
Uruguay 545 < 25 3.60 (0.33–39.8) history of BC, age at menarche, number
2004–2009 Hospital; non-hospitalized 25–30 5.40 (1.77–16.6) of live births, age at first delivery,
women aged 23–69 yr; ≥ 30 0.84 (0.33–2.12) months of breastfeeding
age-matched, with normal
mammography
Bandera et al. (2013a) 978 postmenopausal women of Current BMI Age, ethnicity, country of origin,
USA, New York City and African ancestry < 25 74 1.00 education level, family history of BC,
New Jersey 958 25–29.99 131 0.93 (0.60–1.44) history of benign breast disease, age at
NR Population; random-digit ≥ 30 304 0.98 (0.66–1.45) menarche, age at menopause, parity,
dialling [Ptrend] [0.94] breastfeeding, age at first birth, HRT
use, OC use

Source of controls
John et al. (2013) 1389 of 2571 Current BMI All non-users of HRT
USA 1644 of 2706 < 25.0 208 1.00
Hispanic cases: 1995–2002 Hispanic: 25.0–29.9 278 0.95 (0.74–1.21)
African American cases: 1119 ≥ 30 312 0.94 (0.74–1.20)
1995–1999 1462 [Ptrend] [0.64]
Non-Hispanic White African American:
cases: 1995–1999 543
598
Non-Hispanic White:
596
646
Population; controls randomly
selected and frequency-matched
by race/ethnicity and expected
5-yr age distribution of cases
Noh et al. (2013) 270 BMI 106 Number of live births, family history of
Republic of Korea 540 < 25 69 1.00 BC, age at menarche, smoking, alcohol
1995–2011 Population; women attending ≥ 25 37 2.24 (1.22–4.10) consumption, physical activity, use of
routine health examination, with HRT
no evidence of malignant disease;
matched by age, menopausal
status, and time of visit to Health
Promotion Center

Sangrajrang et al. (2013) 1126 Current BMI
Thailand 1135 < 18.5 27 1.94 (0.98–3.85)
May 2002–March 2004; Hospital/population; visitors 18.5–24.9 248 1.00
August 2005–August 2006 of hospital patients admitted ≥ 25.0 203 1.67 (1.24–2.25)
for conditions other than BC or
ovarian cancer
Singh & Jangra (2013) 128 aged 20–80 yr BMI [No CIs provided]
India 128 < 18.5 4 0.217
August 2009–July 2010 Hospital; enrolled from the 18.5–23.0 34 1.00
general surgical ward, without 23.0–25.0 14 1.647
history of any type of cancer, 25.0–30.0 21 1.647
matched to cases within 2-yr age > 30.0 6 2.118
interval [Ptrend] [0.016]
253
254


Source of controls
Troisi et al. (2013) Women aged < 85 yr Pre-pregnancy BMI (after 1992) Age at delivery, race/ethnicity, parity at
USA 22 646 with primary in situ or Aged ≥ 50 yr at diagnosis: 144 index birth, year of index birth
1974–2009 invasive cancer < 18.5 3 0.62 (0.19–2.06)
224 721 18.5– < 25 105 1.00
Population; frequency-matched 25– < 30 19 0.60 (0.36–1.01)
to cases by parity, age, calendar ≥ 30 17 0.84 (0.48–1.46)
year of delivery, and race/ [Ptrend] [0.33]
ethnicity
Amadou et al. (2014) 1000 BMI Age, health-care system, region, SES,
Mexico (Mexico City, 1074 < 25 89 1.00 breastfeeding, family history of BC,
Monterrey, Veracruz) Population 25–29.0 239 0.96 (0.64–1.44) alcohol consumption, physical activity,
2004–2007 ≥ 30 257 0.75 (0.51–1.12) total energy intake, height, current BMI
[Ptrend] [0.068]
Elkum et al. (2014) Arab women BMI None
Saudi Arabia 534 18.5–24.9 60 1.00
2007–2012 638 25–29.9 70 1.25 (0.73–2.15)
Population; unmatched, ≥ 30 137 1.66 (1.02–2.70)
randomly selected from primary BMI Age, BMI, marital status, HRT use, age
health care visitors; free of BC 18.5–24.9 1.00 at menarche, breastfeeding, education
≥ 25 2.22 (1.32–3.72) level
Minatoya et al. (2014) 66 BMI Age at menarche, smoking, alcohol
Japan 66 < 19.1 4 0.28 (0.07–1.11) consumption, parity, OC/HRT use
September 2012–July 2013 Hospital; hospitalized for CVD, ≥ 19.1– < 22.3 15 1.00 Ptrend based on χ2 test of log-transformed
hypertension, arrhythmia, ≥ 22.5 25 1.39 (0.50–3.86) continuous variables
nephritis, nephrosis; no BC or [Ptrend] [0.043]
diabetes; matched by age ± 3 yr
and menopausal status
Trentham-Dietz et al. Women aged < 75 yr BMI 16 517 Age, state of residence, study period,
(2014) 23 959 < 18.5 0.75 (0.64–0.88) family history of BC, alcohol
USA 28 304 18.5–24.9 1.00 consumption, age at menarche, parity,
Pooled analysis of 5 case– Population 25–29.9 1.11 (1.06–1.17) age at first pregnancy, OC use, smoking
control studies ≥ 30 1.32 (1.24–1.40) status
1988–2008
BC, breast cancer; BMI, body mass index (in kg/m 2); CI, confidence interval; CVD, cardiovascular diseases; FFTP, first full-term pregnancy; FTP, full-term pregnancy; HRT, hormone
replacement therapy; MET, metabolic equivalent; mo, month or months; NR, not reported; OC, oral contraceptive; OR, odds ratio; SES, socioeconomic status; yr, year or years
a In this table, the study population describes the population of the entire study, and the numbers of cases and controls refer to the number of women in the study, not necessarily the
number of postmenopausal women.

Table 2.2.9g Case–control studies of body mass index and cancer of the breast in postmenopausal women, by hormone
receptor status

Study location Total number of cases cases (95% CI)
Source of controls
Enger et al. 760 BMI Age at reference year, SES, number of
(2000) 1091 ER+PR+: FTPs, months of breastfeeding, age at
USA Population; matched by < 21.7 71 1.00 menopause, HRT use, family history
1997–1989 age, race (Hispanic/non- 21.7–23.6 101 1.36 (0.96–1.94) of BC, alcohol consumption, physical
Hispanic), parity, and 23.7–27.0 127 1.78 (1.26–2.51) activity
residential neighbourhood ≥ 27.1 151 2.45 (1.73–3.47) Results available for BMI at age 18 yr
[Ptrend] [0.0001]
ER+PR−:
< 21.7 34 1.00
21.7–23.6 38 1.12 (0.68–1.85)
23.7–27.0 46 1.35 (0.83–2.20)
≥ 27.1 41 1.29 (0.78–2.15)
[Ptrend] [0.24]
ER−PR−:
< 21.7 31 1.00
21.7–23.6 36 1.19 (0.71–1.99)
23.7–27.0 25 0.80 (0.45–1.40)
≥ 27.1 35 1.20 (0.70–2.05)
[Ptrend] [0.85]
Huang et al. 862 BMI Age at selection, race, age at menarche,
(2000) 790 ER+PR+: 213 nulliparity/age at FFTP, breastfeeding,
USA Population < 23 1.0 abortion or miscarriage, WHR, OC use,

1993–1996 23–31 1.1 (0.7–1.8) HRT use, first-degree family history
> 31 1.6 (0.9–3.0) of BC, medical radiation to the chest,
ER−PR−: 111 cigarette smoking, alcohol consumption,
< 23 1.0 education level, and the offset term
23–31 1.0 (0.6–1.9)
> 31 0.8 (0.4–1.7)
Yoo et al. (2001) Women aged ≥ 25 yr BMI
Japan 1154, no previous history per 1 kg/m2
1988–1992 of cancer ER+ 1.09 (1.05–1.13)
21 714 ER− 1.05 (0.99–1.12)
Hospital PR+ 1.09 (1.04–1.14)
PR− 1.07 (1.02–1.11)
255
256

Table 2.2.9g (continued)

Source of controls
Cotterchio et al. 1867 BMI P heterogeneity = 0.29
(2003) 2452 ER+PR+:
Canada Population; frequency- < 20 45 0.72 (0.43–1.21)
ECSS study: April matched to cases within 20–25 489 1.00
1995–March 1996 5-yr age groups 25.1–27 208 1.10 (0.84–1.43)
WHS study: July > 27 631 1.61 (1.32–1.98)
1996–September ER−PR−:
1998 < 20 25 1.34 (0.72–2.49)
20–25 172 1.00
25.1–27 72 1.09 (0.74–1.61)
> 27 190 1.48 (1.09–1.99)
Rusiecki et al. Women aged 40–80 yr BMI
(2005) 420, no prior BC or benign ER+PR+: 104 Age, age at menarche, parity (nulliparous
USA breast disease < 25.0 1.0 treated separately), age at FFTP, lifetime
January 1994– 406 25.0–29.99 0.7 (0.4–1.2) lactation, ever use of estrogen, alcohol
December 1997 Hospital; no BC or benign ≥ 30.0 1.0 (0.6–1.9) consumption, smoking, family history of
breast disease or incident ER+PR−: 65 BC, race
fibroadenoma, or atypical < 25.0 1.0
hyperplasia, no previous 25.0–29.99 0.7 (0.4–1.4)
cancer disease except for ≥ 30.0 0.8 (0.4–1.8)
non-melanoma of the
ER−PR+: 41
skin, frequency-matched
< 25.0 1.0
by age within 5-yr
25.0–29.99 1.3 (0.6–2.8)
intervals
≥ 30.0 0.9 (0.3–2.3)
ER−PR−: 107
< 25.0 1.0
25.0–29.99 1.3 (0.7–2.1)
≥ 30.0 1.3 (0.7–2.3)
ER+PR+ vs ER−PR−: 104 P for four categories (ER+PR+, ER+PR−,
< 25.0 1.0 ER−PR+, ER−PR−), 0.54
25.0–29.99 0.7 (0.3–1.4)
≥ 30.0 0.7 (0.3–1.4)

Source of controls
Tsakountakis 384 women with primary BMI > 29 vs ≤ 29: Age, residence, menopausal age, OC
et al. (2005) invasive BC HER2/neu+ 4.83 (2.75–8.49) use, HRT use, first-degree family history
Greece 566 HER2/neu− 2.67 (1.56–4.55) of BC, age at FFTP, parity, abortion,
1996–2002 Hospital; women referred Ratio HER2/neu+ to 2.23 (1.20–4.15) lactation, medication to suppress
for breast screening and HER2/neu− lactation, radiation to the chest, BMI,
who did not develop ER+ cases: 180 total benign breast disease
cancer
HER2/neu+ 5.59 (2.58–12.13)
HER2/neu− 2.48 (1.52–5.32)
Ratio HER2/neu+ to NS
HER2/neu−
ER− cases: 197 total
HER2/neu+ 5.33 (2.59–10.94)
HER2/neu− 2.41 (1.15–5.04)
Ratio HER2/neu+ to 2.46 (0.97–6.21)
HER2/neu−
Li et al. (2006) 975 BMI, 65–79 yr Age at diagnosis, reference year, type of
USA 1007 ER+PR+: 615 menopause
1997–1999 Population ≤ 24.9 218 1.0
25.0–29.9 223 1.3 (1.0–1.6)
≥ 30.0 174 1.3 (1.0–1.7)
ER+PR−: 139
≤ 24.9 55 1.0

25.0–29.9 48 1.1 (0.7–1.7)
≥ 30.0 36 1.1 (0.7–1.7)
ER−PR−: 95
≤ 24.9 38 1.0
25.0–29.9 35 1.1 (0.7–1.8)
≥ 30.0 22 0.9 (0.5–1.6)
Rosenberg et al. Women aged 50–74 yr Recent BMI P for ER+PR+ vs ER−PR−, 0.48
(2006) 2643 ER+PR+: Exclusion: women who are ever-users of
Sweden 3065 < 22.2 105 1.0 HRT
October 1993– Population: 22.2–24.0 128 1.3 (1.0–1.7) Adjusted for age, age at first birth
March 1995 frequency-matched to 24.1–25.8 135 1.3 (1.0–1.8)
cases, with no history of 25.9–28.2 176 1.7 (1.3–2.3)
invasive cancer other than ≥ 28.3 228 2.2 (1.7–2.8)
257
non-melanoma of the skin

258


Source of controls
Rosenberg et al. ER+PR−:
(2006) < 22.2 45 1.0
(cont.) 22.2–24.0 35 0.8 (0.5–1.3)
24.1–25.8 40 0.9 (0.6–1.5)
25.9–28.2 37 0.9 (0.5–1.3)
≥ 28.3 45 1.0 (0.7–1.6)
ER−PR+:
< 22.2 7 1.0
22.2–24.0 2 0.3 (0.1–1.5)
24.1–25.8 11 1.7 (0.7–4.5)
25.9–28.2 7 1.1 (0.4–3.1)
≥ 28.3 14 2.2 (0.9–5.6)
ER−PR−:
< 22.2 35 1.0
22.2–24.0 41 1.3 (0.8–2.0)
24.1–25.8 45 1.4 (0.9–2.2)
25.9–28.2 50 1.5 (0.9–2.3)
≥ 28.3 55 1.6 (1.0–2.5)
Phipps et al. 1233 (ductal only), aged BMI Age, reference year
(2008) 65–79 yr at diagnosis HER2-overexpressing cases:
USA (study 1), and 55–74 yr at < 25.0 15 1.0
Study 1: April diagnosis (study 2) 25.0–29.0 11 0.8 (0.4–1.8)
1997–May 1999 (study 1: 975; study 2: ≥ 30.0 13 1.1 (0.5–2.4)
Study 2: January 1044) [Ptrend] [0.78]
2000–March 1447
Triple-negative cases:
2004 (study 1: 1007; study 2:
469) < 25.0 24 1.0
Population; from 25.0–29.0 26 1.2 (0.7–2.1)
Health Care Financing ≥ 30.0 27 1.4 (0.8–2.5)
Administration records, [Ptrend] [0.26]
frequency-matched to BMI at age 30 yr
cases by age HER2-overexpressing cases:
< 20.8 11 1.0
20.8–22.3 9 0.8 (0.3–2.0)
22.4–24.3 6 0.6 (0.2–1.6)
> 24.3 13 1.2 (0.5–2.8)
[Ptrend] [0.74]

Source of controls
Phipps et al. Triple-negative cases:
(2008) < 20.8 21 1.0
(cont.) 20.8–22.3 18 0.9 (0.5–1.7)
22.4–24.3 11 0.6 (0.3–1.2)
> 24.3 27 1.4 (0.8–2.5)
[Ptrend] [0.41]
Dey et al. (2009) 431 BMI Age, religion, education level, SES, age
South India 387 ER+: 170 at menarche, parity, age at marriage,
2002–2005 Population; visitors of ≤ 21.4 1.00 total duration of breastfeeding, physical
non-BC patients, matched 21.4–25.1 1.72 (1.04–2.84) activity per day
to cases by age (5-yr > 25.1 1.34 (0.81–2.23)
groups) and residence type [Ptrend] [0.32]
(urban/rural) ER−: 261
≤ 21.4 1.00
21.4–25.1 1.35 (0.88–2.07)
> 25.1 1.51 (0.98–2.30)
[Ptrend] [0.07]
Bao et al. (2011) 1045 BMI Age, education, history of breast
China 1508 ER+PR+: fibroadenoma, first-degree family
Phase I: 1996– Population; randomly < 21.00 54 1.00 history of BC, regular exercise, years of
1998, Phase II: selected, Shanghai 21.00–23.02 100 1.59 (1.09–2.33) menstruation, history of live birth, parity,
2002–2005 Resident Registry; 23.03–25.15 152 1.93 (1.34–2.79) study phase (I or II)
frequency-matched by ≥ 25.16 215 2.40 (1.65–3.47)

5-yr age groups [Ptrend] [< 0.01]
ER+PR+: 522 ER−PR−:
ER−PR−: 299 < 21.00 46 1.00
21.00–23.02 67 1.10 (0.72–1.68)
23.03–25.15 87 1.06 (0.70–1.60)
≥ 25.16 99 1.00 (0.66–1.53)
[Ptrend] [0.88]
259
260


Source of controls
Barnes et al. 3074 BMI Family history of BC, benign breast
(2011) 6386 ER+PR+: disease, age at menarche, duration of OC
Germany Population; frequency- ≤ 22.4 831 1.00 use, duration of breastfeeding, parity,
2001–2005 matched by year of birth 22.5–24.9 653 1.15 (1.02–1.30) cause of menopause, age at menopause,
and study region 25–29.9 402 1.13 (0.97–1.31) alcohol consumption, HRT use, recent
≥ 30 70 1.06 (0.80–1.42) physical activity, occupational status,
ER+PR−: year of birth, study region, lifetime
≤ 22.4 226 1.00 number of mammograms
22.5–24.9 152 0.96 (0.78–1.20)
25–29.9 90 0.90 (0.69–1.18)
≥ 30 12 0.63 (0.34–1.16)
ER−PR−:
≤ 22.4 252 1.00
22.5–24.9 156 0.87 (0.70–1.07)
25–29.9 110 0.92 (0.72–1.18)
≥ 30 22 0.94 (0.59–1.50)
Dogan et al. 250 BMI, mean Mostly postmenopausal women, but not
(2011) 250 ER+ 1.144 (1.063–1.746) clearly stated
Turkey Hospital PR+ 1.053 (1.095–1.756)
NR Luminal 1.245 (1.023–1.456)
Gaudet et al. 890 BMI treated as ordinal variable Age at diagnosis, age at menarche,
(2011) 3432 Underweight, < 18.5 nulliparity, age at first birth per 5-yr
USA Population; frequency- Normal weight, 18.5– < 25.0 interval, duration of breastfeeding, ever
December 1980– matched, aged ≤ 56 yr Overweight, 25.0– < 30.0 use of OC, benign breast disease, family
December 1982 Obese, ≥ 30.0 history of BC
Luminal A (n = 455) 151 1.16 (0.87–1.54) P for subtype vs luminal A:
Luminal B (n = 72) 18 0.83 (0.36–1.93) 0.58
HER2/neu+ (n = 117) 57 0.93 (0.57–1.52) 0.53
Triple-negative (n = 246) 86 1.02 (0.70–1.48) 0.72
Bandera et al. Postmenopausal women of Current BMI Age, ethnicity, country of origin,
(2013b) African ancestry ER+PR+: education level, family history of BC,
USA 978 < 25 26 1.00 history of benign breast disease, age at
NR 958 25–29.99 49 1.05 (0.55–1.98) menarche, parity, breastfeeding, age at
Population; random-digit ≥ 30 131 1.04 (0.50–2.18) first birth, HRT use, OC use
dialling [Ptrend] [0.95]

Source of controls
Bandera et al. ER−PR−:
(2013b) < 25 20 1.00
(cont.) 25–29.99 34 0.77 (0.38–1.59)
≥ 30 47 0.37 (0.15–0.96)
[Ptrend] [0.03]
John et al. (2013) 1389 of 2571 Current BMI All non-users of HRT
USA 1644 of 2706 ER+PR+: Results available for Hispanic, African
Hispanic cases: Hispanic: < 25.0 98 1.00 American, and non-Hispanic White
1995–2002 1119 25.0–29.9 141 1.09 (0.80–1.49) women separately
African 1462 ≥ 30 175 1.30 (0.95–1.78)
American cases: African American: [Ptrend] [0.09]
1995–1999 543 ER−PR−:
Non-Hispanic 598 < 25.0 34 1.00
White cases: Non-Hispanic White: 25.0–29.9 46 0.75 (0.46–1.22)
1995–1999 596 ≥ 30 54 0.72 (0.45–1.16)
646 [Ptrend] [0.21]
Population; controls
BMI in young adulthood All non-users of HRT
randomly selected and
ER+PR+: Results available for Hispanic, African
frequency-matched
T1: ≤ 21.2 147 1.00 American, and non-Hispanic White
by race/ethnicity and
T2: 21.3–23.7 133 0.87 (0.65–1.15) women separately
expected 5-yr age
distribution of cases T3: > 23.7 116 0.73 (0.54–0.98)
[Ptrend] [0.04]
ER−PR−:

T1: ≤ 21.2 46 1.00
T2: 21.3–23.7 43 0.82 (0.52–1.29)
T3: > 23.7 37 0.61 (0.38–0.97)
[Ptrend] [0.04]
Kawai et al. 1017 BMI Age, smoking, alcohol consumption,
(2013) 2902 ER+PR+: 277 family history of BC, occupation, age at
Japan Hospital; female non- < 18.5 10 1.00 menarche, age at first birth, parity, use
1997–2009 cancer patients = benign 18.5–22.1 54 0.88 (0.41–1.87) of exogenous female hormones or OC,
tumours, cardiovascular 22.1–25.0 84 1.29 (0.61–2.72) year of recruitment, area, referral base
diseases, digestive tract 25.0–30.0 95 1.72 (0.82–3.60) (screening, other), height, time spent
diseases, respiratory ≥ 30.0 34 6.24 (2.68–14.53) exercising
tract disease, urological– [Ptrend] [< 0.0001]
gynaecological disease
261
262


Source of controls
Kawai et al. ER−PR−: 142 P heterogeneity = 0.0002
(2013) < 18.5 5 1.00
(cont.) 18.5–22.1 45 1.49 (0.56–3.96)
22.1–25.0 47 1.43 (0.53–3.80)
25.0–30.0 36 1.19 (0.44–3.21)
≥ 30.0 9 2.43 (0.74–7.95)
[Ptrend] [0.86]
BC, breast cancer; BMI, body mass index (in kg/m 2); CI, confidence interval; ER, estrogen receptor; FFTP, first full-term pregnancy; HER2, human epidermal growth factor receptor 2;
HRT, hormone replacement therapy; NR, not reported; NS, not significant; OC, oral contraceptive; PR, progesterone receptor; SES, socioeconomic status; WHR, waist-to-hip ratio; yr,
year or years

Table 2.2.9i Case–control studies of body mass index and cancer of the breast in postmenopausal women, by ethnicity

Source of controls
Wenten et al. (2002) Women aged 30–70 yr Usual BMI NR Age, first-degree family history of BC,
USA 712 diagnosed with invasive or Hispanic: total METs, parity, OC use, months
January 1992–December in situ breast cancer < 22 1.00 of breastfeeding, age at first full-term
1994 1039 22– < 25 1.53 (0.67–3.50) birth, HRT use, weight at age 18 yr
Hispanic: 25– < 30 1.60 (0.67–3.82) Results also reported for BMI at age
332 ≥ 30 1.32 (0.47–3.72) 18 yr
511 [Ptrend] [0.58]
Non-Hispanic White:
Non-Hispanic White:
380
528 < 22 1.00
Population 22– < 25 0.90 (0.51–1.61)
25– < 30 1.15 (0.53–2.47)
≥ 30 2.77 (0.86–8.89)
[Ptrend] [0.16]
Ziv et al. (2006) Hispanic/Latina women BMI Age, case–control status, grandparents’
USA 357 diagnosed 1997–1999 All Latinas: place of birth, age at migration,
1995–2002 479 < 25 48 1.00 education level, place of birth (born in
Completed interview: 25–29.9 71 1.93 (1.38–2.69) USA vs foreign-born)
324 ≥ 30 115 1.51 (1.12–2.04)
421
Latinas born in USA: 106 total Age, case–control status, grandparents’
Provided blood sample:
< 25 1.00 place of birth, education level
241
333 25–29.9 1.25 (0.79–1.96)
Population; matched to cases ≥ 30 1.26 (0.83–1.92)

by ethnicity and 5-yr age Foreign-born Latinas: 128 total Age, case–control status, grandparents’
groups < 25 1.00 place of birth, age at migration,
25–29.9 3.44 (1.97–5.99) education level
≥ 30 1.95 (1.24–3.06)
263
264

Table 2.2.9i (continued)

Source of controls
Slattery et al. (2007) Hispanic women living in non- BMI in reference year, no recent hormone exposure Age, height, physical activity, energy
USA reservations and non-Hispanic Non-Hispanic White: intake, parity, alcohol consumption,
1999–2004 White women < 25 146 1.00 age at first pregnancy, age at
2325 2525 25–29.9 122 1.60 (1.06–2.40) menopause, centre
Non-Hispanic White: ≥ 30 112 1.61 (1.05–2.45) Analyses of BMI at age 18 yr also
1527 [Ptrend] [0.03] reported
1601 Hispanic:
Hispanic: < 25 43 1.00
798 25–29.9 91 0.68 (0.38–1.24)
924 ≥ 30 104 0.80 (0.44–1.45)
Population; matched by [Ptrend] [0.61]
ethnicity, age in 5-yr classes,
BMI in reference year, recent hormone exposure
random selection
Non-Hispanic White:
< 25 306 1.00
25–29.9 194 1.02 (0.79–1.32)
≥ 30 202 0.72 (0.54–0.96)
[Ptrend] [0.04]
Hispanic:
< 25 92 1.00
25–29.9 120 0.91 (0.60–1.38)
≥ 30 114 0.74 (0.47–1.15)
[Ptrend] [0.17]
Berstad et al. (2010) 4575 BMI at age 18 yr Age, race, education level, study site,
USA: Atlanta (Georgia), 4682 Caucasian: 1261 first-degree family history of BC,
Seattle (Washington), Caucasian: < 20 682 1.00 parity, age at menopause, HRT use,
Detroit (Michigan), 2953 20–24 517 0.89 (0.75–1.05) BMI at the other time point
Philadelphia 3021 ≥ 25 62 0.70 (0.49–1.00) Results available by hormonal status
(Pennsylvania), Los African American: [Ptrend] [0.03] for BMI by age 18 yr and 5 yr before
Angeles (California); 1622 reference date, for each ethnic group
African American: 639
July 1994–April 1998 1661
Population < 20 297 1.00
20–24 286 0.91 (0.72–1.14)
≥ 25 56 0.80 (0.54–1.19)
[Ptrend] [0.22]

Source of controls
Berstad et al. (2010) BMI 5 yr before reference date
(cont.) Caucasian:
< 25 733 1.00
25–29 333 0.93 (0.77–1.12)
30–34 127 0.94 (0.72–1.24)
≥ 35 68 0.75 (0.53–1.06)
[Ptrend] [0.13]
African American:
< 25 185 1.00
25–29 246 1.05 (0.80–1.37)
30–34 127 0.98 (0.71–1.35)
≥ 35 81 1.26 (0.85–1.85)
[Ptrend] [0.44]
Bandera et al. (2013a) Postmenopausal women BMI at age 20 yr Age, ethnicity (Hispanic/non-
USA of African and Caucasian African American: Hispanic), country of origin, family
New York City: 2002–2008 ancestry < 25 392 1.00 history of BC, history of benign breast
New Jersey: 2006–2012 1751 25–29.9 52 1.01 (0.65–1.58) disease, age at menarche, parity,
1673 ≥ 30 17 0.88 (0.43–1.81) breastfeeding status, age at first birth,
African American: [Ptrend] [0.82] HRT use, OC use, height and weight at
979 menarche
European American:
958
European American: < 25 342 1.00
772 25–29.9 17 0.82 (0.38–1.77)

715 ≥ 30 4 0.15 (0.04–0.60)
Population [Ptrend] [0.01]
265
266


Source of controls
John et al. (2013) 1389 of 2571 Current BMI Non-users of HRT
USA 1644 of 2706 Hispanic: Results available for ER+PR+ tumours
Hispanic cases: 1995–2002 Hispanic: < 25.0 81 1.00 (for both current BMI and BMI in
African American cases: 1119 25.0–29.9 133 0.78 (0.54–1.14) young adulthood, separated by race)
1995–1999 1462 ≥ 30 161 0.77 (0.53–1.12)
Non-Hispanic White African American: [Ptrend] [0.24]
cases: 1995–1999 543
African American:
598
Non-Hispanic White: < 25.0 51 1.00
596 25.0–29.9 90 1.19 (0.74–1.94)
646 ≥ 30 101 1.07 (0.66–1.73)
Population; controls randomly [Ptrend] [0.88]
selected and frequency- Non-Hispanic White:
matched by race/ethnicity and < 25.0 76 1.00
expected 5-yr age distribution 25.0–29.9 55 0.90 (0.56–1.43)
of cases ≥ 30 50 1.19 (0.72–1.99)
[Ptrend] [0.58]
BMI in young adulthood
Hispanic:
T1: ≤ 21.2 109 1.00
T2: 21.3–23.7 122 0.85 (0.60–1.20)
T3: > 23.7 115 0.63 (0.45–0.90)
[Ptrend] [0.01]
African American:
T1: ≤ 21.2 93 1.00
T2: 21.3–23.7 77 1.17 (0.76–1.79)
T3: > 23.7 67 0.93 (0.59–1.45)
[Ptrend] [0.80]
Non-Hispanic White:
T1: ≤ 21.2 84 1.00
T2: 21.3–23.7 60 0.65 (0.41–1.02)
T3: > 23.7 34 0.52 (0.30–0.90)
[Ptrend] [0.01]

Source of controls
Robinson et al. (2014) Women aged 20–74 yr Measured BMI Age, age squared, family history of BC,
USA 1783 Black: alcohol consumption, menarche, parity,
1993–2001 1536 < 25 74 1.00 age at FFTP composite, lactation,
Black: 25–30 121 0.61 (0.38–0.98) education level, smoking
788 30–35 118 0.77 (0.47–1.28) Data also reported for BMI at age 18
718 ≥ 35 113 0.58 (0.35–0.94) yr, 35 yr, and one yr before interview,
White: [Ptrend] [0.11] by ethnicity; all of these associations
995 White: were null
818 < 25 212 1.00
Population; frequency-matched 25–30 165 0.91 (0.67–1.25)
to cases by 5-yr age group 30–35 69 0.83 (0.55–1.25)
≥ 35 30 0.61 (0.35–1.06)
[Ptrend] [0.08]
John et al. (2015b) 4271 ER+PR+:
USA 4713 Current BMI Age, study, ethnicity/English language
2 population-based case– Population Hispanic: 294 acculturation, education level, first-
control studies degree family history of BC, age at
per 5 kg/m2 0.81 (0.65–1.01)
San Francisco Bay Area menarche, number of FTPs, age at
Study FFTP, lifetime months of breastfeeding,
4-Corners Breast Cancer average alcohol consumption
Study Non-Hispanic White: 292 Age, study, ethnicity, education level,
Hispanic: 1995–2002 first-degree family history of BC, age
per 5 kg/m2 0.94 (0.74–1.19)
Non-Hispanic White: at menarche, number of FTPs, age at

1995–2004 FFTP, lifetime months of breastfeeding,
average alcohol consumption
ER−PR−:
Current BMI Age, study, ethnicity/English language
Hispanic: 153 acculturation, first-degree family
per 5 kg/m2 0.76 (0.57–1.01) history of BC, age at menarche, HRT
use
Non-Hispanic White: Age, study, ethnicity, first-degree
per 5 kg/m2 0.63 (0.43–0.92) family history of BC, age at menarche,
HRT use
267
268


Source of controls
Sanderson et al. (2015) Women aged 25–75 yr BMI Age, education level, first-degree family
USA 2614 with primary ductal Black: history of BC, OC use, age at menarche
2001–2011 carcinoma in situ or invasive < 25.0 75 1.0 Pinteraction = 0.43
breast cancer 25.0–29.9 129 1.0 (0.6–1.7)
2306 30.0–34.9 123 1.2 (0.7–2.0)
Population; matched by 5-yr ≥ 35 113 1.0 (0.6–1.7)
age groups, race, and county of [Ptrend] [0.90]
residence White:
< 25.0 493 1.0
25.0–29.9 433 1.1 (0.9–1.3)
30.0–34.9 223 1.1 (0.9–1.4)
≥ 35 121 0.8 (0.6–1.1)
[Ptrend] [0.67]
BC, breast cancer; BMI, body mass index (in kg/m 2); CI, confidence interval; ER, estrogen receptor; HRT, hormone replacement therapy; MET, metabolic equivalent; NR, not reported;
OC, oral contraceptive; PR, progesterone receptor; yr, year or years

Table 2.2.9k Case–control studies of waist circumference and cancer of the breast in postmenopausal women
Reference, study Study populationa Exposure Exposed cases Relative risk Covariates
location and period Total number of cases categories (95% CI)
Total number of controls (cm, unless
Source of controls otherwise stated)
Friedenreich et al. 771 < 75.6 1533 1.00 Current age, total energy intake,
(2002) 762 ≥ 75.6– < 82.8 175 0.89 (0.66–1.20) total lifetime physical activity,
Canada Population-based using Waksberg ≥ 82.8– < 91.5 159 1.06 (0.79–1.42) education level, ever use of HRT,
1995–1997 method; frequency-matched to cases ≥ 91.5 187 1.30 (0.97–1.73) ever diagnosed with benign breast
by age, 5-yr intervals, and place of [Ptrend] 242 [0.07] disease, first-degree family history
residence (urban/rural) of BC, ever alcohol consumption,
current smoking
Slattery et al. (2007) Hispanic women living in non- WC (in), no recent hormone exposure Age, height, physical activity, energy
USA reservations and non-Hispanic White Non-Hispanic White: intake, parity, alcohol consumption,
1999–2004 women < 35 197 1.00 age at first pregnancy, age at
Non-Hispanic White: 35–40 95 1.73 (1.16–2.58) menopause, centre
858 > 40 83 1.29 (0.83–1.99)
1008 [Ptrend] [0.11]
Hispanic:
Hispanic:
399
522 < 35 80 1.00
Population; matched by ethnicity, age 35–40 83 0.98 (0.59–1.63)
in 5-yr classes, random selection > 40 71 0.81 (1.47–1.39)
[Ptrend] [0.45]
WC (in), recent hormone exposure
Non-Hispanic White:
< 35 393 1.00
35–40 180 0.99 (0.76–1.28)
> 40 115 0.88 (0.64–1.21)

[Ptrend] [0.48]
Hispanic:
< 35 148 1.00
35–40 108 1.18 (0.80–1.75)
> 40 65 0.86 (0.53–1.38)
[Ptrend] [0.74]
269
270

Table 2.2.9k (continued)
Tian et al. (2007) 102 aged 22–87 yr ≤ 81.00 54 1.00 Age at enrolment, fasting status,
Taiwan 103 > 81.00 48 2.02 (1.05–3.91) levels of adiponectin
2004–2005 Hospital; recruited from health
examination clinics at the same
hospital and time, free for cancer
history, matched by menopausal
status, date of enrolment, duration of
fasting
Mathew et al. (2008) 968 ≤ 85 57 1.00 Age, centre, religion, marital status,
India 691 > 85 380 1.61 (1.22–2.12) education level, SES, residence status,
2002–2005 Accompanying persons to cancer Unknown 31 2.88 (0.76– parity, age at first birth, duration of
cases; matched by age ± 5 yr and 10.90) breastfeeding, physical activity
residence type (urban/rural)
Nemesure et al. (2009) Women of African descent aged ≥ 21 yr Aged ≥ 50 yr: Current age, HRT use, parity, family
Barbados 222 < 80 18 1.00 history of BC, history of benign
2002–2006 454 80–101 88 1.35 (0.57–3.18) breast disease, age at first pregnancy,
Population; Barbados Statistical ≥ 101 38 2.98 (0.91–9.71) age at menarche, physical activity,
Services; frequency-matched by 5-yr other body size variable
age group
Rosato et al. (2011) Postmenopausal women < 88 869 1.00 Age, study centre, study
Italy, Switzerland 1747 ≥ 88 878 1.17 (1.02–1.35) period, education level, alcohol
1983–1994 1935 consumption, age at menarche, age
(Italy), 1991–2007 Hospital; admitted for acute, non- at first birth, age at menopause, HRT
(Switzerland) neoplastic diseases, not related use, family history of BC
to gynaecological or hormonal
conditions, matched by age and study
centre
Bandera et al. (2013b) Postmenopausal women of African ≤ 87.88 87 1.00 BMI, age, ethnicity, country of
USA ancestry 87.89–97.75 119 1.13 (0.73–1.76) origin, education level, family
NR 978 97.76–110.25 154 1.51 (0.92–2.48) history of BC, history of benign
958 > 110.25 140 1.23 (0.64–2.34) breast disease, age at menarche, age
Population; random-digit dialling [Ptrend] [0.48] at menopause, parity, breastfeeding,
ER+PR+: age at first birth, HRT use, OC use
≤ 87.88 36 1.00
87.89–97.75 39 0.88 (0.48–1.60)
97.76–110.25 56 1.30 (0.68–2.48)
> 110.25 74 1.55 (0.68–3.55)
[Ptrend] [0.20]
ER−PR−:
≤ 87.88 23 1.00
87.89–97.75 25 0.93 (0.45–1.92)
97.76–110.25 25 1.11 (0.48–2.57)
> 110.25 27 1.08 (0.35–3.31)
[Ptrend] [0.83]
John et al. (2013) 1389 postmenopausal women All: All non-users of HRT
USA 1644 ≤ 85.0 198 1.00
1995–2002 Population; controls randomly 85.1–96.4 214 0.99 (0.77–1.27)
selected and frequency-matched by > 96.4 293 1.32 (1.03–1.69)
race/ethnicity and expected 5-yr age [Ptrend] [0.02]
distribution of cases ER+PR+:

≤ 85.0 95 1.00
85.1–96.4 106 1.11 (0.80–1.54)
> 96.4 162 1.76 (1.28–2.41)
[Ptrend] [< 0.01]
ER−PR−:
≤ 85.0 28 1.00
85.1–96.4 40 1.13 (0.67–1.89)
> 96.4 48 1.24 (0.75–2.06)
[Ptrend] [0.41]
271
272

Sangrajrang et al. 470 < 80 199 1.00
(2013) 385 ≥ 80 271 1.18 (0.89–1.57)
Thailand Hospital/population; female visitors
May 2002–March of hospital patients admitted for
2004; August 2005– conditions other than BC or ovarian
August 2006 cancer
Amadou et al. (2014) 585 < 93 187 1.00 Age, health care system, region,
Mexico 598 93–103 218 0.96 (0.70–1.32) SES, breastfeeding, family history of
2004–2007 Population ≥ 103 180 0.62 (0.44–0.85) BC, alcohol consumption, physical
[Ptrend] [0.003] activity, total energy intake, height,
current BMI
Robinson et al. (2014) Women aged 20–74 yr Black: Age, age squared, family history of
USA 911 ≤ 88 113 1.00 BC, alcohol consumption, menarche,
1993–2001 825 > 88 321 1.39 (0.92–2.10) parity, age at FFTP composite,
Black: [Ptrend] [0.11] lactation, education level, smoking,
434 White: reference BMI
380
≤ 88 314 1.00
White:
> 88 163 1.31 (0.88–1.95)
477
[Ptrend] [0.18]
445
Population; frequency-matched to
cases by 5-yr age group
BC, breast cancer; BMI, body mass index (in kg/m 2); CI, confidence interval; ER, estrogen receptor; FFTP, first full-term pregnancy; HRT, hormone replacement therapy; NR, not
reported; OC, oral contraceptive; PR, progesterone receptor; SES, socioeconomic status; WC, waist circumference (in cm); yr, year or years

Table 2.2.9m Case–control studies of change in body mass index or weight and cancer of the breast in postmenopausal
women
Reference Study populationa Exposure categories Exposed cases Relative risk Covariates
Study location Total number of cases (95% CI)
Source of controls
BMI change
Hirose et al. (2001) 1584 BMI change from age 20 yr, without family history of BC Age, age at menarche, menstrual
Japan 15 331 < 0 127 0.69 (0.52–0.92) regularity in the 20s, age at first
1988–1997 First visit outpatients 0–1.24 89 1.00 birth, parity
(screening) without any 1.25–2.99 137 1.02 (0.77–1.40)
previous diagnosis of cancer ≥ 3 238 1.34 (1.00–1.70)
[Ptrend] [< 0.001]
BMI change from age 20 yr, with family history of BC
< 0 9 1.56 (0.44–5.60)
0–1.24 4 1.00
1.25–2.99 13 2.74 (0.82–9.10)
≥ 3 17 2.19 (0.68–7.00)
[Ptrend] [0.26]
Robinson et al. (2014) 1783 women aged 20–74 yr BMI change, ages 18–35 yr Age, age squared, family history
USA 1536 Black: of BC, alcohol consumption,
1993–2001 Black: < 1.77 103 1.0 menarche, parity, age at FFTP
788 1.77–4.44 151 1.47 (0.98–2.18) composite, lactation, education
718 ≥ 4.44 161 1.14 (0.76–1.70) level, smoking, reference BMI
White: [Ptrend] [0.63]
995
White:
818
Population; frequency- < 1.77 194 1.0

matched to cases by 5-yr age 1.77–4.44 172 1.17 (0.85–1.60)
group ≥ 4.44 98 1.33 (0.88–2.02)
[Ptrend] [0.16]
Weight change
Li et al. (2000) 479 Weight change (lb), age 18 yr to reference date, 50–64 yr Age, height, weight at age 18 yr,
USA 435 < −10 14 0.9 (0.4–1.9) family history of BC, parity, HRT
January 1988–June Population; Caucasian −10 to 10 113 1.0 use, OC use
1990 women 11–30 153 1.1 (0.7–1.5)
31–50 100 1.2 (0.8–1.7)
51–70 43 1.3 (0.7–2.1)
> 70 55 2.7 (1.5–4.9)
273
274

Table 2.2.9m (continued)
Source of controls
Trentham-Dietz et al. Postmenopausal women Weight loss (kg), overall Parity, age at FFTP, family
(2000) aged 50–79 yr 0.0 1690 1.0 history of BC, recent alcohol
USA 5031 0.1–4.9 1637 1.1 (1.0–1.2) consumption, education level,
January 1992– 5255 5.0–9.9 809 1.0 (0.9–1.2) age at menopause, height, highest
December 1994 Population; matched by age ≥ 10.0 668 1.0 (0.9–1.2) weight and age at highest weight
and state [Ptrend] [0.1] Analyses of weight loss since age
11–45 yr and since age > 45 yr
gave similar results to weight loss
overall
Weight gain (kg), overall Parity, age at FFTP, family
0–5.0 730 1.0 history of BC, recent alcohol
5.1–10.0 853 1.1 (0.9–1.3) consumption, education level,
10.1–15.0 872 1.1 (1.0–1.3) age at menopause, height, lowest
15.1–25.0 1409 1.4 (1.2–1.6) weight and time since lowest
> 25.0 1008 1.7 (1.5–2.0) weight
[Ptrend] [< 0.001] Analyses of weight gain since age
20, since age 21–30 yr and since
age > 30 yr gave similar results to
weight gain overall
de Vasconcelos et al. 177 Weight change (kg) since age 18 yr Age, parity, age at menarche,
(2001) 377 > 22.3 31 1.00 family history of BC, weight and
Brazil Hospital/population; 13.11–22.3 38 1.39 (0.75–2.59) height at 18 yr
May 1995–February visitors at hospital; 27 0–13.10 28 1.24 (0.62–2.50) Analyses of weight change from
1996 relatives of breast cancer Weight loss 12 2.05 (0.75–5.59) age 18 yr to age 30 yr and weight
patients [Ptrend] [0.24] change since age 30 yr gave similar
results
Source of controls
Shu et al. (2001) Women aged 25–64 yr Weight gain (kg) since age 20 yr Age, education level, family
China 1459 of 1602 < 1.15 20.4% 1.0 history of BC, ever had
August 1996–March 1556 of 1724 1.15–3.41 31.7% 1.4 (1.0–2.1) fibroadenoma, age at menarche,
1998 Population; randomly 3.42–5.64 26.6% 1.3 (0.9–1.9) age at first live birth, exercise, age
selected from female ≥ 5.65 21.3% 2.7 (1.7–4.2) at menopause
residents of Shanghai [Ptrend] [< 0.001]
(Shanghai Resident Weight gain (kg) during past 10 yr
Registry), matched to cases
< 1.15 37.1% 1.0
by age, 5-yr interval
1.15–3.41 19.8% 1.6 (1.1–2.2)
3.42–5.64 14.3% 1.2 (0.8–1.8)
≥ 5.65 28.8% 1.5 (1.1–2.1)
[Ptrend] [0.03]
Friedenreich et al. 1233 Weight gain (kg) since age 20 yr Current age, total energy intake,
(2002) 1241 < 7.80 181 1.00 total lifetime physical activity,
Canada Population-based using ≥ 7.80– < 15.7 173 1.02 (0.75–1.37) education level, ever use of HRT,
1995–1997 Waksberg method; ≥ 15.7– < 25.0 182 1.08 (0.80–1.45) ever diagnosed with benign breast
frequency-matched to cases ≥ 25.0 231 1.35 (1.01–1.81) disease, first-degree family history
by age, 5-yr interval, and [Ptrend] [0.05] of BC, ever alcohol consumption,
place of residence (urban/ Difference, maximum − minimum weight (kg) over adult lifetime current smoking
rural)
< 9.07 161 1.00
≥ 9.07– < 15.4 161 0.94 (0.69–1.28)
≥ 15.4– < 22.7 184 1.21 (0.89–1.64)
≥ 22.7 265 1.56 (1.16–2.08)

[Ptrend] [0.0007]
275
276

Source of controls
Wenten et al. (2002) 712 women aged 30–70 yr Weight change (kg), age 18 yr to usual adult weight Age, first-degree family history
USA diagnosed with invasive or Hispanic: of BC, total METs, parity, OC
January 1992– in situ breast cancer < 4 1.00 use, months of breastfeeding, age
December 1994 1039 4–7 2.48 (0.89–6.93) at first full-term birth, HRT use,
Hispanic: 8–14 2.04 (0.73–5.68) weight at age 18 yr
332 > 14 2.46 (0.98–6.17)
511 [Ptrend] [0.14]
Non-Hispanic White:
Non-Hispanic White:
380
528 < 4 1.00
Population 4–7 1.34 (0.66–2.74)
8–14 1.33 (0.63–2.77)
> 14 2.27 (1.09–4.73)
[Ptrend] [0.04]
Carpenter et al. Caucasian (including Weight change (%), age 18 yr to reference date (1 yr before diagnosis) Age at FFTP, ages at menarche
(2003) Hispanic), born in Canada, Negative change to no change 229 1.00 and menopause, family history
Canada, USA, USA, or western Europe > 0–16.9% 573 1.16 (0.92–1.47) of BC, interviewer, average MET
western Europe 1883 diagnosed at age 17.0–29.1% 404 1.13 (0.88–1.45) hours per week of lifetime exercise
Group I: March 55–64 yr (Group I), age ≥ 29.2% 677 1.36 (1.08–1.73) activity
1987–December 1989 55–69 yr (Group II), or age [Ptrend] [0.01]
Group II: January 55–72 yr (Group III)
1992–December 1992 1628
Group III: September Population; matched to
1995–April 1996 cases by neighbourhood
Eng et al. (2005) 1006 Weight change (kg), age 20 yr to 1 yr before reference date Age at reference date, number of
USA 990 −44.91 to −3.01 36 0.55 (0.32–0.96) pregnancies, months of HRT use,
August 1996–July Population; frequency- −3.00 to 3.00 103 1.00 history of BC in a first-degree
1997 matched by 5-yr age group 3.01–7.71 141 1.03 (0.70–1.50) relative, history of benign breast
7.71–8.15 241 1.18 (0.84–1.74) disease, BMI at age 20 yr
8.16–14.96 209 1.21 (0.84–1.74)
14.97–87.09 256 1.58 (1.11–2.26)
[Ptrend] [0.0001]
Source of controls
Eng et al. (2005) Weight change (kg), age 50 yr to 1 yr before reference date Age at reference date, number of
(cont.) −68.04 to −0.01 157 1.19 (0.85–1.67) pregnancies, months of HRT use,
0.00 167 1.00 history of BC in a first-degree
0.01–2.71 133 1.19 (0.84–1.69) relative, history of benign breast
2.72–4.98 124 0.96 (0.68–1.37) disease, BMI at age 50 yr
4.99–11.33 195 1.58 (1.14–2.23)
11.34–62.14 171 1.62 (1.14–2.30)
[Ptrend] [0.003]
Han et al. (2006) 1166 Weight change (kg), age 20 yr to 1 yr before study enrolment Age, education level, previous
USA 2105 ≤ 0 841 0.90 (0.56–1.45) benign disease, age at menarche,
1996–2001 Population; frequency- 0–9.1 47 1.00 age at first birth, family history of
matched by age, race, and 9.1–17.7 137 1.45 (1.06–1.96) BC, age at menopause, HRT use,
county of residence 17.7–27.3 208 1.53 (1.12–2.08) BMI residuals
> 27.3 227 1.71 (1.23–2.37) Weight change (kg) from age
[Ptrend] 222 [0.05] at first pregnancy to age at
menopause also showed a positive
association with breast cancer risk
(Ptrend = 0.01)
Wu et al. (2006) Asian American women Weight gain (kg) since age 18 yr (recent weight − weight at age 18 yr) Age, ethnicity, duration of
USA 1277 aged 25–74 yr at ≤ 10 319 1.00 residence in the USA, education
1995–2001 diagnosis > 10– ≤ 15 138 1.24 (0.90–1.72) level, age at menarche, number
1160 > 15– ≤ 20 95 1.10 (0.75–1.62) of live births, age at menopause,
Chinese: > 20 95 1.66 (1.09–2.53) intake of tea and soy during

450 [Ptrend] [0.036] adolescence and adult life, years of
486 Weight gain (kg) since age 30 yr (recent weight − weight at age 30 yr) physical activity, height
Japanese:
≤ 10 518 1.00
352
> 10– ≤ 15 91 1.51 (1.02–2.22)
311
> 15– ≤ 20 44 1.17 (0.70–1.96)
Filipino:
> 20 27 2.23 (1.00–4.94)
475
[Ptrend] [0.023]
363
Population; neighbourhood
controls; frequency-matched
by ethnicity and 5-yr age
group
277
278

Source of controls
Slattery et al. (2007) Hispanic women living in Total weight gain (kg) between age 15 yr and reference year Age, height, physical activity,
USA non-reservations and non- No recent hormone exposure energy intake, parity, alcohol
1999–2004 Hispanic White women Non-Hispanic White: consumption, age at first
2325 ≤ 5.0 57 1.00 pregnancy, age at menopause,
2525 5.1–15.0 99 1.19 (0.67–2.09) centre
Non-Hispanic White: 15.1–25.0 94 1.40 (0.79–2.48)
1527 > 25.0 104 1.75 (1.00–3.05)
1601 [Ptrend] [0.03]
Hispanic:
Hispanic:
798
≤ 5.0 22 1.00
924
5.1–15.0 37 1.14 (0.49–2.67)
Population; matched by
15.1–25.0 79 0.70 (0.32–1.52)
ethnicity, age in 5-yr classes,
> 25.0 78 0.76 (0.35–1.65)
random selection
[Ptrend] [0.25]
Recent hormone exposure
Non-Hispanic White:
≤ 5.0 115 1.00
5.1–15.0 176 1.14 (0.80–1.61)
15.1–25.0 182 1.08 (0.77–1.53)
> 25.0 200 0.95 (0.66–1.35)
[Ptrend] [0.57]
Hispanic:
≤ 5.0 25 1.00
5.1–15.0 77 0.73 (0.37–1.43)
15.1–25.0 98 0.79 (0.41–1.51)
> 25.0 108 0.64 (0.34–1.23)
[Ptrend] [0.26]
Shin et al. (2009) 3452 aged 20–64 yr Weight change (kg) since age 20 yr
China (phase 1), 20–70 yr (phase 2) ≤ 0 141 1.0
1996–1998 (phase 1), 3474 0.1–9.4 383 1.3 (1.0–1.6)
April 2002–February Population; controls 9.5–14.9 307 1.5 (1.1–2.0)
2005 (phase 2) frequency-matched to cases ≥ 15 471 1.8 (1.4–2.4)
by age [Ptrend] [< 0.001]
Source of controls
Berstad et al. (2010) 4575 Weight change (kg) since age 18 yr 1900
USA 4682 ≤ 5 363 1.00 Also adjusted for BMI at age 18 yr
July 1994–April 1998 Caucasian: 5.1–15.0 641 1.10 (0.91–1.32)
2953 15.1–25.0 507 1.01 (0.83–1.23)
3021 ≥ 25.1 389 1.03 (0.84–1.27)
African American: [Ptrend] [0.92]
1622
1661
Population
Cribb et al. (2011) 207 Weight gain (kg) since age 25 yr Parity, OC use, BMI, smoking
Canada 621 > 10 61% 1.34 (0.85–2.12)
1999–2002 Population; women
presenting for routine
mammography screening;
matched by age, menopausal
status, and family history
of BC
Sangaramoorthy Women aged 35–79 yr Relative weight vs peers at age 10 yr Analysis of Hispanic women only
et al. (2011) 931 of 1031 Women not currently using HRT 205 Age, country of birth, education
USA 1050 of 1198 Lighter 114 1.00 level, first-degree family history
1998–2002 Hispanic: Same 61 0.84 (0.55–1.29) of BC, prior biopsy history of
650 Heavier 23 0.68 (0.37–1.25) benign breast disease, number
766 [Ptrend] [0.19] of FTPs, age at FFTP, lifetime

African American: breastfeeding, OC use, adult
134 height, alcohol consumption,
137 average energy intake, BMI
Non-Hispanic White: Measures of relative weight vs
147 peers at 15 yr and 20 yr gave
147 similar results
Population; frequency-
matched by race and age in
5-yr groups, without history
of BC
279
280

Source of controls
Bandera et al. (2013b) Postmenopausal women Weight gain (kg) since age 20 yr, quartiles Age, ethnicity (Hispanic/non-
USA of African and European African American: Hispanic), country of origin,
New York City: ancestry Q1: ≤ 13.82 75 1.00 family history of BC, history
2002–2008 1751 Q2: 13.83–23.72 115 1.35 (0.87–2.10) of benign breast disease, age at
New Jersey: 1673 Q3: 23.73–34.56 110 1.29 (0.80–2.09) menarche, age at menopause,
2006–2012 African American: Q4: > 34.56 139 1.42 (0.80–2.53) parity, breastfeeding status, age
979 [Ptrend] [0.34] at first birth, HRT use, OC use,
958 current BMI
European American:
European American:
772 Q1: ≤ 7.57 75 1.00
715 Q2: 7.58–14.57 77 0.97 (0.56–1.66)
Population Q3: 14.58–24.52 91 0.90 (0.52–1.57)
Q4: > 24.52 90 0.95 (0.46–1.95)
[Ptrend] [0.88]
John et al. (2013) 1389 of 2571 Weight gain (kg) from 20s, all non-users of HRT Subanalysis by race/ethnicity
USA 1644 of 2706 Stable 78 1.00 showed a positive association in
Hispanic cases: Hispanic: 3.0–9.9 180 1.15 (0.82–1.63) White non-Hispanic women only
1995–2002 1119 10.0–19.9 217 1.06 (0.76–1.48)
African American 1462 20.0–29.9 142 1.03 (0.72–1.48)
cases: 1995–1999 African American: ≥ 30.0 111 1.19 (0.81–1.75)
Non-Hispanic White 543 [Ptrend] [0.75]
cases: 1995–1999 598
Non-Hispanic White:
596
646
Population; controls
randomly selected and
frequency-matched by race/
ethnicity and expected 5-yr
age distribution of cases
Source of controls
Troisi et al. (2013) 22 646 women aged < 85 yr, Weight gain (lb), since 1989 Age at delivery, race/ethnicity,
USA with primary in situ or Aged ≥ 50 yr at diagnosis: 299 parity at index birth, year of index
1974–2009 invasive cancer < 25 62 1.00 birth
224 721 25– < 31 99 1.33 (0.95–1.86)
Population; frequency- 31– < 40 72 1.23 (0.86–1.76)
matched to cases by parity, ≥ 40 66 1.06 (0.74–1.54)
age, calendar year of
delivery, and race/ethnicity
Robinson et al. (2014) Women aged 20–74 yr Adult weight gain (lb) since age 18 yr Age, age squared, family history
USA 1783 Black: of BC, alcohol consumption, age
1993–2001 1536 ≤ 25 81 1.00 at menarche, parity, age at FFTP
Black: 26–54 126 0.70 (0.44–1.12) composite, lactation, education
788 ≥ 55 222 0.84 (0.50–1.40) level, smoking, reference BMI
718 [Ptrend] [0.64]
White: White:
995 ≤ 25 185 1.00
818 26–54 184 1.17 (0.82–1.65)
Population; frequency- ≥ 55 101 1.25 (0.70–2.23)
matched to cases by 5-yr age [Ptrend] [0.38]
group
Sanderson et al. 2614 aged 25–75 yr, primary Weight change (lb) since age 18 yr Age, education level, first-degree
(2015) ductal carcinoma in situ or Black: family history of BC, OC use, age
USA invasive breast cancer ≤ 0 23 1.0 at menarche, weight at 18 yr

2001–2011 2306 1–31 79 0.8 (0.3–2.1)
Population; matched by 5-yr 32–60 138 0.9 (0.4–2.3)
age groups, race, and county > 61 200 0.9 (0.4–2.2)
of residence [Ptrend] [0.90]
White:
≤ 0 71 1.0 Pinteraction = 0.62
1–31 406 1.2 (0.8–1.6)
32–60 460 1.3 (0.9–1.9)
> 61 329 1.1 (0.8–1.6)
[Ptrend] [0.76]
BC, breast cancer; BMI, body mass index (in kg/m 2); CI, confidence interval; FFTP, first full-term pregnancy; FTP, full-term pregnancy; HRT, hormone replacement therapy; MET,
metabolic equivalent of task; OC, oral contraceptive; yr, year(s)
281

282

Table 2.2.9o Mendelian randomization studies of body mass index and cancer of the breast
Reference Study population Sample size Exposure assessment Outcome Relative risk
Study (95% CI)
Gao et al. (2016) Women from 11 studies of 33 832 (15 748 cases Adult BMI: Adult BMI:
Genetic Associations individuals of European and 18 084 controls) Increase of 1 SD (equivalent All breast cancer 0.91 (0.88–0.94)
and Mechanisms in ancestry to 4.5 kg/m2) in genetically ER− breast cancer 0.89 (0.84–0.94)
Oncology (GAME- predicted adult BMI
ON) Consortium Increase of 1 SD Childhood BMI:
(~0.073 kg/m2) in genetically All breast cancer 0.71 (0.60–0.80)
predicted childhood BMI ER− breast cancer 0.69 (0.53–0.98)
BMI, body mass index (in kg/m 2); CI, confidence interval; ER, estrogen receptor; SD, standard deviation
References Bandera EV, Chandran U, Zirpoli G, Gong Z, McCann

SE, Hong CC, et al. (2013b). Body fatness and breast
cancer risk in women of African ancestry. BMC
Adebamowo CA, Ogundiran TO, Adenipekun AA, Cancer, 13(1):475. doi:10.1186/1471-2407-13-475
Oyesegun RA, Campbell OB, Akang EU, et al. (2003). PMID:24118876
Obesity and height in urban Nigerian women with Bao PP, Shu XO, Gao YT, Zheng Y, Cai H, Deming SL, et al.
breast cancer. Ann Epidemiol, 13(6):455–61. doi:10.1016/ (2011). Association of hormone-related characteristics
S1047-2797(02)00426-X PMID:12875805 and breast cancer risk by estrogen receptor/proges-
Adegoke OJ, Shu XO, Gao YT, Cai Q, Breyer J, Smith terone receptor status in the Shanghai breast cancer
J, et al. (2004). Genetic polymorphisms in uridine study. Am J Epidemiol, 174(6):661–71. doi:10.1093/aje/
diphospho-glucuronosyltransferase 1A1 (UGT1A1) kwr145 PMID:21768404
and risk of breast cancer. Breast Cancer Res Treat, Barnes BB, Steindorf K, Hein R, Flesch-Janys D, Chang-
85(3):239–45. doi:10.1023/B:BREA.0000025419.26423. Claude J (2011). Population attributable risk of inva-
b8 PMID:15111762 sive postmenopausal breast cancer and breast cancer
Ahn J, Schatzkin A, Lacey JV Jr, Albanes D, Ballard- subtypes for modifiable and non-modifiable risk
Barbash R, Adams KF, et al. (2007). Adiposity, adult factors. Cancer Epidemiol, 35(4):345–52. doi:10.1016/j.
weight change, and postmenopausal breast cancer canep.2010.11.003 PMID:21159569
risk. Arch Intern Med, 167(19):2091–102. doi:10.1001/ Berstad P, Coates RJ, Bernstein L, Folger SG, Malone KE,
archinte.167.19.2091 PMID:17954804 Marchbanks PA, et al. (2010). A case-control study
Alothaimeen A, Ezzat A, Mohamed G, Muammar T, of body mass index and breast cancer risk in white
Al-Madouj A (2004). Dietary fat and breast cancer and African-American women. Cancer Epidemiol
in Saudi Arabia: a case-control study. East Mediterr Biomarkers Prev, 19(6):1532–44. doi:10.1158/1055-
Health J, 10(6):879–86. PMID:16335776 9965.EPI-10-0025 PMID:20501755
Amadou A, Torres Mejia G, Fagherazzi G, Ortega C, Canchola AJ, Anton-Culver H, Bernstein L, Clarke CA,
Angeles-Llerenas A, Chajes V, et al. (2014). Anthro Henderson K, Ma H, et al. (2012). Body size and the
pometry, silhouette trajectory, and risk of breast cancer risk of postmenopausal breast cancer subtypes in
in Mexican women. Am J Prev Med, 46(3 Suppl 1):S52– the California Teachers Study cohort. Cancer Causes
64. doi:10.1016/j.amepre.2013.10.024 PMID:24512931 Control, 23(3):473–85. doi:10.1007/s10552-012-9897-x
Andreotti G, Hou L, Beane Freeman LE, Mahajan R, PMID:22286371
Koutros S, Coble J, et al. (2010). Body mass index, Carpenter CL, Ross RK, Paganini-Hill A, Bernstein L
agricultural pesticide use, and cancer incidence in (2003). Effect of family history, obesity and exercise
the Agricultural Health Study cohort. Cancer Causes on breast cancer risk among postmenopausal women.
Control, 21(11):1759–75. doi:10.1007/s10552-010-9603-9 Int J Cancer, 106(1):96–102. doi:10.1002/ijc.11186
PMID:20730623 PMID:12794763
Attner B, Landin-Olsson M, Lithman T, Noreen Cerne JZ, Ferk P, Leskosek B, Gersak K (2011). Hormone
D, Olsson H (2012). Cancer among patients with replacement therapy and some risk factors for breast
diabetes, obesity and abnormal blood lipids: a popu- cancer among Slovenian postmenopausal women.
lation-based register study in Sweden. Cancer Causes Climacteric, 14(4):458–63. doi:10.3109/13697137.2010.
Control, 23(5):769–77. doi:10.1007/s10552-012-9946-5 541307 PMID:21294673
PMID:22467266 Chang SC, Ziegler RG, Dunn B, Stolzenberg-Solomon
Awatef M, Olfa G, Rim C, Asma K, Kacem M, Makram R, Lacey JV Jr, Huang WY, et al. (2006). Association
H, et al. (2011). Physical activity reduces breast cancer of energy intake and energy balance with postmeno-
risk: a case-control study in Tunisia. Cancer Epidemiol, pausal breast cancer in the Prostate, Lung, Colorectal,
35(6):540–4. doi :10 .1016/j.c a ne p. 2 011. 0 2 . 011 and Ovarian Cancer Screening Trial. Cancer Epidemiol
PMID:21470932 Biomarkers Prev, 15(2):334–41. doi:10.1158/1055-9965.
Bandera EV, Chandran U, Hong CC, Troester MA, Bethea EPI-05-0479 PMID:16492925
TN, Adams-Campbell LL, et al. (2015). Obesity, body Cheraghi Z, Poorolajal J, Hashem T, Esmailnasab N,
fat distribution, and risk of breast cancer subtypes in Doosti Irani A (2012). Effect of body mass index on
African American women participating in the AMBER breast cancer during premenopausal and postmeno-
Consortium. Breast Cancer Res Treat, 150(3):655–66. pausal periods: a meta-analysis. PLoS One, 7(12):e51446.
doi:10.1007/s10549-015-3353-z PMID:25809092 doi:10.1371/journal.pone.0051446 PMID:23236502
Bandera EV, Chandran U, Zirpoli G, Ciupak G, Bovbjerg Chow LW, Lui KL, Chan JC, Chan TC, Ho PK, Lee WY,
DH, Jandorf L, et al. (2013a). Body size in early life and et al. (2005). Association between body mass index
breast cancer risk in African American and European and risk of formation of breast cancer in Chinese
American women. Cancer Causes Control, 24(12):2231– women. Asian J Surg, 28(3):179–84. doi:10.1016/S1015-
43. doi:10.1007/s10552-013-0302-1 PMID:24113797 9584(09)60338-9 PMID:16024311
283
Cotterchio M, Kreiger N, Theis B, Sloan M, Bahl S cancer risk in the EPIC-PANACEA study. Int J Cancer,
(2003). Hormonal factors and the risk of breast 135(12):2887–99. doi:10.1002/ijc.28926 PMID:24771551
cancer according to estrogen- and progesterone-re- Eng SM, Gammon MD, Terry MB, Kushi LH, Teitelbaum
ceptor subgroup. Cancer Epidemiol Biomarkers Prev, SL, Britton JA, et al. (2005). Body size changes in rela-
12(10):1053–60. PMID:14578142 tion to postmenopausal breast cancer among women
Cribb AE, Joy Knight M, Guernsey J, Dryer D, Hender on Long Island, New York. Am J Epidemiol, 162(3):229–
K, Shawwa A, et al. (2011). CYP17, catechol-O-methyl- 37. doi:10.1093/aje/kwi195 PMID:15987723
transferase, and glutathione transferase M1 genetic Enger SM, Ross RK, Paganini-Hill A, Carpenter CL,
polymorphisms, lifestyle factors, and breast cancer Bernstein L (2000). Body size, physical activity, and
risk in women on Prince Edward Island. Breast J, breast cancer hormone receptor status: results from
17(1):24–31. doi:10.1111/j.1524-4741.2010.01025.x two case-control studies. Cancer Epidemiol Biomarkers
PMID:21129090 Prev, 9(7):681–7. PMID:10919738
Dartois L, Fagherazzi G, Baglietto L, Boutron-Ruault Ericson U, Sonestedt E, Gullberg B, Olsson H, Wirfält
MC, Delaloge S, Mesrine S, et al. (2016). Proportion E (2007). High folate intake is associated with
of premenopausal and postmenopausal breast cancers lower breast cancer incidence in postmenopausal
attributable to known risk factors: Estimates from women in the Malmö Diet and Cancer cohort. Am J
the E3N-EPIC cohort. Int J Cancer, 138(10):2415–27. Clin Nutr, 86(2):434–43. doi:10.1093/ajcn/86.2.434
doi:10.1002/ijc.29987 PMID:26756677 PMID:17684216
de Vasconcelos AB, Azevedo e Silva Mendonça G, Sichieri Feigelson HS, Jonas CR, Teras LR, Thun MJ, Calle EE
R (2001). Height, weight, weight change and risk of (2004). Weight gain, body mass index, hormone
breast cancer in Rio de Janeiro, Brazil. Sao Paulo Med replacement therapy, and postmenopausal breast
J, 119(2):62–6. doi:10.1590/S1516-31802001000200005 cancer in a large prospective study. Cancer Epidemiol
PMID:11276168 Biomarkers Prev, 13(2):220–4. doi:10.1158/1055-9965.
Dey S, Boffetta P, Mathews A, Brennan P, Soliman A, EPI-03-0301 PMID:14973094
Mathew A (2009). Risk factors according to estrogen Felix JF, Bradfield JP, Monnereau C, van der Valk RJ,
receptor status of breast cancer patients in Trivandrum, Stergiakouli E, Chesi A, et al.; Bone Mineral Density
South India. Int J Cancer, 125(7):1663–70. doi:10.1002/ in Childhood Study (BMDCS); Early Genetics and
ijc.24460 PMID:19452528 Lifecourse Epidemiology (EAGLE) consortium; Early
Dinger JC, Heinemann LA, Möhner S, Thai M, Assmann Growth Genetics (EGG) Consortium (2016). Genome-
A (2006). Breast cancer risk associated with different wide association analysis identifies three new suscep-
HRT formulations: a register-based case-control study. tibility loci for childhood body mass index. Hum
BMC Womens Health, 6(1):13. doi:10.1186/1472-6874- Mol Genet, 25(2):389–403. doi:10.1093/hmg/ddv472
6-13 PMID:16965641 PMID:26604143
Dogan L, Kalaylioglu Z, Karaman N, Ozaslan C, Atalay Fourkala E-O, Burnell M, Cox C, Ryan A, Salter LC,
C, Altinok M (2011). Relationships between epidemi- Gentry-Maharaj A, et al. (2014). Association of skirt
ological features and tumor characteristics of breast size and postmenopausal breast cancer risk in older
cancer. Asian Pac J Cancer Prev, 12(12):3375–80. women: a cohort study within the UK Collaborative
PMID:22471484 Trial of Ovarian Cancer Screening (UKCTOCS).
Dolle JM, Daling JR, White E, Brinton LA, Doody DR, BMJ Open, 4(9):e005400. doi:10.1136/bmjopen-2014-
Porter PL, et al. (2009). Risk factors for triple-nega- 005400 PMID:25252818
tive breast cancer in women under the age of 45 years. Friedenreich CM, Courneya KS, Bryant HE (2002). Case-
Cancer Epidemiol Biomarkers Prev, 18(4):1157–66. control study of anthropometric measures and breast
doi:10.1158/1055-9965.EPI-08-1005 PMID:19336554 cancer risk. Int J Cancer, 99(3):445–52. doi:10.1002/
Eliassen AH, Colditz GA, Rosner B, Willett WC, ijc.10389 PMID:11992416
Hankinson SE (2006). Adult weight change and risk of Gao C, Patel CJ, Michailidou K, Peters U, Gong J,
postmenopausal breast cancer. JAMA, 296(2):193–201. Schildkraut J, et al.; the Colorectal Transdisciplinary
doi:10.1001/jama.296.2.193 PMID:16835425 Study (CORECT); Discovery, Biology and Risk
Elkum N, Al-Tweigeri T, Ajarim D, Al-Zahrani A, of Inherited Variants in Breast Cancer (DRIVE);
Amer SM, Aboussekhra A (2014). Obesity is a signif- Elucidating Loci Involved in Prostate Cancer
icant risk factor for breast cancer in Arab women. Susceptibility (ELLIPSE); Follow-up of Ovarian Cancer
BMC Cancer, 14(1):788. doi:10.1186/1471-2407-14-788 Genetic Association and Interaction Studies (FOCI);
PMID:25351244 and Transdisciplinary Research in Cancer of the Lung
Emaus MJ, van Gils CH, Bakker MF, Bisschop CN, (TRICL) (2016). Mendelian randomization study of
Monninkhof EM, Bueno-de-Mesquita HB, et al. adiposity-related traits and risk of breast, ovarian,
(2014). Weight change in middle adulthood and breast prostate, lung and colorectal cancer. Int J Epidemiol,
45(3):896–908. doi:10.1093/ije/dyw129 PMID:27427428
284
Gao CM, Tajima K, Ding JH, Tang JH, Wu JZ, Li SP, women. Cancer Causes Control, 12(4):349–58.
et al. (2009). Body size, physical activity and risk of doi:10.1023/A:1011232602348 PMID:11456231
breast cancer – a case control study in Jangsu Province Huang WY, Newman B, Millikan RC, Schell MJ, Hulka
of China. Asian Pac J Cancer Prev, 10(5):877–81. BS, Moorman PG (2000). Hormone-related factors and
PMID:20104981 risk of breast cancer in relation to estrogen receptor
Garmendia ML, Pereira A, Alvarado ME, Atalah E (2007). and progesterone receptor status. Am J Epidemiol,
Relation between insulin resistance and breast cancer 151(7):703–14. doi:10.1093/oxfordjournals.aje.a010265
among Chilean women. Ann Epidemiol, 17(6):403–9. PMID:10752798
doi:10.1016/j.annepidem.2007.01.037 PMID:17531933 IARC (2002). Weight control and physical activity. Lyon,
Gaudet MM, Carter BD, Patel AV, Teras LR, Jacobs EJ, France: IARC Press (IARC Handbooks of Cancer
Gapstur SM (2014). Waist circumference, body mass Prevention, Vol. 6). Available from: http://publications.
index, and postmenopausal breast cancer incidence iarc.fr/376.
in the Cancer Prevention Study-II Nutrition Cohort. Ibarluzea J, Fernández MF, Santa-Marina L, Olea-Serrano
Cancer Causes Control, 25(6):737–45. doi:10.1007/ MF, Rivas AM, Aurrekoetxea JJ, et al. (2004). Breast
s10552-014-0376-4 PMID:24715420 cancer risk and the combined effect of environ-
Gaudet MM, Press MF, Haile RW, Lynch CF, Glaser SL, mental estrogens. Cancer Causes Control, 15(6):591–
Schildkraut J, et al. (2011). Risk factors by molecular 600. doi:10.1023/B:CACO.0000036167.51236.86
subtypes of breast cancer across a population-based PMID:15280638
study of women 56 years or younger. Breast Cancer Res John EM, Phipps AI, Sangaramoorthy M (2013). Body size,
Treat, 130(2):587–97. doi:10.1007/s10549-011-1616-x modifying factors, and postmenopausal breast cancer
PMID:21667121 risk in a multiethnic population: the San Francisco
Ghiasvand R, Bahmanyar S, Zendehdel K, Tahmasebi Bay Area Breast Cancer Study. Springerplus, 2(1):239.
S, Talei A, Adami HO, et al. (2012). Postmenopausal doi:10.1186/2193-1801-2-239 PMID:23762816
breast cancer in Iran; risk factors and their popul- John EM, Sangaramoorthy M, Hines LM, Stern MC,
ation attributable fractions. BMC Cancer, 12(1):414. Baumgartner KB, Giuliano AR, et al. (2015a). Overall
doi:10.1186/1471-2407-12-414 PMID:22992276 and abdominal adiposity and premenopausal breast
Gilani GM, Kamal S (2004). Risk factors for breast cancer cancer risk among Hispanic women: the breast cancer
in Pakistani women aged less than 45 years. Ann Hum health disparities study. Cancer Epidemiol Biomarkers
Biol, 31(4):398–407. doi:10.1080/0301446042000226763 Prev, 24(1):138–47. doi:10.1158/1055-9965.EPI-13-
PMID:15513691 1007-T PMID:25352526
Han D, Nie J, Bonner MR, McCann SE, Muti P, Trevisan John EM, Sangaramoorthy M, Hines LM, Stern MC,
M, et al. (2006). Lifetime adult weight gain, central Baumgartner KB, Giuliano AR, et al. (2015b). Body
adiposity, and the risk of pre- and postmenopausal size throughout adult life influences postmenopausal
breast cancer in the Western New York exposures breast cancer risk among Hispanic women: the Breast
and breast cancer study. Int J Cancer, 119(12):2931–7. Cancer Health Disparities Study. Cancer Epidemiol
doi:10.1002/ijc.22236 PMID:17016824 Biomarkers Prev, 24(1):128–37. doi:10.1158/1055-9965.
Harris HR, Willett WC, Terry KL, Michels KB (2011). EPI-14-0560 PMID:25352523
Body fat distribution and risk of premenopausal breast John EM, Sangaramoorthy M, Phipps AI, Koo J, Horn-
cancer in the Nurses’ Health Study II. J Natl Cancer Inst, Ross PL (2011). Adult body size, hormone receptor
103(3):273–8. doi:10.1093/jnci/djq500 PMID:21163903 status, and premenopausal breast cancer risk in a
Healy LA, Ryan AM, Rowley S, Boyle T, Connolly E, multiethnic population: the San Francisco Bay Area
Kennedy MJ, et al. (2010). Obesity increases the risk of Breast Cancer Study. Am J Epidemiol, 173(2):201–16.
postmenopausal breast cancer and is associated with doi:10.1093/aje/kwq345 PMID:21084558
more advanced stage at presentation but no impact Justenhoven C, Hamann U, Schubert F, Zapatka M, Pierl
on survival. Breast J, 16(1):95–7. doi:10.1111/j.1524- CB, Rabstein S, et al. (2008). Breast cancer: a candidate
4741.2009.00861.x PMID:19929892 gene approach across the estrogen metabolic pathway.
Hirose K, Tajima K, Hamajima N, Takezak T, Inoue M, Breast Cancer Res Treat, 108(1):137–49. doi:10.1007/
Kuroishi T, et al. (2003). Impact of established risk s10549-007-9586-8 PMID:17588204
factors for breast cancer in nulligravid Japanese women. Kabat GC, Xue X, Kamensky V, Lane D, Bea JW, Chen
Breast Cancer, 10(1):45–53. doi:10.1007/BF02967625 C, et al. (2015). Risk of breast, endometrial, colorectal,
PMID:12525763 and renal cancers in postmenopausal women in asso-
Hirose K, Tajima K, Hamajima N, Takezaki T, Inoue ciation with a body shape index and other anthropo-
M, Kuroishi T, et al. (2001). Association of family metric measures. Cancer Causes Control, 26(2):219–29.
history and other risk factors with breast cancer risk doi:10.1007/s10552-014-0501-4 PMID:25430815
among Japanese premenopausal and postmenopausal Kawai M, Kakugawa Y, Nishino Y, Hamanaka Y, Ohuchi
N, Minami Y (2013). Anthropometric factors, physical
285
activity, and breast cancer risk in relation to hormone Lodha RS, Nandeshwar S, Pal DK, Shrivastav A, Lodha
receptor and menopausal status in Japanese women: a KM, Bhagat VK, et al. (2011). Risk factors for breast
case-control study. Cancer Causes Control, 24(5):1033– cancer among women in Bhopal urban agglom-
44. doi:10.1007/s10552-013-0181-5 PMID:23494727 erate: a case-control study. Asian Pac J Cancer Prev,
Kawai M, Malone KE, Tang MT, Li CI (2014). Height, 12(8):2111–5. PMID:22292661
body mass index (BMI), BMI change, and the risk Lukanova A, Björ O, Kaaks R, Lenner P, Lindahl B,
of estrogen receptor-positive, HER2-positive, and Hallmans G, et al. (2006). Body mass index and cancer:
triple-negative breast cancer among women ages 20 results from the Northern Sweden Health and Disease
to 44 years. Cancer, 120(10):1548–56. doi:10.1002/ Cohort. Int J Cancer, 118(2):458–66. doi:10.1002/
cncr.28601 PMID:24500704 ijc.21354 PMID:16049963
Krebs EE, Taylor BC, Cauley JA, Stone KL, Bowman PJ, Lundqvist E, Kaprio J, Verkasalo PK, Pukkala E,
Ensrud KE (2006). Measures of adiposity and risk Koskenvuo M, Söderberg KC, et al. (2007). Co-twin
of breast cancer in older postmenopausal women. control and cohort analyses of body mass index and
J Am Geriatr Soc, 54(1):63–9. doi:10.1111/j.1532- height in relation to breast, prostate, ovarian, corpus
5415.2005.00541.x PMID:16420199 uteri, colon and rectal cancer among Swedish and
Kruk J (2007). Association of lifestyle and other risk Finnish twins. Int J Cancer, 121(4):810–8. doi:10.1002/
factors with breast cancer according to menopausal ijc.22746 PMID:17455257
status: a case-control study in the Region of Western Ma H, Bernstein L, Ross RK, Ursin G (2006). Hormone-
Pomerania (Poland). Asian Pac J Cancer Prev, 8(4):513– related risk factors for breast cancer in women under
24. PMID:18260721 age 50 years by estrogen and progesterone receptor
Kuriyama S, Tsubono Y, Hozawa A, Shimazu T, Suzuki Y, status: results from a case-control and a case-case
Koizumi Y, et al. (2005). Obesity and risk of cancer in comparison. Breast Cancer Res, 8(4):R39. doi:10.1186/
Japan. Int J Cancer, 113(1):148–57. doi:10.1002/ijc.20529 bcr1514 PMID:16846528
PMID:15386435 MacInnis RJ, English DR, Gertig DM, Hopper JL, Giles
Lahmann PH, Hoffmann K, Allen N, van Gils CH, GG (2004). Body size and composition and risk of
Khaw KT, Tehard B, et al. (2004). Body size and breast postmenopausal breast cancer. Cancer Epidemiol
cancer risk: findings from the European Prospective Biomarkers Prev, 13(12):2117–25. PMID:15598769
Investigation into Cancer and Nutrition (EPIC). Magnusson CM, Roddam AW, Pike MC, Chilvers C,
Int J Cancer, 111(5):762–71. doi:10.1002/ijc.20315 Crossley B, Hermon C, et al. (2005). Body fatness and
PMID:15252848 physical activity at young ages and the risk of breast
Li CI, Malone KE, Daling JR (2006). Interactions between cancer in premenopausal women. Br J Cancer, 93(7):817–
body mass index and hormone therapy and postmeno- 24. doi:10.1038/sj.bjc.6602758 PMID:16160699
pausal breast cancer risk (United States). Cancer Causes Mathew A, Gajalakshmi V, Rajan B, Kanimozhi V, Brennan
Control, 17(5):695–703. doi:10.1007/s10552-005-0001-7 P, Mathew BS, et al. (2008). Anthropometric factors
PMID:16633917 and breast cancer risk among urban and rural women
Li CI, Malone KE, Porter PL, Weiss NS, Tang MT, Daling in South India: a multicentric case-control study. Br
JR (2003). Reproductive and anthropometric factors in J Cancer, 99(1):207–13. doi:10.1038/sj.bjc.6604423
relation to the risk of lobular and ductal breast carci- PMID:18542077
noma among women 65-79 years of age. Int J Cancer, Mellemkjaer L, Bigaard J, Tjønneland A, Christensen J,
107(4):647–51. doi:10.1002/ijc.11465 PMID:14520705 Thomsen B, Johansen C, et al. (2006). Body compo-
Li CI, Stanford JL, Daling JR (2000). Anthropometric sition and breast cancer in postmenopausal women:
variables in relation to risk of breast cancer in middle- a Danish prospective cohort study. Obesity (Silver
aged women. Int J Epidemiol, 29(2):208–13. doi:10.1093/ Spring), 14(10):1854–62. doi:10.1038/oby.2006.214
ije/29.2.208 PMID:10817115 PMID:17062817
Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH, Michels KB, Terry KL, Eliassen AH, Hankinson SE,
Day FR, et al.; LifeLines Cohort Study; ADIPOGen Willett WC (2012). Adult weight change and inci-
Consortium; AGEN-BMI Working Group; dence of premenopausal breast cancer. Int J Cancer,
CARDIOGRAMplusC4D Consortium; CKDGen 130(4):902–9. doi:10.1002/ijc.26069 PMID:21413008
Consortium; GLGC; ICBP; MAGIC Investigators; Millikan RC, Newman B, Tse CK, Moorman PG, Conway
MuTHER Consortium; MIGen Consortium; PAGE K, Dressler LG, et al. (2008). Epidemiology of basal-like
Consortium; ReproGen Consortium; GENIE breast cancer. Breast Cancer Res Treat, 109(1):123–39.
Consortium; International Endogene Consortium doi:10.1007/s10549-007-9632-6 PMID:17578664
(2015). Genetic studies of body mass index yield new Minatoya M, Kutomi G, Shima H, Asakura S, Otokozawa
insights for obesity biology. Nature, 518(7538):197–206. S, Ohnishi H, et al. (2014). Relation of serum
doi:10.1038/nature14177 PMID:25673413 adiponectin levels and obesity with breast cancer: a
Japanese case-control study. Asian Pac J Cancer Prev,
286
15(19):8325–30. doi:10.7314/APJCP.2014.15.19.8325 of 424,519 participants. Lancet Oncol, 11(8):741–52.

PMID:25339024 doi:10.1016/S1470-2045(10)70141-8 PMID:20594911
Montazeri A, Sadighi J, Farzadi F, Maftoon F, Vahdaninia Phipps AI, Malone KE, Porter PL, Daling JR, Li CI (2008).
M, Ansari M, et al. (2008). Weight, height, body mass Body size and risk of luminal, HER2-overexpressing,
index and risk of breast cancer in postmenopausal and triple-negative breast cancer in postmeno-
women: a case-control study. BMC Cancer, 8(1):278. pausal women. Cancer Epidemiol Biomarkers Prev,
doi:10.1186/1471-2407-8-278 PMID:18826621 17(8):2078–86. doi:10.1158/1055-9965.EPI-08-0206
Msolly A, Gharbi O, Mahmoudi K, Limem S, Hochlef M, PMID:18664548
Ben Ahmed S (2011). Association between body mass Rapp K, Schroeder J, Klenk J, Stoehr S, Ulmer H, Concin
index and risk of breast cancer in Tunisian women. Ann H, et al. (2005). Obesity and incidence of cancer: a
Saudi Med, 31(4):393–7. doi:10.4103/0256-4947.83211 large cohort study of over 145,000 adults in Austria.
PMID:21808117 Br J Cancer, 93(9):1062–7. doi:10.1038/sj.bjc.6602819
Nemesure B, Wu SY, Hennis A, Leske MC; Barbados PMID:16234822
National Cancer Study Group (2009). Body size and Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D;
breast cancer in a black population – the Barbados Million Women Study Collaboration (2007). Cancer
National Cancer Study. Cancer Causes Control, incidence and mortality in relation to body mass
20(3):387–94. doi:10.10 07/s10552- 0 0 8 -9253 -3 index in the Million Women Study: cohort study. BMJ,
PMID:18987981 335(7630):1134. doi:10.1136/bmj.39367.495995.AE
Neuhouser ML, Aragaki AK, Prentice RL, Manson JE, PMID:17986716
Chlebowski R, Carty CL, et al. (2015). Overweight, Reinier KS, Vacek PM, Geller BM (2007). Risk factors
obesity, and postmenopausal invasive breast cancer risk: for breast carcinoma in situ versus invasive breast
a secondary analysis of the Women’s Health Initiative cancer in a prospective study of pre- and post-meno-
randomized clinical trials. JAMA Oncol, 1(5):611–21. pausal women. Breast Cancer Res Treat, 103(3):343–8.
doi:10.1001/jamaoncol.2015.1546 PMID:26182172 doi:10.1007/s10549-006-9375-9 PMID:17063272
Nichols HB, Trentham-Dietz A, Love RR, Hampton JM, Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen
Hoang Anh PT, Allred DC, et al. (2005). Differences in M (2008). Body-mass index and incidence of cancer:
breast cancer risk factors by tumor marker subtypes a systematic review and meta-analysis of prospec-
among premenopausal Vietnamese and Chinese tive observational studies. Lancet, 371(9612):569–78.
women. Cancer Epidemiol Biomarkers Prev, 14(1):41–7. doi:10.1016/S0140-6736(08)60269-X PMID:18280327
PMID:15668474 Robinson WR, Tse CK, Olshan AF, Troester MA (2014).
Noh HM, Song YM, Park JH, Kim BK, Choi YH (2013). Body size across the life course and risk of premeno-
Metabolic factors and breast cancer risk in Korean pausal and postmenopausal breast cancer in Black
women. Cancer Causes Control, 24(6):1061–8. women, the Carolina Breast Cancer Study, 1993-2001.
doi:10.1007/s10552-013-0183-3 PMID:23504150 Cancer Causes Control, 25(9):1101–17. doi:10.1007/
O’Brien KM, Sun J, Sandler DP, DeRoo LA, Weinberg CR s10552-014-0411-5 PMID:24924530
(2015). Risk factors for young-onset invasive and in situ Ronco AL, De Stefani E, Deneo-Pellegrini H, Quarneti
breast cancer. Cancer Causes Control, 26(12):1771–8. A (2012). Diabetes, overweight and risk of postmeno-
doi:10.1007/s10552-015-0670-9 PMID:26407954 pausal breast cancer: a case-control study in Uruguay.
Ogundiran TO, Huo D, Adenipekun A, Campbell O, Asian Pac J Cancer Prev, 13(1):139–46. doi:10.7314/
Oyesegun R, Akang E, et al. (2010). Case-control study APJCP.2012.13.1.139 PMID:22502657
of body size and breast cancer risk in Nigerian women. Rosato V, Bosetti C, Talamini R, Levi F, Montella M,
Am J Epidemiol, 172(6):682–90. doi:10.1093/aje/kwq180 Giacosa A, et al. (2011). Metabolic syndrome and the
PMID:20716701 risk of breast cancer in postmenopausal women. Ann
Okobia MN, Bunker CH, Zmuda JM, Osime U, Ezeome ER, Oncol, 22(12):2687–92. doi:10.1093/annonc/mdr025
Anyanwu SN, et al. (2006). Anthropometry and breast PMID:21415236
cancer risk in Nigerian women. Breast J, 12(5):462–6. Rosenberg LU, Einarsdóttir K, Friman EI, Wedrén S,
doi:10.1111/j.1075-122X.2006.00304.x PMID:16958966 Dickman PW, Hall P, et al. (2006). Risk factors for
Pan SY, Johnson KC, Ugnat AM, Wen SW, Mao Y; hormone receptor-defined breast cancer in postmeno-
Canadian Cancer Registries Epidemiology Research pausal women. Cancer Epidemiol Biomarkers Prev,
Group (2004). Association of obesity and cancer risk 15(12):2482–8. doi:10.1158/1055-9965.EPI-06-0489
in Canada. Am J Epidemiol, 159(3):259–68. doi:10.1093/ PMID:17164374
aje/kwh041 PMID:14742286 Rosner B, Eliassen AH, Toriola AT, Hankinson SE, Willett
Parr CL, Batty GD, Lam TH, Barzi F, Fang X, Ho SC, WC, Natarajan L, et al. (2015). Short-term weight gain
et al.; Asia-Pacific Cohort Studies Collaboration (2010). and breast cancer risk by hormone receptor classifica-
Body-mass index and cancer mortality in the Asia- tion among pre- and postmenopausal women. Breast
Pacific Cohort Studies Collaboration: pooled analyses
287
Cancer Res Treat, 150(3):643–53. doi:10.1007/s10549- risk defined by estrogen and progesterone receptor
015-3344-0 PMID:25796612 status among Swedish women: a prospective cohort
Rusiecki JA, Holford TR, Zahm SH, Zheng T (2005). study. Int J Cancer, 119(7):1683–9. doi:10.1002/ijc.22034
Breast cancer risk factors according to joint estrogen PMID:16646051
receptor and progesterone receptor status. Cancer Sweeney C, Blair CK, Anderson KE, Lazovich D, Folsom
Detect Prev, 29(5):419–26. doi:10.1016/j.cdp.2005.07.004 AR (2004). Risk factors for breast cancer in elderly
PMID:16185815 women. Am J Epidemiol, 160(9):868–75. doi:10.1093/
Sanderson M, Lipworth L, Shen-Miller D, Nechuta S, aje/kwh276 PMID:15496539
Beeghly-Fadiel A, Shrubsole MJ, et al. (2015). Energy- Tehard B, Clavel-Chapelon F (2006). Several anthropo-
related indicators and breast cancer risk among metric measurements and breast cancer risk: results
white and black women. PLoS One, 10(4):e0125058. of the E3N cohort study. Int J Obes, 30(1):156–63.
doi:10.1371/journal.pone.0125058 PMID:25927530 doi:10.1038/sj.ijo.0803133 PMID:16231021
Sangaramoorthy M, Phipps AI, Horn-Ross PL, Koo J, Teras LR, Goodman M, Patel AV, Diver WR, Flanders WD,
John EM (2011). Early-life factors and breast cancer risk Feigelson HS (2011). Weight loss and postmenopausal
in Hispanic women: the role of adolescent body size. breast cancer in a prospective cohort of overweight and
Cancer Epidemiol Biomarkers Prev, 20(12):2572–82. obese US women. Cancer Causes Control, 22(4):573–9.
doi:10.1158/1055-9965.EPI-11-0848 PMID:22056503 doi:10.1007/s10552-011-9730-y PMID:21327461
Sangrajrang S, Chaiwerawattana A, Ploysawang P, Tian YF, Chu CH, Wu MH, Chang CL, Yang T, Chou
Nooklang K, Jamsri P, Somharnwong S (2013). Obesity, YC, et al. (2007). Anthropometric measures, plasma
diet and physical inactivity and risk of breast cancer in adiponectin, and breast cancer risk. Endocr Relat
Thai women. Asian Pac J Cancer Prev, 14(11):7023–7. Cancer, 14(3):669–77. doi:10.1677/ERC-06-0089
doi:10.7314/APJCP.2013.14.11.7023 PMID:24377643 PMID:17914097
Shi J, Zhang M, Li L, Holman CD, Chen J, Teng Y, et al. Trentham-Dietz A, Newcomb PA, Egan KM, Titus-
(2010). Body mass index and its change in adulthood Ernstoff L, Baron JA, Storer BE, et al. (2000). Weight
and breast cancer risk in China. Asian Pac J Cancer change and risk of postmenopausal breast cancer
Prev, 11(5):1213–8. PMID:21198265 (United States). Cancer Causes Control, 11(6):533–42.
Shin A, Matthews CE, Shu XO, Gao YT, Lu W, Gu K, et al. doi:10.1023/A:1008961931534 PMID:10880035
(2009). Joint effects of body size, energy intake, and Trentham-Dietz A, Sprague BL, Hampton JM, Miglioretti
physical activity on breast cancer risk. Breast Cancer DL, Nelson HD, Titus LJ, et al. (2014). Modification
Res Treat, 113(1):153–61. doi:10.1007/s10549-008- of breast cancer risk according to age and meno-
9903-x PMID:18228135 pausal status: a combined analysis of five popula-
Shu XO, Jin F, Dai Q, Shi JR, Potter JD, Brinton LA, tion-based case-control studies. Breast Cancer Res
et al. (2001). Association of body size and fat distri- Treat, 145(1):165–75. doi:10.1007/s10549-014-2905-y
bution with risk of breast cancer among Chinese PMID:24647890
women. Int J Cancer, 94(3):449–55. doi:10.1002/ijc.1487 Trivers KF, Lund MJ, Porter PL, Liff JM, Flagg EW, Coates
PMID:11745429 RJ, et al. (2009). The epidemiology of triple-nega-
Silvera SA, Jain M, Howe GR, Miller AB, Rohan TE (2006). tive breast cancer, including race. Cancer Causes
Energy balance and breast cancer risk: a prospective Control, 20(7):1071–82. doi:10.1007/s10552-009-9331-1
cohort study. Breast Cancer Res Treat, 97(1):97–106. PMID:19343511
doi:10.1007/s10549-005-9098-3 PMID:16319973 Troisi R, Doody DR, Mueller BA (2013). A linked-reg-
Singh M, Jangra B (2013). Association between body istry study of gestational factors and subsequent breast
mass index and risk of breast cancer among females cancer risk in the mother. Cancer Epidemiol Biomarkers
of north India. South Asian J Cancer, 2(3):121–5. Prev, 22(5):835–47. doi:10.1158/1055-9965.EPI-12-1375
doi:10.4103/2278-330X.114108 PMID:24455581 PMID:23592822
Slattery ML, Sweeney C, Edwards S, Herrick J, Tsakountakis N, Sanidas E, Stathopoulos E, Kafousi
Baumgartner K, Wolff R, et al. (2007). Body size, M, Anogiannaki N, Georgoulias V, et al. (2005).
weight change, fat distribution and breast cancer risk Correlation of breast cancer risk factors with HER-2/
in Hispanic and non-Hispanic white women. Breast neu protein overexpression according to menopausal
Cancer Res Treat, 102(1):85–101. doi:10.1007/s10549- and estrogen receptor status. BMC Womens Health,
006-9292-y PMID:17080310 5(1):1. doi:10.1186/1472-6874-5-1 PMID:15694000
Song YM, Sung J, Ha M (2008). Obesity and risk of Verla-Tebit E, Chang-Claude J (2005). Anthropometric
cancer in postmenopausal Korean women. J Clin factors and the risk of premenopausal breast
Oncol, 26(20):3395–402. doi:10.1200/JCO.2007.15.7867 cancer in Germany. Eur J Cancer Prev, 14(4):419–
PMID:18612154 26. doi:10.1097/00008469-200508000-00016
Suzuki R, Rylander-Rudqvist T, Ye W, Saji S, Wolk A PMID:16030434
(2006). Body weight and postmenopausal breast cancer
288
Vrieling A, Buck K, Kaaks R, Chang-Claude J (2010).

Adult weight gain in relation to breast cancer risk by
estrogen and progesterone receptor status: a meta-ana-
lysis. Breast Cancer Res Treat, 123(3):641–9.
Food, nutrition, physical activity, and the prevention
of breast cancer. Washington (DC), USA: American
Institute for Cancer Research. Available from: https://
www.wcrf.org/sites/default/files/Breast-Cancer-2010-
Report.pdf.
Weiderpass E, Braaten T, Magnusson C, Kumle M, Vainio
H, Lund E, et al. (2004). A prospective study of body
size in different periods of life and risk of premeno-
pausal breast cancer. Cancer Epidemiol Biomarkers
Prev, 13(7):1121–7. PMID:15247122
Wenten M, Gilliland FD, Baumgartner K, Samet JM (2002).
Associations of weight, weight change, and body mass
with breast cancer risk in Hispanic and non-Hispanic
white women. Ann Epidemiol, 12(6):435–4. doi:10.1016/
S1047-2797(01)00293-9 PMID:12160603
White KK, Park SY, Kolonel LN, Henderson BE, Wilkens
LR (2012). Body size and breast cancer risk: the
Multiethnic Cohort. Int J Cancer, 131(5):E705–16.
doi:10.1002/ijc.27373 PMID:22120517
Wrensch M, Chew T, Farren G, Barlow J, Belli F, Clarke C,
et al. (2003). Risk factors for breast cancer in a popul-
ation with high incidence rates. Breast Cancer Res,
5(4):R88–102. doi:10.1186/bcr605 PMID:12817999
Wu AH, Yu MC, Tseng CC, Pike MC (2006). Body
size, hormone therapy and risk of breast cancer in
Asian-American women. Int J Cancer, 120(4):844–52.
doi:10.1002/ijc.22387 PMID:17131315
Yoo K, Tajima K, Park S, Kang D, Kim S, Hirose K, et al.
(2001). Postmenopausal obesity as a breast cancer risk
factor according to estrogen and progesterone receptor
status (Japan). Cancer Lett, 167(1):57–63. doi:10.1016/
S0304-3835(01)00463-3 PMID:11323099
Zhu K, Caulfield J, Hunter S, Roland CL, Payne-Wilks K,
Texter L (2005). Body mass index and breast cancer
risk in African American women. Ann Epidemiol,
15(2):123–8. doi:10.1016/j.annepidem.2004.05.011
PMID:15652717
Ziv E, John EM, Choudhry S, Kho J, Lorizio W, Perez-
Stable EJ, et al. (2006). Genetic ancestry and risk
factors for breast cancer among Latinas in the San
Francisco Bay Area. Cancer Epidemiol Biomarkers
Prev, 15(10):1878–85. doi:10.1158/1055-9965.EPI-06-
0092 PMID:17035394
289
2.2.10 Cancer of the breast in men Reference

Breast cancer in men is uncommon, with an
Brinton LA, Cook MB, McCormack V, Johnson KC,
incidence that is often cited as being less than 1% Olsson H, Casagrande JT, et al.; European Rare Cancer
of that for breast cancer in women. Risk factors Study Group (2014). Anthropometric and hormonal
for breast cancer in men include Klinefelter risk factors for male breast cancer: Male Breast Cancer
syndrome, a rare hereditary condition character- Pooling Project results. J Natl Cancer Inst, 106(3):djt465.
doi:10.1093/jnci/djt465 PMID:24552677
ized by a chromosomal abnormality of 46 XXY
karyotype with associated hormonal alterations,
and gynaecomastia, a condition linked with
excess estrogen. Like for breast cancer in women,
risk is likely to be mediated through hormonal
mechanisms.
The single largest study of the association of
BMI with breast cancer in men is the Male Breast
Cancer Pooling Project, a consortium of 11 case–
control studies and 10 cohort studies involving
2405 cases (1190 from case–control studies and
1215 from cohort studies) and 52 013 controls
(Brinton et al., 2014).
(a) Adult body mass index

BMI at baseline was associated with a small
but significant positive association between
increased BMI and risk of male breast cancer.
However, this association was observed only
with case–control studies (OR per 5 kg/m2, 1.24;
95% CI, 1.12–1.38), whereas the risk estimates
based on cohort studies were not significant (OR
per 5 kg/m2, 1.11; 95% CI, 0.97–1.28).
(b) Body mass index at earlier ages

Recalled BMI at age 18–21 years (based on six
cohort studies) showed no association with risk
of male breast cancer (OR per 5 kg/m2, 1.05; 95%
CI, 0.80–1.38).
290
2.2.11 Cancer of the endometrium has been estimated to be 1.6–1.9 (Renehan et al.,
2008; Yang et al., 2012; Bhaskaran et al., 2014).
Cancer of the endometrium is the sixth most Among those studies that distinguished
common cancer diagnosis in women (WCRF/ endometrial cancers by type (Bjørge et al., 2007;
AICR, 2013). Known risk factors for endometrial McCullough et al., 2008; Yang et al., 2013), all
cancer include exogenous estrogens, as delivered studies showed positive associations with BMI
in menopausal estrogen replacement therapies for both type 1 and type 2, with a stronger asso-
unopposed with progesterone, and diabetes. ciation for type 1 cancers.
Tobacco smoking is associated with reduced risk, The association between BMI and endome-
by mechanisms that are not well understood. trial cancer risk was much stronger in never-users
There are two subtypes of endometrial cancer: of HRT than in ever-users (McCullough et al.,
type 1, which is most common (accounting for 2008; Canchola et al., 2010); in a meta-analysis of
about 80–90% of endometrial cancer), and type 24 studies (Crosbie et al., 2010), the relative risk
2, which is more lethal but much less common per 5 kg/m2 was 1.18 in ever-users compared with
(about 10–20%). 1.90 in never-users.
In 2001, the Working Group of the IARC In the two studies that reported differences
Handbook on weight control and physical activity by smoking status, there was no difference in the
(IARC, 2002) concluded that there was sufficient association of BMI with endometrial cancer risk
evidence for a cancer-preventive effect of avoid- between smokers and never-smokers (Reeves
ance of weight gain for cancer of the endome- et al., 2007; Bhaskaran et al., 2014).
trium. The 2007 WCRF review concluded that In those studies that included measurements
there was convincing evidence of a positive asso- of waist circumference and hip circumference
ciation between body fatness and risk of endo- (Conroy et al., 2009; Canchola et al., 2010; Reeves
metrial cancer (WCRF/AICR, 2007), and this et al., 2011; Kabat et al., 2015), waist circumfer-
was later reaffirmed (WCRF/AICR, 2013). ence and waist-to-hip ratio were less strongly
(a) Cohort studies associated with risk than was BMI.
In those studies that examined the associa-
The scientific evidence since 2000 includes 20 tion between BMI at different ages and subse-
publications from cohort studies (excluding anal- quent risk of endometrial cancer (Jonsson et al.,
yses that were later updated and analyses based 2003; Chang et al., 2007; McCullough et al., 2008;
on fewer than 100 incident cases). Table 2.2.11a Canchola et al., 2010; Park et al., 2010; Yang et al.,
presents those findings by BMI at baseline, with 2012), BMI at earlier times in life was generally
comments on findings according to smoking more weakly related or was not related to risk
status, use of HRT, weight change over the life- of endometrial cancer, compared with BMI at
course, and waist circumference. baseline.
In general, findings are very consistent across
studies, showing a strongly positive association (b) Case–control studies
between BMI and endometrial cancer risk. All A total of 30 case–control studies have been
of the 20 cohort studies showed a statistically published since 2000 on the association between
significant positive association. There is an BMI at diagnosis and endometrial cancer risk,
approximately linear pattern of increasing risk including 21 population-based studies and 9
with increasing BMI. The relative risk per 5 kg/m2 hospital-based studies (Table 2.2.11b). Studies
were conducted in the USA (n = 10), Australia,
291
Canada, China, the Czech Republic, Israel, Italy, 2 kg/m2, 1.20; 95% CI, 1.19–1.21) compared with
Japan, Mexico, Puerto Rico, the Republic of type 2 tumours (RR, 1.12; 95% CI, 1.09–1.14) and
Korea, Switzerland, and the United Kingdom. among endometrioid grade 1 and 2 compared
In most of the studies, a statistically significant with endometrioid grade 3. The heterogeneity
increased risk of endometrial cancer was observed was present when cohort studies and case–
in overweight and obese women compared with control studies were considered separately, or
normal-weight women. when registry-based studies were compared
Among the case–control studies that evalu- with those where cases were further ascertained
ated BMI measured or recalled at different ages through pathology reports.
(Xu et al., 2006; Lucenteforte et al., 2007; Thomas
et al., 2009; Dal Maso et al., 2011; Hosono et al., (d) Mendelian randomization studies
2011; Lu et al., 2011; Nagle et al., 2013), an increased Nead et al. (2015) applied Mendelian random-
risk of endometrial cancer was also observed; the ization to assess the association of markers of
BMI measured or recalled closer to the date of metabolic disease, including BMI, with risk of
diagnosis was usually related to the highest risk. endometrial cancer using 32 genetic variants
Six studies reported associations between as instrumental variables for BMI (Speliotes
waist circumference and endometrial cancer risk, et al., 2010). Mendelian randomization analyses
showing a 2–5-fold increase in risk for women showed that each increase of 1 standard devi-
in the highest category of waist circumference ation in BMI was associated with a significant
versus the lowest. increase in risk of endometrial cancer (OR, 3.86;
95% CI, 2.24–6.64) (Table 2.2.11d).
(c) Pooled analyses and meta-analyses
Several recent pooled analyses and meta-
analyses have been published on the associa-
tion between BMI and endometrial cancer risk
(Dobbins et al., 2013; Felix et al., 2013; Setiawan
et al., 2013; Cote et al., 2015; Jenabi & Poorolajal,
2015; Table 2.2.11c).
A large meta-analysis of 20 case–control
studies reported a relative risk of 1.43 (95% CI,
1.30–1.56) for overweight and of 3.33 (95% CI,
2.87–3.79) for obese women compared with
normal-weight women (Jenabi & Poorolajal,
2015). In a pooled analysis of 7 cohort studies and
14 case–control studies, the risk of endometrial
cancer was similar for obese Black and White
women compared with their normal-weight
counterparts (Cote et al., 2015).
A recent pooled analysis of 10 cohort studies
and 14 case–control studies explored the hetero-
geneity of the association between BMI and
endometrial cancer risk according to tumour
types (Setiawan et al., 2013). They reported
stronger associations among type 1 (RR per
292
Table 2.2.11a Cohort studies of measures of body fatness and cancer of the endometrium
Reference Total number of Subtype Exposure Exposed Relative risk Covariates Comments
Location Incidence/
Calle et al. (2003) 495 477 BMI Age, education level,
Population-based Mortality 18.5–24.9 333 1.00 smoking, physical
cohort 25–29.9 225 1.50 (1.26–1.78) activity, alcohol
1982–1998 35–39.9 25 2.77 (1.83–4.18) status, aspirin use, fat
≥ 40 16 6.25 (3.75–10.42) intake, vegetable intake,
[Ptrend] [< 0.001] HRT use
Jonsson et al. (2003) 14 131 BMI Age Recalled BMI at ages
Swedish Twin Incidence < 18.49 1 0.4 (0.1–3.1) 25 yr and 40 yr gave
Registry 18.5–24.99 69 1.0 RR for BMI ≥ 25.0 vs
Sweden 25–29.99 46 1.3 (0.9–1.9) < 25.0 of 1.9 (1.2–3.0)
1969–1997 ≥ 30 21 3.2 (2.0–5.2) and 2.0 (0.9–4.4),
respectively
Rapp et al. (2005) 78 484 BMI Age, smoking,
Population-based Incidence 18.5–24.9 63 1.0 occupation
cohort 25.0–29.9 59 1.29 (0.90–1.86)
Austria 30–34.9 33 2.13 (1.38–3.27)
1985–2002 ≥ 35 20 3.93 (2.35–6.56)
[Ptrend] [< 0.001]
Lukanova et al. 35 362 BMI Age, tobacco use
(2006) Incidence 18.5–24.9 42 1.0
Population-based 25–29.9 41 1.45 (0.93–2.24)
cohort ≥ 30 35 2.93 (1.85–4.61)

1994–2004
Bjørge et al. (2007) 1 million Type 1 BMI Age, birth cohort Similar association for
Norwegian health Incidence < 18.5 82 0.90 (0.72–1.12) BMI at ages 20–49 yr
surveys 18.5–24.9 2960 1.00 and 50–74 yr
Norway 25–29.9 2361 1.39 (1.32–1.47)
1963–2003 ≥ 30 1761 2.72 (2.56–2.90)
[Ptrend] [< 0.001]
Type 2 BMI Age, birth cohort Similar association for
< 18.5 4 0.42 (0.16–1.13) BMI at ages 20–49 yr
18.5–24.9 366 1.00 and 50–74 yr
25–29.9 369 1.26 (1.09–1.46)
≥ 30 253 1.94 (1.64–2.30)
293
[Ptrend] [< 0.001]
294

Location Incidence/
Chang et al. (2007) 103 882 BMI Age, physical activity, BMI at ages 18 yr,
NIH-AARP cohort Incidence < 25 200 1.0 diabetes, HRT use, age 35 yr, and 50 yr not
USA 25–29.9 181 1.31 (1.07–1.61) at menarche, parity, age associated with risk
1995–2000 ≥ 30 296 3.03 (2.50–3.68) at menopause, OC use,
[Ptrend] [< 0.0001] smoking, race
Friberg et al. (2007) 36 773 BMI Age, physical activity Women without
Swedish Incidence < 30 154 1.0 diabetes
mammography ≥ 30 43 2.49 (1.77–3.51)
cohort
Sweden
1987–2003
Lundqvist et al. 14 017 older twins BMI Smoking, physical
(2007) (mean baseline < 18.5 1 0.3 (0.1–2.5) activity, education level,
Twin cohort studies age, 56 yr) 18.5–24.9 92 1.0 diabetes
Sweden and Finland Incidence 25–29.9 57 1.2 (0.8–1.6)
1961–2004 ≥ 30 30 3.2 (2.1–4.8)
per 1 kg/m2 1.11 (1.06–1.15)
[Ptrend] [< 0.0001]
Reeves et al. (2007) 1.2 million BMI Age, region, SES, Association similar in
Population-based Incidence < 22.5 340 0.84 (0.75–0.93) reproductive history, never-smokers
cohort 22.5–24.9 524 1.00 smoking, alcohol
1996–2001 27.5–29.9 366 1.43 (1.29–1.58) activity, HRT use
≥ 30 911 2.73 (2.55–2.92)
per 10 kg/m2 2.89 (2.62–3.18)
Lindemann et al. 36 761 BMI Age, physical activity, Similar associations
(2008) Incidence < 20 4 0.53 (0.19–1.47) hypertension, alcohol for women aged
HUNT cohort 20–24 64 1.00 consumption < 55 yr and aged
Norway 25–29 90 1.74 (1.25–2.43) ≥ 55 yr
1984–2002 30–34 32 1.66 (1.06–2.59)
35–39 23 4.28 (2.58–7.09)
≥ 40 9 6.36 (3.08–13.16)
[Ptrend] [< 0.0001]
Location Incidence/
McCullough et al. 33 436 BMI Age, age at menarche, Stronger association
(2008) Incidence < 22.5 54 0.92 (0.63–1.34) age at menopause, for never- vs ever-
Cancer Prevention 22.5–24.9 53 1.00 parity, HRT use, users of HRT. Stronger
Study II (CPS II) 25–29.9 91 1.40 (0.99–1.96) smoking, exercise, OC association for type 1
USA 30–34.9 76 3.27 (2.29–4.67) use vs type 2 cancer; null
1992–2003 ≥ 35 44 4.70 (3.12–7.07) association with BMI
[Ptrend] [< 0.0001] at age 18 yr
Song et al. (2008) 107 481 BMI Age, smoking, alcohol
Korean medical Incidence < 18.5 2 1.26 (0.29–5.51) consumption, exercise
insurance cohort 18.5–20.9 6 0.74 (0.29–1.90)
Republic of Korea 21–22.9 16 1.00
1994–2003 23.0–24.9 22 1.20 (0.62–2.32)
25.0–26.7 28 1.61 (0.84–3.09)
27.0–29.9 31 2.70 (1.42–5.13)
≥ 30 7 2.95 (1.20–7.24)
per 1 kg/m2 1.13 (1.07–1.20)
Conroy et al. (2009) 19 917 BMI Age, physical activity, Weaker association
Women’s Health Incidence < 22.5 57 1.00 smoking, alcohol with WC
Study 22.5–24.9 50 0.97 (0.65–1.44) consumption, diet,
USA 25–29.9 68 1.09 (0.75–1.58) parity, HRT use
1992–2007 ≥ 30 89 2.49 (1.73–3.59)
[Ptrend] [< 0.0001]
Epstein et al. (2009) 17 822 BMI Age

Lund cohort Incidence < 25 45 1.0
Sweden 25–29.9 41 1.4 (0.9–2.2)
1990–2007 ≥ 30 36 3.5 (2.2–5.4)
Canchola et al. 28 418 never-users BMI Age, parity, age at first Much weaker
(2010) of HRT < 25 34 1.0 pregnancy, physical association among
California Teachers Incidence 25–29.9 26 1.2 (0.74–2.1) activity, OC use HRT users. Similar
Study Cohort ≥ 30 48 3.5 (2.2–5.5) risk for recalled
USA [Ptrend] [< 0.001] BMI at age 18 yr;
1995–2006 per 1 kg/m2 1.07 (1.04–1.09) association also
observed with WC
295
296

Location Incidence/
Dossus et al. (2010) 370 000 BMI Age, centre
EPIC cohort Incidence < 25 81 1.0
Europe 25–29.9 82 1.23 (0.82–1.84)
1992–2003 ≥ 30 61 2.02 (1.26–3.23)
[Ptrend] [0.005]
Park et al. (2010) 50 376 women BMI at baseline Age, ethnicity, Results available for
Multiethnic Cohort aged 45–75 yr, < 25 175 1.00 education level, BMI at age 21 yr, BMI
USA (California, from 5 racial/ 25– < 30 119 1.36 (1.06–1.75) age at menarche, change since age 21 yr,
Hawaii) ethnic populations ≥ 30 169 3.54 (2.70–4.63) menopausal status, age weight at baseline, and
1993–2004 [Ptrend] [< 0.001] at menopause, HRT use, weight at age 21 yr
OC use, parity, smoking
history, diabetes,
hypertension
Reeves et al. (2011) 86 937 BMI Age, race, education WHR more weakly
Women’s Health Incidence < 25 264 1.0 level, smoking, physical associated, and
Initiative 25–29.9 207 0.84 (0.67–1.05) activity, intake of fruits association disappears
USA ≥ 30 334 1.68 (1.33–2.13) and vegetables, diabetes, with BMI adjustment
1993–NR [Ptrend] [0.0001] dietary fat, fibre intake
Ollberding et al. 46 027 BMI 489 total Age, race, ethnicity,
(2012) Incidence < 25 1.00 hypertension, diabetes,
Multiethnic Cohort 25–29.9 1.38 (1.09–1.74) smoking, HRT use, OC
USA ≥ 30 2.68 (2.10–3.42) use, parity
1993–2007 [Ptrend] [< 0.01]
Yang et al. (2012) 249 791 BMI Age, region, height, (Update of study by
Million Women Incidence < 22.5 139 1.00 age at menarche, age Reeves et al., 2007)
Study 22.5–27.4 465 1.40 (1.27–1.53) at menopause, parity, Body size and BMI at
United Kingdom 27.5–32.4 390 2.63 (2.39–2.91) HRT use, alcohol ages 10 yr and 20 yr
1996–2009 32.5–34.9 158 5.07 (4.33–5.93) consumption, smoking, less associated than
≥ 35 258 7.72 (6.79–8.77) exercise BMI at baseline
per 5 kg/m2 1.87 (1.77–1.96)
Yang et al. (2013) 114 409 Type 1 BMI Age, OC use, HRT use, Most women
NIH-AARP cohort Incidence < 30 708 1.00 parity, age at menarche, postmenopausal at
USA ≥ 30 570 2.93 (2.62–3.28) menopausal status, race, time of study entry
1995–2006 smoking
Location Incidence/
Yang et al. (2013) Type 2 BMI Age, OC use, HRT
(cont.) < 30 86 1.00 use, parity, menarche,
≥ 30 47 1.83 (1.27–2.63) menopause, race,
smoking
Bhaskaran et al. 5.24 million BMI 2758 Age, sex, year, diabetes, Similar association in
(2014) Incidence per 5 kg/m2 1.62 (1.56–1.69) alcohol consumption, never-smokers
Health system smoking, SES
clinical database
United Kingdom
1987–2012
Kabat et al. (2015) 143 901 BMI, quintiles 1157 total Age, alcohol Similar association
Women’s Health Incidence Q1 1.0 consumption, smoking, with WC
Initiative cohort Q2 0.93 (0.76–1.14) parity, HRT use, OC
USA Q3 1.08 (0.89–1.32) use, ethnicity, education
1992–2013 Q4 1.29 (1.06–1.58)
Q5 2.32 (1.93–2.80)
[Ptrend] [< 0.0001]
BMI, body mass index (in kg/m 2); CI, confidence interval; EPIC, European Prospective Investigation into Cancer and Nutrition; HRT, hormone replacement therapy; NIH-AARP,
National Institutes of Health–AARP Diet and Health Study; NR, not reported; OC, oral contraceptive; RR, relative risk; SES, socioeconomic status; WC, waist circumference; WHR,
waist-to-hip ratio; yr, year or years

297
298

Table 2.2.11b Case–control studies of measures of body fatness and cancer of the endometrium
Reference Total number of cases Exposure categories Exposed cases Relative risk Covariates
Study location Total number of (95% CI) Comments
Period controls
Source of controls
McCann et al. (2000) 232 BMI Age
USA 639 < 27.5 112 1.0
1986–1991 Population ≥ 27.5 120 2.6 (1.9–3.6)
Salazar-Martínez et al. 85 BMI Age, an ovulatory index, smoking,
(2000) 668 < 25 21 1.0 physical activity, menopausal status,
Mexico Population 25–30 28 1.1 (0.61–2.1) hypertension, diabetes
1995–1997 > 30 35 2.2 (1.2–4.2)
Benshushan et al. (2001, 128 BMI
2002) 255 < 27 49 1.00
Israel Population ≥ 27 79 2.47 [1.51–4.06]
1989–1992
Newcomer et al. (2001) 740 BMI Age
USA 2372 < 22.55 97 1.0
1991–1994 Population 22.55–25.34 120 1.2 (0.9–1.7)
25.35–29.14 150 1.6 (1.2–2.1)
≥ 29.15 293 3.0 (2.3–3.9)
McElroy et al. (2002) 148 BMI Age
USA 659 < 22.7 13 1.00
1991–1994 Population 22.7–25.5 18 1.52 (0.80–2.88)
25.6–29.0 20 1.60 (0.84–3.03)
≥ 29.1 45 3.72 (2.10–6.57)
Augustin et al. (2003) 410 BMI Age, study centre, education level,
Italy and Switzerland 753 < 20 33 1.0 history of diabetes and hypertension,
1988 –1998 Hospital 20–25 162 1.2 (0.8–2.0) HRT use, total energy intake
25– < 30 131 1.3 (0.8–2.2)
≥ 30 84 2.2 (1.2–3.8)
Dal Maso et al. (2004) 87 BMI Age, education level
Italy 132 < 25 20 1.00
1999–2002 Hospital 25–29 34 1.80 (0.90–3.59)
≥ 30 33 5.87 (2.58–13.38)
Period controls
Source of controls
Xu et al. (2005, 2006) 832 BMI, quartiles Age, education level, years of
China 846 Recent BMI menstruation, OC use, number of
1997–2001 Population > 25.69 vs < 21.03 302 3.3 (2.4–4.5) pregnancies, menopausal status, family
BMI at age 20 yr history of cancer; for recent BMI,
> 21.09 vs < 17.63 205 1.3 (1.0–1.8) additionally adjusted for BMI at age
20 yr
BMI at age 30 yr
> 22.43 vs < 18.81 226 1.5 (1.1–2.0)
BMI at age 40 yr
> 24.00 vs < 19.83 269 2.0 (1.5–2.8)
BMI at age 50 yr
> 25.30 vs < 20.83 217 2.5 (1.7–3.6)
BMI at age 60 yr
> 25.97 vs < 21.48 122 2.9 (1.7–4.9)
Xu et al. (2005) 832 WC (cm) Age, education level, years of
China 846 ≤ 73 102 1.0 menstruation, number of pregnancies,
1997–2001 Population 74–79 157 1.9 (1.4–2.7) BMI
80–86 215 2.6 (1.9–3.6)
> 86 357 4.7 (3.4–6.4)
Okamura et al. (2006) 155 BMI Age
Japan 96 < 20.04 36 1.00
1998–2000 Hospital 20.04–21.63 27 0.47 (0.22–0.99)
21.64–23.92 45 1.24 (0.58–2.67)

≥ 23.93 47 1.92 (0.86–4.30)
Trentham-Dietz et al. 740 BMI Age, age at menarche, parity,
(2006) 2342 14.5–22.6 100 1.00 menopausal status, age at menopause,
USA Population 22.6–25.4 123 1.19 (0.88–1.61) smoking, HRT use, recent physical
1991–1994 25.5–29.2 153 1.62 (1.21–2.18) activity, diabetes
29.1–82.4 313 3.20 (2.42–4.24)
Weiss et al. (2006) 1304 BMI Low tumour HRT use, age, county of residence,
USA 1779 aggressiveness: reference year
1985–1991, 1994–1995, Population < 30.0 374 1.0 Tumours with moderate or high
1997–1999 30.0–34.9 57 1.6 (1.2–2.3) aggressiveness gave very similar results
≥ 35.0 65 5.1 (3.5–7.4)
299
300

Period controls
Source of controls
Lucenteforte et al. (2007) 777 BMI at diagnosis Age, history of diabetes, physical
Italy and Switzerland 1550 < 30 555 1.0 activity, history of hypertension, year of
1988–2006 Hospital ≥ 30 218 2.4 (1.9–3.1) interview, study centre, education level,
BMI at age 30–39 yr parity, menopausal status, OC use, HRT
< 25 532 1.0 use
≥ 25 215 1.6 (1.3–2.0)
Máchová et al. (2007) 87 BMI NR Age, smoking, hypertension, height
Czech Republic 20 776 < 25 1.00
1987–2002 Population ≥ 25– < 30 1.84 (0.95–3.57)
≥ 30 3.25 (1.65–6.37)
Niwa et al. (2007) 110 BMI
Japan 220 < 25.0 75 1.00
2001–2004 Hospital ≥ 25.0 35 2.35 (1.32–4.17)
Wen et al. (2008) 1046 BMI Age at menarche, menopausal status,
China 1035 < 20.92 104 1.1 (0.9–1.5) total years of menstruation, OC use,
1997–2003 Population 20.93–22.68 128 1.0 (0.9–1.1) cancer history in first-degree relatives,
22.69–24.32 190 1.0 and BMI (for WC) or WC (for BMI)
24.33–26.47 214 1.0 (0.9–1.2)
> 26.47 408 1.1 (0.8–1.5)
WC (cm)
< 71 71 0.5 (0.3–0.6)
72–76 141 0.7 (0.6–0.8)
77–80 168 1.0
81–87 282 1.5 (1.3–1.7)
> 87 382 2.3 (1.7–3.1)
Fortuny et al. (2009) 469 BMI Age
USA 467 < 25 118 1.0
2001–2005 Population 25– < 30 127 1.6 (1.1–2.2)
30– < 35 80 2.0 (1.4–3.0)
≥ 35 142 7.6 (4.8–11.8)
Thomas et al. (2009) 421 Adult BMI LMP < 45 yr: Age, race, education level, OC use,
USA 3159 < 25.0 59 1.0 parity, use of estrogen therapy,
1980–1982 Population 25.0–29.9 26 2.9 (1.7–4.8) menopausal status, history of high
30.0–34.9 23 6.0 (3.3–10.7) blood pressure
≥ 35.0 30 21.7 (11.3–41.7)
Period controls
Source of controls
Thomas et al. (2009) Adult BMI LMP ≥ 45 yr: Weaker associations with BMI at age
(cont.) < 25.0 168 1.0 18 yr vs adult BMI for both LMP < 45 yr
25.0–29.9 60 1.5 (1.0–2.1) and LMP ≥ 45 yr
30.0–34.9 31 2.3 (1.4–3.6)
≥ 35.0 24 3.7 (2.0–6.6)
Tong et al. (2009) 125 BMI Age
Republic of Korea 302 < 23 30 1.0
1998–2006 Hospital 23–25 34 1.19 (0.62–2.29)
≥ 25 61 2.65 (1.44–4.89)
Chandran et al. (2010) 424 BMI Age
USA 398 < 25 105 1.00
2001–2005 Population 25–29.9 121 1.93 (1.36–2.75)
30–34.9 68 2.02 (1.32–3.08)
≥ 35 123 8.47 (5.16–13.89)
Charneco et al. (2010) 74 BMI Crude
Puerto Rico 88 ≤ 24.9 6 1.00 Age, education level, employment
2004–2007 Hospital 25.0–29.9 25 4.44 (1.60–12.26) status, poultry consumption, OC use,
≥ 30 43 9.85 (3.61–26.87) diabetes, hypertension
BMI
< 30 31 1.00
≥ 30 43 4.11 (1.76–9.93)
John et al. (2010) 472 BMI Age, race/ethnicity
USA 443 < 25 176 1.00

1996–1999 Population 25–29.9 135 0.92 (0.67–1.26)
≥ 30 184 1.93 (1.39–2.68)
Zhang et al. (2010) 942 BMI
China 1721 18.5–24.9 571 1.00
2004–2008 Population 25.0–29.9 284 1.51 (1.26–1.81)
≥ 30.0 80 6.15 (3.98–9.51)
Dal Maso et al. (2011) 454 ΒΜΙ ≥ 30: Age, study centre, calendar period
Italy 908 BMI at baseline 168 4.08 (2.90–5.74) of interview, years of education,
1992–2006 Hospital BMI at age 30 yr 29 1.78 (1.01–3.14) smoking habits, age at menarche, age at
BMI at age 50 yr 96 3.37 (2.26–5.04) menopause, parity, OC use, HRT use
BMI, 5 kg/m2 increase 1.89 (1.65–2.17)
WC (cm)
≥ 96 vs < 84 127 2.68 (1.78–4.03)
301
302

Period controls
Source of controls
Delahanty et al. (2011) 832 BMI Age, income, education level
China 2049 < 21.7 14.0% 1.00
1996–2005 Population 21.7–24.5 28.3% 1.68 (1.30–2.18)
> 24.5 57.7% 3.13 (2.44–4.01)
Friedenreich et al. (2011) 515 WC (cm) Reference WC not reported
Canada 962 ≥ 88 343 2.32 (1.82–2.96) Age
2002–2006 Population
Hosono et al. (2011) 222 BMI at baseline Age, smoking, alcohol consumption,
Japan 2162 < 25 152 1.00 regular exercise, age at menarche,
2001–2005 Hospital ≥ 25 65 2.22 (1.59–3.09) duration of menstruation, parity,
BMI at age 20 yr diabetes history, history of OC use,
< 25 196 1.00 history of HRT use
≥ 25 17 2.30 (1.29–4.11)
BMI change from age 20 yr to enrolment
≤ 0 57 1.00
0–3 73 1.26 (0.86–1.84)
> 3 82 1.48 (0.95–2.29)
Lu et al. (2011) 668 BMI > 30 vs < 25: Age, ethnic group, education level,
USA 674 current 354 4.76 (3.50–6.49) pregnancy, family history of cancer,
2004–2009 Population 5 yr in the past 321 4.22 (3.05–5.84) estrogen use, OC use, smoking, alcohol
at age 20s 60 1.96 (1.16–3.29) consumption
at age 30s 106 2.19 (1.46–3.28)
at age 40s 150 3.84 (2.62–5.61)
at age 50s 156 5.44 (3.62–8.17)
at age 60s 67 4.09 (2.32–7.21)
Rosato et al. (2011) 454 BMI Age, study centre, year of interview,
Italy 798 ≤ 30 312 1.00 education level, age at menarche, parity,
1992–2006 Hospital > 30 142 3.83 (2.74–5.36) menopausal status, OC use, HRT use
WC (cm)
< 80 vs ≥ 80 266 1.62 (1.00–2.62)
≤ 88 vs > 88 195 1.90 (1.34–2.71)
Friedenreich et al. (2012) 541 BMI Same study/data set as Friedenreich
Canada 961 per 1 kg/m2 increase 1.10 (1.08–1.12) et al. (2011)
2002–2006 Population Adjusted for age
Period controls
Source of controls
Amankwah et al. (2013) 524 BMI Age, residence type (rural or urban),
Canada 1032 < 25 87 1.00 age at menarche, menopausal status/
2002–2006 Population 25– < 30 124 1.26 (0.91–1.73) hormone use, parity/age at first
≥ 30 256 2.81 (2.06–3.84) pregnancy, hypertension
[Ptrend] [< 0.001]
WC (cm)
> 96.0 vs ≤ 76.5 220 4.21 (2.90–6.10)
Becker et al. (2013) 2554 BMI Crude estimates
United Kingdom 15 324 < 25 560 1.00
1995–2012 Population 25–29.9 560 1.49 (1.32–1.68)
30–59.9 877 3.18 (2.82–3.57)
King et al. (2013) 424 BMI Age
USA 398 < 25 105 1.00
2001–2005 Population 25–29.9 121 1.93 (1.36–2.75)
30–34.9 68 2.02 (1.32–3.08)
≥ 35 123 8.47 (5.16–13.89)
Nagle et al. (2013) 1398 Recent BMI Age, age at menarche, parity, duration of
Australia 1538 ≥ 40 vs < 25 192 7.98 (5.41–11.77) OC use, HRT use ≥ 3 months, smoking
2005–2007 Population Maximum BMI status, diabetes
≥ 40 vs < 25 257 6.62 (4.72–9.29)
BMI at age 20 yr
≥ 30 vs < 25 72 0.75 (0.43–1.33)
BMI change from age 20 yr

Always overweight vs 203 3.60 (2.62–4.95)
always normal
Change from maximum to recent BMI
Always ≥ 30 vs always < 25 637 3.71 (2.96–4.67)
BMI, body mass index (in kg/m 2); CI, confidence interval; HRT, hormone replacement therapy; LMP, last menstrual period; NR, not reported; OC, oral contraceptive; WC, waist
circumference; yr, year or years
303
304

Table 2.2.11c Pooled analyses and meta-analyses of measures of body fatness and cancer of the endometrium
Reference Number and type Population size and Exposure Exposed cases Relative risk Adjustments
of studies type categories (95% CI) Comments
Crosbie et al. Meta-analysis of 24 17 710 cases BMI P heterogeneity = 0.215
(2010) studies published 27 1.22 (1.19–1.24)
1966–2009 32 2.09 (1.94–2.26)
37 4.36 (3.75–5.10)
42 9.11 (7.26–11.51)
per 5 kg/m2 1.60 (1.52–1.68)
Dobbins et al. Meta-analysis of 16 Obese vs 1.85 (1.30–2.65) P heterogeneity = 0.00001
(2013) cohort and case– normal-weight
control studies
Felix et al. Pooled analysis of 8096 cases (primarily BMI Age, race, age at menarche, parity,
(2013) 13 studies endometrioid < 25 2675 1.00 menopausal status, menopausal
(E2C2) endometrial 25–30 2246 1.37 (1.28–1.46) estrogen plus progestin, menopausal
carcinomas) and ≥ 30 2479 3.03 (2.82–3.26) estrogen use, OC use, smoking status,
28 829 controls [Ptrend] [0.0001] history of diabetes, site
Setiawan et al. Pooled analysis of 14 069 cases and BMI Type 1: Age, study, race/ethnicity, age
(2013) 10 cohort studies 35 312 controls 18– < 25 4602 1.00 at menarche, parity, OC use,
and 14 case– 25– < 30 3718 1.45 (1.37–1.53) menopausal status, menopausal HRT
control studies in 30– < 35 2294 2.52 (2.35–2.69) use, smoking status
China, Europe, and 35– < 40 1247 4.45 (4.05–4.89)
North America ≥ 40 992 7.14 (6.33–8.06)
(E2C2) [Ptrend] [< 0.0001]
BMI Type 2:
18– < 25 330 1.00
25– < 30 253 1.16 (0.98–1.38)
30– < 35 159 1.73 (1.40–2.12)
35– < 40 65 2.15 (1.60–2.88)
≥ 40 47 3.11 (2.19–4.44)
[Ptrend] [< 0.0001]
Cote et al. Pooled analysis of 2011 Black women (516 BMI Black women: Age, smoking, OC use, diabetes,
(2015) 7 cohort studies cases and 1495 controls) 18.5–24.9 76 1.00 study site, age at menarche, parity as a
and 4 case–control 25–29.9 129 1.37 (0.97–1.94) continuous variable
studies ≥ 30 300 2.93 (2.11–4.07)
19 297 White women BMI White women:
(5693 cases and 13 604 18.5–24.9 1950 1.00
controls) 25–29.9 1541 1.43 (1.32–1.56)
≥ 30 2107 2.99 (2.74–3.26)
Reference Number and type Population size and Exposure Exposed cases Relative risk Adjustments
of studies type categories (95% CI) Comments
Jenabi & Meta-analysis of 20 32 281 242 participants BMI
Poorolajal cohort studies total Normal 1.00 P heterogeneity:
(2015) Overweight 1.34 (1.20–1.48) Overweight: P = 0.001
Obese 2.54 (2.27–2.81) Obesity: P = 0.001
Meta-analysis of BMI
20 case–control Normal 1.00 P heterogeneity:
studies Overweight 1.43 (1.30–1.56) Overweight: P = 0.017
Obese 3.33 (2.87–3.79) Obesity: P = 0.001
BMI, body mass index (in kg/m 2); CI, confidence interval; E2C2, Epidemiology of Endometrial Cancer Consortium; OC, oral contraceptive; yr, year or years

305
306

Table 2.2.11d Mendelian randomization studies of measures of body fatness and cancer of the endometrium
Reference Characteristics of study population Sample size Exposure Outcome Odds ratio (95% CI)
with each SD increase in
exposure
Nead et al. Cases were from the Australian National Endometrial Cancer Study 9560 (1287 BMI Endometrial 3.86 (2.24–6.64)
(2015) (ANECS) or the Studies of Epidemiology and Risk Factors in Cancer cases and cancer
Heredity study (SEARCH), United Kingdom 8273 controls)
Control participants were from the Wellcome Trust Case Control
Consortium (WTCCC), and Australian control participants were from
parents of twins in the Brisbane Adolescent Twin Study and from the
Hunter Community Study
BMI, body mass index (in kg/m 2); CI, confidence interval; SD, standard deviation; yr, year or years
References Cancer Epidemiol Biomarkers Prev, 16(4):723–30.

doi:10.1158/1055-9965.EPI-06-0675 PMID:17416763
Charneco E, Ortiz AP, Venegas-Ríos HL, Romaguera J,
Amankwah EK, Friedenreich CM, Magliocco AM, Brant Umpierre S (2010). Clinic-based case-control study of
R, Courneya KS, Speidel T, et al. (2013). Anthropometric the association between body mass index and endome-
measures and the risk of endometrial cancer, overall trial cancer in Puerto Rican women. P R Health Sci J,
and by tumor microsatellite status and histological 29(3):272–8. PMID:20799515
subtype. Am J Epidemiol, 177(12):1378–87. doi:10.1093/ Conroy MB, Sattelmair JR, Cook NR, Manson JE, Buring
aje/kws434 PMID:23673247 JE, Lee IM (2009). Physical activity, adiposity, and
Augustin LS, Gallus S, Bosetti C, Levi F, Negri E, risk of endometrial cancer. Cancer Causes Control,
Franceschi S, et al. (2003). Glycemic index and glycemic 20(7):1107–15. doi:10.1007/s10552-009-9313-3
load in endometrial cancer. Int J Cancer, 105(3):404–7. PMID:19247787
doi:10.1002/ijc.11089 PMID:12704677 Cote ML, Alhajj T, Ruterbusch JJ, Bernstein L, Brinton
Becker C, Jick SS, Meier CR, Bodmer M (2013). Metformin LA, Blot WJ, et al. (2015). Risk factors for endometrial
and the risk of endometrial cancer: a case-control cancer in black and white women: a pooled analysis
analysis. Gynecol Oncol, 129(3):565–9. doi:10.1016/j. from the Epidemiology of Endometrial Cancer
ygyno.2013.03.009 PMID:23523618 Consortium (E2C2). Cancer Causes Control, 26(2):287–
Benshushan A, Paltiel O, Brzezinski A, Tanos V, Barchana 96. doi:10.1007/s10552-014-0510-3 PMID:25534916
M, Shoshani O, et al. (2001). Ovulation induction and Crosbie EJ, Zwahlen M, Kitchener HC, Egger M, Renehan
risk of endometrial cancer: a pilot study. Eur J Obstet AG (2010). Body mass index, hormone replacement
Gynecol Reprod Biol, 98(1):53–7. doi:10.1016/S0301- therapy, and endometrial cancer risk: a meta-analysis.
2115(01)00344-X PMID:11516800 Cancer Epidemiol Biomarkers Prev, 19(12):3119–30.
Benshushan A, Paltiel O, Rojansky N, Brzezinski A, doi:10.1158/1055-9965.EPI-10-0832 PMID:21030602
Laufer N (2002). IUD use and the risk of endometrial Dal Maso L, Augustin LS, Karalis A, Talamini R,
cancer. Eur J Obstet Gynecol Reprod Biol, 105(2):166–9. Franceschi S, Trichopoulos D, et al. (2004). Circulating
doi:10.1016/S0301-2115(02)00153-7 PMID:12381481 adiponectin and endometrial cancer risk. J Clin
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, Endocrinol Metab, 89(3):1160–3. doi:10.1210/jc.2003-
Leon DA, Smeeth L (2014). Body-mass index and risk 031716 PMID:15001602
of 22 specific cancers: a population-based cohort study Dal Maso L, Tavani A, Zucchetto A, Montella M,
of 5.24 million UK adults. Lancet, 384(9945):755–65. Ferraroni M, Negri E, et al. (2011). Anthropometric
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 measures at different ages and endometrial cancer risk.
Bjørge T, Engeland A, Tretli S, Weiderpass E (2007). Body Br J Cancer, 104(7):1207–13. doi:10.1038/bjc.2011.63
size in relation to cancer of the uterine corpus in 1 PMID:21386846
million Norwegian women. Int J Cancer, 120(2):378– Delahanty RJ, Beeghly-Fadiel A, Xiang YB, Long J, Cai
83. doi:10.1002/ijc.22260 PMID:17066451 Q, Wen W, et al. (2011). Association of obesity-related
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ genetic variants with endometrial cancer risk: a report
(2003). Overweight, obesity, and mortality from cancer from the Shanghai Endometrial Cancer Genetics
in a prospectively studied cohort of U.S. adults. N Engl Study. Am J Epidemiol, 174(10):1115–26. doi:10.1093/
J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423 aje/kwr233 PMID:21976109
PMID:12711737 Dobbins M, Decorby K, Choi BC (2013). The associa-
Canchola AJ, Chang ET, Bernstein L, Largent JA, tion between obesity and cancer risk: a meta-analysis
Reynolds P, Deapen D, et al. (2010). Body size and the of observational studies from 1985 to 2011. ISRN
risk of endometrial cancer by hormone therapy use in Prev Med, 2013:680536. doi:10.5402/2013/680536
postmenopausal women in the California Teachers PMID:24977095
Study cohort. Cancer Causes Control, 21(9):1407–16. Dossus L, Rinaldi S, Becker S, Lukanova A, Tjonneland A,
doi:10.1007/s10552-010-9568-8 PMID:20431936 Olsen A, et al. (2010). Obesity, inflammatory markers,
Chandran U, Bandera EV, Williams-King MG, Sima C, and endometrial cancer risk: a prospective case-control
Bayuga S, Pulick K, et al. (2010). Adherence to the study. Endocr Relat Cancer, 17(4):1007–19. doi:10.1677/
dietary guidelines for Americans and endometrial ERC-10-0053 PMID:20843938
cancer risk. Cancer Causes Control, 21(11):1895–904. Epstein E, Lindqvist PG, Olsson H (2009). A popula-
doi:10.1007/s10552-010-9617-3 PMID:20652737 tion-based cohort study on the use of hormone treat-
Chang S-C, Lacey JV Jr, Brinton LA, Hartge P, Adams ment and endometrial cancer in southern Sweden.
K, Mouw T, et al. (2007). Lifetime weight history Int J Cancer, 125(2):421–5. doi:10.1002/ijc.24284
and endometrial cancer risk by type of menopausal PMID:19326453
hormone use in the NIH-AARP Diet and Health Study.
307
Felix AS, Cook LS, Gaudet MM, Rohan TE, Schouten King MG, Chandran U, Olson SH, Demissie K, Lu SE,
LJ, Setiawan VW, et al. (2013). The etiology of uterine Parekh N, et al. (2013). Consumption of sugary foods
sarcomas: a pooled analysis of the epidemiology of endo- and drinks and risk of endometrial cancer. Cancer
metrial cancer consortium. Br J Cancer, 108(3):727–34. Causes Control, 24(7):1427–36. doi:10.1007/s10552-
doi:10.1038/bjc.2013.2 PMID:23348519 013-0222-0 PMID:23657460
Fortuny J, Sima C, Bayuga S, Wilcox H, Pulick K, Lindemann K, Vatten LJ, Ellstrøm-Engh M, Eskild A (2008).
Faulkner S, et al. (2009). Risk of endometrial cancer Body mass, diabetes and smoking, and endometrial
in relation to medical conditions and medication use. cancer risk: a follow-up study. Br J Cancer, 98(9):1582–5.
Cancer Epidemiol Biomarkers Prev, 18(5):1448–56. doi:10.1038/sj.bjc.6604313 PMID:18362938
doi:10.1158/1055-9965.EPI-08-0936 PMID:19383893 Lu L, Risch H, Irwin ML, Mayne ST, Cartmel B, Schwartz
Friberg E, Mantzoros CS, Wolk A (2007). Diabetes P, et al. (2011). Long-term overweight and weight gain
and risk of endometrial cancer: a population-based in early adulthood in association with risk of endome-
prospective cohort study. Cancer Epidemiol Biomarkers trial cancer. Int J Cancer, 129(5):1237–43. doi:10.1002/
Prev, 16(2):276–80. doi:10.1158/1055-9965.EPI-06-0751 ijc.26046 PMID:21387312
PMID:17301260 Lucenteforte E, Bosetti C, Talamini R, Montella M,
Friedenreich CM, Biel RK, Lau DC, Csizmadi I, Courneya Zucchetto A, Pelucchi C, et al. (2007). Diabetes
KS, Magliocco AM, et al. (2011). Case-control study of and endometrial cancer: effect modification by
the metabolic syndrome and metabolic risk factors body weight, physical activity and hypertension.
for endometrial cancer. Cancer Epidemiol Biomarkers Br J Cancer, 97(7):995–8. doi:10.1038/sj.bjc.6603933
Prev, 20(11):2384–95. doi:10.1158/1055-9965.EPI-11- PMID:17912243
0715 PMID:21921255 Lukanova A, Björ O, Kaaks R, Lenner P, Lindahl B,
Friedenreich CM, Langley AR, Speidel TP, Lau DC, Hallmans G, et al. (2006). Body mass index and cancer:
Courneya KS, Csizmadi I, et al. (2012). Case-control results from the Northern Sweden Health and Disease
study of markers of insulin resistance and endome- Cohort. Int J Cancer, 118(2):458–66. doi:10.1002/
trial cancer risk. Endocr Relat Cancer, 19(6):785–92. ijc.21354 PMID:16049963
doi:10.1530/ERC-12-0211 PMID:23033315 Lundqvist E, Kaprio J, Verkasalo PK, Pukkala E,
Hosono S, Matsuo K, Hirose K, Ito H, Suzuki T, Kawase Koskenvuo M, Söderberg KC, et al. (2007). Co-twin
T, et al. (2011). Weight gain during adulthood and body control and cohort analyses of body mass index and
weight at age 20 are associated with the risk of endome- height in relation to breast, prostate, ovarian, corpus
trial cancer in Japanese women. J Epidemiol, 21(6):466– uteri, colon and rectal cancer among Swedish and
73. doi:10.2188/jea.JE20110020 PMID:21986192 Finnish twins. Int J Cancer, 121(4):810–8. doi:10.1002/
IARC (2002). Weight control and physical activity. Lyon, ijc.22746 PMID:17455257
France: IARC Press (IARC Handbooks of Cancer Máchová L, Cízek L, Horáková D, Koutná J, Lorenc J,
Prevention, Vol. 6). Available from: http://publications. Janoutová G, et al. (2007). Association between obesity
iarc.fr/376. and cancer incidence in the population of the District
Jenabi E, Poorolajal J (2015). The effect of body mass Sumperk, Czech Republic. Onkologie, 30(11):538–42.
index on endometrial cancer: a meta-analysis. Public doi:10.1159/000108284 PMID:17992023
Health, 129(7):872–80. doi:10.1016/j.puhe.2015.04.017 McCann SE, Freudenheim JL, Marshall JR, Brasure JR,
PMID:26026348 Swanson MK, Graham S (2000). Diet in the epide-
John EM, Koo J, Horn-Ross PL (2010). Lifetime miology of endometrial cancer in western New York
physical activity and risk of endometrial cancer. (United States). Cancer Causes Control, 11(10):965–74.
Cancer Epidemiol Biomarkers Prev, 19(5):1276–83. doi:10.1023/A:1026551309873 PMID:11142531
doi:10.1158/1055-9965.EPI-09-1316 PMID:20406960 McCullough ML, Patel AV, Patel R, Rodriguez C, Feigelson
Jonsson F, Wolk A, Pedersen NL, Lichtenstein P, Terry HS, Bandera EV, et al. (2008). Body mass and endo-
P, Ahlbom A, et al. (2003). Obesity and hormone-de- metrial cancer risk by hormone replacement therapy
pendent tumors: cohort and co-twin control studies and cancer subtype. Cancer Epidemiol Biomarkers
based on the Swedish Twin Registry. Int J Cancer, Prev, 17(1):73–9. doi:10.1158/1055-9965.EPI-07-2567
106(4):594–9. doi:10.1002/ijc.11266 PMID:12845658 PMID:18187388
Kabat GC, Xue X, Kamensky V, Lane D, Bea JW, Chen McElroy JA, Newcomb PA, Trentham-Dietz A, Hampton
C, et al. (2015). Risk of breast, endometrial, colorectal, JM, Kanarek MS, Remington PL (2002). Endometrial
and renal cancers in postmenopausal women in asso- cancer incidence in relation to electric blanket use.
ciation with a body shape index and other anthropo- Am J Epidemiol, 156(3):262–7. doi:10.1093/aje/kwf020
metric measures. Cancer Causes Control, 26(2):219–29. PMID:12142261
doi:10.1007/s10552-014-0501-4 PMID:25430815
308
Nagle CM, Marquart L, Bain CJ, O’Brien S, Lahmann Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen
PH, Quinn M, et al.; Australian National Endometrial M (2008). Body-mass index and incidence of cancer:
Cancer Study Group (2013). Impact of weight change a systematic review and meta-analysis of prospec-
and weight cycling on risk of different subtypes of tive observational studies. Lancet, 371(9612):569–78.
endometrial cancer. Eur J Cancer, 49(12):2717–26. doi:10.1016/S0140-6736(08)60269-X PMID:18280327
doi:10.1016/j.ejca.2013.03.015 PMID:23583438 Rosato V, Zucchetto A, Bosetti C, Dal Maso L, Montella
Nead KT, Sharp SJ, Thompson DJ, Painter JN, Savage DB, M, Pelucchi C, et al. (2011). Metabolic syndrome and
Semple RK, et al.; Australian National Endometrial endometrial cancer risk. Ann Oncol, 22(4):884–9.
Cancer Study Group (ANECS) (2015). Evidence of a doi:10.1093/annonc/mdq464 PMID:20937645
causal association between insulinemia and endome- Salazar-Martínez E, Lazcano-Ponce EC, Lira-Lira GG,
trial cancer: a Mendelian randomization analysis. J Escudero-De los Rios P, Salmerón-Castro J, Larrea F,
Natl Cancer Inst, 107(9):djv178. doi:10.1093/jnci/djv178 et al. (2000). Case-control study of diabetes, obesity,
PMID:26134033 physical activity and risk of endometrial cancer among
Newcomer LM, Newcomb PA, Trentham-Dietz A, Mexican women. Cancer Causes Control, 11(8):707–11.
Storer BE (2001). Hormonal risk factors for endo- doi:10.1023/A:1008913619107 PMID:11065007
metrial cancer: modification by cigarette smoking Setiawan VW, Yang HP, Pike MC, McCann SE, Yu H,
(United States). Cancer Causes Control, 12(9):829–35. Xiang YB, et al.; Australian National Endometrial
doi:10.1023/A:1012297905601 PMID:11714111 Cancer Study Group (2013). Type I and II endome-
Niwa Y, Ito H, Matsuo K, Hirose K, Ito N, Mizuno M, et al. trial cancers: have they different risk factors? J Clin
(2007). Lymphotoxin-alpha polymorphisms and the Oncol, 31(20):2607–18. doi:10.1200/JCO.2012.48.2596
risk of endometrial cancer in Japanese subjects. Gynecol PMID:23733771
Oncol, 104(3):586–90. doi:10.1016/j.ygyno.2006.09.007 Song Y-M, Sung J, Ha M (2008). Obesity and risk of
PMID:17045328 cancer in postmenopausal Korean women. J Clin
Okamura C, Tsubono Y, Ito K, Niikura H, Takano T, Oncol, 26(20):3395–402. doi:10.1200/JCO.2007.15.7867
Nagase S, et al. (2006). Lactation and risk of endo- PMID:18612154
metrial cancer in Japan: a case-control study. Tohoku Speliotes EK, Willer CJ, Berndt SI, Monda KL, Thorleifsson
J Exp Med, 208(2):109–15. doi:10.1620/tjem.208.109 G, Jackson AU, et al.; MAGIC; Procardis Consortium
PMID:16434833 (2010). Association analyses of 249,796 individ-
Ollberding NJ, Lim U, Wilkens LR, Setiawan VW, uals reveal 18 new loci associated with body mass
Shvetsov YB, Henderson BE, et al. (2012). Legume, soy, index. Nat Genet, 42(11):937–48. doi:10.1038/ng.686
tofu, and isoflavone intake and endometrial cancer risk PMID:20935630
in postmenopausal women in the Multiethnic Cohort Thomas CC, Wingo PA, Dolan MS, Lee NC, Richardson
Study. J Natl Cancer Inst, 104(1):67–76. doi:10.1093/ LC (2009). Endometrial cancer risk among younger,
jnci/djr475 PMID:22158125 overweight women. Obstet Gynecol, 114(1):22–7.
Park SL, Goodman MT, Zhang ZF, Kolonel LN, Henderson doi:10.1097/AOG.0b013e3181ab6784 PMID:19546754
BE, Setiawan VW (2010). Body size, adult BMI gain Tong SY, Ha SY, Ki KD, Lee JM, Lee SK, Lee KB, et al.
and endometrial cancer risk: the Multiethnic Cohort. (2009). The effects of obesity and HER-2 polymor-
Int J Cancer, 126(2):490–9. doi:10.1002/ijc.24718 phisms as risk factors for endometrial cancer in Korean
PMID:19585578 women. BJOG, 116(8):1046–52. doi:10.1111/j.1471-
Rapp K, Schroeder J, Klenk J, Stoehr S, Ulmer H, Concin 0528.2009.02186.x PMID:19438491
H, et al. (2005). Obesity and incidence of cancer: a Trentham-Dietz A, Nichols HB, Hampton JM, Newcomb
large cohort study of over 145,000 adults in Austria. PA (2006). Weight change and risk of endometrial
Br J Cancer, 93(9):1062–7. doi:10.1038/sj.bjc.6602819 cancer. Int J Epidemiol, 35(1):151–8. doi:10.1093/ije/
PMID:16234822 dyi226 PMID:16278243
Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D; WCRF/AICR (2007). Food, nutrition, physical activity,
Million Women Study Collaboration (2007). Cancer and the prevention of cancer: a global perspective.
incidence and mortality in relation to body mass Washington (DC), USA: American Institute for Cancer
index in the Million Women Study: cohort study. BMJ, Research. Available from: http://www.aicr.org/assets/
335(7630):1134. doi:10.1136/bmj.39367.495995.AE docs/pdf/reports/Second_Expert_Report.pdf.
PMID:17986716 WCRF/AICR (2013). Continuous Update Project Report.
Reeves KW, Carter GC, Rodabough RJ, Lane D, Food, nutrition, physical activity, and the prevention of
McNeeley SG, Stefanick ML, et al. (2011). Obesity in endometrial cancer. Washington (DC), USA: American
relation to endometrial cancer risk and disease char- Institute for Cancer Research. Available from: http://
acteristics in the Women’s Health Initiative. Gynecol w w w.aicr.org/assets/docs/pdf/reports/2013-cup-
Oncol, 121(2):376–82. doi:10.1016/j.ygyno.2011.01.027 endometrial-cancer.pdf.
PMID:21324514
309
Weiss JM, Saltzman BS, Doherty JA, Voigt LF, Chen C,

Beresford SA, et al. (2006). Risk factors for the incidence
of endometrial cancer according to the aggressiveness
of disease. Am J Epidemiol, 164(1):56–62. doi:10.1093/
aje/kwj152 PMID:16675538
Wen W, Cai Q, Xiang YB, Xu WH, Ruan ZX, Cheng J,
et al. (2008). The modifying effect of C-reactive
protein gene polymorphisms on the association
between central obesity and endometrial cancer
risk. Cancer, 112(11):2409–16. doi:10.1002/cncr.23453
PMID:18383516
Xu WH, Matthews CE, Xiang YB, Zheng W, Ruan ZX,
Cheng JR, et al. (2005). Effect of adiposity and fat
distribution on endometrial cancer risk in Shanghai
women. Am J Epidemiol, 161(10):939–47. doi:10.1093/
aje/kwi127 PMID:15870158
Xu WH, Xiang YB, Zheng W, Zhang X, Ruan ZX, Cheng
JR, et al. (2006). Weight history and risk of endome-
trial cancer among Chinese women. Int J Epidemiol,
35(1):159–66. doi:10.1093/ije/dyi223 PMID:16258056
Yang HP, Wentzensen N, Trabert B, Gierach GL, Felix
AS, Gunter MJ, et al. (2013). Endometrial cancer risk
factors by 2 main histologic subtypes: the NIH-AARP
Diet and Health Study. Am J Epidemiol, 177(2):142–51.
doi:10.1093/aje/kws200 PMID:23171881
Yang TYO, Cairns BJ, Allen N, Sweetland S, Reeves GK,
Beral V; Million Women Study (2012). Postmenopausal
endometrial cancer risk and body size in early
life and middle age: prospective cohort study. Br
J Cancer, 107(1):169–75. doi:10.1038/bjc.2012.229
PMID:22644298
Zhang Y, Liu Z, Yu X, Zhang X, Lü S, Chen X, et al.
(2010). The association between metabolic abnormality
and endometrial cancer: a large case-control study
in China. Gynecol Oncol, 117(1):41–6. doi:10.1016/j.
ygyno.2009.12.029 PMID:20096921
310
2.2.12 Cancer of the cervix References

Cancer of the cervix is the fourth most
common cancer in women. Human papilloma- Leon DA, Smeeth L (2014). Body-mass index and risk
virus (HPV) infection, which is present in almost of 22 specific cancers: a population-based cohort study
all cases of cervical cancer, is not related to of 5.24 million UK adults. Lancet, 384(9945):755–65.
adiposity (Wee et al., 2008). In 2001, the Working doi:10.1016/S0140-6736(14)60892-8 PMID:25129328
Brinton LA, Herrero R, Reeves WC, de Britton RC, Gaitan
Group of the IARC Handbook on weight control E, Tenorio F (1993). Risk factors for cervical cancer by
and physical activity (IARC, 2002) concluded histology. Gynecol Oncol, 51(3):301–6. doi:10.1006/
that the evidence of an association between gyno.1993.1294 PMID:8112636
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ
avoidance of weight gain and cervical cancer was (2003). Overweight, obesity, and mortality from cancer
inadequate. in a prospectively studied cohort of U.S. adults. N Engl
J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423
(a) Cohort studies PMID:12711737
Cusimano R, Dardanoni G, Dardanoni L, La Rosa M,
Since 2001, at least eight cohort studies of Pavone G, Tumino R, et al. (1989). Risk factors of female
cervical cancer and body weight (Wolk et al., cancers in Ragusa population (Sicily) – 1. Endometrium
2001; Calle et al., 2003; Rapp et al., 2005; Reeves and cervix uteri cancers. Eur J Epidemiol, 5(3):363–71.
doi:10.1007/BF00144839 PMID:2792311
et al., 2007; Song et al., 2008; Ulmer et al., 2012; IARC (2002). Weight control and physical activity. Lyon,
Lee et al., 2013; Bhaskaran et al., 2014) and one France: IARC Press (IARC Handbooks of Cancer
pooled analysis of 39 cohort studies (Parr et al., Prevention, Vol. 6). Available from: http://publications.
2010) have been published (Table 2.2.12a; web iarc.fr/376.
Lacey JV Jr, Swanson CA, Brinton LA, Altekruse SF,
only; available at: http://publications.iarc.fr/570). Barnes WA, Gravitt PE, et al. (2003). Obesity as a poten-
Although some studies reported statistically tial risk factor for adenocarcinomas and squamous cell
significant increases, the data overall remained carcinomas of the uterine cervix. Cancer, 98(4):814–21.
inconsistent. doi:10.1002/cncr.11567 PMID:12910527
Lee JK, So KA, Piyathilake CJ, Kim MK (2013). Mild
obesity, physical activity, calorie intake, and the risks
(b) Case–control studies of cervical intraepithelial neoplasia and cervical
The five case–control studies assessing the cancer. PLoS One, 8(6):e66555. doi:10.1371/journal.
pone.0066555 PMID:23776686
association between body fatness and cervical Máchová L, Cízek L, Horáková D, Koutná J, Lorenc J,
cancer (Cusimano et al., 1989; Brinton et al., 1993; Janoutová G, et al. (2007). Association between obesity
Ursin et al., 1996; Lacey et al., 2003; Máchová and cancer incidence in the population of the District
et al., 2007) had relatively small sample sizes Sumperk, Czech Republic. Onkologie, 30(11):538–42.
doi:10.1159/000108284 PMID:17992023
(< 150 cases), and the results are inconsistent Parr CL, Batty GD, Lam TH, Barzi F, Fang X, Ho SC,
(Table 2.2.12b; web only; available at: http:// et al.; Asia-Pacific Cohort Studies Collaboration (2010).
publications.iarc.fr/570). Body-mass index and cancer mortality in the Asia-
of 424,519 participants. Lancet Oncol, 11(8):741–52.
doi:10.1016/S1470-2045(10)70141-8 PMID:20594911
Rapp K, Schroeder J, Klenk J, Stoehr S, Ulmer H, Concin
large cohort study of over 145,000 adults in Austria.
PMID:16234822
Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D;
Million Women Study Collaboration (2007). Cancer
incidence and mortality in relation to body mass
311
index in the Million Women Study: cohort study. BMJ,

335(7630):1134. doi:10.1136/bmj.39367.495995.AE
PMID:17986716
Song YM, Sung J, Ha M (2008). Obesity and risk of
cancer in postmenopausal Korean women. J Clin
Oncol, 26(20):3395–402. doi:10.1200/JCO.2007.15.7867
PMID:18612154
Ulmer H, Bjørge T, Concin H, Lukanova A, Manjer J,
Hallmans G, et al. (2012). Metabolic risk factors and
cervical cancer in the Metabolic Syndrome and Cancer
Project (Me-Can). Gynecol Oncol, 125(2):330–5.
doi:10.1016/j.ygyno.2012.01.052 PMID:22330614
Ursin G, Pike MC, Preston-Martin S, d’Ablaing G 3rd,
Peters RK (1996). Sexual, reproductive, and other risk
factors for adenocarcinoma of the cervix: results from
a population-based case-control study (California,
United States) Cancer Causes Control, 7(3):391–401.
doi:10.1007/BF00052946 PMID:8734834
Wee CC, Huang A, Huskey KW, McCarthy EP (2008).
Obesity and the likelihood of sexual behavioral risk
factors for HPV and cervical cancer. Obesity (Silver
Spring), 16(11):2552–5. doi:10.1038/oby.2008.394
PMID:18719677
Control, 12(1):13–21. doi:10.1023/A:1008995217664
PMID:11227921
312
2.2.13 Cancer of the ovary women and of 23% in obese women compared
with women of normal BMI (Liu et al., 2015).
Cancer of the ovary accounts for about 4% of Aune et al. (2015), in a meta-analysis including
all cancer diagnoses in women. Risk of ovarian 25 prospective studies, found a summary relative
cancer is known to be reduced with use of oral risk per 5 kg/m2 increase in BMI of 1.07 (95% CI,
contraceptives, and increased with BRCA gene 1.03–1.11) [moderate heterogeneity (54%) across
mutations and use of estrogen (unopposed) studies was reported] (Aune et al., 2015).
HRT. There are histologically distinct subtypes The association is stronger in never-users of
of ovarian cancer, including serous, mucinous, HRT (Leitzmann et al., 2009). The Collaborative
clear cell, endometrioid, and other/mixed types Group on Epidemiological Studies of Ovarian
(Jayson et al., 2014). Cancer found the relative risk per 5 kg/m2
In 2001, the Working Group of the IARC increase in BMI to be 1.10 (95% CI, 1.07–1.13;
Handbook on weight control and physical activity Ptrend = 0.02) in never-users of HRT, but 0.95
(IARC, 2002) concluded that the evidence of an (95% CI, 0.92–0.99; Ptrend = 0.02) in ever-users of
association between avoidance of weight gain HRT (Collaborative Group on Epidemiological
and ovarian cancer was inadequate. The 2007 Studies of Ovarian Cancer, 2012).
WCRF review did not draw any conclusions The Collaborative Group on Epidemiological
regarding body fatness and ovarian cancer risk Studies of Ovarian Cancer (2012) also examined
(WCRF/AICR, 2007). On the basis of many more the relationship between BMI and ovarian cancer
studies, including pooled analyses, the WCRF risk separately by histological type. The associ-
Continuous Update Project in 2014 concluded ation was broadly similar across the common
that there was a small but convincing positive histological subtypes of ovarian cancer, except
association between BMI and ovarian cancer risk, for serous tumours of borderline malignancy, for
but limited and inconsistent evidence regarding which the association was considerably greater
waist circumference (WCRF/AICR, 2014). than for the other tumour subtypes.
Table 2.2.13a, Table 2.2.13b, and Table 2.2.13c There was no consistency in the evidence for
present the findings from cohort studies, case– whether BMI earlier in life is more or less predic-
control studies, and meta-analyses, respectively, tive of ovarian cancer than is BMI at a later age.
published since 2000. Findings are presented The systematic review by Aune et al. (2015) and
by BMI at baseline, with comments on findings a twin cohort study by Lundqvist and collabo-
according to weight change over the life-course rators (Lundqvist et al., 2007) found marginally
and waist circumference. stronger associations with BMI in early adult-
(a) Cohort studies hood than with BMI later in life. However, a
pooled analysis including 13 548 cases found the
The evidence published since 2000 includes opposite (Olsen et al., 2013). Two cohort studies
15 cohort studies (excluding analyses that were examining weight gain from age 18–20 years
later updated and analyses based on fewer than reported positive associations (Ma et al., 2013
100 incident cases) (Table 2.2.13a) and several based on 152 cases; Ptrend = 0.05; Canchola et al.,
meta-analyses of cohort studies (Table 2.2.13c). 2010), whereas the meta-analyses by Aune et al.
In general, findings were consistent across (2015) based on 6 cohort studies and 1338 cases
studies, suggesting a modest positive association did not find evidence of this association [signif-
between baseline BMI and ovarian cancer risk. A icant heterogeneity was reported in this study;
meta-analysis including 13 cohort studies found P heterogeneity = 0.01].
significant increases in risk of 7% in overweight
313
In three of the four cohorts that included the associations were stronger among premeno-
measurements of waist circumference, this was pausal women who had never used HRT.
found to be less associated with ovarian cancer In the few studies that examined the rela-
risk than was BMI (Chionh et al., 2010; Lahmann tionship between BMI and ovarian cancer risk
et al., 2010; Ma et al., 2013); one study showed separately by histological type, the associa-
significant positive associations stronger than tions seemed to be confined to non-serous and
those reported with BMI (Canchola et al., 2010). low-grade serous tumours (Olsen et al., 2013).
An earlier pooled analysis of 10 case–control
(b) Case–control studies studies found no association for serous cancers,
A total of 35 case–control studies (including but there was an association for all other ovarian
7 hospital-based studies) from Asia, Australia, cancer types (Kurian et al., 2005). The risk was
Canada, Europe, and the USA and several significantly increased in both invasive and
meta-analyses including case–control studies borderline ovarian cancer subtypes, with a
have been published since 2000 on the associ- somewhat stronger association with borderline
ation between BMI at diagnosis and ovarian tumours (Olsen et al., 2013).
cancer risk (Table 2.2.13b and Table 2.2.13c). An Among the 10 studies that reported on the
increase in risk was generally observed, although association between BMI in young adulthood
estimates were not statistically significant in and ovarian cancer risk, 7 observed a non-
most individual studies. However, a meta-anal- significant increase in risk, two observed a
ysis including 13 case–control studies and significant increase in risk (Lubin et al., 2003;
presenting low heterogeneity (I2 = 11.3%) found Olsen et al., 2013), and one observed a significant
significant increased risk of ovarian cancer in decrease in risk (Kuper et al., 2002). Four studies
overweight women (RR, 1.09; 95% CI, 1.00–1.19) evaluated BMI change between early adulthood
and in obese women (RR, 1.31; 95% CI, 1.12–1.54) and diagnosis and showed no significant associ-
compared with women of normal BMI (Liu et al., ation with ovarian cancer risk (Lubin et al., 2003;
2015). Another meta-analysis of 47 epidemio- Zhang et al., 2005; Greer et al., 2006; Peterson
logical studies, which included 30 case–control et al., 2006).
studies, showed a significant 5% increase in risk
in those studies with population-based controls (c) Mendelian randomization studies
(n = 17) and a significant 8% decrease in risk One large-scale Mendelian randomization
in those studies with hospital-based controls study has been conducted to assess the associ-
(n = 13) [the decreased risk in hospital-based ation of childhood and adult BMI with ovarian
studies is probably due to selection bias related to cancer risk, separated into histological subtypes
BMI] (Collaborative Group on Epidemiological including clear cell, endometrioid, and serous
Studies of Ovarian Cancer, 2012). cancer (Gao et al., 2016; Table 2.2.13d). With each
When stratifying by menopausal status 1 kg/m2 increase in adult BMI (assuming that a
or HRT use, the Collaborative Group on standard deviation was equivalent to 4.5 kg/m2),
Epidemiological Studies of Ovarian Cancer there was evidence for an increased risk of all
(2012) reported a significant interaction with ovarian cancer (OR, 1.07; 95% CI, 1.01–1.13;
HRT use, with evidence of a 10% increased risk P = 0.02) and weak, not statistically significant,
only among never-users of HRT (n = 11 456 evidence for an increased risk of clear cell ovarian
cases). A pooled analysis from 15 case–control cancer (OR, 1.12; 95% CI, 0.96–1.31; P = 0.14)
studies (Olsen et al., 2013) also reported that and serous ovarian cancer (OR, 1.06; 95% CI,
0.99–1.13; P = 0.09). There was no evidence for
314
statistically significant associations between

childhood BMI and risk of any ovarian cancer
types.
In sensitivity analyses exploring the validity
of the genetic variants used, there was evidence
for negative pleiotropy in the association between
adult BMI and endometrioid ovarian cancer [thus
suggesting that the positive association may have
been underestimated in the main analyses].
315
316

Table 2.2.13a Cohort studies of measures of body fatness and cancer of the ovary
Reference Total number of Exposure Exposed Relative risk Covariates Comments

Cohort women categories cases (95% CI)
Location Incidence/
Calle et al. (2003) 495 477 BMI Age, education level, Women who had either a
Population-based cohort Mortality 18.5–24.9 873 1.00 smoking, physical activity, hysterectomy or ovarian
USA 25–29.9 437 1.15 (1.02–1.29) alcohol consumption, surgery were excluded
1982–1998 30–34.9 126 1.16 (0.96–1.40) marital status, aspirin use,
35–39.9 49 1.51 (1.12–2.02) fat intake, vegetable intake,
[Ptrend] [0.001] HRT use
Rapp et al. (2005) 78 484 BMI Age, smoking, occupation
Population-based cohort Incidence 18.5–24.9 61 1.0
Austria 25.0–29.9 39 1.03 (0.68–1.56)
1985–2002 ≥ 30 21 1.25 (0.75–2.08)
[Ptrend] [0.44]
Lacey et al. (2006) 46 026 BMI Age, race, menopausal
Breast Cancer Detection Incidence < 18.5 7 0.95 (0.45–2.01) status, parity, OC use, HRT
Demonstration Project 18.5–24.9 219 1.00 use
Follow-Up Study 25.0–29.9 83 1.00 (0.78–1.29)
USA 30–34.9 20 0.94 (0.59–1.48)
1973–1997 ≥ 35 11 1.55 (0.84–2.84)
per 1 kg/m2 1.01 (0.98–1.03)
Lundqvist et al. (2007) 14 058 twins BMI at baseline Age, country, smoking,
Twin cohort studies (mean age, 56 yr) < 18.5 1 0.4 (0.1–2.6) physical activity, education
Sweden and Finland Incidence 18.5–24.9 86 1.0 level, diabetes, parity
1961–2004 25–29.9 57 1.2 (0.8–1.6)
≥ 30 7 0.7 (0.3–1.5)
[Ptrend] [0.95]
22 432 twins BMI at baseline Age, smoking, physical
(mean age, 30 yr) < 18.5 8 0.7 (0.3–1.4) activity, education level,
Incidence 18.5–24.9 120 1.0 diabetes, parity
25–29.9 31 1.5 (1.0–2.3)
≥ 30 3 0.8 (0.2–2.6)
[Ptrend] [0.01]

Location Incidence/
Reeves et al. (2007) 1.2 million BMI Age, region, SES,
Million Women Study Incidence < 22.5 478 0.98 (0.89–1.07) reproductive history,
United Kingdom 22.5–24.9 631 1.00 smoking, alcohol
1996–2001 25.0–27.4 510 0.99 (0.91–1.08) consumption, physical
27.5–29.9 349 1.13 (1.02–1.25) activity, HRT use
≥ 30 438 1.12 (1.02–1.23)
per 10 kg/m2 1.14 (1.03–1.27)
Schouten et al. (2008) 531 583 BMI Postmenopausal:
Pooling Project of Incidence < 23 426 1.0
Prospective Studies of Diet 23–24.9 291 0.91 (0.78–1.06)
and Cancer (12 cohorts 25.0–26.9 222 0.95 (0.80–1.13)
pooled) 27–29.9 206 0.96 (0.80–1.14)
North America and ≥ 30 191 1.07 (0.87–1.33)
western Europe [Ptrend] [0.53]
Follow-up varied by cohort BMI Premenopausal:
< 23 64 1.0
23–24.9 34 1.29 (0.83–2.00)
25.0–26.9 14 0.95 (0.50–1.81)
27–29.9 14 1.28 (0.59–2.79)
≥ 30 22 1.72 (1.02–2.29)
[Ptrend] [0.13]
Song et al. (2008) 107 481, BMI Age, smoking, alcohol Ovary and other
Korean medical insurance postmenopausal < 18.5 3 0.98 (0.29–3.24) consumption, physical unspecified female genital

cohort Incidence 18.5–20.9 13 0.85 (0.43–1.68) exercise, income level at organs
Republic of Korea 21–22.9 30 1.00 study entry
1994–2003 23.0–24.9 53 1.63 (1.01–2.63)
25.0–26.7 42 1.62 (0.98–2.67)
27.0–29.9 30 1.57 (0.91–2.73)
≥ 30 5 0.93 (0.32–2.67)
per 1 kg/m2 1.04 (0.99–1.09)
Leitzmann et al. (2009) 94 525 BMI Never-users of HRT: Age, race/ethnicity, family
NIH-AARP cohort Incidence < 25 39 1.00 history, OC use, physical
USA 25–29.9 43 1.39 (0.89–2.14) activity
1996–2003 ≥ 30 43 1.83 (1.18–2.84)
[Ptrend] [0.007]
317
318


Location Incidence/
Leitzmann et al. (2009) BMI Ever-users of HRT:
(cont.) < 25 102 1.00
25–29.9 43 0.68 (0.48–0.98)
≥ 30 33 0.96 (0.65–1.43)
[Ptrend] [0.53]
Canchola et al. (2010) 56 091 BMI Race, OC use, parity, wine Weight gain from age
California Teachers Study Never-users of < 25 57 1.0 intake, physical activity, 18 yr to baseline positively
Cohort HRT 25–29.9 29 1.1 (0.71–1.8) smoking, tubal ligation associated
USA Incidence ≥ 30 21 1.2 (0.72–2.0)
1995–2007 WC (in)
< 35 32 1.0
≥ 35 29 1.8 (1.1–3.0)
Chionh et al. (2010) 18 700 BMI Country of birth,
Melbourne Collaborative Incidence < 25 39 1.00 education level, age at
Cohort Study 25–29.9 40 1.05 (0.66–1.65) menarche, parity, OC use,
Australia ≥ 30 34 1.58 (0.96–2.62) hysterectomy, tobacco use,
1990–2008 per 5 kg/m2 1.22 (1.00–1.48) physical activity, energy
[Ptrend] [0.06] intake from diet
WC, quartiles
Q1 24 1.00
Q2 27 0.97 (0.56–1.69)
Q3 30 1.03 (0.59–1.78)
Q4 32 0.96 (0.54–1.69)
[Ptrend] [0.71]
Kotsopoulos et al. (2010) 182 700 BMI Age, age at menarche,
Nurses’ Health Study 1 Incidence < 21 125 1.00 parity, OC use, tubal
and 2 21–22.9 155 0.97 (0.77–1.23) ligation, height, family
USA 23–24.9 168 1.02 (0.81–1.29) history of breast or ovarian
1976–2006 25.0–29.9 242 0.96 (0.77–1.19) cancer, caffeine intake,
≥ 30 177 1.12 (0.89–1.42) hysterectomy; for WC,
[Ptrend] [0.29] additionally adjusted for
BMI

Location Incidence/
Kotsopoulos et al. (2010) WC (in)
(cont.) < 28 67 1.0
28–29.9 65 0.91 (0.64–1.29)
30–31.9 56 0.89 (0.61–1.30)
32–34.9 68 0.90 (0.61–1.33)
≥ 35 79 1.00 (0.62–1.88)
[Ptrend] [0.65]
Lahmann et al. (2010) 226 798 BMI Age, parity, age at Stronger association in
EPIC cohort Incidence < 25 287 1.00 menarche, smoking, OC use postmenopausal women
Europe 25–29.9 211 1.14 (0.94–1.37) than in premenopausal
1992–2007 ≥ 30 113 1.33 (1.05–1.68) women
[Ptrend] [0.02]
WC, quartiles Similar association in
Q1 122 1.00 premenopausal and
Q2 155 1.03 (0.81–1.31) postmenopausal women
Q3 175 1.10 (0.87–1.41)
Q4 159 1.12 (0.86–1.45)
[Ptrend] [0.32]
Yang et al. (2012) 169 391 BMI Age, OC use, HRT use, Stronger association with
NIH-AARP cohort Incidence < 30 617 1.00 parity endometrioid histological
USA ≥ 30 197 1.15 (0.98–1.35) subtype
1995–2006
Ma et al. (2013) 70 258 BMI Age, education level Weight gain from age

Shanghai Women’s Health Incidence < 18.5 7 1.73 (0.80–3.75) 20 yr also positively
Study (SWHS) (population- 18.5–24.9 75 1.00 associated with risk
based cohort) 25.0–29.9 55 1.49 (1.05–2.13)
Shanghai, China ≥ 30 15 2.42 (1.37–4.28)
1996–2009 [Ptrend] [0.008]
WC, quartiles Age, education level
Q1 27 1.00
Q2 34 1.36 (0.82–2.26)
Q3 41 1.50 (0.92–2.46)
Q4 50 1.61 (0.98–2.64)
[Ptrend] [0.06]
319
320


Location Incidence/
Bhaskaran et al. (2014) 2 864 658 BMI 3684 Age, sex, year, diabetes, Similar association in
Health system clinical Incidence per 5 kg/m2 1.09 (1.04–1.14) alcohol consumption, never-smokers
database smoking, SES
United Kingdom
1987–2012
Gay et al. (2015) 28 234 BMI Age, housing, family
Singapore Breast Cancer Incidence < 18.5 6 1.96 (0.64–5.97) history of breast cancer
Screening Project (SBCSP) 18.5–22.9 28 1.00
Singapore 23–27.4 56 1.34 (0.69–2.58)
1994–2012 ≥ 27.5 17 0.55 (0.19–1.55)
[Ptrend] [0.22]
BMI, body mass index (in kg/m 2); CI, confidence interval; EPIC, European Prospective Investigation into Cancer and Nutrition; HRT, hormone replacement therapy; NIH-AARP,
National Institutes of Health–AARP Diet and Health Study; OC, oral contraceptive; SES, socioeconomic status; WC, waist circumference; yr, year or years
Table 2.2.13b Case–control studies of measures of body fatness and cancer of the ovary
Period Source of
controls
Greggi et al. 440 BMI Age, education level,
(2000) Hospital < 22.5 118 1.0 parity, OC use, family
Italy 22.5–26 129 (0.8–1.5) history of ovarian cancer
1998 > 26 140 (0.8–1.6)
Purdie et al. 775 BMI, percentiles 518 total Age, age squared, Stronger risks were
(2001) Population < 15th 1.0 (0.7–1.6) geographical location, observed in premenopausal
Australia 15th–35th 1.5 (1.0–2.2) education level, parity, women above the 65th
1990–1993 35th–65th 1.0 duration of OC use, percentile
65th–85th 1.3 (0.9–1.9) smoking history, ever-
≥ 85th 1.7 (1.1–2.6) use of talc in the perineal
[Ptrend] [0.12] region, tubal sterilization,
hysterectomy, history of
breast or ovarian cancer in
a first-degree relative
Dal Maso et al. 1031 BMI Age, education level, A significant association
(2002) Hospital < 21 143 1.00 parity, OC use was observed with waist-
Italy 21– < 25 406 0.99 (0.77–1.27) to-hip ratio. No association
1992–1999 25– < 30 299 0.76 (0.58–0.99) was observed with
≥ 30 173 1.07 (0.79–1.44) increased body weight
[Ptrend] [0.53]
Kuper et al. 563 BMI Age, site, parity, OC use, In stratified analyses, a
(2002) Population < 20 67 1.00 family history of breast, higher risk with BMI and
USA ≥ 20– < 25 255 0.97 (0.64–1.45) ovarian, or prostate cancer weight was observed in

1992–1997 ≥ 25– < 30 138 1.02 (0.65–1.60) in a first-degree relative, premenopausal women
≥ 30 104 1.24 (0.77–2.01) tubal ligation, education
level, marital status
Lubin et al. 1269 BMI at age 18 yr
(2003) Population < 19.1 1.00
Israel 19.1–20.9 1.16 (0.89–1.51)
1994–1999 21.0–22.8 1.13 (0.87–1.48)
22.9–35.2 1.42 (1.08–1.85)
[Ptrend] [0.009]
321
322

Period Source of
controls
Lubin et al. BMI change from age 18 yr
(2003) < 0.73 1.00
(cont.) 0.73–2.70 0.82 (0.63–1.06)
2.71–5.71 0.79 (0.60–1.03)
≥ 5.72 0.91 (0.69–1.20)
[Ptrend] [0.50]
Yen et al. (2003) 86 BMI Age, income during
Taiwan, China Hospital < 25 63 1.00 marriage, education level
1993–1998 ≥ 25 23 0.77 (0.45–1.33)
Pan et al. (2004) 442 BMI 442 total 5-year age group,
Canada Population < 25 1.00 province of residence,
1994–1997 25– < 30 1.16 (0.90–1.50) education level, pack-
≥ 30 1.95 (1.44–2.64) years of smoking, alcohol
consumption, total energy
intake, vegetable intake,
dietary fibre intake,
recreational physical
activity, menopausal
status, number of live
births, age at menarche,
age at end of first
pregnancy
Pike et al. (2004) 477 BMI Ethnicity, age, education
USA Population < 25 261 1.00 level, SES, family history
1992–1998 25–29 120 0.97 (0.71–1.33) of ovarian cancer, tubal
30–34 56 1.29 (0.83–1.99) legation, use of talc in the
≥ 35 40 1.46 (0.87–2.44) genital area, nulliparity,
age at last birth, number
of births, number of
incomplete pregnancies,
OC use, menopausal
status, age at natural
menopause, age at surgical
menopause, HRT use
Period Source of
controls
Riman et al. 655 BMI 1 yr ago Age, parity, and age at Stronger associations were
(2004) Population < 22 122 1.00 menopause as categorized observed for the mucinous
Sweden 22– < 25 197 0.99 (0.77–1.28) variables, duration of OC histological subgroup, and
1993–1995 25– < 27 127 1.06 (0.80–1.40) use, ever-use of HRT no associations for the
27– < 30 115 1.10 (0.83–1.46) serous and endometrioid
≥ 30 93 1.37 (1.01–1.85) types
Hoyo et al. 593 BMI Race, age, parity, history Positive non-significant
(2005) Population < 25 230 1.0 of ovarian cancer, associations with weight
USA 25–29.99 158 1.0 (0.7–1.3) history of breast cancer, gain from age 18 yr (3rd
1999–2003 ≥ 30 192 1.4 (1.0–1.8) hysterectomy, OC use, tertile, 204 cases) and with
menstrual status WC (3rd tertile, 213 cases).
In stratified analyses,
associations with recent
BMI were only significant
among Whites (vs African
Americans)
Kurian et al. 1834 cases BMI Serous: 241 Parity, OC use
(2005) with invasive < 24 1.00
Pooled analysis epithelial ≥ 24 0.72 (0.59–0.88)
of 10 case– ovarian cancer BMI Mucinous: 57
control studies of Serous: 1067 < 24 1.0
ovarian cancer in Mucinous: 254 ≥ 24 1.3 (0.88–2.0)
the USA Endometrioid: BMI Endometrioid: 82

373 < 24 1.0
Clear cell: 140 ≥ 24 1.3 (0.95–1.9)
Controls: 7
BMI Clear cell: 28
population, 3
< 24 1.0
hospital
≥ 24 0.9 (0.55–1.6)
Zhang et al. 254 BMI at diagnosis Age at diagnosis, locality, No significant associations
(2005) Hospital < 18.5 93 1.60 (0.91–2.83) tobacco smoking, alcohol were observed with body
China 18.5–21.9 28 1.00 consumption, parity, weight at diagnosis or with
1999–2000 22.0–24.9 86 0.98 (0.69–1.41) menopausal status, HRT, BMI/weight change.
≥ 25.0 47 0.88 (0.57–1.34) OC use, ovarian cancer in Statistically significant
[Ptrend] [0.19] first-degree relatives, total associations with BMI and
energy intake weight were observed 5 yr
before diagnosis
323
324

Period Source of
controls
Zhang et al. BMI at age 21 yr
(2005) < 18.5 134 0.94 (0.62–1.45)
(cont.) 18.5–21.9 41 1.00
22.0–24.9 66 1.04 (0.73–1.50)
≥ 25.0 11 1.20 (0.56–2.56)
[Ptrend] [0.37]
Beehler et al. 427 ΒΜΙ Age, geographical area,
(2006) Hospital ≤ 24.9 229 1.00 year of study participation
USA 25.0–29.9 116 1.02 (0.77–1.36)
1982–1998 ≥ 30.0 82 1.17 (0.84–1.65)
Greer et al. 762 BMI, quartiles Age, race, number of live Highest BMI (4th quartile,
(2006) Population Q1 173 1.00 births, family history 69 cases) and adult weight
USA Q2 196 1.10 (0.85–1.44) of ovarian cancer, tubal gain were associated with
1994–1998 Q3 192 1.14 (0.87−1.49) ligation, OC use increased ovarian cancer
Q4 201 1.24 (0.95–1.63) risk among nulliparous
[Ptrend] [0.12] women only
Huusom et al. 202 BMI Age, childbirth, number Significant associations
(2006) Population < 22 67 1.00 of additional births, age at with BMI among the serous
Denmark 22–24 52 0.76 (0.51–1.14) first birth, breastfeeding, histological subgroup only
1995–1999 25–26 29 1.06 (0.64–1.74) duration of OC use,
27–29 29 1.33 (0.80–2.19) smoking, intake of milk
≥ 30 24 1.09 (0.64–1.84)
Peterson et al. 700 Recent BMI Age, state, enrolment Positive, non-significant
(2006) Population < 18.5 13 1.12 (0.62–2.03) period, education level, association with recent
USA 18.5–24.9 304 1.00 family history of breast or weight was reported
1993–2001 25.0–29.9 232 1.23 (0.67–2.23) ovarian cancer, OC use,
30.0 151 1.29 (0.70–2.37) parity, history of bilateral
[Ptrend] [0.15] tubal ligation
Weight change (kg)
Loss 45 1.00 (0.68–1.48)
0–9.06 gain 93 1.00
9.07–15.87 gain 121 0.89 (0.66–1.20)
15.88–23.58 gain 90 0.90 (0.65–1.24)
23.59 gain 85 0.77 (0.56–1.06)
[Ptrend] [0.14]
Period Source of
controls
Rossing et al. 355 BMI 5 yr before diagnosis or reference date Age, race, study site, Similar associations were
(2006) Population < 25 130 1.0 number of full-term observed for BMI and for
USA 25– < 30 96 1.2 (0.9–1.7) births, duration of OC weight at ages 18 yr and
1994–1998 ≥ 30 127 1.5 (0.9–2.4) use, weight/BMI 30 yr
Máchová et al. 174 BMI 174 total Age, smoking,
(2007) Population 18.5– < 25 1.00 hypertension, height
Czech Republic ≥ 25– < 30 1.05 (0.68–1.61)
1987−2002 ≥ 30 1.38 (0.87–2.20)
Olsen et al. Meta-analysis BMI at age 17–20 yr Overall:
(2007) Population ≥ 25 vs < 25 1.22 (1.02–1.45)
Meta-analysis Case–control:
(Australia, North ≥ 25 vs < 25 1.21 (0.97–1.52)
America, western
Europe)
Soegaard et al. 554 BMI at age 30–39 yr, quartiles Age, pregnancy, additional Associations seemed
(2007) Population Q1 124 1.00 pregnancies, duration of somewhat stronger
Denmark Q2 153 1.31 (0.98–1.73) OC use in mucinous and
1995–1999 Q3 114 1.00 (0.74–1.36) endometrioid tumours; no
Q4 138 1.23 (0.92–1.65) association with BMI ≥ 25
in adulthood
Lurie et al. (2008) 274 BMI
USA Population ≤ 18.5 6 1.00
1993–2006 18.5– < 25 141 1.72 (0.64–4.75)

25– < 30 64 1.44 (0.50–4.09)
≥ 30 64 1.63 (0.57–4.71)
Nagle et al. Endometrioid: BMI 1 yr before diagnosis Age, education level,
(2008) 142 Endometrioid: parity, OC use
Australia Clear cell: 90 < 18 2 0.9 (0.2–4.0)
NR Controls: 1508 18.5–24.9 52 1.0
Population 25–29.9 46 1.3 (0.8–2.0)
≥ 30 30 1.2 (0.7–1.9)
[Ptrend] [0.41]
325
326

Period Source of
controls
Nagle et al. Clear cell:
(2008) < 18 3 2.9 (0.8–11.1)
(cont.) 18.5–24.9 23 1.0
25–29.9 27 1.7 (0.9–3.0)
≥ 30 25 2.2 (1.2–4.1)
[Ptrend] [0.01]
Boyce et al. 72 BMI Age, race This study investigated
(2009) Population 20–24.9 14 1.00 granulosa cell tumours
USA 25–29.9 15 1.72 (0.82–3.59)
1988–2008 30–39.9 22 5.02 (2.52–10.0)
> 40 5 6.60 (2.19–19.8)
Delort et al. 55 (with BMI Age BMI at age 20 yr not
(2009) no BRCA < 20 10 1.00 significantly associated
Auvergne, France mutation) 20–25 29 0.88 (0.62–1.26) with increased risk. WC
1996–1999, Mammographic 25.1–30 9 0.78 (0.38–1.60) significantly associated
2005–2006 screening centre > 30 6 0.69 (0.24–2.02) with increased risk
Moorman et al. African BMI White: Age
(2009) American: < 25 312 1.00
USA 143/189 25– < 30 212 0.96 (0.76–1.22)
1999–2008 White: 943/868 30– < 35 114 1.08 (0.80–1.45)
Population ≥ 35 83 1.04 (0.75–1.45)
BMI African American:
< 25 17 1.00
25– < 30 26 0.84 (0.39–1.78)
30– < 35 22 0.94 (0.43–2.07)
≥ 35 42 1.62 (0.79–3.35)
Reis & 217 BMI Not specified
Kizilkayabeji Hospital 18.5–24.99 86 1.00
(2010) ≥ 25 131 1.96 (1.41–2.72)
Turkey [Ptrend] [< 0.001]
2002–2003
Bandera et al. 205 BMI Age
(2011) Population 18.5–25 90 1.00
USA 25–29.9 54 1.07 (0.69–1.65)
2004–2008 30–34.9 36 1.39 (0.83–2.32)
≥ 35 24 1.54 (0.81–2.89)
Period Source of
controls
Bodmer et al. 1611 BMI
(2011) Hospital < 25 562 1.00
United Kingdom 25–29.9 453 1.08 (0.94–1.23)
1995–2009 ≥ 30 293 1.11 (0.95–1.29)
Su et al. (2012) 500 BMI 5 yr ago All: Age, OC use, parity, Asian population cut-offs
China Hospital ≤ 18.49 36 1.00 menopausal status, used for BMI
2006–2008 18.5–22.9 348 1.15 (0.72–1.85) ovarian and/or breast Significant associations
≥ 23 116 1.77 (1.04–3.02) cancer in a first-degree were observed for weight
BMI 5 yr ago Serous: relative, age at menarche, (kg), especially in the
≤ 18.49 15 1.00 smoking status, alcohol serous ovarian cancer
18.5–22.9 175 1.43 (0.77–2.69) consumption; for weight, subtype
≥ 23 60 2.26 (1.13–4.52) additional adjustment for
BMI 5 yr ago Mucinous: height
≤ 18.49 8 1.00
18.5–22.9 58 0.87 (0.38–1.98)
≥ 23 14 1.00 (0.38–2.61)
Su et al. (2012) 500 BMI 5 yr ago, tertiles Age, OC use, parity,
China Hospital vs T1: ≤ 20.00 All: menopausal status,
2006–2008 T2: 20.01–21.88 158 1.24 (0.89–1.72) ovarian or breast
T3: ≥ 21.89 221 1.75 (1.28–2.40) cancer in a first-degree
Serous: relative, age at menarche,
T2: 20.01–21.88 83 1.47 (0.97–2.22) smoking status, alcohol
T3: ≥ 21.89 112 1.98 (1.33–2.95) consumption

Mucinous:
T2: 20.01–21.88 26 1.31 (0.69–2.49)
T3: ≥ 21.89 35 1.84 (1.00–3.38)
Weight (kg), tertiles
vs T1: ≤ 50 All:
T3: ≥ 55.1 187 1.84 (1.34–2.54)
Serous:
T3: ≥ 55.1 100 2.23 (1.50–3.33)
Mucinous:
T3: ≥ 55.1 27 1.67 (0.91–3.06)
327
328

Period Source of
controls
King et al. (2013) 205 BMI Age
USA Population < 25 91 1.00
2001–2008 25–29.9 54 1.07 (0.69–1.65)
30–34.9 36 1.39 (0.83–2.32)
≥ 35 24 1.54 (0.82–2.89)
Olsen et al. 13 548 cases BMI Invasive: Age, parity, OC use, BMI in early adulthood
(2013) Invasive: 8763 < 18.5 183 1.08 (0.84–1.39) family history of breast was significantly associated
Pooled analyses Borderline: 18.5–24.9 4020 1.00 or ovarian cancer in with 8% and 15% increased
of 15 case– 2465 25.0–29.9 2500 1.00 (0.92–1.09) a first-degree relative, risk of invasive and
control studies 1 study 30–34.5 1166 1.06 (0.97–1.16) race/ethnicity where borderline ovarian cancer
hospital -based, 35–39.9 511 1.21 (1.07–1.38) appropriate subtypes, respectively
14 studies ≥ 40 383 1.22 (1.05–1.41)
population- per 5 kg/m2 1.04 (1.00–1.08)
based BMI Borderline:
< 18.5 57 1.13 (0.82–1.55)
18.5–24.9 1080 1.00
25.0–29.9 662 1.23 (1.09–1.39)
30–34.5 379 1.61 (1.40–1.85)
35–39.9 150 1.68 (1.37–2.06)
≥ 40 137 1.96 (1.57–2.46)
per 5 kg/m2 1.18 (1.14–1.23)
Le et al. (2014) 608 BMI Age
Canada Population < 25 330 1.00
2001–2007 25–30 180 0.80 (0.59–1.09)
30–35 57 0.87 (0.54–1.41)
≥ 35 41 0.91 (0.53–1.58)
Schildkraut et al. 403 BMI Age, months of OC use, Study in African American
(2014) Population < 24.9 54 1.00 parity women
USA 25–29.9 95 1.31 (0.86–1.99)
2010–2014 30–34.9 107 1.50 (0.99–2.27)
≥ 35 113 1.27 (0.85–1.91)
Period Source of
controls
Burghaus et al. 289 BMI, tertiles (median) NR Low vs medium: Age, OC use,
(2015) Hospital Low (21.7) 0.99 (0.83–1.17) pregnancies, self-reported
Germany Medium (25.0) High vs medium: endometriosis
2002–2013 High (30.1) 1.26 (1.09–1.46)
High vs low:
1.28 (0.95–1.72)
BMI, body mass index (in kg/m 2); CI, confidence interval; HRT, hormone replacement therapy; NR, not reported; OC, oral contraceptive; WC, waist circumference; yr, year or years

329
330

Table 2.2.13c Meta-analyses of measures of body fatness and cancer of the ovary
Reference Total number of studies Exposure categories Relative risk Adjustment for Comments
Total number of cases (95% CI) confounding
Olsen et al. (2007) 16 studies for adult BMI Adult BMI In adult BMI, no difference
(8 case–control and 8 18.5–24.9 1.00 was observed when
cohort) and 9 for BMI in 25.0–29.9 1.16 (1.01–1.32) stratifying by study design
early adulthood (5 case– ≥ 30 1.30 (1.12–1.50) type
control and 4 cohort) BMI at age 17–20 yr
NR 18.5–24.9 1.00
≥ 25 1.22 (1.02–1.45)
Guh et al. (2009) 9 cohort studies BMI Unadjusted RRs
NR 18.5–24.9 1.00
25.0–29.9 1.18 (1.12–1.23)
≥ 30 1.28 (1.20–1.36)
Collaborative Group on 47 studies (17 prospective BMI Study, age at diagnosis, In stratified analyses,
Epidemiological Studies and 30 case–control) < 22.5 1.00 (0.95–1.05) parity, menopausal associations were only
of Ovarian Cancer 25 157 cases 22.5–24.9 1.05 (1.00–1.11) status/hysterectomy, OC significant among never-
(2012) 25–27.4 1.08 (1.02–1.13) use, HRT use, height users of HRT (RR, ~1.1 for
27.5–29.9 1.07 (0.99–1.17) overweight; ~1.2 for obesity)
≥ 30 1.13 (1.06–1.20)
[Ptrend] [0.01]
Poorolajal et al. (2014) 10 cohort studies and 9 BMI Case–control: NR In stratified analysis
case–control studies 18.5–24.9 1.00 by menopausal status,
NR 25.0–29.9 1.08 (0.90–1.31) stronger associations were
≥ 30 1.27 (1.19–1.35) found in all cases in the
BMI Cohort: premenopausal period
18.5–24.9 1.00
25.0–29.9 1.26 (0.97–1.63)
≥ 30 1.26 (1.06–1.50)
Aune et al. (2015) 25 studies BMI Maximally adjusted HR, Non-linearity, with risk
19 825 cases per 5 kg/m2 increase 1.07 (1.03–1.11) RR, or OR were used increasing significantly
(covariates NR) from BMI above 28 kg/m2;
relatively stronger risk
with BMI increase in
early adulthood, based on
6 studies (RR, 1.12); no
association with weight gain
Reference Total number of studies Exposure categories Relative risk Adjustment for Comments
Total number of cases (95% CI) confounding
Liu et al. (2015) 26 studies (13 case– BMI Case–control: No associations with
control and 13 cohort) 18.5–24.9 1.00 BMI were found in
12 963 cases 25.0–29.9 1.09 (1.00–1.18) postmenopausal women
≥ 30 1.31 (1.21–1.54)
BMI Cohort:
18.5–24.9 1.00
25.0–29.9 1.07 (1.01–1.13)
≥ 30 1.23 (1.10–1.39)
BMI Overall:
18.5–24.9 1.00
25.0–29.9 1.07 (1.02–1.12)
≥ 30 1.28 (1.16–1.41)
BMI, body mass index (in kg/m 2); CI, confidence interval; HR, hazard ratio; HRT, hormone replacement therapy; NR, not reported; OC, oral contraceptive; OR, odds ratio; RR, relative
risk; yr, year or years

331
332

Table 2.2.13d Mendelian randomization studies of measures of body fatness and cancer of the ovary
Reference Characteristics of Sample size Exposure (unit) Odds ratio Comments

Study study population (95% CI)
Ptrend
Gao et al. (2016) Women from 13 492 (4369 Increase of 1 SD in genetically Childhood BMI: Similar associations were found for adult
Genetic 3 studies of cases and predicted childhood BMI or 1.07 (0.82–1.39) BMI with serous ovarian cancer, and
Associations and individuals of 9123 controls) adult BMI Ptrend = 0.62 moderate but not statistically significant
Mechanisms European ancestry Adult BMI: with clear cell and endometrioid histological
in Oncology 1.07 (1.01–1.13) subtypes. No associations were observed
(GAME-ON) Ptrend = 0.02 between childhood BMI and subtypes of
Consortium ovarian cancer
BMI, body mass index (in kg/m 2); CI, confidence interval; SD, standard deviation
References Collaborative Group on Epidemiological Studies of

Ovarian Cancer (2012). Ovarian cancer and body size:
individual participant meta-analysis including 25,157
Aune D, Navarro Rosenblatt DA, Chan DS, Abar L, women with ovarian cancer from 47 epidemiological
Vingeliene S, Vieira AR, et al. (2015). Anthropometric studies. PLoS Med, 9(4):e1001200. doi:10.1371/journal.
factors and ovarian cancer risk: a systematic review pmed.1001200 PMID:22606070
and nonlinear dose-response meta-analysis of prospec- Dal Maso L, Franceschi S, Negri E, Conti E, Montella M,
tive studies. Int J Cancer, 136(8):1888–98. doi:10.1002/ Vaccarella S, et al. (2002). Body size indices at different
ijc.29207 PMID:25250505 ages and epithelial ovarian cancer risk. Eur J Cancer,
Bandera EV, King M, Chandran U, Paddock LE, 38(13):1769–74. doi:10.1016/S0959-8049(02)00155-7
Rodriguez-Rodriguez L, Olson SH (2011). Phyto PMID:12175694
estrogen consumption from foods and supplements Delort L, Kwiatkowski F, Chalabi N, Satih S, Bignon
and epithelial ovarian cancer risk: a population-based YJ, Bernard-Gallon DJ (2009). Central adiposity as a
case control study. BMC Womens Health, 11(40):40. major risk factor of ovarian cancer. Anticancer Res,
doi:10.1186/1472-6874-11-40 PMID:21943063 29(12):5229–34. PMID:20044641
Beehler GP, Sekhon M, Baker JA, Teter BE, McCann SE, Gao C, Patel CJ, Michailidou K, Peters U, Gong J, Schildkraut
Rodabaugh KJ, et al. (2006). Risk of ovarian cancer J, et al.; the Colorectal Transdisciplinary Study
associated with BMI varies by menopausal status. J (CORECT); Discovery, Biology and Risk of Inherited
Nutr, 136(11):2881–6. PMID:17056817 Variants in Breast Cancer (DRIVE); Elucidating Loci
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, Involved in Prostate Cancer Susceptibility (ELLIPSE);
Leon DA, Smeeth L (2014). Body-mass index and risk Follow-up of Ovarian Cancer Genetic Association and
of 22 specific cancers: a population-based cohort study Interaction Studies (FOCI); and Transdisciplinary
of 5.24 million UK adults. Lancet, 384(9945):755–65. Research in Cancer of the Lung (TRICL) (2016).
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 Mendelian randomization study of adiposity-related
Bodmer M, Becker C, Meier C, Jick SS, Meier CR (2011). traits and risk of breast, ovarian, prostate, lung and
Use of metformin and the risk of ovarian cancer: a colorectal cancer. Int J Epidemiol, 45(3):896–908.
case-control analysis. Gynecol Oncol, 123(2):200–4. doi:10.1093/ije/dyw129 PMID:27427428
doi:10.1016/j.ygyno.2011.06.038 PMID:21802715 Gay GMW, Lim JS, Chay WY, Chow KY, Tan MH, Lim
Boyce EA, Costaggini I, Vitonis A, Feltmate C, Muto M, WY (2015). Reproductive factors, adiposity, breast-
Berkowitz R, et al. (2009). The epidemiology of ovarian feeding and their associations with ovarian cancer in
granulosa cell tumors: a case-control study. Gynecol an Asian cohort. Cancer Causes Control, 26(11):1561–
Oncol, 115(2):221–5. doi:10.1016/j.ygyno.2009.06.040 73. doi:10.1007/s10552-015-0649-6 PMID:26342607
PMID:19664811 Greer JB, Modugno F, Ness RB, Allen GO (2006).
Burghaus S, Häberle L, Schrauder MG, Heusinger K, Thiel Anthropometry and the risk of epithelial ovarian
FC, Hein A, et al. (2015). Endometriosis as a risk factor cancer. Cancer, 106(10):2247–57. doi:10.1002/cncr.21830
for ovarian or endometrial cancer - results of a hospi- PMID:16596653
tal-based case-control study. BMC Cancer, 15(1):751. Greggi S, Parazzini F, Paratore MP, Chatenoud L, Legge F,
doi:10.1186/s12885-015-1821-9 PMID:26487094 Mancuso S, et al. (2000). Risk factors for ovarian cancer
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ in central Italy. Gynecol Oncol, 79(1):50–4. doi:10.1006/
(2003). Overweight, obesity, and mortality from cancer gyno.2000.5909 PMID:11006030
in a prospectively studied cohort of U.S. adults. N Engl Guh DP, Zhang W, Bansback N, Amarsi Z, Birmingham
J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423 CL, Anis AH (2009). The incidence of co-morbidi-
PMID:12711737 ties related to obesity and overweight: a systematic
Canchola AJ, Chang ET, Bernstein L, Largent JA, review and meta-analysis. BMC Public Health, 9(1):88.
Reynolds P, Deapen D, et al. (2010). Body size and doi:10.1186/1471-2458-9-88 PMID:19320986
the risk of ovarian cancer by hormone therapy use in Hoyo C, Berchuck A, Halabi S, Bentley RC, Moorman
the California Teachers Study cohort. Cancer Causes P, Calingaert B, et al. (2005). Anthropometric meas-
Control, 21(12):2241–8. doi:10.1007/s10552-010-9647-x urements and epithelial ovarian cancer risk in
PMID:20924664 African-American and White women. Cancer Causes
Chionh F, Baglietto L, Krishnan K, English DR, MacInnis Control, 16(8):955–63. doi:10.1007/s10552-005-3205-y
RJ, Gertig DM, et al. (2010). Physical activity, body size PMID:16132804
and composition, and risk of ovarian cancer. Cancer Huusom LD, Frederiksen K, Høgdall EV, Glud E,
Causes Control, 21(12):2183–94. doi:10.1007/s10552- Christensen L, Høgdall CK, et al. (2006). Association of
010-9638-y PMID:20827504 reproductive factors, oral contraceptive use and selected
lifestyle factors with the risk of ovarian borderline
tumors: a Danish case-control study. Cancer Causes
333
Control, 17(6):821–9. doi:10.1007/s10552-006-0022-x Am J Epidemiol, 157(2):113–20. doi:10.1093/aje/kwf184

PMID:16783610 PMID:12522018
IARC (2002). Weight control and physical activity. Lyon, Lundqvist E, Kaprio J, Verkasalo PK, Pukkala E,
France: IARC Press (IARC Handbooks of Cancer Koskenvuo M, Söderberg KC, et al. (2007). Co-twin
Prevention, Vol. 6). Available from: http://publications. control and cohort analyses of body mass index and
iarc.fr/376. height in relation to breast, prostate, ovarian, corpus
Jayson GC, Kohn EC, Kitchener HC, Ledermann JA uteri, colon and rectal cancer among Swedish and
(2014). Ovarian cancer. Lancet, 384(9951):1376–88. Finnish twins. Int J Cancer, 121(4):810–8. doi:10.1002/
King MG, Olson SH, Paddock L, Chandran U, Demissie ijc.22746 PMID:17455257
K, Lu SE, et al. (2013). Sugary food and beverage Lurie G, Wilkens LR, Thompson PJ, McDuffie KE, Carney
consumption and epithelial ovarian cancer risk: a ME, Terada KY, et al. (2008). Genetic polymorphisms
population-based case-control study. BMC Cancer, in the Paraoxonase 1 gene and risk of ovarian epithe-
13(1):94 doi:10.1186/1471-2407-13-94 PMID:23442818 lial carcinoma. Cancer Epidemiol Biomarkers Prev,
Kotsopoulos J, Baer HJ, Tworoger SS (2010). 17(8):2070–7. doi:10.1158/1055-9965.EPI-08-0145
Anthropometric measures and risk of epithelial ovarian PMID:18708400
cancer: results from the Nurses’ Health Study. Obesity Ma X, Beeghly-Fadiel A, Shu XO, Li H, Yang G, Gao YT,
(Silver Spring), 18(8):1625–31. doi:10.1038/oby.2009.461 et al. (2013). Anthropometric measures and epithelial
PMID:20035276 ovarian cancer risk among Chinese women: results
Kuper H, Cramer DW, Titus-Ernstoff L (2002). Risk from the Shanghai Women’s Health Study. Br J Cancer,
of ovarian cancer in the United States in relation 109(3):751–5. doi:10.1038/bjc.2013.384 PMID:23860524
to anthropometric measures: does the associa- Máchová L, Cízek L, Horáková D, Koutná J, Lorenc J,
tion depend on menopausal status? Cancer Causes Janoutová G, et al. (2007). Association between obesity
Control, 13(5):455–63. doi:10.1023/A:1015751105039 and cancer incidence in the population of the District
PMID:12146850 Sumperk, Czech Republic. Onkologie, 30(11):538–42.
Kurian AW, Balise RR, McGuire V, Whittemore AS (2005). PMID:17992023
Histologic types of epithelial ovarian cancer: have they Moorman PG, Palmieri RT, Akushevich L, Berchuck
different risk factors? Gynecol Oncol, 96(2):520–30. A, Schildkraut JM (2009). Ovarian cancer risk
doi:10.1016/j.ygyno.2004.10.037 PMID:15661246 factors in African-American and white women. Am
Lacey JV Jr, Leitzmann M, Brinton LA, Lubin JH, Sherman J Epidemiol, 170(5):598–606. doi:10.1093/aje/kwp176
ME, Schatzkin A, et al. (2006). Weight, height, and PMID:19605513
body mass index and risk for ovarian cancer in a cohort Nagle CM, Olsen CM, Webb PM, Jordan SJ, Whiteman DC,
study. Ann Epidemiol, 16(12):869–76. doi:10.1016/j. Green AC; Australian Cancer Study Group; Australian
annepidem.2006.07.011 PMID:17027285 Ovarian Cancer Study Group (2008). Endometrioid
Lahmann PH, Cust AE, Friedenreich CM, Schulz M, and clear cell ovarian cancers: a comparative analysis of
Lukanova A, Kaaks R, et al. (2010). Anthropometric risk factors. Eur J Cancer, 44(16):2477–84. doi:10.1016/j.
measures and epithelial ovarian cancer risk in the ejca.2008.07.009 PMID:18707869
European Prospective Investigation into Cancer Olsen CM, Green AC, Whiteman DC, Sadeghi S,
and Nutrition. Int J Cancer, 126(10):2404–15. Kolahdooz F, Webb PM (2007). Obesity and the risk
PMID:19821492 of epithelial ovarian cancer: a systematic review
Le ND, Leung A, Brooks-Wilson A, Gallagher RP, and meta-analysis. Eur J Cancer, 43(4):690–709.
Swenerton KD, Demers PA, et al. (2014). Occupational doi:10.1016/j.ejca.2006.11.010 PMID:17223544
exposure and ovarian cancer risk. Cancer Causes Olsen CM, Nagle CM, Whiteman DC, Ness R, Pearce
Control, 25(7):829–41. doi:10.1007/s10552-014-0384-4 CL, Pike MC, et al.; Australian Cancer Study (Ovarian
PMID:24728670 Cancer); Australian Ovarian Cancer Study Group;
Leitzmann MF, Koebnick C, Danforth KN, Brinton LA, Ovarian Cancer Association Consortium (2013).
Moore SC, Hollenbeck AR, et al. (2009). Body mass Obesity and risk of ovarian cancer subtypes: evidence
index and risk of ovarian cancer. Cancer, 115(4):812– from the Ovarian Cancer Association Consortium.
22. doi:10.1002/cncr.24086 PMID:19127552 Endocr Relat Cancer, 20(2):251–62. doi:10.1530/
Liu Z, Zhang TT, Zhao JJ, Qi SF, Du P, Liu DW, et al. ERC-12-0395 PMID:23404857
(2015). The association between overweight, obesity Pan SY, Johnson KC, Ugnat AM, Wen SW, Mao Y;
and ovarian cancer: a meta-analysis. Jpn J Clin Oncol, Canadian Cancer Registries Epidemiology Research
45(12):1107–15. PMID:26491203 Group (2004). Association of obesity and cancer risk
Lubin F, Chetrit A, Freedman LS, Alfandary E, Fishler in Canada. Am J Epidemiol, 159(3):259–68. doi:10.1093/
Y, Nitzan H, et al. (2003). Body mass index at age 18 aje/kwh041 PMID:14742286
years and during adult life and ovarian cancer risk.
334
Peterson NB, Trentham-Dietz A, Newcomb PA, Chen Z, Soegaard M, Jensen A, Høgdall E, Christensen L, Høgdall
Gebretsadik T, Hampton JM, et al. (2006). Relation C, Blaakaer J, et al. (2007). Different risk factor profiles
of anthropometric measurements to ovarian cancer for mucinous and nonmucinous ovarian cancer: results
risk in a population-based case-control study from the Danish MALOVA study. Cancer Epidemiol
(United States). Cancer Causes Control, 17(4):459–67. Biomarkers Prev, 16(6):1160–6. doi:10.1158/1055-9965.
doi:10.1007/s10552-005-0416-1 PMID:16596298 EPI-07-0089 PMID:17548679
Pike MC, Pearce CL, Peters R, Cozen W, Wan P, Wu Song Y-M, Sung J, Ha M (2008). Obesity and risk of
AH (2004). Hormonal factors and the risk of inva- cancer in postmenopausal Korean women. J Clin
sive ovarian cancer: a population-based case-con- Oncol, 26(20):3395–402. doi:10.1200/JCO.2007.15.7867
trol study. Fertil Steril, 82(1):186–95. doi:10.1016/j. PMID:18612154
fertnstert.2004.03.013 PMID:15237010 Su D, Pasalich M, Binns CW, Lee AH (2012). Is body size
Poorolajal J, Jenabi E, Masoumi SZ (2014). Body mass associated with ovarian cancer in southern Chinese
index effects on risk of ovarian cancer: a meta-analysis. women? Cancer Causes Control, 23(12):1977–84.
Asian Pac J Cancer Prev, 15(18):7665–71. doi:10.7314/ doi:10.1007/s10552-012-0075-y PMID:23065073
APJCP.2014.15.18.7665 PMID:25292044 WCRF/AICR (2007). Food, nutrition, physical activity,
Purdie DM, Bain CJ, Webb PM, Whiteman DC, and the prevention of cancer: a global perspective.
Pirozzo S, Green AC (2001). Body size and ovarian Washington (DC), USA: American Institute for Cancer
cancer: case-control study and systematic review Research. Available from: http://www.aicr.org/assets/
(Australia). Cancer Causes Control, 12(9):855–63. docs/pdf/reports/Second_Expert_Report.pdf.
doi:10.1023/A:1012267619561 PMID:11714114 WCRF/AICR (2014). Continuous Update Project Report:
Rapp K, Schroeder J, Klenk J, Stoehr S, Ulmer H, Concin ovarian cancer 2014. Food, nutrition, physical activity,
H, et al. (2005). Obesity and incidence of cancer: a and the prevention of ovarian cancer. Washington
large cohort study of over 145,000 adults in Austria. (DC), USA: American Institute for Cancer Research.
Br J Cancer, 93(9):1062–7. doi:10.1038/sj.bjc.6602819 Available from: http://www.wcrf.org/sites/default/files/
PMID:16234822 Ovarian-Cancer-2014-Report.pdf.
Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D; Yang HP, Trabert B, Murphy MA, Sherman ME, Sampson
Million Women Study Collaboration (2007). Cancer JN, Brinton LA, et al. (2012). Ovarian cancer risk factors
incidence and mortality in relation to body mass by histologic subtypes in the NIH-AARP Diet and
index in the Million Women Study: cohort study. BMJ, Health Study. Int J Cancer, 131(4):938–48. doi:10.1002/
335(7630):1134. doi:10.1136/bmj.39367.495995.AE ijc.26469 PMID:21960414
PMID:17986716 Yen ML, Yen BL, Bai CH, Lin RS (2003). Risk factors for
Reis N, Kizilkayabeji N (2010). Risk factors for ovarian ovarian cancer in Taiwan: a case-control study in a
cancer: results from a hospital-based case-control low-incidence population. Gynecol Oncol, 89(2):318–24.
study. Turkiye Klinikleri J Med Sci, 30(1):79–87. doi:10.1016/S0090-8258(03)00088-X PMID:12713998
Riman T, Dickman PW, Nilsson S, Nordlinder H, Zhang M, Xie X, Holman CD (2005). Body weight and
Magnusson CM, Persson IR (2004). Some life-style body mass index and ovarian cancer risk: a case-con-
factors and the risk of invasive epithelial ovarian cancer trol study in China. Gynecol Oncol, 98(2):228–34.
in Swedish women. Eur J Epidemiol, 19(11):1011–9. doi:10.1016/j.ygyno.2005.04.026 PMID:15979697
doi:10.1007/s10654-004-1633-8 PMID:15648594
Rossing MA, Tang MT, Flagg EW, Weiss LK, Wicklund
KG, Weiss NS (2006). Body size and risk of epithe-
lial ovarian cancer (United States). Cancer Causes
Control, 17(5):713–20. doi:10.1007/s10552-006-0010-1
PMID:16633919
Schildkraut JM, Alberg AJ, Bandera EV, Barnholtz-
Sloan J, Bondy M, Cote ML, et al. (2014). A multi-
center population-based case-control study of ovarian
cancer in African-American women: the African
American Cancer Epidemiology Study (AACES).
BMC Cancer, 14(1):688. doi:10.1186/1471-2407-14-688
PMID:25242549
Schouten LJ, Rivera C, Hunter DJ, Spiegelman D, Adami
HO, Arslan A, et al. (2008). Height, body mass index,
and ovarian cancer: a pooled analysis of 12 cohort
studies. Cancer Epidemiol Biomarkers Prev, 17(4):902–
12. doi:10.1158/1055-9965.EPI-07-2524 PMID:18381473
335
2.2.14 Cancer of the prostate (a) Cohort studies

Cancer of the prostate is the fourth most The IARC Handbook on weight control and
commonly diagnosed cancer worldwide, and physical activity (IARC, 2002), in the evaluation
one of the most frequent causes of cancer-related of prostate cancer risk and measures of body
mortality in developed countries. fatness, included 13 prospective cohort studies
The relationship between body weight and with at least 100 cases (not shown in Table 2.2.14a).
prostate cancer risk is complex, for several Of those, four found a positive association and
reasons. First, prostate cancer-specific mortality nine found no association. Notably, across all
(death attributed to the underlying cancer) is a prospective studies, the highest category of BMI
proxy for incidence in some studies, whereas it was overweight (25–29.9 kg/m2) but not obese
is a primary end-point in other studies, along (≥ 30 kg/m2).
with different types of prostate cancer incidence Since 2000, associations of body fatness
defined by tumour characteristics. However, assessed at baseline with total prostate cancer
prostate cancer-specific mortality may be over- incidence have been examined in numerous
represented in patients who die with but not of individual prospective studies with at least 100
the disease. This is a particular concern if, for cases and in at least two meta-analyses. In most
example, obese patients with prostate cancer have studies, neither BMI nor weight was associated
other comorbid disease and more regular contact with risk (Habel et al., 2000; Schuurman et al.,
with the health-care system; the cancer may be 2000; Lee et al., 2001; Jonsson et al., 2003; Rapp et
more prominent in their management and may al., 2005; Gong et al., 2006; Lukanova et al., 2006;
be recorded on the death certificate, even if heart Tande et al., 2006; Fujino et al., 2007; Giovannucci
disease is the underlying cause of death. Second, et al., 2007; Littman et al., 2007; Máchová et al.,
detection bias could also be a concern in studies 2007; Rodriguez et al., 2007; Pischon et al., 2008;
of prostate cancer incidence; because obese men Wallström et al., 2009; Andreotti et al., 2010;
have lower levels of prostate-specific antigen Stocks et al., 2010; Bassett et al., 2012). However,
(PSA), their tumours are more difficult to detect, in some studies statistically significant positive
and they are less likely to undergo a biopsy (Allot associations (or trends) between BMI at base-
et al., 2013). However, potential biological mech- line and prostate cancer incidence were found
anisms have also been proposed to explain a (Engeland et al., 2003; Samanic et al., 2004, 2006;
lower risk of early-stage prostate cancer in men Jee et al., 2008; Barrington et al., 2015), and four
who are overweight or obese (see Section 4.3.1d). prospective studies found lower risk of prostate
In 2001, the Working Group of the IARC cancer with increasing BMI (Wright et al., 2007;
Handbook on weight control and physical activity Bhaskaran et al., 2014; Møller et al., 2015). In a
(IARC, 2002) concluded that the evidence of an meta-analysis of 27 prospective studies, there
association between avoidance of weight gain was a statistically significant positive associ-
and prostate cancer was inadequate. Since then, ation with prostate cancer incidence (RR per
numerous prospective studies with at least 100 5 kg/m2 increase in BMI, 1.03; 95% CI, 1.00–1.07)
cases (Table 2.2.14a) and case–control studies (Renehan et al., 2008).
(Table 2.2.14b) have been published, as well as Associations of body fatness at baseline with
several meta-analyses of observational studies stage of the disease were examined in several
addressing different measures of body fatness studies. Regarding the incidence of localized,
(Table 2.2.14c). low-grade, or non-aggressive disease, although
five studies found no association (Schuurman et
336
al., 2000; Giovannucci et al., 2007; Pischon et al., 2007; Stocks et al., 2010 Bassett et al., 2012) and
2008; Wallström et al., 2009; Bassett et al., 2012), a large pooled analysis of 57 prospective studies
at least seven other studies found an inverse asso- from Europe, Japan, and the USA, reporting a
ciation of BMI and/or weight with the incidence relative risk of mortality per 5 kg/m2 increase in
of non-aggressive (Littman et al., 2007; Stocks et BMI of 1.13 (95% CI, 1.02–1.24) across the BMI
al., 2010), non-metastatic low- to moderate-grade range of 15–50 kg/m2 (Whitlock et al., 2009).
(Gong et al., 2006; Rodriguez et al., 2007; Møller However, at least six other individual prospec-
et al., 2016 for BMI at age 21 years), or local- tive studies found no association between BMI
ized (Wright et al., 2007; Discacciati et al., 2011; at baseline and death from prostate cancer
Hernandez et al., 2009 for BMI at age 21 years) (Batty et al., 2005; Fujino et al., 2007; Burton et
prostate cancer. In the Selenium and Vitamin al., 2010 for BMI at age < 30 years; Discacciati et
E Cancer Prevention Trial (SELECT), there was al., 2011; Meyer et al., 2015; Møller et al., 2015).
evidence of a significant inverse trend between Similarly, BMI was not associated with prostate
BMI and the incidence of low-grade prostate cancer mortality in a pooled analysis from the
cancer in non-Hispanic White men, and a statis- Asia Cohort Consortium (Fowke et al., 2015).
tically significant positive association in African [The Working Group noted that in this analysis,
American men (Barrington et al., 2015). the reference group was men with a BMI of
Nine prospective studies found no associa- 22.5–24.9 kg/m2, compared with men with a
tions of BMI and/or weight with the incidence of BMI of 25–50 kg/m2. A possible effect of obesity
regional or distant prostate cancer (Habel et al., (BMI > 30 kg/m2) on prostate cancer mortality
2000), advanced, high-grade, or moderately to might have been missed in this study.]
poorly differentiated prostate cancer (Schuurman At least six prospective studies found no asso-
et al., 2000; Pischon et al., 2008; Discacciati et ciations between BMI or weight at younger ages of
al., 2011; Møller et al., 2015), aggressive pros- adulthood and risk of prostate cancer (total, local-
tate cancer (Littman et al., 2007; Wallström et ized, advanced, or fatal) (Giovannucci et al., 1997;
al., 2009; Stocks et al., 2010), or extraprostatic Jonsson et al., 2003; Fujino et al., 2007; Hernandez
prostate cancer (Wright et al., 2007). However, et al., 2009; Burton et al., 2010; Discacciati et al.,
five other studies found positive associations or 2011; Bassett et al., 2012), whereas in two other
trends of BMI and/or weight with the incidence studies higher BMI (Schuurman et al., 2000) or
of high-grade or advanced prostate cancer (Gong weight (Littman et al., 2007) in young adulthood
et al., 2006; Giovannucci et al., 2007; Rodriguez was significantly associated with increased total
et al., 2007; Hernandez et al., 2009 for BMI at age prostate cancer incidence. In the NIH-AARP
21 years; Bassett et al., 2012; Barrington et al., cohort, both BMI and weight at age 18 years
2015). A meta-analysis combining data from 24 were not associated with the incidence of total
prospective studies found a statistically signifi- prostate cancer or extraprostatic prostate cancer,
cant positive association between BMI and risk whereas inverse associations with localized pros-
of advanced, high-grade, or fatal prostate cancer tate cancer were reported (Ptrend = 0.04) (Wright
(RR per 5 kg/m2 increase in BMI, 1.08; 95% CI, et al., 2007). Similarly, in the Multiethnic Cohort
1.04–1.12) (WCRF/AICR, 2014). Study and the Health Professionals Follow-up
There is considerable evidence of a positive Study, BMI at age 21 years was inversely asso-
association of BMI with prostate cancer mortality, ciated with the incidence of total, localized, and
based on findings from both individual prospec- low- and moderate-grade prostate cancer and
tive studies (Rodriguez et al., 2001; Calle et al., was not associated with the incidence of high-
2003; Giovannucci et al., 2007; Wright et al., grade or fatal prostate cancer (Hernandez et al.,
337
2009; Møller et al., 2016). Similarly, in the study 2006; Tande et al., 2006; Pischon et al., 2008;
by Littman et al. (2007), the positive association Wallström et al., 2009; Møller et al., 2015). On
with weight in young adulthood (ages 18, 30, or the basis of four prospective studies, the WCRF
45 years) was restricted to the aggressive type. Continuous Update Project summary (WCRF/
In a meta-analysis of nine prospective studies, AICR, 2014) found no dose–response association
Robinson et al. (2008) found a positive associ- between waist circumference and risk of total
ation between BMI in early life (i.e. < 29 years) or non-advanced prostate cancer, but a statisti-
and prostate cancer incidence or mortality (RR cally significant positive association with risk of
per 5 kg/m2 increase in BMI, 1.08). advanced or fatal prostate cancer (RR per 10 cm
In at least four individual prospective studies, increase, 1.12; 95% CI, 1.04–1.21).
change in neither BMI nor weight during adult-
hood was associated with prostate cancer inci- (b) Case–control studies
dence (Jonsson et al., 2003; Samanic et al., Case–control studies of BMI and other
2006; Rodriguez et al., 2007; Rapp et al., 2008). adiposity indices in relation to prostate cancer
Similarly, a meta-analysis of four prospective risk are presented in Table 2.2.14b. In the IARC
studies also found no associations of adult weight Handbook on weight control and physical activity
gain [after adjustment for age and baseline BMI (IARC, 2002), 15 case–control studies of BMI and
or weight in all studies] with total, localized, or prostate cancer were reviewed (not shown here).
advanced prostate cancer incidence (Keum et Since then, at least 35 case–control studies and
al., 2015). However, in the Netherlands Cohort 5 meta-analyses including case–control study
Study, there was suggestive evidence of an designs, focused on the association between
inverse trend between increase in BMI from age weight, BMI, or waist circumference and pros-
20 years to baseline (≥ 6 kg/m2) and total prostate tate cancer, have been conducted in Asia (China,
cancer incidence (Ptrend = 0.07), and this associa- India, Japan, and Pakistan), the Caribbean
tion was statistically significant for poorly differ- (Barbados and Jamaica), Europe, the Islamic
entiated or undifferentiated prostate tumours Republic of Iran, Nigeria, North America, and
(Schuurman et al., 2000). In the Vitamins and Oceania (Australia and New Zealand). In all of
Lifestyle (VITAL) cohort, both weight loss and these studies, BMI was assessed on the basis of
weight gain were associated with a lower risk of self-reported height and body weight, or body
non-aggressive prostate cancer, but there was weight and height verified at the time of a hospital
no association with aggressive prostate cancer consultation.
(Littman et al., 2007). In the NIH-AARP cohort, Positive associations between high BMI and
weight gain from age 18 years to baseline was not total prostate cancer incidence were reported
associated with prostate cancer incidence (total, in six of the case–control studies. Bashir et al.
localized, or extraprostatic), but was associated (2014), in a hospital-based case–control study in
with prostate cancer mortality (Ptrend = 0.009) Pakistan with 140 cases and 280 controls, found a
(Wright et al., 2007). significant increase in the risk of prostate cancer
The association between waist circumfer- for men with BMI > 25 kg/m2 (OR, 5.78; 95% CI,
ence and total prostate cancer incidence was 2.67–12.6). In a multicentre hospital-based case–
examined in at least eight individual prospective control study in Italy, Dal Maso et al. (2004) iden-
studies, and no study found evidence of statisti- tified a dose–response relationship between BMI
cally significant associations with total prostate at age 30 years and prostate cancer risk, based on
cancer incidence (Giovannucci et al., 1997; Lee 1257 cases (Ptrend = 0.004). Ganesh et al. (2011)
et al., 2001; MacInnis et al., 2003; Gong et al., reported a 2-fold greater risk of prostate cancer
338
in Indian men with BMI ≥ 25 kg/m2 (OR, 2.1; 95% with adult weight was observed for high-risk (RR,
CI, 1.1–4.4). A hospital-based case–control study 1.13; 95% CI, 1.00–1.28) and fatal (RR, 1.58; 95%
in France found a positive association between CI, 1.01–2.47) prostate cancer subtypes (Chen et
BMI > 29 kg/m2 and risk of prostate cancer al., 2016).
(OR, 2.47; 95% CI, 1.41–4.34) (Irani et al., 2003). Six case–control studies differentiated pros-
Similarly, a study in Canada reported a signif- tate cancer by grade, stage, or aggressiveness, and
icant 27% increase in risk of prostate cancer in generally reported positive associations of BMI,
men with BMI ≥ 30 kg/m2 compared with those waist circumference, or waist-to-hip ratio with
with BMI < 25 kg/m2 (Pan et al., 2004). prostate cancers with higher Gleason scores.
An inverse association between BMI and Fowke et al. (2012) analysed 809 hospital-based
prostate cancer has also been reported in cases and 1057 controls in the USA by Gleason
several studies. Beebe-Dimmer et al. (2009), in score. On the basis of 135 cases, BMI and waist
a hospital-based case–control study in the USA, circumference were marginally associated with
found an inverse relationship between high increased risk of high-grade prostate cancer
BMI (≥ 30 kg/m2) and prostate cancer risk in (OR per 1 kg/m2 increase in BMI, 1.04; 95% CI,
Caucasian men, based on 494 cases (OR, 0.51; 1.00–1.08 and OR per 1 cm increase in waist
95% CI, 0.33–0.80), but not in African American circumference, 1.01; 95% CI, 0.99–1.03). Jackson
men. Similarly, a study in Canada found a statis- et al. (2010) separated patients with high-grade
tically significant inverse relationship between prostate cancer in their hospital-based case–
BMI ≥ 30 kg/m2 and prostate cancer risk (OR, control study (243 cases and 275 controls) in
0.72; 95% CI, 0.60–0.87), but no associations with Jamaica. Waist circumference and waist-to-hip
waist circumference or waist-to-hip ratio were ratio were positively associated with high-grade
found (Boehm et al., 2015). A population-based prostate cancer after adjustment for BMI. A
case–control study in the Islamic Republic of dose–response relationship was also observed
Iran (Hosseini et al., 2010), with 137 cases and 137 for waist circumference, and no association was
controls, also found a significant inverse relation- found with BMI. A case–control study in Italy
ship between high BMI (≥ 25 kg/m2) and prostate observed significant positive associations of BMI
cancer risk (OR, 0.4; 95% CI, 0.2–0.8). Finally, and prostate cancer of Gleason score 7–10 only
Agalliu et al. (2015) conducted a small hospi- (Ptrend < 0.01) (Dal Maso et al., 2004). Liu et al.
tal-based case–control study in Nigeria, with 50 (2005) conducted a population-based sibling
cases and 50 controls. Inverse associations were case–control study in the USA with 439 cases and
reported for weight (OR per kg increase, 0.97; 479 controls and found no association of aggres-
95% CI, 0.94–1.00) and waist circumference (OR sive prostate cancer (defined as Gleason score ≥ 7
per cm increase, 0.91; 95% CI, 0.87–0.96). or tumour stage T2C or greater) with increased
One additional case–control study found an BMI, whereas an inverse association was observed
increased risk of total prostate cancer in men for lean body mass (Ptrend = 0.02). Nemesure et
with an increased waist circumference (Beebe- al. (2012) conducted a population-based case–
Dimmer et al., 2007). control study in Barbados with 963 cases and
Three meta-analyses that included case– 941 controls and reported a positive association
control studies suggested a small increase in risk of waist circumference with all prostate cancers
of prostate cancer associated with higher BMI (OR for highest versus lowest quartiles, 1.84; 95%
(Bergström et al., 2001; MacInnis & English, CI, 1.19–2.85), which did not hold when strati-
2006; Robinson et al., 2008). In one additional fying by disease grade. Robinson et al. (2005) in
meta-analysis, a significant positive association the USA reported an inverse association between
339
BMI > 30 kg/m2 at age 20–29 years and advanced (c) Mendelian randomization studies
prostate cancer [based on 12 cases]. Three Mendelian randomization studies have
Several studies assessed BMI and body weight been conducted in this context (Table 2.2.14d).
at different ages, and BMI/weight change. In a Lewis et al. (2010) showed that each additional
population-based case–control study in Sweden, A allele of the FTO rs9939609 SNP was associ-
Gerdtsson et al. (2015) investigated several ated with an increase of 0.56 kg/m2 (P = 0.007)
anthropometric measures, including BMI and in BMI across all groups (cases and controls).
weight, at multiple time points in life. Weight Estimates obtained from Mendelian random-
increase in adolescence (age 16–22 years) was ization analyses provided odds ratios of 0.77
associated with increased risk of prostate cancer (95% CI, 0.52–1.15; P = 0.20) for prostate cancer
(OR per 5 kg increase in weight, 1.05; 95% CI, and 1.35 (95% CI, 0.90–2.03; P = 0.14) for high-
1.01–1.09), and increase in BMI and weight in grade versus low-grade cancer with each 1 kg/m2
middle age (age 44–50 years) was associated increase in BMI.
with increased mortality from prostate cancer, Davies et al. (2015) extended this work by
and with increased metastasis. Weight gain using a genetic risk score based on 32 SNPs
of 10.0–14.9 kg in adulthood was significantly associated with BMI (Speliotes et al., 2010) as
associated with a 3–4-fold greater risk of pros- an instrument for BMI within a much larger
tate cancer in a population-based case–control sample size. Each increase of 1 standard devia-
study in Japan (Mori et al., 2011). In the same tion in genetically predicted BMI was associated
study, BMI of 23.0–24.9 kg/m2 at age 20 years was on average with a nonsignificant 2% reduction in
associated with a reduced risk of prostate cancer risk (95% CI, 0.96–1.00; P = 0.07) in any prostate
(OR, 0.47; 95% CI, 0.22–0.98) (Mori et al., 2011) cancer diagnosis.
[based on 11 cases only]. In contrast, a total of In Mendelian randomization analyses that
16 case–control studies conducted in Australia, used genetic risk scores based on 77 SNPs for
Canada, the Czech Republic, Italy, Japan, New adult BMI (Locke et al., 2015) and 15 SNPs for
Zealand, Spain, Sweden, Switzerland, the United childhood BMI (Felix et al., 2016), Gao et al.
Kingdom, and the USA reported no associations (2016) found no strong evidence for associations
between risk of total prostate cancer and BMI or of childhood or adult BMI with either total or
other adiposity indices at different ages (Putnam aggressive prostate cancer risk.
et al., 2000; Sharpe & Siemiatycki, 2001; Giles [Although results from Lewis et al. (2010)
et al., 2003; Friedenreich et al., 2004; Porter & and Davies et al. (2015) point towards an inverse
Stanford, 2005; Robinson et al., 2005; Wuermli association between BMI and prostate cancer
et al., 2005; Cox et al., 2006; Gallus et al., 2007; risk, this association was not significant and was
Máchová et al., 2007; Nagata et al., 2007; Magura not consistently found in all three studies.]
et al., 2008; Dimitropoulou et al., 2011; Pelucchi
et al., 2011; Möller et al., 2013; Alvarez-Cubero
et al., 2015; Zhang et al., 2015) or BMI change or
weight gain from early adulthood (Putnam et al.,
2000; Giles et al., 2003; Friedenreich et al., 2004).
340
Table 2.2.14a Cohort studies of measures of body fatness and cancer of the prostate
Reference Total number Organ site or Exposure Exposed Relative risk Covariates Comments
Cohort of subjects cancer subtype categories cases (95% CI)
Location Incidence/ (ICD code)
Giovannucci et al. 47 781 Prostate, BMI at age 21 yr Age, height
(1997) Incidence advanced < 20 81 1.00
Health 20–21.9 117 0.91 (0.69–1.22)
Professionals 22–22.9 59 0.88 (0.62–1.24)
Follow-up Study 23–23.9 56 0.77 (0.54–1.10)
USA 24–25.9 60 0.71 (0.50–1.02)
1986–1994 ≥ 26 26 0.53 (0.33–0.86)
[Ptrend] [< 0.006]
Prostate, all BMI at age 21 yr WC also not associated
< 20 229 1.00 with increased risk
20–21.9 353 0.98 (0.83–1.16)
22–22.9 188 1.00 (0.82–1.22)
23–23.9 200 1.03 (0.84–1.26)
24–25.9 223 1.00 (0.82–1.22)
≥ 26 104 0.87 (0.67–1.12)
[Ptrend] [0.60]
Habel et al. (2000) 70 712 Prostate BMI 2079 total Age, race, year of birth Weight also not
Kaiser Incidence < 22.7 1.00 associated with increased
Permanente 22.7–24.3 1.09 (0.93–1.27) risk
USA 24.4–25.9 1.04 (0.89–1.21) No associations were
1964–1973 to 1996 26–27.9 1.04 (0.90–1.21) observed in results
> 27.9 0.99 (0.85–1.15) stratified by race
Prostate, BMI 578 total

regional/distant < 22.7 1.00
22.7–24.3 0.84 (0.62–1.13)
24.4–25.9 1.05 (0.80–1.39)
26–27.9 1.04 (0.79–1.37)
> 27.9 0.91 (0.69–1.20)
Schuurman et al. 58 279 Prostate BMI at baseline Age, family history of
(2000) Incidence < 22 63 1.00 prostate cancer, SES;
Netherlands 22–23 164 1.20 (0.84–1.73) BMI change results
Cohort Study 24–25 236 1.35 (0.95–1.90) also adjusted for BMI
The Netherlands 26–27 150 1.26 (0.87–1.83) at age 20 yr
1986–1982 ≥ 28 62 0.89 (0.58–1.37)
[Ptrend] [0.73]
per 2 kg/m2 1.00 (0.92–1.07)
341
342

Schuurman et al. 58 279 BMI at age 20 yr
(2000) Incidence < 19 57 1.00
(cont.) 19–20.9 122 1.06 (0.72–1.56)
21–22.9 176 1.09 (0.76–1.58)
23–24.9 119 1.39 (0.93–2.06)
≥ 25 44 1.33 (0.81–2.19)
[Ptrend] [0.02]
per 2 kg/m2 1.08 (0.99–1.18)
BMI change
−9.2 to < 0 47 1.19 (0.74–1.90)
0–1.9 120 1.00
2–3.9 176 1.32 (0.98–1.79)
4–5.9 113 1.04 (0.74–1.47)
6–7.9 43 0.83 (0.52–1.31)
≥ 8 19 0.67 (0.36–1.23)
[Ptrend] [0.07]
per 2 kg/m2 0.93 (0.84–1.03)
Prostate, BMI, per 2 kg/m2 239 total
localized BMI at baseline 0.96 (0.86–1.06)
TNM: T0–2, M0 BMI at age 20 yr 1.18 (1.04–1.35)
BMI change 0.87 (0.74–1.02)
advanced BMI at baseline 1.01 (0.90–1.13)
TNM: T3–4, BMI at age 20 yr 1.03 (0.91–1.18)
M0; T0–4, M1
BMI change 0.93 (0.80–1.08)
Schuurman et al. Prostate, well- BMI, per 2 kg/m2 194 total
(2000) differentiated BMI at baseline 0.92 (0.82–1.04)
(cont.) BMI at age 20 yr 1.09 (0.94–1.26)
BMI change 0.77 (0.65–0.92)
moderately BMI at baseline 1.02 (0.93–1.13)
differentiated BMI at age 20 yr 1.15 (1.01–1.31)
BMI change 0.97 (0.83–1.13)
Prostate, poorly BMI, per 2 kg/m2 174 total
differentiated or BMI at baseline 1.01 (0.89–1.14)
undifferentiated BMI at age 20 yr 0.97 (0.83–1.13)
BMI change 0.68 (0.58–0.81)
Lee et al. (2001) 8922 Prostate BMI at baseline Age, smoking, alcohol WC also not associated
Harvard Alumni Incidence < 22.5 87 1.00 consumption, paternal with increased risk
Health Study 22.5–24.9 172 1.27 (0.94–1.71) history of prostate BMI at age 18 yr
USA 25.0–27.4 134 1.26 (0.92–1.72) cancer (available for 92% of the
1988–1993 27.5 46 1.02 (0.68–1.53) men) also not associated
[Ptrend] [0.71] with increased risk
Rodriguez et al. 381 638 Prostate BMI Age, race, height,
(2001) Mortality ICD-7: 177 < 25 782 1.00 education level,
Cancer 25–29.99 698 1.02 (0.92–1.14) exercise, smoking
Prevention Study ≥ 30 110 1.27 (1.04–1.56) status, family history

I (CPS I) [Ptrend] [0.06] of prostate cancer
USA
1959–1972
Calle et al. (2003) 404 576 Prostate BMI Age, education level,
Cancer Mortality 18.5–24.9 1681 1.00 smoking, physical
Prevention Study 25–29.9 1971 1.08 (1.01–1.15) activity, alcohol
II (CPS II) 30–34.9 311 1.20 (1.06–1.36) consumption, marital
USA ≥ 35 41 1.34 (0.98–1.83) status, race, aspirin
1982–1998 [Ptrend] [< 0.001] use, fat consumption,
vegetable consumption
343
344

Engeland et al. 951 466 Prostate BMI Age at BMI In stratified analyses by
(2003) Incidence ICD-7: 177 < 18.5 147 0.92 (0.78–1.08) measurement, birth age at BMI measurement,
Norwegian 18.5–24.9 16 720 1.00 cohort no differences in risk by
clinical 25–29.9 14 524 1.07 (1.05–1.09) age strata were observed
population ≥ 30 1923 1.09 (1.04–1.15)
Norway [Ptrend] [0.001]
1963–1999 to 2001
Jonsson et al. 8998 Prostate BMI at baseline Age; BMI at age No associations were
(2003) Incidence ICD-7: 177 < 18.5 6 1.4 (0.6–3.1) 25 yr and 40 yr also observed in stratified
Swedish Twin 18.5–24.9 355 1.0 controlled for BMI at analyses by age at
Registry 25.0–29.9 248 1.0 (0.8–1.2) baseline diagnosis (≥ 70 yr vs
Sweden ≥ 30 22 1.0 (0.6–1.5) < 70 yr)
1969–2003 BMI at age 25 yr
< 18.5 4 0.5 (0.2–1.5)
18.5–24.9 436 1.0
≥ 25 64 1.0 (0.7–1.3)
BMI at age 40 yr
< 18.5 6 2.5 (1.1–5.5)
18.5–24.9 368 1.0
25.0–29.9 155 0.9 (0.7–1.1)
≥ 30 13 0.9 (0.5–1.6)
Adult weight change (kg)
< 0 96 0.9 (0.7–1.2)
0–5 178 1.0
6–10 114 1.0 (0.8–1.3)
11–20 95 0.9 (0.7–1.2)
≥ 21 21 1.1 (0.8–1.8)
Samanic et al. 4 500 700 Prostate Obesity Age, calendar year Obesity defined as
(2004) Incidence ICD-9: 185 Black men: discharge diagnosis of
United States Non-obese 15 272 1.00 obesity: ICD-8: 277;
Veterans cohort Obese 815 1.12 (1.04–1.20) ICD-9: 278.0
USA White men:
1969–1996 Non-obese 45 901 1.00
Obese 3206 1.19 (1.15–1.24)
Batty et al. (2005) 18 403 Prostate BMI Age, employment
Whitehall Study Mortality 18.5–24.9 243 1.00 grade, physical activity,
United Kingdom 25.0–29.9 175 0.92 (0.75–1.13) smoking, marital
1967–2002 ≥ 30 13 0.91 (0.51–1.63) status, prevalent
[Ptrend] [0.45] disease, past-year
weight loss, BP
medication, height,
skinfold thickness,
systolic BP, plasma
cholesterol, glucose
intolerance, diabetes
Rapp et al. (2005) 67 447 Prostate BMI Age, smoking status,
Vorarlberg Incidence ICD-9: 185 18.5–24.9 446 1.00 occupation
VHM&PP 25–29.9 583 1.03 (0.91–1.17)
Austria 30–34.9 99 0.82 (0.66–1.03)
1985–2001 ≥ 35 10 0.73 (0.39–1.37)
[Ptrend] [0.16]
Gong et al. (2006) 10 258 Prostate BMI 1936 total Age, race, treatment, Analyses of the
Prostate Cancer Incidence < 25 1.00 diabetes, family history association of WC with
Prevention Trial 25–26.9 0.91 (0.79–1.05) of prostate cancer total prostate, and low-
(PCPT) 27–29.9 0.96 (0.83–1.10) grade and high-grade
USA ≥ 30 0.96 (0.83–1.10) subtypes also reported
N/A–2003 [Ptrend] [0.67]

Prostate, low- BMI 1300 total
grade < 25 1.00
25–26.9 0.88 (0.74–1.04)
27–29.9 0.88 (0.75–1.04)
≥ 30 0.82 (0.69–0.98)
[Ptrend] [0.03]
Prostate, high- BMI 521 total
grade < 25 1.00
25–26.9 0.97 (0.75–1.27)
27–29.9 1.09 (0.85–1.40)
≥ 30 1.29 (1.01–1.67)
[Ptrend] [0.04]
345
346

Lukanova et al. 33 424 Prostate BMI Age, calendar year,
(2006) Incidence/ 18.5–23.4 93 1.00 smoking
Northern Sweden mortality 23.5–25.3 114 1.00 (0.76–1.32)
Health and 25.4–27.6 129 0.96 (0.74–1.26)
Disease Cohort ≥ 27.1 125 0.89 (0.68–1.16)
(NSHDC) [Ptrend] [0.31]
1985–2003
Samanic et al. 362 552 Prostate BMI Attained age, calendar
(2006) Incidence ICD-7: 177 18.5–24.9 3003 1.00 year, smoking
Swedish 25–29.9 3160 1.06 (1.01–1.12)
Construction ≥ 30 528 1.09 (0.99–1.19)
Sweden 107 815 (in 6-yr BMI change
1958–1999 BMI change −4% to 4.9% 1281 1.00
analysis) 5–9.9% 417 1.09 (0.98–1.22)
Incidence 10–14.9% 97 0.93 (0.75–1.14)
≥ 15% 22 0.75 (0.49–1.15)
[Ptrend] [> 0.5]
Tande et al. 6332 Prostate BMI Age, race WC also not associated
(2006) Incidence < 24.7 94 1.00 with increased risk
Atherosclerosis 24.7–26.9 99 1.17 (0.88–1.55) Men with metabolic
Risk in 27.0–29.7 91 0.97 (0.72–1.29) syndrome were 27% less
Communities ≥ 29.8 101 1.14 (0.86–1.50) likely to develop prostate
(ARIC) Study cancer
USA
1987–2000
Fujino et al. NR Prostate BMI Age, area of study [No information
(2007) Mortality < 18.5 17 1.39 (0.83–2.34) reported on follow-up
Japan 18.5–24 107 1.00 period or total number of
Collaborative 25–29 31 1.56 (1.04–2.34) participants included in
Cohort Study ≥ 30 1 0.87 (0.12–6.29) the study]
for Evaluation of Weight at baseline and
Cancer (JACC) at age 20 yr also not
Japan associated with increased
NR mortality
Giovannucci et al. 47 750 Prostate BMI 3544 total Age, time period, [CI provided only for the
(2007) Incidence < 21 1.00 BMI at age 21 yr, last BMI category]
Health 21–22.9 1.21 height, pack-years No association was
Professionals 23–24.9 1.36 of smoking, physical observed with BMI for
Follow-up Study 25–27.4 1.24 activity, family low-grade or high-grade
USA 27.5–29.9 1.24 history of prostate prostate cancer (based on
1986–2002 ≥ 30 1.13 (0.91–1.41) cancer, diabetes, race, Gleason score)
Updated [Ptrend] [0.84] energy intake, intake
follow-up from Prostate, BMI 523 total of processed meat,
Giovannucci et al. advanced < 21 1.00 fish, α-linolenic acid,
(1997) TNM: T3b or T4 ≥ 30 1.34 (0.79–2.26) tomato sauce, vitamin
or N1 or M1 [Ptrend] [≤ 0.05] E supplements
47 750 Prostate BMI 323 total
Mortality 21–22.9 1.00
23–24.9 1.44
25–27.4 1.30
27.5–29.9 1.43
≥ 30 1.80 (1.10–2.93)
Littman et al. 34 754 Prostate BMI at baseline Age, family history of BMI at ages 18 yr, 30 yr,
(2007) Incidence < 25 218 1.0 prostate cancer, race, and 45 yr also not
Vitamins and 25–29.9 435 1.1 (0.97–1.4) baseline BMI, recent associated with increased
Lifestyle (VITAL) ≥ 30 155 0.87 (0.71–1.1) PSA screening risk
cohort [Ptrend] [0.13]

USA Prostate, non- BMI at baseline BMI at ages 18 yr, 30 yr,
2000–2004 aggressive < 25 129 1.0 and 45 yr also not
Gleason score 25–29.9 222 0.99 (0.79–1.2) associated with increased
< 7 ≥ 30 73 0.69 (0.52–0.93) risk
[Ptrend] [0.01]
Prostate, BMI at baseline BMI at ages 18 yr, 30 yr,
aggressive < 25 85 1.0 and 45 yr also not
Gleason score 25–29.9 209 1.4 (1.1–1.8) associated with increased
7–10 ≥ 30 179 1.1 (0.83–1.6) risk
[Ptrend] [0.69]
347
348

Littman et al. 34 754 Prostate Weight (lb) at age 18 yr Age, family history of For non-aggressive
(2007) Incidence < 139 166 1.0 prostate cancer, race, prostate cancer, weight at
(cont.) 139–154 203 1.2 (0.96–1.5) baseline BMI, recent age 18 yr and 30 yr was
155–170 198 1.1 (0.93–1.4) PSA screening not associated with an
≥ 171 231 1.2 (1.0–1.5) increased risk
[Ptrend] [0.08]
Weight (lb) at age 30 yr
< 154 174 1.0
154–169 192 1.2 (0.95–1.4)
170–184 188 1.1 (0.93–1.4)
≥ 185 241 1.3 (1.0–1.6)
[Ptrend] [0.03]
< 165 194 1.0
165–179 182 1.0 (0.82–1.2)
180–199 224 1.1 (0.91–1.3)
≥ 200 200 1.1 (0.87–1.3)
[Ptrend] [0.46]
Weight (lb) at baseline
< 173 211 1.0
174–189 181 1.0 (0.83–1.2)
190–214 233 0.99 (0.82–1.2)
≥ 215 192 0.92 (0.75–1.1)
[Ptrend] [0.35]
Prostate, non- Weight (lb) at baseline Weight gain of ≥ 30 lb
aggressive < 173 130 1.00 since age 18 yr associated
Gleason score 174–189 90 0.82 (0.62–1.1) with 33% lower risk of
< 7 190–214 116 0.81 (0.63–1.1) incidence
≥ 215 92 0.71 (0.54–0.93)
[Ptrend] [0.02]
Prostate, Weight (lb) at age 18 yr Age, family history of
aggressive < 139 71 1.00 prostate cancer, race,
Gleason score 139–154 94 1.3 (0.92–1.7) baseline BMI, recent
7–10 155–170 89 1.2 (0.86–1.6) PSA screening
≥ 171 117 1.4 (1.0–1.9)
[Ptrend] [0.04]
Littman et al. 34 754 Weight (lb) at age 30 yr
(2007) Incidence < 154 72 1.0
(cont.) 154–169 84 1.2 (0.90–1.7)
170–184 93 1.4 (0.99–1.9)
≥ 185 119 1.5 (1.1–2.0)
[Ptrend] [0.01]
< 165 72 1.0
165–179 86 1.3 (0.93–1.8)
180–199 111 1.5 (1.1–2.0)
≥ 200 102 1.4 (1.1–2.0)
[Ptrend] [0.032]
Weight (lb) at baseline Weight gain since age
< 173 78 1.0 18 yr not associated with
174–189 87 1.3 (0.96–1.8) risk of incidence
190–214 115 1.3 (0.97–1.7)
≥ 215 98 1.3 (0.93–1.7)
[Ptrend] [0.23]
Máchová et al. 17 334 Prostate BMI 338 total Age, smoking,
(2007) Incidence ICD-10: C61 18.5–24.9 1.00 hypertension, height
National Cancer 25–29.9 1.05 (0.72–1.39)
Registry ≥ 30 0.97 (0.66–1.41)
Nested case–

control study in
the population
of the Šumperk
District
Czech Republic
1987–2002
349
350

Rodriguez et al. 69 991 Prostate BMI Age, race, education
(2007) Incidence < 25 1935 1.00 level, family history of
Cancer 25–27.4 1742 1.02 (0.96–1.09) prostate cancer, energy
Prevention 27.5–29.9 920 0.98 (0.90–1.06) intake, smoking status,
Study II (CPS II) 30–34.9 556 0.94 (0.85–1.04) PSA testing, diabetes,
Nutrition Cohort ≥ 35 99 0.91 (0.75–1.12) physical activity;
USA [Ptrend] [0.14] Weight change also
1992–2003 Weight change (lb), 1982–1992 adjusted for BMI in When stratifying by
≥ 21 loss 113 0.84 (0.69–1.02) 1982 and height subtype, weight change
11–20 loss 349 0.84 (0.75–0.95) also not associated with
6−19 loss 541 0.98 (0.89–1.08) increased risk for any
5 loss to 5 gain 2450 1.00 subtype
6–10 gain 751 0.98 (0.90–1.06)
11–20 gain 687 0.97 (0.89–1.05)
≥ 21 gain 322 0.89 (0.79–1.00)
Prostate, non- BMI
metastatic, < 25 1544 1.00
low-grade 25–27.4 1409 1.03 (0.96–1.10)
TNM: T1–3, N0, 27.5–29.9 700 0.92 (0.84–1.01)
M0 30–34.9 412 0.86 (0.77–0.97)
Gleason score ≥ 35 73 0.84 (0.66–1.06)
≤ 8 [Ptrend] [0.002]
Prostate, non- BMI
metastatic high- < 25 239 1.00
grade 25–27.4 180 0.87 (0.72–1.06)
TNM: T1–3, N0, 27.5–29.9 140 1.23 (1.00–1.53)
M0 ≥ 30 103 1.22 (0.96–1.55)
Gleason score [Ptrend] [0.03]
> 8
69 991 Prostate, BMI
Incidence or metastatic or < 25 92 1.00
mortality fatal 25–27.4 104 1.41 (1.06–1.87)
TNM: T4, Nx, 27.5–29.9 46 1.14 (0.79–1.63)
Mx or Tx, N1–2, ≥ 30 46 1.54 (1.06–2.23)
Mx or Tx, Nx, [Ptrend] [0.05]
M1
Wright et al. 172 961 Prostate BMI Age, race, smoking
(2007) Incidence ICD-9: 185 < 25 3076 1.00 status, education level,
NIH-AARP ICD-10: C61 25–29.9 5054 1.00 (0.95–1.04) diabetes, family history
cohort 30–34.9 1532 0.97 (0.91–1.03) of prostate cancer
USA 35–39.9 269 0.84 (0.74–0.95) For BMI at age 18 yr,
1995–2000 ≥ 40 55 0.65 (0.50–0.85) also BMI at baseline,
BMI at age 18 yr
< 18.5 723 0.95 (0.87–1.04)
18.5–20.9 1787 1.00
21–22.9 1510 1.01 (0.95–1.09)
23–24.9 775 0.90 (0.83–0.98)
≥ 25 641 0.93 (0.84–1.02)
[Ptrend] [0.17]
Weight (kg) at age 18 yr, quintiles Age, race, smoking
< 58.6 1004 1.0 status, education
58.7–64.5 1338 1.01 (0.93–1.10) level, diabetes, family
64.6–69.9 1043 0.99 (0.91–1.09) history of prostate
70–76.7 1138 0.99 (0.91–1.09) cancer, BMI, height
> 76.7 1071 0.92 (0.84–1.02)
[Ptrend] [0.08]
Weight (kg) at baseline, quintiles Weight at baseline
< 74.5 1126 1.0 also not associated

74.6–81.3 1224 1.02 (0.93–1.11) with increased risk
81.4–87.2 1204 1.01 (0.92–1.10) for localized and with
87.3–97.2 1157 1.00 (0.91–1.09) metastatic prostate
> 97.2 1014 0.91 (0.82–1.00) cancer subtypes
[Ptrend] [0.99]
351
352

Wright et al. 172 961 Weight change (kg), age 18 yr to baseline Weight change also not
(2007) Incidence < −4 161 1.00 (0.83–1.19) associated with increased
(cont.) −4 to 3.9 430 1.0 risk for localized and for
4–9.9 936 1.04 (0.93–1.17) extraprostatic prostate
10–19.9 1896 1.12 (1.00–1.24) cancer subtypes
20–29.9 1425 1.12 (1.00–1.26)
30–39.9 469 0.99 (0.87–1.14)
≥ 40 277 1.03 (0.88–1.20)
[Ptrend] [0.81]
Prostate, BMI Age, race, smoking
localized < 25 2652 1.00 status, education level,
TNM: T1a to 25–29.9 4328 0.99 (0.94–1.04) diabetes, family history
T2b, N0, M0 30–34.9 1277 0.94 (0.88–1.01) of prostate cancer
35–39.9 236 0.86 (0.75–0.98) For BMI at age 18 yr,
≥ 40 48 0.67 (0.50–0.89) also BMI at baseline,
BMI at age 18 yr
< 18.5 633 0.95 (0.86–1.04)
18.5–20.9 1570 1.0
21–22.9 1317 1.01 (0.94–1.09)
23–24.9 653 0.87 (0.80–0.96)
≥ 25 535 0.89 (0.80–0.99)
[Ptrend] [0.04]
< 58.6 881 0.95 (0.86–1.04) status, education
58.7–64.5 1185 1.00 level, diabetes, family
> 76.7 891 0.89 (0.80–0.99)
[Ptrend] [0.04]
Prostate, BMI Age, race, smoking
extraprostatic < 25 424 1.0 status, education level,
TNM: T3 or T4, 25–29.9 726 1.03 (0.91–1.16) diabetes, family history
N1, or M1 30–34.9 255 1.14 (0.97–1.33) of prostate cancer
≥ 35 40 0.68 (0.49–0.94) For BMI at age 18 yr,
[Ptrend] [0.64] also BMI, height
Wright et al. 172 961 BMI at age 18 yr
(2007) Incidence < 18.5 90 0.98 (0.77–1.26)
(cont.) 18.5–20.9 217 1.00
21–22.9 193 1.04 (0.86–1.27)
23–24.9 122 1.11 (0.88–1.39)
≥ 25 106 1.15 (0.90–1.47)
[Ptrend] [0.18]
< 58.6 123 1.0 status, education
58.7–64.5 153 0.95 (0.74–1.20) level, diabetes, family
> 76.7 180 1.18 (0.91–1.54)
[Ptrend] [0.13]
Wright et al. Mortality Prostate BMI Age, race, smoking Weight at baseline also
(2007) ICD-9: 185 < 25 44 1.0 status, education level, associated with increased
NIH-AARP ICD-10: C61 25–29.9 87 1.25 (0.87–1.80) diabetes, family history risk
cohort 30–34.9 31 1.46 (0.92–2.33) of prostate cancer
USA ≥ 35 11 2.12 (1.08–4.15) For BMI at age 18 yr,
1995–2000 [Ptrend] [0.02] also BMI at baseline,
BMI at age 18 yr height Weight (kg) at age 18 yr
< 18.5 13 1.67 (0.82–3.42) also not associated with
18.5–20.9 18 1.0 increased mortality

21–22.9 25 1.65 (0.90–3.02)
23–24.9 16 1.71 (0.86–3.39)
≥ 25 11 1.35 (0.62–2.95)
[Ptrend] [0.73]
Weight change (kg), age 18 yr to baseline Age, race, smoking
< −4 3 1.18 (0.29–4.74) status, education
−4 to 3.9 6 1.0 level, diabetes, family
4–9.9 12 1.06 (0.40–2.83) history of prostate
20–29.9 24 1.74 (0.69–4.40)
30–39.9 10 2.05 (0.72–5.90)
40 8 2.98 (0.99–9.04)
[Ptrend] [0.009]
353
354

Jee et al. (2008) 770 556 Prostate BMI Age, smoking
National Health Incidence < 20.0 265 0.67 (0.56–0.80)
Insurance 20.0–22.9 896 0.87 (0.77–0.98)
Corporation 23.0–24.9 747 1.00
(NHIC) medical 25.0–29.9 638 0.95 (0.83–1.08)
evaluation ≥ 30.0 23 1.39 (0.90–2.17)
Republic of Korea [Ptrend] [< 0.0001]
1992–2006
Pischon et al. 129 502 Prostate BMI, quintiles 2446 total Study centre, age, Also examined hip
(2008) Incidence ICD-10: C61 < 23.6 1.00 smoking status, circumference and waist-
EPIC cohort 23.6–25.3 1.06 (0.93–1.20) education level, alcohol to-hip ratio
8 European 25.4–27 1.08 (0.95–1.23) consumption, physical WC also not associated
countries, 27.1–29.3 0.95 (0.83–1.09) activity, height with increased risk
1992–2000 (8.5 yr ≥ 29.4 0.99 (0.86–1.13)
follow-up on [Ptrend] [0.37]
average) per 5 kg/m2 0.96 (0.90–1.02)
Prostate, BMI, quintiles 991 total Study centre, age, WC also not associated
localized < 23.6 1.00 smoking status, with increased risk
TNM: T0–T2 23.6–25.3 1.09 (0.89–1.34) education level, alcohol
and N0/Nx, M0 25.4–27 1.02 (0.83–1.25) consumption, physical
27.1–29.3 0.88 (0.71–1.10) activity, height
≥ 29.4 0.95 (0.77–1.18)
[Ptrend] [0.22]
continuous 0.92 (0.84–1.01)
Prostate, BMI 499 total Study centre, age, WC also not associated
advanced < 23.6 1.00 smoking status, with increased risk
TNM: T3–T4 23.6–25.3 1.05 (0.78–1.40) education level, alcohol
and/or N1–N3 25.4–27 1.25 (0.94–1.66) consumption, physical
and/or M1 27.1–29.3 1.08 (0.81–1.46) activity, height
≥ 29.4 1.17 (0.86–1.58)
[Ptrend] [0.34]
continuous 1.09 (0.96–1.24)
Pischon et al. 129 502 Prostate, low- BMI 841 total Study centre, age, WC also not associated
(2008) Incidence grade < 23.6 1.00 smoking status, with increased risk
(cont.) Gleason score 23.6–25.3 0.97 (0.78–1.21) education level, alcohol
< 7 25.4–27 0.95 (0.77–1.19) consumption, physical
27.1–29.3 0.83 (0.66–1.04) activity, height
≥ 29.4 0.84 (0.66–1.06)
[Ptrend] [0.06]
continuous 0.88 (0.79–0.98)
Prostate, high- BMI 580 total Study centre, age, WC also not associated
grade < 23.6 1.00 smoking status, with increased risk
Gleason score 23.6–25.3 1.26 (0.96–1.65) education level, alcohol
≥ 7 25.4–27 1.34 (1.02–1.76) consumption, physical
27.1–29.3 1.16 (0.87–1.54) activity, height
≥ 29.4 1.23 (0.92–1.65)
[Ptrend] [0.37]
continuous 1.04 (0.92–1.18)
Rapp et al. (2008) 28 711 Prostate BMI change, annual Age, smoking status,
VHM&PP Incidence ICD-10: C61 < −0.1 164 0.96 (0.79–1.16) blood glucose,
Austria −0.1– < 0.1 317 1.00 occupational group,
1985–2002 0.1– < 0.3 231 1.00 (0.85–1.19) BMI at baseline
0.3– < 0.5 72 1.01 (0.78–1.31)
≥ 0.5 12 0.43 (0.24–0.76)
[Ptrend] [0.06]

Hernandez et al. 83 879 Prostate, BMI at age 21 yr No associations were
(2009) Incidence advanced < 18.5 41 0.96 (0.69–1.35) observed with high grade
Multiethnic 18.5–24.9 475 1.00 either
Cohort ≥ 25.0 86 1.09 (0.85–1.40) Inverse associations were
USA [Ptrend] [0.46] observed with localized
1993/1996– and with low-grade
2002/2005 subtypes
355
356

Wallström et al. 11 063 Prostate BMI Age, height, WC also not associated
(2009) Incidence ICD-9: 185 < 18.5 8 2.29 (1.13–4.63) cohabitation with increased risk
Malmö Diet and 18.5–24.9 287 1.00 status, SES, alcohol
Cancer Study 25–29.9 417 1.02 (0.88–1.19) consumption,
Sweden ≥ 30 105 1.06 (0.84–1.33) smoking, prevalent
1991–2005 [Ptrend] [0.15] diabetes, physical
Prostate, BMI activity, country of WC also not associated
aggressive < 18.5 4 3.15 (1.15–8.62) birth, total intake of with increased risk
TNM: T3–T4, 18.5–24.9 102 1.00 eicosapentaenoic acid,
or N1 or M1, or 25–29.9 140 0.99 (0.76–1.29) docosahexaenoic acid,
Gleason score ≥ 30 35 1.02 (0.69–1.52) red meat, calcium
≥ 8, or PSA [Ptrend] [0.16]
> 50 ng/mL
Prostate, non- BMI WC also not associated
aggressive < 18.5 4 0.84 (0.63–1.11) with increased risk
Not stage T3– 18.5–24.9 183 1.00
T4, or N1 or M1, 25–29.9 274 1.16 (0.89–1.50)
or Gleason score ≥ 30 69 1.11 (0.85–1.44)
≥ 8, or PSA [Ptrend] [0.65]
> 50 ng/mL
Whitlock et al. 894 576 Prostate BMI, per 5 kg/m2 Study, sex, age,
(2009) Mortality ICD-9: 185 For BMI 15–25 578 1.00 (0.75–1.32) smoking
Prospective For BMI 25–50 665 1.09 (0.91–1.31)
Studies For BMI 15–50 1.13 (1.02–1.24)
Collaboration
(pooled analysis
of 57 cohorts from
Europe, Japan,
and the USA)
Follow-up varied
by cohort
Andreotti et al. 39 628 Prostate BMI Race, smoking status,
(2010) Incidence < 18.5 0 – exercise, family history
Agricultural 18.5–24.9 308 1.00 of prostate cancer
Health Study 25–29.9 696 1.06 (0.89–1.27)
USA 30–34.9 226 0.89 (0.71–1.13)
1993–2005 ≥ 35 44 0.94 (0.61–1.44)
[Ptrend] [0.56]
Burton et al. 9549 Prostate BMI, young adult (age < 30 yr) Smoking, SES, height
(2010) Incidence ICD-9: 185 < 19 25 1.30 (0.84–1.99)
Glasgow Alumni ICD-10: C61 19–22.9 125 1.00
Cohort 23–24.9 33 1.14 (0.78–1.68)
United Kingdom ≥ 25 14 1.18 (0.68–2.06)
1948–1968 to per 1 kg/m2 1.00 (0.93–1.06)
2009 [Ptrend] [0.89]
9549 Prostate BMI, young adult (age < 30 yr)
Mortality ICD-9: 185 < 19 14 1.58 (0.88–2.83)
ICD-10: C61 19–22.9 59 1.00
23–24.9 21 1.52 (0.92–2.50)
≥ 25 8 1.43 (0.68–3.00)
per 1 kg/m2 1.02 (0.93–1.11)
[Ptrend] [0.74]
Stocks et al. (2010) 336 159 Prostate BMI Birth cohort, smoking No association of
Swedish Mortality ICD-7: 177 < 21.9 230 1.00 BMI with incidence

Construction 21.9– < 23.5 383 1.17 (1.00–1.39) of prostate (total), or
Worker Cohort 23.5– < 25 476 1.09 (0.93–1.27) aggressive prostate
Sweden 25– < 27 702 1.26 (1.08–1.46) cancer subtypes.
1971–2004 ≥ 27 810 1.28 (1.11–1.49) Significant negative
[Ptrend] [0.0004] association observed
between BMI and
incidence for non-
aggressive prostate
cancer subtype
357
358

Discacciati et al. 36 959 Prostate, BMI at baseline BMI at age 30 yr, age,
(2011) Incidence localized < 21 62 0.78 (0.54–1.13) energy intake, physical
Sweden TNM: T1–2 21–22.9 245 1.00 activity, education
1998–2008 and NX–0 and 23–24.9 401 1.00 (0.94–1.06) level, smoking, family
MX–0 or PSA 25–27.4 467 0.95 (0.86–1.05) history of prostate
< 20 ng/mL or 27.5–29.9 204 0.88 (0.76–1.02) cancer, diabetes
Gleason score ≥ 30 124 0.71 (0.53–0.94)
< 7 BMI at age 30 yr
< 21 287 1.01 (0.91–1.12)
21–22.9 539 1.00
23–24.9 467 0.99 (0.94–1.05)
25–27.4 154 0.99 (0.89–1.10)
27.5–29.9 41 0.98 (0.82–1.16)
≥ 30 15 0.96 (0.69–1.34)
per 5 kg/m2 0.98 (0.87–1.12)
Prostate, BMI at baseline BMI at age 30 yr, age,
advanced < 21 27 0.97 (0.85–1.10) energy intake, physical
TNM: T3–4 21–22.9 72 1.00 activity, education
and NX–1 and 23–24.9 163 1.02 (0.95–1.08) level, smoking, family
MX–1 or PSA 25–27.4 150 1.03 (0.90–1.18) history of prostate
> 100 ng/mL or 27.5–29.9 79 1.05 (0.85–1.31) cancer, diabetes
Gleason score ≥ 30 47 1.11 (0.73–1.68)
> 7 per 5 kg/m2 1.04 (0.88–1.22)
BMI at age 30 yr
< 21 108 1.09 (0.92–1.29)
21–22.9 185 1.00
23–24.9 164 0.96 (0.88–1.04)
25–27.4 69 0.91 (0.77–1.09)
27.5–29.9 8 0.87 (0.65–1.15)
≥ 30 4 0.76 (0.44–1.30)
per 5 kg/m2 0.90 (0.73–1.11)
Discacciati et al. 36 959 Prostate BMI at baseline BMI at age 30 yr, age, BMI at age 30 yr also not
(2011) Mortality < 21 11 0.91 (0.75–1.11) energy intake, physical associated with increased
(cont.) 21–22.9 35 1.00 activity, education risk
23–24.9 62 1.05 (0.95–1.16) level, smoking, family
25–27.4 59 1.11 (0.89–1.36) history of prostate
27.5–29.9 29 1.16 (0.83–1.63) cancer, diabetes
≥ 30 23 1.34 (0.70–2.55)
per 5 kg/m2 1.12 (0.87–1.43)
Bassett et al. 16 525 Prostate BMI at baseline Country of birth, No associations were
(2012) Incidence ICD-9: 185 < 18.5 111 0.73 (0.59–0.91) education level observed between weight
Melbourne ICD-10: C61 18.5–22.9 259 1.00 at baseline, BMI or
Collaborative 23–24.9 757 0.98 (0.85–1.12) weight (kg) at age 18 yr,
Cohort Study ≥ 25 247 0.96 (0.80–1.15) or WC, and prostate
(MCCS) per 5 kg/m2 1.06 (0.97–1.17) cancer risk (incidence)
Australia [Ptrend] [0.19]
1990–2004 Prostate, non- BMI at baseline Country of birth, No associations were
Same cohort as aggressive < 18.5 83 0.73 (0.56–0.94) education level observed between
MacInnis et al. Not Gleason 18.5–22.9 194 1.00 weight at baseline,
(2003) score > 7, stage 23–24.9 527 0.91 (0.77–1.08) BMI or weight (kg) at
4, or death from ≥ 25 160 0.83 (0.67–1.03) age 18 yr, or WC, and
prostate cancer per 5 kg/m2 0.99 (0.89–1.10) non-aggressive prostate
[Ptrend] [0.83] cancer risk (incidence)
Prostate, BMI at baseline Country of birth, No associations were

aggressive < 18.5 28 0.74 (0.47–1.15) education level observed between weight
Gleason score 18.5–22.9 65 1.00 at baseline, BMI or
> 7, stage 4, 23–24.9 230 1.17 (0.89–1.54) weight (kg) at age 18 yr,
or death from ≥ 25 87 1.33 (0.96–1.84) or WC, and aggressive
prostate cancer per 5 kg/m2 1.27 (1.08–1.49) prostate cancer risk
[Ptrend] [0.004] (incidence)
16 525 Prostate BMI at baseline Country of birth, Weight at baseline also
Mortality ICD-9: 185 < 18.5 7 0.53 (0.23–1.24) education level associated with increased
ICD-10: C61 18.5–22.9 23 1.00 mortality
23–24.9 71 0.95 (0.59–1.53) No association was
≥ 25 38 1.52 (0.89–2.58) observed with BMI or
per 5 kg/m2 1.49 (1.11–2.00) weight at age 18 yr and
[Ptrend] [0.01] mortality
359
360

Bhaskaran et al. 2 379 320 Prostate BMI 24 901 total Age, diabetes, No differences were
(2014) Incidence ICD-10: C61 per 5 kg/m2 0.98 (0.95–1.00) smoking, alcohol found in non-smokers
Clinical Practice [Ptrend] [0.0042] consumption, SES, only
Research Datalink calendar year, sex
United Kingdom
1987–2012
Barrington et al. 26 035 Prostate BMI Non-Hispanic White: Age, education level, For African Americans,
(2015) Incidence < 25.0 289 1.00 diabetes, smoking, BMI < 25.0 in Non-
Participants in 25.0–27.5 438 1.12 (0.97–1.30) family history of Hispanic Whites was
the Selenium and 27.5–29.9 333 1.04 (0.89–1.22) prostate cancer, study taken as reference
Vitamin E cancer 30–34.9 299 0.96 (0.82–1.13) arm
Prevention Trial 35–50 94 0.94 (0.74–1.19)
(SELECT) [Ptrend] [0.63]
USA BMI African American:
2001–2008
< 25.0 39 1.28 (0.91–1.80)
25.0–27.5 63 1.67 (1.27–2.21)
27.5–29.9 57 1.64 (1.23–2.19)
30–34.9 74 1.68 (1.29–2.18)
35–50 37 1.90 (1.34–2.70)
[Ptrend] [0.03]
26 035 Prostate, low- BMI Non-Hispanic White: Age, education level,
Incidence grade < 25.0 182 1.00 diabetes, smoking,
Gleason score 25.0–27.5 293 1.18 (0.98–1.42) family history of
2–6 27.5–29.9 202 1.00 (0.82–1.22) prostate cancer, study
30–34.9 170 0.86 (0.70–1.06) arm
35–50 51 0.80 (0.58–1.09)
[Ptrend] [0.02]
< 25.0 16 0.80 (0.48–1.43)
25.0–27.5 37 1.47 (1.03–2.10)
27.5–29.9 35 1.52 (1.05–2.20)
30–34.9 37 1.27 (0.83–1.82)
35–50 23 1.77 (1.14–2.76)
[Ptrend] [0.05]
Barrington et al. 26 035 Prostate, high- BMI Non-Hispanic White: Age, education level,
(2015) Incidence grade < 25.0 84 1.00 diabetes, smoking,
(cont.) Gleason score 25.0–27.5 115 1.03 (0.78–1.37) family history of
7–10 27.5–29.9 101 1.11 (0.83–1.49) prostate cancer, study
30–34.9 104 1.18 (0.88–1.58) arm
35–50 37 1.33 (0.90–1.97)
[Ptrend] [0.01]
< 25.0 11 1.32 (0.70–2.51)
25.0–27.5 19 1.94 (1.17–3.22)
27.5–29.9 17 1.87 (1.10–3.16)
30–34.9 29 2.53 (1.64–3.90)
35–50 12 2.39 (1.29–4.43)
[Ptrend] [0.02]
Fowke et al. 522 736 Prostate BMI Age, education level, Similar results were
(2015) Mortality 12–19.9 142 0.98 (0.78–1.23) population density, observed in stratified
Pooled analysis 20–22.4 188 0.92 (0.75–1.13) marital status, history analyses by region
in Asia Cohort 22.5–24.9 184 1.00 of severe cancer, heart
Consortium 25–50 120 1.08 (0.85–1.36) disease, or stroke at
(ACC) [Ptrend] [0.58] baseline
Different Asian
countries
(1963–2001) to

2006
Meyer et al. (2015) 35 703 in Prostate BMI 170 total Age, survey, alcohol Those who were
Population-based cohort, number ICD-8: 185 < 25 1.00 consumption, physical overweight and who also
Swiss cohort of men NR ICD-10: C61 25–29.9 1.45 (1.03–2.04) activity, civil status, smoked (ever smoking)
study Mortality ≥ 30 1.54 (0.93–2.55) years of education, had a higher risk
Switzerland nationality, diet
1977–2008
361
362

Møller et al. 26 044 Prostate BMI NR WC showed no
(2015) Incidence 15.4–24.9 649 1.00 association with total
Diet, Cancer and 25–29.9 920 0.94 (0.85–1.04) prostate cancer incidence
Health Study 30–52.7 244 0.86 (0.74–0.99) Inverse associations
Denmark [Ptrend] [0.03] were observed with the
1993–2011 upper quartile of body
fat percentage (15%
decreased risk)
Prostate BMI NR WC also no associated
Stage 3–4 15.4–24.9 208 1.00 with advanced prostate
25–29.9 314 1.00 (0.84–1.19) cancer incidence
30–52.7 104 1.14 (0.90–1.44) Positive associations
[Ptrend] [0.37] were observed with the
upper quartile of body
fat percentage (31%
increased risk)
26 044 Prostate BMI Stage at diagnosis WC also not associated
Mortality 15.4–24.9 92 1.00 with increased mortality
25–29.9 147 1.10 (0.85–1.43) A positive association
30–52.7 51 1.27 (0.90–1.80) was observed with
[Ptrend] [0.19] increasing body fat
percentage
Møller et al. 47 491 Prostate BMI at age 21 yr Age, calendar time, When analysing
(2016) Incidence and < 20 825 0.99 (0.90–1.08) ethnicity, physical cumulative BMI average,
Health mortality 20–21.9 1546 1.00 activity, energy intake, the significant decrease
Professionals 22–23.9 1852 0.98 (0.91–1.05) smoking, diabetes, in risk persisted only in
Follow-up Study 24–25.9 1132 0.92 (0.85–1.00) family history of those younger than 65 yr
USA ≥ 26 588 0.89 (0.80–0.98) prostate cancer, PSA
1986–2010 [Ptrend] [0.01] testing
per 5 kg/m2 0.94 (0.89–0.98)
Møller et al. 47 491 Prostate, fatal BMI at age 21 yr BMI at age 21 yr also
(2016) Incidence and < 20 94 0.83 (0.64–1.07) not associated with
(cont.) mortality 20–21.9 181 1.00 lethal subtypes (incident
22–23.9 177 0.92 (0.74–1.14) cases and deaths due to
24–25.9 88 0.74 (0.57–0.97) prostate cancer or distant
≥ 26 51 0.77 (0.56–1.07) metastases at diagnosis
[Ptrend] [0.20] or during follow-up)
per 5 kg/m2 0.88 (0.75–1.02)
Prostate, high- BMI at age 21 yr
grade < 20 85 0.82 (0.63–1.07)
Gleason score 20–21.9 181 1.00
8–10 22–23.9 204 0.93 (0.75–1.15)
24–25.9 130 0.91 (0.72–1.16)
≥ 26 79 1.10 (0.83–1.45)
[Ptrend] [0.27]
per 5 kg/m2 1.03 (0.90–1.19)
Prostate, BMI at age 21 yr Age, calendar time,
moderate-grade < 20 233 0.98 (0.83–1.15) ethnicity, physical
Gleason score 7 20–21.9 446 1.00 activity, energy intake,
22–23.9 548 0.98 (0.86–1.11) smoking, diabetes,
24–25.9 333 0.90 (0.78–1.04) family history of
≥ 26 159 0.77 (0.64–0.93) prostate cancer, PSA
[Ptrend] [0.01] testing

per 5 kg/m2 0.87 (0.80–0.95)
Prostate, low- BMI at age 21 yr
grade < 20 333 1.01 (0.88–1.16)
Gleason score 20–21.9 620 1.00
2–6 22–23.9 735 0.94 (0.84–1.05)
24–25.9 465 0.90 (0.79–1.02)
≥ 26 236 0.88 (0.75–1.03)
[Ptrend] [0.03]
per 5 kg/m2 0.93 (0.87–1.01)
BMI, body mass index (in kg/m 2); BP, blood pressure; CI, confidence interval; EPIC, European Prospective Investigation into Cancer and Nutrition; ICD, International Classification of
Diseases; N/A, not applicable; NIH-AARP, National Institutes of Health–AARP Diet and Health Study; NR, not reported; PSA, prostate-specific antigen; SD, standard deviation; SES,
socioeconomic status; TNM, tumour–node–metastasis; VHM&PP, Vorarlberg Health Monitoring and Prevention Program; WC, waist circumference; yr, year or years
363
364

Table 2.2.14b Case–control studies of measures of body fatness and cancer of the prostate
Period Source of
controls
Putnam et al. 101 BMI Age
(2000) Population < 24.1 27 1.0
USA 24.1–26.6 31 1.0 (0.6–1.7)
1986–1989 > 26.6 38 1.3 (0.8- 2.2)
BMI change (%) from age 20 yr
> 5% loss 1 0.2 (0.02–1.5)
5% loss to 5% gain 12 1.0
5.1–10.0% gain 15 1.3 (0.6–2.7)
10.1–15.0% gain 14 1.0 (0.5–1.9)
> 15.0% gain 51 1.3 (0.8–2.2)
Weight (kg)
< 74.8 22 1.0
74.8–83.9 41 1.4 (0.8–2.3)
> 83.9 33 1.2 (0.7–2.1)
Sharpe & 399 BMI Age, ethnicity,
Siemiatycki Population < 24.05 127 0.87 (0.6–1.22) respondent status,
(2001) 24.05–26.66 128 1.00 family income, alcohol
Canada > 26.66 141 1.14 (0.81–1.61) consumption
1979–1985
Giles et al. (2003) 1476 BMI at age 21 yr Age, country of birth, No associations were
Australia Population < 20.5 353 1.00 family history of prostate observed for weight or
1994–1998 20.5–22.1 372 0.99 (0.79–1.23) cancer, study centre, WC at age 21 yr
22.2–23.9 337 0.96 (0.76–1.20) calendar year
> 23.9 332 1.10 (0.88–1.39)
Irani et al. (2003) 194 BMI NR Age
France Hospital < 29 1.00
1993–1999 > 29 2.47 (1.41–4.34)
Period Source of
controls
Dal Maso et al. 1294 BMI at baseline Age, study centre, No associations were
(2004) Hospital < 24.22 301 1.00 education level, physical observed between
Italy 24.22–26.18 346 1.18 (0.95–1.47) activity, family history of weight (kg), waist-to-
1991–2002 26.18–28.41 324 1.12 (0.89–1.40) prostate cancer hip ratio, or lean body
≥ 28.41 319 1.18 (0.94–1.47) mass and prostate
[Ptrend] [0.23] cancer. When stratified
BMI at age 30 yr by grade, associations
< 22.65 406 1.00 of BMI at diagnosis
22.65–24.69 437 1.33 (1.09–1.62) were only significant
≥ 24.69 414 1.22 (1.01–1.48) with prostate cancer of
[Ptrend] [0.004] Gleason score 7–10 (384
cases, Ptrend < 0.01)
Friedenreich et 988 BMI, quartiles Age, region, education
al. (2004) Population Q1 252 1.00 level, average lifetime
Canada Q2 236 0.95 (0.74–1.23) total alcohol intake, first-
1997–2000 Q3 245 0.98 (0.76–1.26) degree family history of
Q4 254 1.07 (0.83–1.38) prostate cancer, number
[Ptrend] [0.57] of times had PSA test
Weight, quartiles done, number of digital
Q1 268 1.00 rectal exams, total
Q2 233 0.93 (0.72–1.21) lifetime physical activity
Q3 262 1.00 (0.78–1.28)
Q4 224 0.91 (0.70–1.18)

[Ptrend] [0.18]
Weight gain (kg) since age 20 yr
< 4.54 241 1.00
4.54–13.6 286 1.14 (0.89–1.47)
13.6–20.4 238 1.05 (0.82–1.36)
≥ 20.4 215 0.91 (0.70–1.19)
[Ptrend] [0.26]
365
366

Period Source of
controls
Pan et al. (2004) 1801 BMI Age group, province of
Canada Population < 25 1.00 residence, education level,
1994–1997 25–30 1.16 (0.94–1.43) pack-years of smoking,
≥ 30 1.27 (1.09–1.47) alcohol consumption,
[Ptrend] [0.026] total energy intake,
vegetable intake, dietary
fibre intake, recreational
physical activity
Liu et al. (2005) 439 BMI, quartiles Age, education, calorie Results are presented
USA Population Q1 106 1.00 intake for high-aggressiveness
NR (sibling-based) Q2 112 1.57 (0.85–2.89) prostate cancer (Gleason
Q3 110 1.43 (0.78–2.61) score ≥ 7, or tumour
Q4 106 0.91 (0.49–1.70) stage T2C or greater)
[Ptrend] [0.73]
LBM, quartiles LBM > 66.3:
Q1 113 1.00
Q2 104 0.58 (0.31–1.08)
Q3 114 0.43 (0.22–0.81)
Q4 103 0.41 (0.20–0.84)
[Ptrend] [0.02]
Porter & 753 BMI Age, race, education
Stanford (2005) Population 18–24.4 195 1.00 level, smoking, family
USA 24.4–26.5 202 1.04 (0.78–1.39) history of prostate cancer,
1993–1996 26.5–29.1 178 0.85 (0.64–1.14) prostate cancer screening,
29.1–55 178 0.91 (0.66–1.21) dietary fat, energy intake
[Ptrend] [0.04]
Weight (kg)
< 77.2 175 1.00
77.2–85.8 222 0.96 (0.70–1.30)
85.9–95.3 193 0.77 (0.56–1.06)
> 95.3 163 0.74 (0.53–1.03)
[Ptrend] [0.03]
Robinson et al. 568 BMI at age 20–29 yr Age, race, family history This study evaluated
(2005) Population < 25.0 361 1.00 of prostate cancer, the association with
USA 25.0–29.9 191 1.13 (0.87–1.47) saturated fat intake advanced prostate
1997–2000 ≥ 30.0 12 0.40 (0.20–0.81) cancer
Period Source of
controls
Wuermli et al. 504 BMI NR Age, BMI, diabetes, lipid-
(2005) Hospital < 30 1.00 lowering drugs
Switzerland > 30 0.97 (0.93–1.01)
1997–2002
Cox et al. (2006) 550 BMI 5 yr before interview, quintiles Age No associations were
New Zealand Population Q1 50 1.0 observed between BMI
1996–1998 Q2 40 0.9 (0.5–1.6) or weight at age 20 yr
Q3 105 0.8 (0.6–1.2) and prostate cancer
Q4 122 0.9 (0.6–1.3)
Q5 233 0.9 (0.6–1.3)
Beebe-Dimmer 139 WC (cm) 59 Age, smoking history
et al. (2007) Population ≤ 102 1.00
USA (community- > 102 1.84 (1.17–2.91)
1996–2002 based)
Gallus et al. 219 BMI Age, education
(2007) Hospital < 24.84 69 1.0 level, study centre,
Italy 24.84–27.76 80 1.3 (0.8–2.0) occupational physical
1991–2002 ≥ 27.77 70 1.2 (0.8–1.9) activity, family history of
[Ptrend] [0.38] prostate cancer
Máchová et al. 338 BMI NR Age, smoking,
(2007) Population 18.5–< 25 1.00 hypertension, height
Czech Republic 25–30 1.05 (0.72–1.39)
1987–2002 ≥ 30 0.97 (0.66–1.41)

Nagata et al. 200 BMI 1 yr before diagnosis Smoking BMI at age 40–45 yr
(2007) Hospital < 23.0 81 1.00 not associated with
Japan 23.0–24.9 60 1.28 (0.87–1.87) increased risk of
1996–2003 > 25.0 59 1.06 (0.72–1.55) prostate cancer
[Ptrend] [0.65]
Magura et al. 312 BMI Age, family history of
(2008) Hospital < 25 30 1.00 prostate cancer, type 2
USA ≥ 25 282 1.04 (0.58–1.85) diabetes, smoking, use
2004–2006 of multivitamins, use of
statins
367
368

Period Source of
controls
Beebe-Dimmer 637 BMI Age, PSA screening Inverse association
et al. (2009) Hospital < 30 – 1.00 history, hypertension, was observed only in
USA ≥ 30 208 0.51 (0.33–0.80) diabetes, low HDL, high Caucasians (n = 494).
2001–2004 triglycerides No association observed
in African Americans
(n = 381)
Hosseini et al. 137 BMI Age, family history of [Discrepancy in the
(2010) Population ≤ 25 105 1.0 prostate cancer, history number of reported
Islamic Republic > 25 35 0.4 (0.2–0.8) of other cancers, history cases]
of Iran of prostatitis, alcohol
2005–2008 consumption, smoking,
physical activity
Jackson et al. 243 BMI, quartiles NR BMI: age, education level, Results are presented
(2010) Hospital Q4 vs Q1 (ref) 0.90 (0.42–1.91) medical history, first- for high-grade cancer
Jamaica [Ptrend] [0.28] degree family history of (Gleason score ≥ 7)
2005–2007 WC, tertiles prostate cancer, smoking, 12% of the cases were
T3 vs T1 (ref) 5.57 (1.43–18.63) physical activity obese
[Ptrend] [0.008] WC and waist-to-hip
Waist-to-hip ratio ratio: age, height and BMI
< 0.95 1.00 as continuous; education
≥ 0.95 2.94 (1.34–6.38) level, current smoker,
physical activity
Dimitropoulou 960 BMI Age, family history of
et al. (2011) Population < 25.0 264 1.00 prostate cancer
United Kingdom 25.0–29.9 481 0.98 (0.82–1.16)
2001–2008 > 30.0 174 0.83 (0.67–1.03)
[Ptrend] [0.097]
WC, tertiles
T1 385 1.00
T2 286 1.01 (0.85–1.20)
T3 289 0.94 (0.80–1.12)
[Ptrend] [0.517]
Ganesh et al. 123 BMI Age, religion, education
(2011) Hospital < 25 41 1.0 level, hypertension
India ≥ 25 76 2.1 (1.1–4.4)
1999–2001
Period Source of
controls
Mori et al. (2011) 117 BMI Dietary intake, physical BMI of 23–25 at age
Japan Population < 21.0 14 1.00 activity, smoking, alcohol 20 yr associated with a
2007–2008 21.0–22.9 29 1.05 (0.50–2.21) consumption 53% reduced risk (based
23.0–24.9 41 1.63 (0.77–3.45) on 11 cases)
≥ 25.0 33 1.39 (0.66–2.96) No associations between
[Ptrend] [0.07] body weight at age 20 yr
Weight (kg) and prostate cancer risk
< 55 7 1.00
55.0–64.9 52 1.49 (0.57–3.85)
65.0–74.9 45 1.74 (0.65–4.64)
≥ 75.0 13 1.64 (0.55–4.91)
Weight gain (kg) in adult life
< 5 18 1.00
5.0–9.9 24 1.22 (0.58–2.55)
10.0–14.9 43 3.55 (1.71–7.39)
≥ 15 32 1.73 (0.83–3.59)
Pelucchi et al. 1294 BMI Age, study centre,
(2011) Hospital < 28 909 1.00 education level, smoking,
Italy ≥ 28 381 0.98 (0.83–1.17) alcohol consumption,
1991–2002 WC (cm) physical activity, family
< 94 242 1.00 history of prostate cancer,
≥ 94 730 1.13 (0.91–1.40) non-alcohol energy
Abdominal obesity (combined WC, BMI) intake

No 470 1.00
Yes 820 1.02 (0.86–1.21)
Fowke et al. 809 BMI Age, PSA, prostate Results are presented
(2012) Hospital per 1 kg/m2 increase 135 1.04 (1.00–1.08) volume, race, family for high-grade (Gleason
USA WC history of prostate cancer, score 8–10) prostate
NR per 1 cm increase 135 1.01 (0.99–1.03) current treatment for cancer
diabetes, benign prostatic
hyperplasia, CVD, or
hyperlipidaemia
369
370

Period Source of
controls
Nemesure et al. 963 WC (cm), quartiles NR Age, marital status, Study in African
(2012) Population Q1: < 84 1.00 religion, occupation, Barbadian population.
Barbados Q2: 84–92 1.36 (1.01–1.85) smoking, family history When stratifying by
2002–2011 Q3: 92–99 1.67 (1.14–2.44) of prostate cancer, BMI high-grade (n = 434)
Q4: ≥ 99 1.84 (1.19–2.85) vs low-grade (n = 480)
prostate cancer, the
associations were not
significant in either
group
Möller et al. 1499 BMI Age, region of residence, No associations with
(2013) Population < 22.5 382 1.00 time span between first BMI when stratifying by
Sweden 22.5– < 25 655 0.94 (0.76–1.15) and last recalled weight low- and intermediate-
2001–2002 25– < 27.5 295 0.90 (0.71–1.15) grade vs high-grade
≥ 27.5 120 0.96 (0.69–1.33) prostate cancer
per 5 kg/m2 0.98 (0.83–1.16) No significant
[Ptrend] [0.54] associations with BMI at
age 20 yr
Bashir et al. 140 BMI Age, lifestyle (physical
(2014) Hospital ≤ 25 66 1.00 activity), family history
Pakistan > 25 74 5.78 (2.67–12.6) of prostate cancer,
2012–2013 smoking, diet
Agalliu et al. 50 BMI Age
(2015) Hospital < 25 21 1
Nigeria 25–29.9 21 1.39 (0.59–3.28)
2011–2012 ≥ 30 8 1.35 (0.42–4.36)
Weight (kg)
per kg increase 0.97 (0.94–1.00)
WC (cm)
per cm increase 0.91 (0.87–0.96)
Alvarez-Cubero 100 BMI 31 Age, residential area,
et al. (2015) Hospital ≥ 30 vs < 30 1.65 (0.36–7.57) family history of prostate
Spain cancer
2011–2014
Period Source of
controls
Boehm et al. 1933 BMI Age, ancestry, first-degree No associations were
(2015) Population < 25 649 1.00 family history of prostate observed with waist-to-
Canada 25–29.9 922 0.87 (0.74–1.01) cancer, annual physician hip ratio
2005–2012 ≥ 30 351 0.72 (0.60–0.87) visits, number of PSA
WC (cm) tests within 5 yr before
< 102 1073 1.00 index date
≥ 102 711 1.03 (0.89–1.19)
Gerdtsson et al. 1355 Weight at age 16–22 yr Incidence: No associations were
(2015) Population per 5 kg increase 1.05 (1.01–1.09) observed with BMI or
Sweden BMI at age 44–50 yr Mortality: weight at age 44–50 yr
1974–1996 per 5 kg increase 1.08 (1.03–1.13) and prostate cancer risk
Weight at age 44–50 yr Mortality: BMI and weight at age
per 5 kg increase 1.11 (1.03–1.19) 44–50 yr also associated
with metastasis
Zhang et al. 101 BMI WC, BP, triglyceride
(2015) Hospital < 24 35 1.00 levels, free blood glucose
China ≥ 24 66 2.51 (0.18–9.52)
2013–2014
BMI, body mass index (in kg/m 2); BP, blood pressure; CI, confidence interval; CVD, cardiovascular disease; HDL, high-density lipoprotein; LBM, lean body mass; NR, not reported;
PSA, prostate-specific antigen; SD, standard deviation; WC, waist circumference; yr, years or years

371
372

Table 2.2.14c Meta-analyses of measures of body fatness and cancer of the prostate
Reference Total number of Organ site or cancer Exposure Relative risk Adjustment for Comments
studies subtype categories (95% CI) confounding
Total number of
cases
Bergström et al. 6 observational Prostate BMI Different
(2001) studies (4 cohort and per 1 kg/m2 1.01 (1.00–1.02) adjustment by study,
2 case–control) increase some non-adjusted
4592
MacInnis & English 43 observational Prostate BMI Different No associations were
(2006) studies (22 cohort per 5 kg/m2 1.05 (1.01–1.08) adjustment by study found with WC
and 21 case–control) increase
(9 studies for WC)
68 753
Renehan et al. (2008) 27 prospective Prostate BMI Between-study
studies per 5 kg/m2 1.03 (1.00–1.07) heterogeneity of
70 421 increase I2 = 73%
No differences in the
results were observed
by region (Asia-Pacific,
Australia, Europe,
North America)
Robinson et al. 9 cohort studies Prostate BMI before age Cohort: Age for all; other
(2008) and 7 case–control 29 yr, 1.08 (0.97–1.19) factors depending
studies per 5 kg/m2 Case–control: on the study
NR increase 1.07 (0.98–1.17)
Guh et al. (2009) 7 cohort studies Prostate BMI NR
NR Normal 1.00
Overweight 1.14 (1.00–1.31)
Obesity 1.05 (0.85–1.30)
Esposito et al. (2013) 13 observational Prostate BMI NR [Cut-off values differ by
studies (cohort and High vs low 1.05 (0.97–1.15) study]
case–control)
4634
WCRF/AICR (2014) 24 prospective Prostate, advanced BMI NR Advanced prostate
Continuous Update studies for BMI, 4 per 5 kg/m2 1.08 (1.04–1.12) cancer includes
Project for WC increase advanced, high-grade,
11 149 WC and fatal prostate
per 10 cm 1.12 (1.04–1.21) cancers
increase
Reference Total number of Organ site or cancer Exposure Relative risk Adjustment for Comments
studies subtype categories (95% CI) confounding
Total number of
cases
Keum et al. (2015) 4 prospective studies Prostate Weight gain Age and baseline
6882 per 5 kg increase 0.98 (0.94–1.02) BMI or weight in
Prostate, localized Weight gain all, and different
per 5 kg increase 0.96 (0.92–1.00) additional
Prostate, advanced Weight gain covariates
per 5 kg increase 1.04 (0.99–1.09) depending on the
study
WC
per 10 cm 1.03 (0.99–1.07)
increase
Chen et al. (2016) 9 observational All Adult weight 1.01 (0.94–1.08) Age (in all studies
studies (5 cohort, 1 Low- and intermediate-grade per 5 kg increase 0.97 (0.87–1.07) except one) and
nested case–control, High-grade 1.13 (1.00–1.28) different covariates
and 3 case–control) Fatal 1.58 (1.01–2.47) depending on the
22 338 study
BMI, body mass index (in kg/m 2); CI, confidence interval; NR, not reported; WC, waist circumference; WCRF/AICR, World Cancer Research Fund/American Institute for Cancer
Research; yr, years or years

373
374

Table 2.2.14d Mendelian randomization studies of measures of body fatness and cancer of the prostate
Reference Characteristics of Sample size Exposure (unit) Odds ratio (95% CI) and P Adjustment for
Study study population value (with each unit increase confounding
in exposure) of the association
between the exposure and
outcome(s)
Lewis et al. (2010) Men aged 50–69 yr 4540 (1550 BMI All: Age, centre
Prostate Testing for Cancer from 300 general cases and 2990 per 1 kg/m2 increase 0.77 (0.52–1.15)
and Treatment Study practices across controls) P = 0.20
(ProtecT) 9 regions in the High-grade vs low-grade:
United Kingdom per 1 kg/m2 increase 1.35 (0.90–2.03)
P = 0.15
Davies et al. (2015) 19 independent 41 062 (20 848 Increase of 1 SD in 0.98 (0.96–1.00) 8 principal
Prostate Cancer studies of cases and genetically predicted BMI P = 0.07 components
Association Group to individuals of 20 214 controls) of population
Investigate Cancer- European descent stratification
Associated Alterations in
the Genome (PRACTICAL)
Consortium
Gao et al. (2016) 6 studies of 26 884 Increase of 1 SD in genetically predicted BMI (~0.073 kg/m2) N/A
Genetic Associations and individuals of (14 160 cases Childhood BMI: All:
Mechanisms in Oncology European ancestry and 12 724 1.01 (0.83–1.22)
(GAME-ON) Consortium controls) P = 0.91
Aggressive:
1.10 (0.83–1.45)
P = 0.49
Adult BMI: All:
1.00 (0.96–1.04)
P = 0.97
Aggressive:
1.02 (0.96–1.08)
P = 0.44
BMI, body mass index (in kg/m 2); CI, confidence interval; N/A, not applicable; SD, standard deviation; vs, versus; yr, years or years
References doi:10.1002/1097-0215(200002)9999:9999<::AID-
IJC1053>3.0.CO;2-T PMID:11169969
Agalliu I, Adebiyi AO, Lounsbury DW, Popoola O, Jinadu Leon DA, Smeeth L (2014). Body-mass index and risk
K, Amodu O, et al. (2015). The feasibility of epidemi- of 22 specific cancers: a population-based cohort study
ological research on prostate cancer in African men of 5.24 million UK adults. Lancet, 384(9945):755–65.
in Ibadan, Nigeria. BMC Public Health, 15(1):425. doi:10.1016/S0140-6736(14)60892-8 PMID:25129328
doi:10.1186/s12889-015-1754-x PMID:25927535 Boehm K, Sun M, Larcher A, Blanc-Lapierre A, Schiffmann
Allot EH, Masko EM, Freedland SJ (2013). Obesity and J, Graefen M, et al. (2015). Waist circumference,
prostate cancer: weighing the evidence. Eur Urol, waist-hip ratio, body mass index, and prostate cancer
63(5):178–9. PMID:23218490 risk: results from the North-American case-control
Alvarez-Cubero MJ, Pascual-Geler M, Rivas A, Martinez- study Prostate Cancer & Environment Study. Urol
Gonzalez LJ, Saiz M, Lorente JA, et al. (2015). Lifestyle Oncol, 33(11):e1–7. doi:10.1016/j.urolonc.2015.07.006
and dietary factors in relation to prostate cancer risk. PMID:26278366
Int J Food Sci Nutr, 66(7):805–10. doi:10.3109/09637486 Burton A, Martin R, Galobardes B, Davey Smith G,
.2015.1077786 PMID:26327471 Jeffreys M (2010). Young adulthood body mass index
Andreotti G, Hou L, Beane Freeman LE, Mahajan R, and risk of cancer in later adulthood: historical
Koutros S, Coble J, et al. (2010). Body mass index, cohort study. Cancer Causes Control, 21(12):2069–77.
agricultural pesticide use, and cancer incidence in doi:10.1007/s10552-010-9625-3 PMID:20680433
the Agricultural Health Study cohort. Cancer Causes Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ
Control, 21(11):1759–75. doi:10.1007/s10552-010-9603-9 (2003). Overweight, obesity, and mortality from cancer
PMID:20730623 in a prospectively studied cohort of U.S. adults. N Engl
Barrington WE, Schenk JM, Etzioni R, Arnold KB, J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423
Neuhouser ML, Thompson IM Jr, et a l. (2015). Difference PMID:12711737
in association of obesity with prostate cancer risk Chen Q, Chen T, Shi W, Zhang T, Zhang W, Jin Z, et al.
between US African American and non-Hispanic (2016). Adult weight gain and risk of prostate cancer: a
white men in the Selenium and Vitamin E Cancer dose-response meta-analysis of observational studies.
Prevention Trial (SELECT). JAMA Oncol, 1(3):342–9. Int J Cancer, 138(4):866–74. doi:10.1002/ijc.29846
doi:10.1001/jamaoncol.2015.0513 PMID:26181184 PMID:26356247
Bashir MN, Ahmad MR, Malik A (2014). Risk factors of Cox B, Sneyd MJ, Paul C, Skegg DC (2006). Risk factors
prostate cancer: a case-control study in Faisalabad, for prostate cancer: a national case-control study.
Pakistan. Asian Pac J Cancer Prev, 15(23):10237–40. Int J Cancer, 119(7):1690–4. doi:10.1002/ijc.22022
doi:10.7314/APJCP.2014.15.23.10237 PMID:25556453 PMID:16646073
Bassett JK, Severi G, Baglietto L, MacInnis RJ, Hoang Dal Maso L, Zucchetto A, La Vecchia C, Montella M, Conti
HN, Hopper JL, et al. (2012). Weight change and E, Canzonieri V, et al. (2004). Prostate cancer and body
prostate cancer incidence and mortality. Int J Cancer, size at different ages: an Italian multicentre case-con-
131(7):1711–9. doi:10.1002/ijc.27414 PMID:22213024 trol study. Br J Cancer, 90(11):2176–80. PMID:15150581
Batty GD, Shipley MJ, Jarrett RJ, Breeze E, Marmot MG, Davies NM, Gaunt TR, Lewis SJ, Holly J, Donovan JL,
Smith GD (2005). Obesity and overweight in relation Hamdy FC, et al.; PRACTICAL consortium (2015). The
to organ-specific cancer mortality in London (UK): effects of height and BMI on prostate cancer incidence
findings from the original Whitehall study. Int J Obes and mortality: a Mendelian randomization study in
(Lond), 29(10):1267–74. doi:10.1038/sj.ijo.0803020 20,848 cases and 20,214 controls from the PRACTICAL
PMID:15997248 consortium. Cancer Causes Control, 26(11):1603–16.
Beebe-Dimmer JL, Dunn RL, Sarma AV, Montie JE, doi:10.1007/s10552-015-0654-9 PMID:26387087
Cooney KA (2007). Features of the metabolic syndrome Dimitropoulou P, Martin RM, Turner EL, Lane JA, Gilbert
and prostate cancer in African-American men. Cancer, R, Davis M, et al. (2011). Association of obesity with
109(5):875–81. doi:10.1002/cncr.22461 PMID:17265528 prostate cancer: a case-control study within the popu-
Beebe-Dimmer JL, Nock NL, Neslund-Dudas C, Rundle lation-based PSA testing phase of the ProtecT study.
A, Bock CH, Tang D, et al. (2009). Racial differences Br J Cancer, 104(5):875–81. doi:10.1038/sj.bjc.6606066
in risk of prostate cancer associated with meta- PMID:21266978
bolic syndrome. Urology, 74(1):185–90. doi:10.1016/j. Discacciati A, Orsini N, Andersson SO, Andrén O,
urology.2009.03.013 PMID:19428088 Johansson JE, Wolk A (2011). Body mass index in
Bergström A, Pisani P, Tenet V, Wolk A, Adami early and middle-late adulthood and risk of localised,
HO (2001). Overweight as an avoidable cause advanced and fatal prostate cancer: a population-based
of cancer in Europe. Int J Cancer, 91(3):421–30. prospective study. Br J Cancer, 105(7):1061–8.
doi:10.1038/bjc.2011.319 PMID:21847119
375
Engeland A, Tretli S, Bjørge T (2003). Height, body adiposity-related traits and risk of breast, ovarian,
mass index, and prostate cancer: a follow-up of prostate, lung and colorectal cancer. Int J Epidemiol,
950000 Norwegian men. Br J Cancer, 89(7):1237–42. 45(3):896–908. doi:10.1093/ije/dyw129 PMID:27427428
doi:10.1038/sj.bjc.6601206 PMID:14520453 Gerdtsson A, Poon JB, Thorek DL, Mucci LA, Evans MJ,
Esposito K, Chiodini P, Capuano A, Bellastella G, Scardino P, et al. (2015). Anthropometric measures
Maiorino MI, Parretta E, et al. (2013). Effect of meta- at multiple times throughout life and prostate cancer
bolic syndrome and its components on prostate cancer diagnosis, metastasis, and death. Eur Urol, 68(6):1076–
risk: meta-analysis. J Endocrinol Invest, 36(2):132–9. 82. doi:10.1016/j.eururo.2015.03.017 PMID:25794458
doi:10.1007/BF03346748 PMID:23481613 Giles GG, Severi G, English DR, McCredie MR, MacInnis
Felix JF, Bradfield JP, Monnereau C, van der Valk RJ, R, Boyle P, et al. (2003). Early growth, adult body size
Stergiakouli E, Chesi A, et al.; Bone Mineral Density and prostate cancer risk. Int J Cancer, 103(2):241–5.
in Childhood Study (BMDCS); Early Genetics and doi:10.1002/ijc.10810 PMID:12455039
Lifecourse Epidemiology (EAGLE) consortium; Early Giovannucci E, Liu Y, Platz EA, Stampfer MJ, Willett
Growth Genetics (EGG) Consortium; Bone Mineral WC (2007). Risk factors for prostate cancer incidence
Density in Childhood Study BMDCS (2016). Genome- and progression in the Health Professionals Follow-up
wide association analysis identifies three new suscep- Study. Int J Cancer, 121(7):1571–8. doi:10.1002/ijc.22788
tibility loci for childhood body mass index. Hum PMID:17450530
Mol Genet, 25(2):389–403. doi:10.1093/hmg/ddv472 Giovannucci E, Rimm EB, Stampfer MJ, Colditz GA,
PMID:26604143 Willett WC (1997). Height, body weight, and risk of
Fowke JH, McLerran DF, Gupta PC, He J, Shu XO, prostate cancer. Cancer Epidemiol Biomarkers Prev,
Ramadas K, et al. (2015). Associations of body mass 6(8):557–63. PMID:9264267
index, smoking, and alcohol consumption with pros- Gong Z, Neuhouser ML, Goodman PJ, Albanes D, Chi
tate cancer mortality in the Asia Cohort Consortium. C, Hsing AW, et al. (2006). Obesity, diabetes, and risk
Am J Epidemiol, 182(5):381–9. doi:10.1093/aje/kwv089 of prostate cancer: results from the Prostate Cancer
PMID:26243736 Prevention Trial. Cancer Epidemiol Biomarkers Prev,
Fowke JH, Motley SS, Concepcion RS, Penson DF, 15(10):1977–83. doi:10.1158/1055-9965.EPI-06-0477
Barocas DA (2012). Obesity, body composition, and PMID:17035408
prostate cancer. BMC Cancer, 12(1):23. doi:10.1186/1471- Guh DP, Zhang W, Bansback N, Amarsi Z, Birmingham
2407-12-23 PMID:22257467 CL, Anis AH (2009). The incidence of co-morbidi-
Friedenreich CM, McGregor SE, Courneya KS, Angyalfi ties related to obesity and overweight: a systematic
SJ, Elliott FG (2004). Case-control study of anthropo- review and meta-analysis. BMC Public Health, 9(1):88.
metric measures and prostate cancer risk. Int J Cancer, doi:10.1186/1471-2458-9-88 PMID:19320986
110(2):278–83. doi:10.1002/ijc.20110 PMID:15069694 Habel LA, Van Den Eeden SK, Friedman GD
Fujino Y; Japan Collaborative Cohort Study for Evaluation (2000). Body size, age at shaving initiation, and
of Cancer (2007). Anthropometry, development history prostate cancer in a large, multiracial cohort.
and mortality in the Japan Collaborative Cohort Study Prostate, 43(2):136–43. doi:10.1002/(SICI)1097-
for Evaluation of Cancer (JACC). Asian Pac J Cancer 0045(20000501)43:2<136::AID-PROS8>3.0.CO;2-L
Prev, 8(Suppl):105–12. PMID:18260709 PMID:10754529
Gallus S, Foschi R, Talamini R, Altieri A, Negri E, Hernandez BY, Park SY, Wilkens LR, Henderson BE,
Franceschi S, et al. (2007). Risk factors for prostate Kolonel LN (2009). Relationship of body mass,
cancer in men aged less than 60 years: a case-control height, and weight gain to prostate cancer risk in the
study from Italy. Urology, 70(6):1121–6. doi:10.1016/j. Multiethnic Cohort. Cancer Epidemiol Biomarkers
urology.2007.07.020 PMID:18158031 Prev, 18(9):2413–21. doi:10.1158/1055-9965.EPI-09-
Ganesh B, Saoba SL, Sarade MN, Pinjari SV (2011). Risk 0293 PMID:19723920
factors for prostate cancer: an hospital-based case-con- Hosseini M, SeyedAlinaghi S, Mahmoudi M, McFarland
trol study from Mumbai, India. Indian J Urol, 27(3):345– W (2010). A case-control study of risk factors for
50. doi:10.4103/0970-1591.85438 PMID:22022057 prostate cancer in Iran. Acta Med Iran, 48(1):61–6.
Gao C, Patel CJ, Michailidou K, Peters U, Gong J, PMID:21137672
Schildkraut J, et al.; the Colorectal Transdisciplinary IARC (2002). Weight control and physical activity. Lyon,
Study (CORECT); Discovery, Biology and Risk France: IARC Press (IARC Handbooks of Cancer
of Inherited Variants in Breast Cancer (DRIVE); Prevention, Vol. 6). Available from: http://publications.
Elucidating Loci Involved in Prostate Cancer iarc.fr/376.
Susceptibility (ELLIPSE); Follow-up of Ovarian Cancer Irani J, Lefebvre O, Murat F, Dahmani L, Doré B (2003).
Genetic Association and Interaction Studies (FOCI); Obesity in relation to prostate cancer risk: compar-
and Transdisciplinary Research in Cancer of the Lung ison with a population having benign prostatic
(TRICL) (2016). Mendelian randomization study of
376
hyperplasia. BJU Int, 91(6):482–4. doi:10.1046/j.1464- Máchová L, Cízek L, Horáková D, Koutná J, Lorenc J,
410X.2003.04133.x PMID:12656898 Janoutová G, et al. (2007). Association between obesity
Jackson MD, Walker SP, Simpson CM, McFarlane- and cancer incidence in the population of the District
Anderson N, Bennett FI, Coard KC, et al. (2010). Body Sumperk, Czech Republic. Onkologie, 30(11):538–42.
size and risk of prostate cancer in Jamaican men. Cancer doi:10.1159/000108284 PMID:17992023
Causes Control, 21(6):909–17. doi:10.1007/s10552-010- MacInnis RJ, English DR (2006). Body size and compo-
9520-y PMID:20157773 sition and prostate cancer risk: systematic review
Jee SH, Yun JE, Park EJ, Cho ER, Park IS, Sull JW, et al. and meta-regression analysis. Cancer Causes Control,
(2008). Body mass index and cancer risk in Korean men 17(8):989–1003. doi:10.1007/s10552-006-0049-z
and women. Int J Cancer, 123(8):1892–6. doi:10.1002/ PMID:16933050
ijc.23719 PMID:18651571 MacInnis RJ, English DR, Gertig DM, Hopper JL, Giles
Jonsson F, Wolk A, Pedersen NL, Lichtenstein P, Terry GG (2003). Body size and composition and pros-
P, Ahlbom A, et al. (2003). Obesity and hormone-de- tate cancer risk. Cancer Epidemiol Biomarkers Prev,
pendent tumors: cohort and co-twin control studies 12(12):1417–21. PMID:14693731
based on the Swedish Twin Registry. Int J Cancer, Magura L, Blanchard R, Hope B, Beal JR, Schwartz GG,
106(4):594–9. doi:10.1002/ijc.11266 PMID:12845658 Sahmoun AE (2008). Hypercholesterolemia and pros-
Keum N, Greenwood DC, Lee DH, Kim R, Aune D, Ju W, tate cancer: a hospital-based case-control study. Cancer
et al. (2015). Adult weight gain and adiposity-related Causes Control, 19(10):1259–66. doi:10.1007/s10552-
cancers: a dose-response meta-analysis of prospective 008-9197-7 PMID:18704722
observational studies. J Natl Cancer Inst, 107(2):djv088. Meyer J, Rohrmann S, Bopp M, Faeh D; Swiss National
doi:10.1093/jnci/djv088 PMID:25757865 Cohort Study Group (2015). Impact of smoking and
Lee IM, Sesso HD, Paffenbarger RS Jr (2001). A excess body weight on overall and site-specific cancer
prospective cohort study of physical activity and mortality risk. Cancer Epidemiol Biomarkers Prev,
body size in relation to prostate cancer risk (United 24(10):1516–22. doi:10.1158/1055-9965.EPI-15-0415
States). Cancer Causes Control, 12(2):187–93. PMID:26215293
doi:10.1023/A:1008952528771 PMID:11246848 Möller E, Adami HO, Mucci LA, Lundholm C, Bellocco
Lewis SJ, Murad A, Chen L, Davey Smith G, Donovan J, R, Johansson JE, et al. (2013). Lifetime body size and
Palmer T et al. (2010). Associations between an obesity prostate cancer risk in a population-based case-control
related genetic variant (FTO rs9939609) and prostate study in Sweden. Cancer Causes Control, 24(12):2143–
cancer risk. PLoS ONE, 5(10):e13485. doi:10.1371/ 55. doi:10.1007/s10552-013-0291-0 PMID:24048969
journal.pone.0013485 PMID:20976066 Møller E, Wilson KM, Batista JL, Mucci LA, Balter K,
Littman AJ, White E, Kristal AR (2007). Anthropometrics Giovannucci E (2016). Body size across the life course
and prostate cancer risk. Am J Epidemiol, 165(11):1271–9. and prostate cancer in the Health Professionals
doi:10.1093/aje/kwm013 PMID:17395597 Follow-up Study. Int J Cancer, 138(4):853–65.
Liu X, Rybicki BA, Casey G, Witte JS (2005). Relationship PMID:26355806
between body size and prostate cancer in a sibling Møller H, Roswall N, Van Hemelrijck M, Larsen SB,
based case-control study. J Urol, 174(6):2169–73. Cuzick J, Holmberg L, et al. (2015). Prostate cancer inci-
doi:10.1097/01.ju.0000181207.02213.06 PMID:16280757 dence, clinical stage and survival in relation to obesity:
Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH, a prospective cohort study in Denmark. Int J Cancer,
Day FR, et al.; LifeLines Cohort Study; ADIPOGen 136(8):1940–7. doi:10.1002/ijc.29238 PMID:25264293
Consortium; AGEN-BMI Working Group; Mori M, Masumori N, Fukuta F, Nagata Y, Sonoda T,
CARDIOGRAMplusC4D Consortium; CKDGen Miyanaga N, et al. (2011). Weight gain and family
Consortium; GLGC; ICBP; MAGIC Investigators; history of prostate or breast cancers as risk factors for
MuTHER Consortium; MIGen Consortium; PAGE prostate cancer: results of a case-control study in Japan.
Consortium; ReproGen Consortium; GENIE Asian Pac J Cancer Prev, 12(3):743–7. PMID:21627376
Consortium; International Endogene Consortium Nagata Y, Sonoda T, Mori M, Miyanaga N, Okumura K,
(2015). Genetic studies of body mass index yield new Goto K, et al. (2007). Dietary isoflavones may protect
insights for obesity biology. Nature, 518(7538):197–206. against prostate cancer in Japanese men. J Nutr,
doi:10.1038/nature14177 PMID:25673413 137(8):1974–9. PMID:17634273
Lukanova A, Björ O, Kaaks R, Lenner P, Lindahl B, Nemesure B, Wu SY, Hennis A, Leske MC; Prostate Cancer
Hallmans G, et al. (2006). Body mass index and cancer: in a Black Population (PCBP) Study Group (2012).
results from the Northern Sweden Health and Disease Central adiposity and prostate cancer in a Black popul-
Cohort. Int J Cancer, 118(2):458–66. doi:10.1002/ ation. Cancer Epidemiol Biomarkers Prev, 21(5):851–8.
ijc.21354 PMID:16049963 doi:10.1158/1055-9965.EPI-12-0071 PMID:22402288
377
Pan SY, Johnson KC, Ugnat AM, Wen SW, Mao Y; Rodriguez C, Patel AV, Calle EE, Jacobs EJ, Chao A,
Canadian Cancer Registries Epidemiology Research Thun MJ (2001). Body mass index, height, and prostate
Group (2004). Association of obesity and cancer risk cancer mortality in two large cohorts of adult men in
in Canada. Am J Epidemiol, 159(3):259–68. doi:10.1093/ the United States. Cancer Epidemiol Biomarkers Prev,
aje/kwh041 PMID:14742286 10(4):345–53. PMID:11319175
Pelucchi C, Serraino D, Negri E, Montella M, Dellanoce Samanic C, Chow WH, Gridley G, Jarvholm B,
C, Talamini R, et al. (2011). The metabolic syndrome Fraumeni JF Jr (2006). Relation of body mass index
and risk of prostate cancer in Italy. Ann Epidemiol, to cancer risk in 362,552 Swedish men. Cancer Causes
21(11):835–41. doi:10.1016/j.annepidem.2011.07.007 Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9
PMID:21982487 PMID:16841257
Pischon T, Boeing H, Weikert S, Allen N, Key T, Johnsen Samanic C, Gridley G, Chow WH, Lubin J, Hoover
NF, et al. (2008). Body size and risk of prostate cancer RN, Fraumeni JF Jr (2004). Obesity and cancer
in the European Prospective Investigation into Cancer risk among white and black United States veterans.
and Nutrition. Cancer Epidemiol Biomarkers Prev, Cancer Causes Control, 15(1):35–43. doi:10.1023/B:-
17(11):3252–61. doi:10.1158/1055-9965.EPI-08-0609 CACO.0000016573.79453.ba PMID:14970733
PMID:18990768 Schuurman AG, Goldbohm RA, Dorant E, van den
Porter MP, Stanford JL (2005). Obesity and the risk of Brandt PA (2000). Anthropometry in relation to pros-
prostate cancer. Prostate, 62(4):316–21. doi:10.1002/ tate cancer risk in the Netherlands Cohort Study. Am
pros.20121 PMID:15389806 J Epidemiol, 151(6):541–9. doi:10.1093/oxfordjournals.
Putnam SD, Cerhan JR, Parker AS, Bianchi GD, Wallace aje.a010241 PMID:10733035
RB, Cantor KP, et al. (2000). Lifestyle and anthropo- Sharpe CR, Siemiatycki J (2001). Joint effects of smoking
metric risk factors for prostate cancer in a cohort of and body mass index on prostate cancer risk.
Iowa men. Ann Epidemiol, 10(6):361–9. doi:10.1016/ Epidemiology, 12(5):546–51. doi:10.1097/00001648-
S1047-2797(00)00057-0 PMID:10964002 200109000-00014 PMID:11505174
Rapp K, Klenk J, Ulmer H, Concin H, Diem G, Speliotes EK, Willer CJ, Berndt SI, Monda KL, Thorleifsson
Oberaigner W, et al. (2008). Weight change and cancer G, Jackson AU, et al.; MAGIC; Procardis Consortium
risk in a cohort of more than 65,000 adults in Austria. (2010). Association analyses of 249,796 individ-
Ann Oncol, 19(4):641–8. doi:10.1093/annonc/mdm549 uals reveal 18 new loci associated with body mass
PMID:18056917 index. Nat Genet, 42(11):937–48. doi:10.1038/ng.686
Rapp K, Schroeder J, Klenk J, Stoehr S, Ulmer H, PMID:20935630
Concin H, et al. (2005). Obesity and incidence of Stocks T, Hergens MP, Englund A, Ye W, Stattin P (2010).
cancer: a large cohort study of over 145,000 adults Blood pressure, body size and prostate cancer risk in
in Austria. Br J Cancer, 93(9):1062–7. doi:10.1038/ the Swedish Construction Workers cohort. Int J Cancer,
sj.bjc.6602819 PMID:16234822 127(7):1660–8. doi:10.1002/ijc.25171 PMID:20087861
Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen M Tande AJ, Platz EA, Folsom AR (2006). The metabolic
(2008). Body-mass index and incidence of cancer: syndrome is associated with reduced risk of prostate
a systematic review and meta-analysis of prospec- cancer. Am J Epidemiol, 164(11):1094–102. doi:10.1093/
tive observational studies. Lancet, 371(9612):569–78. aje/kwj320 PMID:16968859
doi:10.1016/S0140-6736(08)60269-X PMID:18280327 Wallström P, Bjartell A, Gullberg B, Olsson H, Wirfält E
Robinson WR, Poole C, Godley PA (2008). Systematic (2009). A prospective Swedish study on body size, body
review of prostate cancer’s association with body size composition, diabetes, and prostate cancer risk. Br J
in childhood and young adulthood. Cancer Causes Cancer, 100(11):1799–805. doi:10.1038/sj.bjc.6605077
Control, 19(8):793–803. doi:10.1007/s10552-008-9142-9 PMID:19436298
PMID:18347923 WCRF/AICR (2014). Continuous Update Project Report.
Robinson WR, Stevens J, Gammon MD, John EM (2005). Diet, nutrition, physical activity, and prostate cancer.
Obesity before age 30 years and risk of advanced Washington (DC), USA: American Institute for Cancer
prostate cancer. Am J Epidemiol, 161(12):1107–14. Research. Available from: https://www.wcrf.org/sites/
doi:10.1093/aje/kwi150 PMID:15937019 default/files/Prostate-Cancer-2014-Report.pdf.
Rodriguez C, Freedland SJ, Deka A, Jacobs EJ, McCullough Whitlock G, Lewington S, Sherliker P, Clarke R, Emberson
ML, Patel AV, et al. (2007). Body mass index, weight J, Halsey J, et al.; Prospective Studies Collaboration
change, and risk of prostate cancer in the Cancer (2009). Body-mass index and cause-specific mortality
Prevention Study II Nutrition Cohort. Cancer Epidemiol in 900 000 adults: collaborative analyses of 57 prospec-
Biomarkers Prev, 16(1):63–9. doi:10.1158/1055-9965. tive studies. Lancet, 373(9669):1083–96. doi:10.1016/
EPI-06-0754 PMID:17179486 S0140-6736(09)60318-4 PMID:19299006
378
Wright ME, Chang SC, Schatzkin A, Albanes D,

Kipnis V, Mouw T, et al. (2007). Prospective study
of adiposity and weight change in relation to prostate
cancer incidence and mortality. Cancer, 109(4):675–84.
doi:10.1002/cncr.22443 PMID:17211863
Wuermli L, Joerger M, Henz S, Schmid HP, Riesen WF,
Thomas G, et al. (2005). Hypertriglyceridemia as a
possible risk factor for prostate cancer. Prostate Cancer
Prostatic Dis, 8(4):316–20. doi:10.1038/sj.pcan.4500834
PMID:16158078
Zhang JQ, Geng H, Ma M, Nan XY, Sheng BW (2015).
Metabolic syndrome components are associated with
increased prostate cancer risk. Med Sci Monit, 21:2387–
96. doi:10.12659/MSM.893442 PMID:26275075
379
2.2.15 Cancer of the testis men with BMI > 27.4 kg/m2 (n = 26) compared
with men with BMI ≤ 23.15 kg/m2 (OR, 0.42;
Cancer of the testis is a rare malignancy, 95% CI, 0.24–0.75). One study showed that
accounting for 1% of incident cases of cancer in high BMI in men aged 18–29 years was signifi-
men, but the testis is the most common cancer cantly more frequent in testicular cancer cases
site for men aged 15–44 years in developed coun- than in controls (Dieckmann et al., 2009). In
tries. To date, the most important identified risk one study, analysis by subtype yielded an odds
factor for testicular cancer is an undescended ratio of 3.66 (95% CI, 1.87–7.15) for obese men
testicle. Increased risk has also been associated (BMI > 31 kg/m2) with non-seminoma testicular
with family history of testicular cancer, various cancer (n = 11) (Garner et al., 2003).
genetic factors, and several perinatal risk factors. A meta-analysis of the earlier cohort study
In 2001, the Working Group of the IARC and 10 case–control studies showed an inverse
Handbook on weight control and physical activity association between overweight and testicular
(IARC, 2002) concluded that the evidence of an cancer (OR, 0.92; 95% CI, 0.86–0.98), which was
association between avoidance of weight gain not significant for obesity (OR, 0.93; 95% CI,
and testicular cancer was inadequate. 0.75–1.15) (Lerro et al., 2010).
(a) Cohort studies
Since 2000, only one cohort study of excess
body weight in relation to risk of testicular
cancer has been published: a Norwegian cohort
of approximately 600 000 men aged 14–44 years
(Bjørge et al., 2006). For overweight and obesity
compared with normal BMI, the relative risks
were 0.89 (95% CI, 0.77–1.03) and 0.83 (95%
CI, 0.58–1.17), respectively, and the relative risk
per 1 kg/m2 increase in BMI was 0.97 (95% CI,
0.95–1.00). There was no statistically significant
heterogeneity of results between histological
subtypes of testicular cancer.

A total of seven population- or hospital-based
case–control studies published after 2000
focused on the association between BMI and
weight and testicular cancer (Table 2.2.15; web
only; available at: http://publications.iarc.fr/570).
In four studies, there was no overall significant
association with BMI for all testicular cancer
cases (Dieckmann & Pichlmeier, 2002; Richiardi
et al., 2003; Pan et al., 2004; McGlynn et al.,
2007). Giannandrea et al. (2012) found an inverse
association with all testicular cancer cases for
380
References
Bjørge T, Tretli S, Lie AK, Engeland A (2006). The impact
of height and body mass index on the risk of testic-
ular cancer in 600,000 Norwegian men. Cancer Causes
Control, 17(7):983–7. doi:10.1007/s10552-006-0032-8
PMID:16841265
Dieckmann KP, Hartmann JT, Classen J, Diederichs M,
Pichlmeier U (2009). Is increased body mass index
associated with the incidence of testicular germ
cell cancer? J Cancer Res Clin Oncol, 135(5):731–8.
doi:10.1007/s00432-008-0504-1 PMID:19002497
Dieckmann KP, Pichlmeier U (2002). Is risk of testicular
cancer related to body size? Eur Urol, 42(6):564–9.
doi:10.1016/S0302-2838(02)00467-0 PMID:12477651
Garner MJ, Birkett NJ, Johnson KC, Shatenstein B,
Ghadirian P, Krewski D; Canadian Cancer Registries
Epidemiology Research Group (2003). Dietary
risk factors for testicular carcinoma. Int J Cancer,
106(6):934–41. doi:10.1002/ijc.11327 PMID:12918073
Giannandrea F, Paoli D, Lombardo F, Lenzi A, Gandini
L (2012). Case-control study of anthropometric meas-
ures and testicular cancer risk. Front Endocrinol
(Lausanne), 3:144. doi:10.3389/fendo.2012.00144
PMID:23189072
iarc.fr/376.
Lerro CC, McGlynn KA, Cook MB (2010). A systematic
review and meta-analysis of the relationship between
body size and testicular cancer. Br J Cancer, 103(9):1467–
74. doi:10.1038/sj.bjc.6605934 PMID:20978513
McGlynn KA, Sakoda LC, Rubertone MV, Sesterhenn
IA, Lyu C, Graubard BI, et al. (2007). Body size, dairy
consumption, puberty, and risk of testicular germ cell
tumors. Am J Epidemiol, 165(4):355–63. doi:10.1093/
aje/kwk019 PMID:17110638
Pan SY, Johnson KC, Ugnat AM, Wen SW, Mao Y;
Canadian Cancer Registries Epidemiology Research
Group (2004). Association of obesity and cancer risk
in Canada. Am J Epidemiol, 159(3):259–68. doi:10.1093/
aje/kwh041 PMID:14742286
Richiardi L, Askling J, Granath F, Akre O (2003). Body
size at birth and adulthood and the risk for germ-cell
testicular cancer. Cancer Epidemiol Biomarkers Prev,
12(7):669–73. PMID:12869410
381
2.2.16 Cancer of the kidney (renal cell The findings are remarkably consistent across
carcinoma) studies, showing increasing risk of kidney cancer
with increasing BMI. The association is approxi-
Cancer of the kidney accounts for about 2% mately linear with increasing BMI. A meta-ana-
of all cancers diagnosed. Established epidemi- lysis of 21 cohort studies concluded that there
ological risk factors for kidney cancer include was consistency of the association across sexes
tobacco smoking, which can double the risk of the and world regions, with a relative risk for obesity
disease in smokers compared with non-smokers. compared with normal weight of 1.63 (95% CI,
Other established risk factors, which are closely 1.50–1.77) in men and 1.95 (95% CI, 1.81–2.10) in
associated with obesity, are high blood pressure women (Wang & Xu, 2014).
and pre-existing diabetes mellitus. Some investigators have assessed the associ-
The two most common types of kidney ation between BMI at different ages and subse-
cancer are renal cell carcinoma (RCC) and tran- quent risk of kidney cancer (Nicodemus et al.,
sitional cell carcinoma (also known as urothelial 2004; van Dijk et al., 2004; Adams et al., 2008).
cell carcinoma) of the renal pelvis. About 90% of In general, the strong positive association
kidney cancers are RCCs. Histological subtypes between baseline BMI and kidney cancer risk
of RCC include clear cell tumours (about 70% of was also seen for BMI in middle adulthood, but
RCCs), papillary tumours (also called chromo- much less so for BMI in early adulthood (ages
philic RCC; about 10% of RCCs), and chromo- 18–20 years).
phobe RCC (about 5% of RCCs). Various rarer Five cohort studies reported on the asso-
types of RCC exist, each representing less than ciation between measures of waist circumfer-
1% of RCCs. ence and kidney cancer risk (Nicodemus et al.,
In 2001, the Working Group of the IARC 2004; Pischon et al., 2006; Adams et al., 2008;
Handbook on weight control and physical activity Sanfilippo et al., 2014; Kabat et al., 2015). In all
(IARC, 2002) concluded that there was sufficient of the studies, measures of waist circumference
evidence for a cancer-preventive effect of avoid- were associated with kidney cancer risk similarly
ance of weight gain for RCC. The 2007 WCRF to BMI.
review concluded that there was convincing
evidence of a positive association between body (b) Case–control studies
fatness and kidney cancer risk (WCRF/AICR, Since 2000, a total of nine case–control
2007). In 2015, the WCRF Continuous Update studies in China, Europe, and North America
Project reaffirmed the 2007 conclusions (WCRF/ have reported on the association of BMI with
AICR, 2015). risk of RCC (Table 2.2.16b). In all of the studies
(a) Cohort studies except one (Wang et al., 2012), BMI was assessed
through self-reports by patients with RCC and
Since 2000, 19 cohort studies of anthropo- control subjects, with reference to a variable time
morphic measures and risk of kidney cancer frame before cancer diagnosis and an equivalent
have been published (excluding analyses that time frame for the controls. Of the nine studies,
were later updated and analyses based on fewer seven adjusted for smoking and two did not.
than 100 incident cases). Table 2.2.16a shows Other possible confounding factors considered
those findings by BMI at baseline, with comments and adjusted for in some studies included use
on findings according to other anthropometric of artificial sweeteners, pre-existing diabetes
measures of body fatness and weight changes mellitus, use of anti-hypertensive drugs, and
over the life-course.
382
exposures to pesticides, herbicides, or certain (d) Mendelian randomization study

industrial exposures. There has been one Mendelian randomiza-
Most of the studies showed an increased risk tion study, which used the FTO rs9939609 SNP,
of RCC with higher BMI, in men, women, or both robustly associated with BMI (Frayling et al.,
sexes, although this positive association was not 2007; Scuteri et al., 2007; Peeters et al., 2008),
statistically significant in all studies. In all the to estimate the causal association between BMI
larger studies, including the earlier studies, there and kidney cancer, among other cancer types
was a statistically significant trend of increasing (Brennan et al., 2009; Table 2.2.16d). Those
RCC risk with increasing BMI, up to an approx- with the FTO AA genotype had a higher BMI
imately 2–3-fold increased risk for the highest than controls with the TT genotype (difference,
versus the lowest BMI categories, both in men 1.14 kg/m2; 95% CI, 0.66–1.61; P < 0.00001).
and in women. In several studies, RCC risk was Mendelian randomization analyses showed that
also found to be positively associated with BMI each 1 kg/m2 increase in BMI was weakly associ-
at younger ages (20–40 years) (Brock et al., 2007; ated with an increased risk of kidney cancer (OR,
Dal Maso et al., 2007; Beebe-Dimmer et al., 2012). 1.11; 95% CI: 0.91–1.37; P = 0.31), which was more
Purdue et al. (2013) combined the data from pronounced in those younger than 50 years (OR,
a large case–control study in the USA (Beebe- 1.90; 95% CI, 1.16–2.27; P = 0.0002).
Dimmer et al., 2012) and a multicentre study
in central and eastern Europe (Brennan et al.,
2008) to examine the association of BMI with
different histological subtypes of RCC and found
a positive association of BMI with risk of clear
cell RCC (n = 1524; OR per 5 kg/m2, 1.2; 95%
CI, 1.1–1.3) and chromophobe RCC (n = 80; OR
per 5 kg/m2, 1.2; 95% CI, 1.1–1.4), but not papil-
lary RCC (n = 237; OR per 5 kg/m2, 1.1; 95% CI,
1.0–1.2) or RCC not otherwise specified (n = 367;
OR per 5 kg/m2, 1.0; 95% CI, 0.7–1.4).
(c) Meta-analyses
Several meta-analyses of cohort and/or case–
control studies assessed the association between
BMI and kidney cancer risk (Table 2.2.16c).
Bergström et al. (2001) combined data from 14
studies in men and 14 studies in women, and
reported a summary relative risk of RCC of 1.07
per 1 kg/m2 increase in BMI in both men and
women. Two more recent meta-analyses reported
summary relative risks for cohort studies and
case–control studies separately, for women
(Mathew et al., 2009) and for men (Ildaphonse
et al., 2009) respectively, all in the range of 1.05
to 1.07.
383
384

Table 2.2.16a Cohort studies of measures of body fatness and cancer of the kidney
Reference Total number of Exposure categories Exposed Relative risk Covariates Comments
Cohort subjects cases (95% CI)
Location Sex
mortality
Calle et al. (2003) 404 576 BMI Age, education level,
Population-based Men 18.5–24.9 305 1.00 smoking, physical
cohort Mortality 25–29.9 437 1.18 (1.02–1.37) activity, alcohol
1982–1998 35–39.9 14 1.70 (0.99–2.92) status, aspirin, fat
[Ptrend] [0.002] intake, vegetable intake
495 477 BMI Age, education level,
Women 18.5–24.9 243 1.00 smoking, physical
Mortality 25–29.9 153 1.33 (1.08–1.63) activity, alcohol
30–34.9 55 1.66 (1.23–2.24) consumption, marital
35–39.9 12 1.70 (0.94–3.05) status, aspirin, fat
≥ 40 10 4.75 (2.50–9.04) intake, vegetable
[Ptrend] [< 0.001] intake, HRT
Bjørge et al. (2004) 1 037 788 BMI Age Association weaker in current
Population-based Women 18.5–24.9 1061 1.00 and former smokers than in
cohort Incidence 25–29.9 977 1.32 (1.21–1.45) never-smokers
Norway ≥ 30 568 1.85 (1.66–2.06)
1963–2001 [Ptrend] [< 0.001]
963 442 BMI Age Association weaker in current
Men 18.5–24.9 1908 1.00 and former smokers than in
Incidence 25–29.9 1638 1.18 (1.11–1.26) never-smokers
≥ 30 267 1.55 (1.36–1.76)
[Ptrend] [< 0.001]
Nicodemus et al. 34 637 BMI Age Postmenopausal women.
(2004) Women < 22.9 16 1.00 Weight at ages 30 yr, 40 yr,
Iowa Women’s Health Incidence 22.9–25.0 13 0.80 (0.38–1.65) and 50 yr (but not at 18 yr)
Study 25.0–27.4 24 1.46 (0.77–2.74) associated similarly. WC also
USA 27.4–30.6 31 1.87 (1.02–3.41) associated
1986–2000 > 30.6 40 2.49 (1.39–4.44)
[Ptrend] [< 0.0001]
Location Sex
mortality
van Dijk et al. (2004) 120 852 BMI Age, sex No association with BMI at
Netherlands Cohort Women and men 23–24.9 83 1.00 age 20 yr
Study Incidence 25–26.9 54 0.92 (0.61–1.38)
The Netherlands 27–29.9 62 1.46 (0.97–2.21)
1986–1995 ≥ 30 16 1.04 (0.54–1.99)
[Ptrend] [0.04]
per 1 kg/m2 1.07 (1.02–1.12)
Flaherty et al. (2005) 118 191 BMI Age, hypertension, RR for BMI ≥ 30 adjusted for
Nurses’ Health Study Women < 22.0 40 1.0 smoking age only
USA Incidence 22.0–24.9 47 1.3 (0.9–2.0)
1976–2000 25.0–27.9 27 1.6 (0.9–2.5)
28.0–29.9 14 2.2 (1.2–4.1)
≥ 30 26 2.7 (1.6–4.4)
[Ptrend] [< 0.001]
Flaherty et al. (2005) 48 953 BMI Age, hypertension,
Health Professionals Men < 22.0 4 1.0 smoking
Follow-Up Study Incidence 22.0–24.9 37 2.1 (0.7–5.9)
USA 25.0–27.9 45 2.4 (0.9–6.8)
1986–1998 28.0–29.9 12 2.1 (0.7–6.6)
≥ 30 10 2.1 (0.7–6.8)
[Ptrend] [0.19]
Rapp et al. (2005) 67 447 BMI Age, smoking,

Population-based Men 18.5–24.9 46 1.00 occupation
cohort Incidence 25–29.9 70 1.19 (0.82–1.74)
Austria ≥ 30 21 1.46 (0.87–2.46)
1985–2002 [Ptrend] [0.14]
78 484 BMI Age, smoking,
Women 18.5–24.9 32 1.00 occupation
Incidence 25–29.9 44 1.81 (1.13–2.89)
≥ 30 12 1.14 (0.58–2.24)
[Ptrend] [0.3]
385
386

Location Sex
mortality
Pischon et al. (2006) 218 889 BMI, quintiles Smoking, education WC also associated
EPIC cohort Women < 21.8 12 1.00 level, alcohol
Europe Incidence 21.8–23.7 22 1.48 (0.73–3.01) consumption, physical
1992–2004 23.8–25.9 24 1.39 (0.69–2.80) activity
26.0–28.9 37 1.99 (1.03–3.88)
> 29.0 37 2.25 (1.14–4.44)
[Ptrend] [0.009]
129 660 BMI, quintiles
Men < 23.6 29 1.00
Incidence 23.6–25.3 35 1.07 (0.65–1.77)
25.4–27.0 23 0.67 (0.39–1.18)
27.1–29.3 28 0.84 (0.49–1.43)
> 29.4 40 1.22 (0.74–2.03)
[Ptrend] [0.51]
Samanic et al. (2006) 362 552 BMI Age, year, smoking,
Swedish Construction Men 18.5–24.9 444 1.00 hypertension
Worker Cohort Incidence 25–29.9 448 1.23 (1.08–1.42)
Sweden ≥ 30 94 1.61 (1.27–2.04)
1971–1999 [Ptrend] [< 0.001]
Reeves et al. (2007) 1.2 million BMI Age, region, SES, Association slightly weaker in
Million Women Study Women < 22.5 119 0.95 (0.79–1.14) reproductive history, never-smokers
United Kingdom Incidence 22.5–24.9 165 1.00 (0.86–1.17) smoking, alcohol
1995–2005 25.0–27.4 155 1.10 (0.94–1.28) consumption, physical
27.5–29.9 106 1.19 (0.99–1.44) activity, HRT use
≥ 30 178 1.52 (1.31–1.77)
per 10 kg/m2 1.53 (1.27–1.84)
Setiawan et al. (2007) 85 964 BMI Age, ethnicity,
Multiethnic Cohort Women < 25 38 1.00 smoking, alcohol
USA Incidence 25–29.9 52 2.03 (1.31–3.15) consumption,
1993–2002 ≥ 30 37 2.27 (1.37–3.74) hypertension, physical
[Ptrend] [0.001] activity
75 172 BMI
Men < 25 77 1.00
Incidence 25–29.9 93 1.14 (0.84–1.55)
≥ 30 50 1.76 (1.20–2.58)
[Ptrend] [0.005]
Location Sex
mortality
Adams et al. (2008) 214 906 BMI Age, smoking, physical Similar association with BMI
NIH-AARP cohort Women 18.5–22.5 17 1.00 activity, protein intake, at age 50 yr; no association
USA Incidence 22.5–24.9 33 1.66 (0.92–2.98) diabetes, hypertension at age 18 yr or 35 yr. WC also
1995–2003 25–27.5 46 2.44 (1.39–4.26) associated
27.5–29.9 27 2.27 (1.23–4.20)
≥ 30 64 2.67 (1.53–4.66)
[Ptrend] [0.002]
312 500 BMI Similar association with BMI
Men 18.5–22.5 28 1.00 at age 50 yr; no association
Incidence 22.5–24.9 88 1.12 (0.73–1.72) at age 18 yr or 35 yr. WC also
25–27.5 169 1.51 (1.01–2.26) associated
27.5–29.9 127 1.74 (1.15–2.63)
≥ 30 152 1.87 (1.24–2.82)
[Ptrend] [< 0.0005]
Jee et al. (2008) 443 273 BMI Age, smoking
Cohort from National Women < 20 22 0.48 (0.28–0.82)
Health Insurance Incidence 20–22.9 95 0.70 (0.49–0.99)
Corporation 23–24.9 100 1.00
Republic of Korea 25–29.9 100 0.92 (0.64–1.31)
1992–2007 ≥ 30 14 1.21 (0.58–2.53)
[Ptrend] [0.0042]
770 556 BMI Age, smoking Association weaker in ever-

Men < 20 97 0.64 (0.49–0.84) smokers than in non-smokers
Incidence 20–22.9 430 0.67 (0.56–0.79)
23–24.9 425 1.00
25–29.9 392 1.11 (0.93–1.31)
≥ 30 16 1.38 (0.76–2.52)
[Ptrend] [< 0.0001]
Song et al. (2008) 170 481 BMI Age, height, smoking,
Korean medical Women 21.0–22.9 18 1.00 alcohol consumption,
insurance cohort Incidence 23.0–24.9 34 1.74 (0.94–3.22) physical activity, pay
Republic of Korea 25.0–26.9 29 1.74 (0.92–3.29) grade
1993–2003 27.0–29.9 14 1.37 (0.66–2.84)
≥ 30.0 7 2.61 (1.06–6.41)
[Ptrend] [< 0.05]
387
388

Location Sex
mortality
Wilson et al. (2009) 27 111 BMI Age, energy intake
ATBC cohort Men < 23.7 41 1.00
Finland Incidence 23.7–26.0 70 1.8 (1.3–2.7)
1985–2002 26.0–28.5 65 1.8 (1.2–2.7)
≥ 28.5 69 2.1 (1.4–3.1)
[Ptrend] [< 0.001]
Sawada et al. (2010) 46 837 BMI Age, area, tobacco use, Analysis of data in women
Population sample of Men < 21 22 1.86 (1.01–3.45) alcohol consumption, (n = 52 625) was based on
Japan Incidence 21.0–22.9 20 1.16 (0.62–2.16) physical activity, very small number of cases;
Japan 23.0–24.9 21 1.00 hypertension, diabetes association unclear
1990–2006 25.0–26.9 18 1.39 (0.73–2.63)
≥ 27.0 20 1.99 (1.04–3.81)
Häggström et al. 281 468 BMI, quintiles Age, time of
(2013) Women Q1 24 1.00 measurement
3 cohorts Incidence Q2 28 0.95 (0.52–1.74)
Austria, Norway, Q3 61 1.84 (1.08–3.13)
Sweden Q4 66 1.74 (1.02–2.94)
1994–2006 Q5 84 2.21 (1.32–3.70)
[Ptrend] [0.0002]
278 920 BMI, quintiles
Men Q1 89 1.00
Incidence Q2 108 1.11 (0.81–1.52)
Q3 100 0.94 (0.68–1.29)
Q4 139 1.28 (0.95–1.73)
Q5 156 1.51 (1.13–2.03)
[Ptrend] [0.001]
Macleod et al. (2013) 77 260 BMI Age, sex, race,
Population-based Women and men < 25 59 1.00 smoking, alcohol
cohort Incidence 25–29.9 104 1.23 (0.88–1.72) consumption,
USA 30–34.9 47 1.20 (0.81–1.78) hypertension, diabetes
2000–2009 ≥ 35 28 1.71 (1.06–2.79)
Bhaskaran et al. (2014) 5.24 million BMI 1906 total Age, year, sex, Similar findings for never-
Clinical Practice Women and men per 5 kg/m2 1.25 (1.17–1.33) diabetes, SES, alcohol smokers
Research Datalink Incidence consumption, tobacco
United Kingdom use
1987–2012
Location Sex
mortality
Sanfilippo et al. (2014) 156 774 BMI Age, race/ethnicity, (See also Kabat et al., 2015)
Women’s Health Women 18.5–24.9 108 1.00 diastolic blood WC also associated with
Initiative cohort Incidence 25–29.9 144 1.32 (1.03–1.70) pressure increased risk
USA 30–34.9 83 1.47 (1.10–1.96)
1993–1998 35–39.9 45 1.91 (1.33–2.75)
≥ 40 27 2.48 (1.61–3.80)
Kabat et al. (2015) 143 901 BMI, quintiles 376 total Age, alcohol WC also associated with risk
Women’s Health Women Q1 1.00 consumption,
Initiative cohort Incidence Q2 0.89 (0.61–1.28) smoking, physical
USA Q3 1.21 (0.86–1.71) activity, age at
1992–2013 Q4 1.36 (0.96–1.91) menarche, age at first
Q5 1.73 (1.24–2.42) birth, parity, HRT
[Ptrend] [< 0.0001] use, family history
of kidney cancer,
ethnicity, education
level
ATBC, Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study; BMI, body mass index (in kg/m 2); CI, confidence interval; EPIC, European Prospective Investigation into Cancer
and Nutrition; HRT, hormone replacement therapy; NIH-AARP, National Institutes of Health–AARP Diet and Health Study; RR, relative risk; SES, socioeconomic status; WC, waist
circumference; yr, year or years

389
390

Table 2.2.16b Case–control studies of measures of body fatness and cancer of the kidney
Reference Total number of cases Exposure categories Exposed Relative risk Adjustment for confounding
Study location Total number of controls cases (95% CI) Comments
Shapiro et al. (1999) 238 (155 men, 83 women) Median BMI Women: Age, diabetes mellitus,
USA (western 616 (261 men, 355 women) < 22.20 5 1.0 hypertension
Washington state) Population 22.20–24.85 16 3.3 (1.1–9.7) Median BMI calculated using
1980–1995 24.86–28.25 20 3.6 (1.3–10.3) median weight recorded in
> 28.25 29 4.1 (1.5–11.8) medical records during the 5-yr
Top 10% (> 32.99) 6.0 (1.9–18.8) period immediately before the
Median BMI Men: reference date (2 yr before date
< 24.59 23 1.0 of diagnosis and corresponding
24.59–26.39 27 1.1 (0.5–2.1) index date for controls)
26.40–28.88 26 1.0 (0.5–2.0)
> 28.88 45 1.8 (0.9–3.5)
Top 10% (> 31.85) 2.2 (1.0–5.0)
Hu et al. (2003) 1279 (691 men, 588 women) BMI 2 yr before study entry Women: 10-year age group, province,
Canada (8 provinces) 5370 (2696 men, 2674 women) < 18.5–24.9 221 1.0 education level, pack-years of
1994–1997 Population 25.0–29.9 200 1.5 (1.20–1.90) smoking, alcohol consumption,
30.0–34.9 100 2.5 (1.90–3.40) total intake of meat, vegetables,
35.0–39.9 31 2.7 (1.70–4.40) and fruit
≥ 40.00 33 3.8 (2.30–6.40)
BMI 2 yr before study entry Men:
< 18.5–24.9 147 1.0
25.0–29.9 369 2.20 (1.70–2.70)
30.0–34.9 144 2.80 (2.20–3.80)
35.0–39.9 21 1.90 (1.10–3.30)
≥ 40.00 8 3.70 (1.50–9.40)
Chiu et al. (2006) 406 (261 men, 145 women) BMI in 60s Men: All respondents: age, total energy
USA 2434 (1601 men, 833 women) ≤ 23.48 49 1.0 intake, intake of red meat, intake
1986–1990 Population 23.49–25.17 33 0.6 (0.3–1.1) of vegetables, hypertension,
25.18–27.35 34 0.6 (0.3–1.1) education level, smoking, family
27.36–30.07 27 0.8 (0.4–1.7) history of kidney cancer, proxy
≥ 30.08 20 0.4 (0.2–1.0) status; women only: marital
[Ptrend] [0.2] status
BMI in 60s Women: Analyses for BMI at age 20 yr and
≤ 22.20 23 1.0 age 40 yr gave very similar results
22.21–24.32 18 0.5 (0.2–1.4) to BMI at age 60 yr
24.33–27.31 20 1.0 (0.4–2.5)
27.33–30.13 13 0.7 (0.3–2.1)
≥ 30.14 21 2.3 (0.9–6.0)
[Ptrend] [0.1]
Brock et al. (2007) 406 (261 men, 145 women) BMI at age 20 yr Age, sex, proxy status, pack-years
USA (Iowa) 2434 (1601 men, 833 women) < 25 271 1.00 of smoking
1985–1989 Population 25−30 62 1.54 (1.10–2.17) Analysis also reported for men
≥ 30 21 2.75 (1.51–5.01) and women separately
BMI at age 40 yr
< 25 180 1.00
25−30 130 1.36 (1.04–1.79)
≥ 30 51 2.08 (1.39–3.12)
BMI at age 60 yr

< 25 111 1.00
25−30 93 1.12 (0.81–1.55)
≥ 30 39 1.46 (0.94–2.28)
Dal Maso et al. 767 (494 men, 273 women) BMI at age 30 yr Calendar period of interview,
(2007) 1534 (988 men, 546 women) < 25 492 1.00 years of education, smoking
Italy Hospital 25– < 30 194 1.17 (0.95–1.45) habits, family history of kidney
1992–2004 ≥ 30 38 1.46 (0.95–2.25) cancer
[Ptrend] [0.04]
BMI at age 50 yr
< 25 256 1.00
25– < 30 265 1.17 (0.94–1.45)
≥ 30 89 1.48 (1.07–2.03)
[Ptrend] [0.02]
391
392

Dal Maso et al. BMI 1 yr before diagnosis
(2007) < 25 281 1.00
(cont.) 25– < 30 347 0.95 (0.78–1.16)
≥ 30 136 1.29 (0.99–1.69)
[Ptrend] [0.16]
By smoking status
Never-smokers:
< 25 39 1.00
25– < 30 62 1.25 (0.74–2.09)
≥ 30 82 1.83 (1.10–3.04)
Ever-smokers:
< 25 87 1.00
25– < 30 93 0.96 (0.66–1.41)
≥ 30 112 1.37 (0.95–1.98)
By histological type
Clear cell subtype:
< 25 71 1.00
25– < 30 89 0.99 (0.68–1.44)
≥ 30 121 1.40 (0.98–1.99)
Other subtype:
< 25 23 1.00
25– < 30 38 1.30 (0.73–2.30)
≥ 30 41 1.62 (0.92–2.85)
Brennan et al. (2008) 1097 (648 men, 449 women) BMI 2 yr before interview Men: Age, smoking, history of
Czech Republic, 1476 (952 men, 524 women) < 25 191 1.00 hypertension, country
Poland, Romania, Hospital 25–27.5 166 1.19 (0.91–1.56)
Russian Federation 27.5–29.99 125 1.32 (0.98–1.79)
(7 centres) 30–35 133 1.70 (1.25–2.31)
1998–2003 > 35 32 1.72 (1.01–2.94)
[Ptrend] [0.001]
BMI 2 yr before interview Women:
< 25 136 1.00
25–27.5 87 0.86 (0.60–1.25)
27.5–29.99 98 1.16 (0.80–1.70)
30–35 98 0.95 (0.66–1.38)
> 35 30 0.85 (0.49–1.48)
[Ptrend] [0.68]
Beebe-Dimmer et al. 1214 (720 men, 494 women) BMI 5 yr before interview Age, education level,
(2012) 1234 (689 men, 545 women) < 25.0 240 1.0 hypertension, family history of
USA Population 25.0–29.9 436 1.2 (0.9–1.5) renal cancer, smoking history,
2002–2007 30.0–34.9 298 1.5 (1.2–2.1) study centre
≥ 35 230 1.6 (1.1–2.2) Analysis of BMI at age 21 yr gave
per 1 kg/m2 1.02 (1.01–1.04) similar results
[Ptrend] [0.0013]
Wang et al. (2012) 250 Current BMI Univariate analysis
China 299 < 25 157 1.00
2007–2009 Hospital ≥ 25 93 1.94 (1.34–2.81)
Purdue et al. (2013) 2314 BMI a few years before interview Study centre, age, sex, race,
USA (Detroit and 2711 Clear cell: 1524 education level, BMI, smoking
Chicago; USKC Population (USKC), hospital per 5 kg/m2 1.2 (1.1–1.3) status, history of diagnosed
study) and Europe (CEERCC) Papillary: 237 hypertension, family history of
(Czech Republic, per 5 kg/m2 1.1 (1.0–1.2) kidney cancer
Poland, Romania, Time before interview: 5 yr
Chromophobe: 80
Russian Federation; (USKC), 2 yr (CEERCC)
per 5 kg/m2 1.2 (1.1–1.4)
CEERCC study)
2002–2007 Other/NOS: 367
per 5 kg/m2 1.0 (0.7–1.4)
BMI, body mass index (in kg/m ); CEERCC, Central and Eastern European Renal Cell Cancer Study; CI, confidence interval; NOS, not otherwise specified; USKC, United States Kidney
2
Cancer; yr, year or years

393
394

Table 2.2.16c Meta-analyses of measures of body fatness and cancer of the kidney
Reference Total number of studies Exposure categories Relative risk Heterogeneity values
Sex (95% CI)
Bergström et al. (2001) 28 studies (6 cohort studies, 22 case– BMI, per 1 kg/m2
control studies; 16 population-based, 6 All 1.07 (1.05–1.09) P heterogeneity = 0.03
hospital-based) Men 1.07 (1.04–1.09) P heterogeneity = 0.08
Men: 14 studies Women 1.07 (1.05–1.09) P heterogeneity = 0.24
Women: 14 studies
Mathew et al. (2009) 28 studies (15 cohort studies, 13 case– BMI, per 1 kg/m2
control studies) Cohort studies 1.06 (1.05–1.07) P heterogeneity = 0.081
Women Case–control studies 1.07 (1.06–1.08) P heterogeneity = 0.0643
Ildaphonse et al. (2009) 27 studies (13 cohort studies, 14 case– BMI, per 1 kg/m2
control studies) Cohort studies 1.05 (1.04–1.06) P heterogeneity = 0.78
Men Case–control studies 1.08 (1.06–1.09) P heterogeneity = 0.4238
Wang & Xu (2014) 21 cohort studies BMI, vs normal weight BMI in adults was classified
Men and women All: as follows: normal weight, 18.50–
Pre-obesity 1.28 (1.24–1.33) 24.99; pre-obesity, 25.00–29.99;
Obesity 1.77 (1.68–1.87) obesity, ≥ 30.00
Men:
Pre-obesity 1.22 (1.17–1.28)
Obesity 1.63 (1.50–1.77)
Women:
Pre-obesity 1.38 (1.29–1.47)
Obesity 1.95 (1.81–2.10)
BMI, body mass index (in kg/m 2); CI, confidence interval
Table 2.2.16d Mendelian randomization studies of measures of body fatness and cancer of the kidney
Reference Characteristics of study population Sample size Exposure Outcome Odds ratio (95% CI); P value (with each
(unit) unit increase in exposure) of the association
between the exposure and outcome
Brennan et al. Men and women from 15 centres in 6 7067 (4015 cases BMI (kg/m2) Kidney All subjects:
(2009) countries in central and eastern Europe and 3052 controls) cancer 1.11 (0.91–1.37); P = 0.31
(Czech Republic, Hungary, Poland, Subjects aged < 50 yr:
Romania, Russian Federation, and 1.90 (1.16–2.27); P = 0.0002
Slovakia)
BMI, body mass index (in kg/m 2); CI, confidence interval; OR, odds ratio; yr, year or years

395
References Flaherty KT, Fuchs CS, Colditz GA, Stampfer MJ, Speizer
FE, Willett WC, et al. (2005). A prospective study of
body mass index, hypertension, and smoking and the
Adams KF, Leitzmann MF, Albanes D, Kipnis V, Moore risk of renal cell carcinoma (United States). Cancer
SC, Schatzkin A, et al. (2008). Body size and renal Causes Control, 16(9):1099–106. doi:10.1007/s10552-
cell cancer incidence in a large US cohort study. Am J 005-0349-8 PMID:16184476
Epidemiol, 168(3):268–77. PMID:18544571 Frayling TM, Timpson NJ, Weedon MN, Zeggini E,
Beebe-Dimmer JL, Colt JS, Ruterbusch JJ, Keele GR, Freathy RM, Lindgren CM, et al. (2007). A common
Purdue MP, Wacholder S, et al. (2012). Body mass variant in the FTO gene is associated with body mass
index and renal cell cancer: the influence of race index and predisposes to childhood and adult obesity.
and sex. Epidemiology, 23(6):821–8. doi:10.1097/ Science, 316(5826):889–94. doi:10.1126/science.1141634
EDE.0b013e31826b7fe9 PMID:23007040 PMID:17434869
Bergström A, Hsieh CC, Lindblad P, Lu CM, Cook NR, Häggström C, Rapp K, Stocks T, Manjer J, Bjørge T, Ulmer
Wolk A (2001). Obesity and renal cell cancer – a quan- H, et al. (2013). Metabolic factors associated with
titative review. Br J Cancer, 85(7):984–90. doi:10.1054/ risk of renal cell carcinoma. PLoS One, 8(2):e57475.
bjoc.2001.2040 PMID:11592770 doi:10.1371/journal.pone.0057475 PMID:23468995
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, Hu J, Mao Y, White K; Canadian Cancer Registries
Leon DA, Smeeth L (2014). Body-mass index and risk Epidemiology Research Group (2003). Overweight and
of 22 specific cancers: a population-based cohort study obesity in adults and risk of renal cell carcinoma in
of 5.24 million UK adults. Lancet, 384(9945):755–65. Canada. Soz Praventivmed, 48(3):178–85. doi:10.1007/
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 s00038-003-2046-2 PMID:12891869
Bjørge T, Tretli S, Engeland A (2004). Relation of IARC (2002). Weight control and physical activity. Lyon,
height and body mass index to renal cell carcinoma France: IARC Press (IARC Handbooks of Cancer
in two million Norwegian men and women. Am J Prevention, Vol. 6). Available from: http://publications.
Epidemiol, 160(12):1168–76. doi:10.1093/aje/kwh345 iarc.fr/376.
PMID:15583369 Ildaphonse G, George PS, Mathew A (2009). Obesity and
Brennan P, McKay J, Moore L, Zaridze D, Mukeria A, kidney cancer risk in men: a meta-analysis (1992–2008).
Szeszenia-Dabrowska N, et al. (2009). Obesity and Asian Pac J Cancer Prev, 10(2):279–86. PMID:19537897
cancer: Mendelian randomization approach utilizing Jee SH, Yun JE, Park EJ, Cho ER, Park IS, Sull JW, et al.
the FTO genotype. Int J Epidemiol, 38(4):971–5. (2008). Body mass index and cancer risk in Korean men
doi:10.1093/ije/dyp162 PMID:19542184 and women. Int J Cancer, 123(8):1892–6. doi:10.1002/
Brennan P, van der Hel O, Moore LE, Zaridze D, Matveev ijc.23719 PMID:18651571
V, Holcatova I, et al. (2008). Tobacco smoking, body Kabat GC, Xue X, Kamensky V, Lane D, Bea JW, Chen C,
mass index, hypertension, and kidney cancer risk in et al. (2015). Risk of breast, endometrial, colorectal,
central and eastern Europe. Br J Cancer, 99(11):1912–5. and renal cancers in postmenopausal women in asso-
doi:10.1038/sj.bjc.6604761 PMID:19034282 ciation with a body shape index and other anthropo-
Brock KE, Gridley G, Lynch CF, Ershow AG, Cantor KP metric measures. Cancer Causes Control, 26(2):219–29.
(2007). Obesity and hypertension interact to increase doi:10.1007/s10552-014-0501-4 PMID:25430815
risk of renal cell carcinoma in Iowa, USA. Obes Res Macleod LC, Hotaling JM, Wright JL, Davenport MT,
Clin Pract, 1(2):I–II. doi:10.1016/j.orcp.2007.02.004 Gore JL, Harper J, et al. (2013). Risk factors for renal
PMID:24351456 cell carcinoma in the VITAL study. J Urol, 190(5):1657–
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ 61. doi:10.1016/j.juro.2013.04.130 PMID:23665301
(2003). Overweight, obesity, and mortality from cancer Mathew A, George PS, Ildaphonse G (2009). Obesity
in a prospectively studied cohort of U.S. adults. N Engl and kidney cancer risk in women: a meta-analysis
J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423 (1992–2008). Asian Pac J Cancer Prev, 10(3):471–8.
PMID:12711737 PMID:19640194
Chiu BC, Gapstur SM, Chow WH, Kirby KA, Lynch CF, Nicodemus KK, Sweeney C, Folsom AR (2004). Evaluation
Cantor KP (2006). Body mass index, physical activity, of dietary, medical and lifestyle risk factors for incident
and risk of renal cell carcinoma. Int J Obes, 30(6):940–7. kidney cancer in postmenopausal women. Int J Cancer,
doi:10.1038/sj.ijo.0803231 PMID:16446746 108(1):115–21. doi:10.1002/ijc.11532 PMID:14618625
Dal Maso L, Zucchetto A, Tavani A, Montella M, Peeters A, Beckers S, Verrijken A, Roevens P, Peeters P,
Ramazzotti V, Talamini R, et al. (2007). Renal cell Van Gaal L, et al. (2008). Variants in the FTO gene are
cancer and body size at different ages: an Italian multi- associated with common obesity in the Belgian popul-
center case-control study. Am J Epidemiol, 166(5):582– ation. Mol Genet Metab, 93(4):481–4. doi:10.1016/j.
91. doi:10.1093/aje/kwm108 PMID:17591592 ymgme.2007.10.011 PMID:18055244
396
Pischon T, Lahmann PH, Boeing H, Tjønneland A, van Dijk BA, Schouten LJ, Kiemeney LA, Goldbohm RA,
Halkjaer J, Overvad K, et al. (2006). Body size and risk van den Brandt PA (2004). Relation of height, body
of renal cell carcinoma in the European Prospective mass, energy intake, and physical activity to risk of
Investigation into Cancer and Nutrition (EPIC). renal cell carcinoma: results from the Netherlands
Int J Cancer, 118(3):728–38. doi:10.1002/ijc.21398 Cohort Study. Am J Epidemiol, 160(12):1159–67.
PMID:16094628 doi:10.1093/aje/kwh344 PMID:15583368
Purdue MP, Moore LE, Merino MJ, Boffetta P, Colt JS, Wang F, Xu Y (2014). Body mass index and risk of
Schwartz KL, et al. (2013). An investigation of risk renal cell cancer: a dose-response meta-analysis of
factors for renal cell carcinoma by histologic subtype in published cohort studies. Int J Cancer, 135(7):1673–86.
two case-control studies. Int J Cancer, 132(11):2640–7. doi:10.1002/ijc.28813 PMID:24615287
doi:10.1002/ijc.27934 PMID:23150424 Wang G, Hou J, Ma L, Xie J, Yin J, Xu D, et al. (2012). Risk
Rapp K, Schroeder J, Klenk J, Stoehr S, Ulmer H, Concin factor for clear cell renal cell carcinoma in Chinese
H, et al. (2005). Obesity and incidence of cancer: a population: a case-control study. Cancer Epidemiol,
large cohort study of over 145,000 adults in Austria. 36(2):177–82. doi:10.1016/j.canep.2011.09.006
Br J Cancer, 93(9):1062–7. doi:10.1038/sj.bjc.6602819 PMID:22000673
PMID:16234822 WCRF/AICR (2007). Food, nutrition, physical activity,
Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D; and the prevention of cancer: a global perspective.
Million Women Study Collaboration (2007). Cancer Washington (DC), USA: American Institute for Cancer
incidence and mortality in relation to body mass Research. Available from: http://www.aicr.org/assets/
index in the Million Women Study: cohort study. docs/pdf/reports/Second_Expert_Report.pdf.
BMJ, 335(7630):1134. doi:10.1136/bmj.39367.495995. WCRF/AICR (2015). Continuous Update Project Report.
AE PMID:17986716 Diet, nutrition, physical activity and kidney cancer.
Samanic C, Chow WH, Gridley G, Jarvholm B, Washington (DC), USA: American Institute for
Fraumeni JF Jr (2006). Relation of body mass index Cancer Research. Available from: http://wcrf.org/
to cancer risk in 362,552 Swedish men. Cancer Causes kidney-cancer-2015.
Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9 Wilson RT, Wang J, Chinchilli V, Richie JP, Virtamo J,
PMID:16841257 Moore LE, et al. (2009). Fish, vitamin D, and flavo-
Sanfilippo KM, McTigue KM, Fidler CJ, Neaton JD, Chang noids in relation to renal cell cancer among smokers.
Y, Fried LF, et al. (2014). Hypertension and obesity and Am J Epidemiol, 170(6):717–29. doi:10.1093/aje/kwp178
the risk of kidney cancer in 2 large cohorts of US men PMID:19651663
and women. Hypertension, 63(5):934–41. doi:10.1161/
HYPERTENSIONAHA.113.02953 PMID:24637660
Sawada N, Inoue M, Sasazuki S, Iwasaki M, Yamaji T,
Shimazu T, et al.; JPHC Study Group (2010). Body
mass index and subsequent risk of kidney cancer: a
prospective cohort study in Japan. Ann Epidemiol,
20(6):466–72. doi:10.1016/j.annepidem.2010.03.008
PMID:20470974
Scuteri A, Sanna S, Chen W-M, Uda M, Albai G, Strait J,
et al. (2007). Genome-wide association scan shows
genetic variants in the FTO gene are associated with
obesity-related traits. PLoS Genet, 3(7):e115. doi:10.1371/
journal.pgen.0030115 PMID:17658951
Setiawan VW, Stram DO, Nomura AM, Kolonel LN,
Henderson BE (2007). Risk factors for renal cell cancer:
the Multiethnic Cohort. Am J Epidemiol, 166(8):932–
40. doi:10.1093/aje/kwm170 PMID:17656615
Shapiro JA, Williams MA, Weiss NS (1999). Body mass
index and risk of renal cell carcinoma. Epidemiology,
10(2):188–91. doi:10.1097/00001648-199903000-00019
PMID:10069258
Song Y-M, Sung J, Ha M (2008). Obesity and risk of
Oncol, 26(20):3395–402. doi:10.1200/JCO.2007.15.7867
PMID:18612154
397
2.2.17 Cancer of the urinary bladder Two studies did show a positive association
between BMI and risk of urinary bladder cancer.
Cancer of the urinary bladder accounts for The NIH-AARP cohort (Koebnick et al., 2008)
approximately 3% of all cancers and is the ninth reported significantly increased associations
most common cancer worldwide. The incidence with overweight (RR, 1.16; 95% CI, 1.03–1.29),
of urinary bladder cancer in men is approxi- obesity I (RR, 1.23; 95% CI, 1.06–1.43), and
mately 4 times that in women. The average age of obesity II (RR, 1.30; 95% CI, 1.04–1.63) in men
diagnosis is after age 70 years. Globally, incidence and women combined, compared with normal
rates are highest in Europe and North America weight; stratified analysis indicated that these
and lowest in Asia and Latin America. positive associations were limited to men. The
The strongest risk factor is smoking, as EPIC study (Roswall et al., 2014) found a small
was established several decades ago (IARC, but significant association for BMI in men only
1986). Compared with never-smokers, smokers (RR per 2 kg/m2, 1.05; 95% CI, 1.02–1.08), with
have a 6-fold increase in the risk of developing a strong dose–response relationship. Findings
urinary bladder cancer (WCRF/AICR, 2015). from the Iowa Women’s Health Study (Tripathi
Other risk factors include occupational expo- et al., 2002) demonstrated a statistically marginal
sure to aromatic amines and polyaromatic inverse association between BMI and urinary
hydrocarbons. bladder cancer incidence also in men only
About 90% of urinary bladder cancers are (Ptrend = 0.06 after adjustments).
transitional cell carcinoma; the remainder are Almost all studies adjusted for smoking.
squamous cell carcinoma, adenocarcinoma, and Stratified analyses suggested that the associations
small cell carcinoma. were stronger in former smokers than in never-
(a) Cohort studies smokers. Four studies (Calle et al., 2003; Reeves
et al., 2007; Koebnick et al., 2008; Bhaskaran
See Table 2.2.17a (web only; available at: et al., 2014) specifically stratified by never versus
http://publications.iarc.fr/570). ever smoking status and statistically tested for
A total of 23 prospective cohorts were iden- interactions. None of those interactions were
tified that evaluated associations between BMI significant.
and either urinary bladder cancer incidence (19 Several studies reported on the associations
studies) (Tulinius et al., 1997; Nagano et al., 2000; between BMI and urinary bladder cancer in
Tripathi et al., 2002; Samanic et al., 2004, 2006; Asian populations (Nagano et al., 2000; Oh et al.,
Oh et al., 2005; Rapp et al., 2005; Cantwell et al., 2005; Fujino et al., 2007; Jee et al., 2008; Parr
2006; Holick et al., 2007; Reeves et al., 2007; Jee et al., 2010). No pattern of difference compared
et al., 2008; Koebnick et al., 2008; Larsson et al., with European or North American populations
2008; Prentice et al., 2009; Andreotti et al., 2010; was noted.
Häggström et al., 2011; Bhaskaran et al., 2014; From a large meta-analysis for the associa-
Roswall et al., 2014; Song et al., 2014) or urinary tion between BMI and urinary bladder cancer
bladder cancer-related mortality (5 studies) risk, based on 22 prospective cohort studies,
(Calle et al., 2003; Batty et al., 2005; Fujino et the summary risk estimate was 1.03 (95% CI,
al., 2007; Reeves et al., 2007; Parr et al., 2010) 0.97–1.09) (WCRF/AICR, 2015). Two additional
as the end-point. The large majority of these meta-analyses, of 11 cohort studies (Qin et al.,
studies reported no significant association with 2013) and 15 cohort studies (Sun et al., 2015),
urinary bladder cancer incidence or mortality. reported summary risk estimates of positive
398
associations between BMI and urinary bladder Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I,
cancer. [These differences in part reflect varia- Leon DA, Smeeth L (2014). Body-mass index and risk
tions in study inclusion. In the meta-analysis by of 5.24 million UK adults. Lancet, 384(9945):755–65.
Sun et al., (2015), the summary estimate may doi:10.1016/S0140-6736(14)60892-8 PMID:25129328
have been disproportionally influenced by an Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ
(2003). Overweight, obesity, and mortality from cancer
incorrect data extraction of risk estimates from in a prospectively studied cohort of U.S. adults. N Engl
the FINRISK study (Song et al., 2014).] J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423
Three studies evaluated the relationship PMID:12711737
between waist circumference and urinary Cantwell MM, Lacey JV Jr, Schairer C, Schatzkin A,
Michaud DS (2006). Reproductive factors, exogenous
bladder cancer risk. Two studies (Tripathi et al., hormone use and bladder cancer risk in a prospective
2002; Larsson et al., 2008) found no significant study. Int J Cancer, 119(10):2398–401. doi:10.1002/
association; the third study, based on the EPIC ijc.22175 PMID:16894568
cohort (Roswall et al., 2014), found a small but Fujino Y; Japan Collaborative Cohort Study for Eval
uation of Cancer (2007). Anthropometry, develop-
significant association with waist circumfer- ment history and mortality in the Japan Collaborative
ence in men only (RR per 5 cm, 1.04; 95% CI, Cohort Study for Evaluation of Cancer (JACC). Asian
1.01–1.08). Pac J Cancer Prev, 8(Suppl):105–12. PMID:18260709
Häggström C, Stocks T, Rapp K, Bjørge T, Lindkvist B,
(b) Case–control studies Concin H, et al. (2011). Metabolic syndrome and risk
of bladder cancer: prospective cohort study in the
See Table 2.2.17b (web only; available at: Metabolic Syndrome and Cancer Project (Me-Can).
http://publications.iarc.fr/570). Int J Cancer, 128(8):1890–8. doi:10.1002/ijc.25521
PMID:20568111
The four case–control studies that evalu- Holick CN, Giovannucci EL, Stampfer MJ, Michaud DS
ated the relationship between BMI and urinary (2007). Prospective study of body mass index, height,
bladder cancer incidence (Pelucchi et al., 2002; physical activity and incidence of bladder cancer
in US men and women. Int J Cancer, 120(1):140–6.
Lin et al., 2010; MacKenzie et al., 2011; Attner doi:10.1002/ijc.22142 PMID:17036323
et al., 2012) found no significant associations. IARC (1986). Tobacco smoking. IARC Monogr Eval
Carcinog Risk Chem Hum, 38:1–421. Available from:
http://publications.iarc.fr/56.
Jee SH, Yun JE, Park EJ, Cho ER, Park IS, Sull JW, et al.
References (2008). Body mass index and cancer risk in Korean men
and women. Int J Cancer, 123(8):1892–6. doi:10.1002/
Andreotti G, Hou L, Beane Freeman LE, Mahajan R, ijc.23719 PMID:18651571
Koutros S, Coble J, et al. (2010). Body mass index, Koebnick C, Michaud D, Moore SC, Park Y, Hollenbeck
agricultural pesticide use, and cancer incidence in A, Ballard-Barbash R, et al. (2008). Body mass index,
the Agricultural Health Study cohort. Cancer Causes physical activity, and bladder cancer in a large
Control, 21(11):1759–75. doi:10.1007/s10552-010-9603-9 prospective study. Cancer Epidemiol Biomarkers Prev,
PMID:20730623 17(5):1214–21. doi:10.1158/1055-9965.EPI-08-0026
Attner B, Landin-Olsson M, Lithman T, Noreen D, PMID:18483344
Olsson H (2012). Cancer among patients with Larsson SC, Andersson SO, Johansson JE, Wolk A (2008).
diabetes, obesity and abnormal blood lipids: a popu- Diabetes mellitus, body size and bladder cancer
lation-based register study in Sweden. Cancer Causes risk in a prospective study of Swedish men. Eur J
Control, 23(5):769–77. doi:10.1007/s10552-012-9946-5 Cancer, 44(17):2655–60. doi:10.1016/j.ejca.2008.07.012
PMID:22467266 PMID:18707871
Batty GD, Shipley MJ, Jarrett RJ, Breeze E, Marmot MG, Lin J, Wang J, Greisinger AJ, Grossman HB, Forman
Smith GD (2005). Obesity and overweight in rela- MR, Dinney CP, et al. (2010). Energy balance, the
tion to organ-specific cancer mortality in London PI3K-AKT-mTOR pathway genes, and the risk of
(UK): findings from the original Whitehall study. bladder cancer. Cancer Prev Res (Phila), 3(4):505–17.
Int J Obes, 29(10):1267–74. doi:10.1038/sj.ijo.0803020 doi:10.1158/1940-6207.CAPR-09-0263 PMID:20354165
PMID:15997248
399
MacKenzie T, Zens MS, Ferrara A, Schned A, Karagas MR Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9

(2011). Diabetes and risk of bladder cancer: evidence PMID:16841257
from a case-control study in New England. Cancer, Samanic C, Gridley G, Chow WH, Lubin J, Hoover
117(7):1552–6. doi:10.1002/cncr.25641 PMID:21425156 RN, Fraumeni JF Jr (2004). Obesity and cancer
Nagano J, Kono S, Preston DL, Moriwaki H, Sharp risk among white and black United States veterans.
GB, Koyama K, et al. (2000). Bladder-cancer inci- Cancer Causes Control, 15(1):35–43. doi:10.1023/B:-
dence in relation to vegetable and fruit consump- CACO.0000016573.79453.ba PMID:14970733
tion: a prospective study of atomic-bomb survivors. Song X, Pukkala E, Dyba T, Tuomilehto J, Moltchanov
Int J Cancer, 86(1):132–8. doi:10.1002/(SICI)1097- V, Männistö S, et al. (2014). Body mass index
0215(20000401)86:1<132::AID-IJC21>3.0.CO;2-M and cancer incidence: the FINRISK study. Eur J
PMID:10728607 Epidemiol, 29(7):477–87. doi:10.1007/s10654-014-
Oh SW, Yoon YS, Shin SA (2005). Effects of excess 9934-z PMID:24997743
weight on cancer incidences depending on cancer Sun JW, Zhao LG, Yang Y, Ma X, Wang YY, Xiang YB
sites and histologic findings among men: Korea (2015). Obesity and risk of bladder cancer: a dose-re-
National Health Insurance Corporation Study. J Clin sponse meta-analysis of 15 cohort studies. PLoS One,
Oncol, 23(21):4742–54. doi:10.1200/JCO.2005.11.726 10(3):e0119313. doi:10.1371/journal.pone.0119313
PMID:16034050 PMID:25803438
Parr CL, Batty GD, Lam TH, Barzi F, Fang X, Ho SC, Tripathi A, Folsom AR, Anderson KE; Iowa Women’s
et al.; Asia-Pacific Cohort Studies Collaboration (2010). Health Study (2002). Risk factors for urinary bladder
Body-mass index and cancer mortality in the Asia- carcinoma in postmenopausal women. The Iowa
Pacific Cohort Studies Collaboration: pooled analyses Women’s Health Study. Cancer, 95(11):2316–23.
of 424,519 participants. Lancet Oncol, 11(8):741–52. doi:10.1002/cncr.10975 PMID:12436437
doi:10.1016/S1470-2045(10)70141-8 PMID:20594911 Tulinius H, Sigfússon N, Sigvaldason H, Bjarnadóttir
Pelucchi C, La Vecchia C, Negri E, Dal Maso L, Franceschi K, Tryggvadóttir L (1997). Risk factors for malig-
S (2002). Smoking and other risk factors for bladder nant diseases: a cohort study on a population of
cancer in women. Prev Med, 35(2):114–20. doi:10.1006/ 22,946 Icelanders. Cancer Epidemiol Biomarkers Prev,
pmed.2002.1061 PMID:12200095 6(11):863–73. PMID:9367058
Prentice RL, Shaw PA, Bingham SA, Beresford SA, Caan WCRF/AICR (2015). Continuous Update Project Report.
B, Neuhouser ML, et al. (2009). Biomarker-calibrated Diet, nutrition, physical activity and bladder cancer.
energy and protein consumption and increased Washington (DC), USA: American Institute for Cancer
cancer risk among postmenopausal women. Am J Research. Available from: https://www.wcrf.org/sites/
Epidemiol, 169(8):977–89. doi:10.1093/aje/kwp008 default/files/Bladder-Cancer-2015-Report.pdf.
PMID:19258487
Qin Q, Xu X, Wang X, Zheng XY (2013). Obesity and risk
of bladder cancer: a meta-analysis of cohort studies.
Asian Pac J Cancer Prev, 14(5):3117–21. doi:10.7314/
APJCP.2013.14.5.3117 PMID:23803089
PMID:16234822
335(7630):1134. doi:10.1136/bmj.39367.495995.AE
PMID:17986716
Roswall N, Freisling H, Bueno-de-Mesquita HB, Ros M,
Christensen J, Overvad K, et al. (2014). Anthropometric
measures and bladder cancer risk: a prospective study
in the EPIC cohort. Int J Cancer, 135(12):2918–29.
doi:10.1002/ijc.28936 PMID:24771290
400
2.2.18 Primary tumours of the brain and brain and central nervous system in terms of
central nervous system incidence or mortality without specifying the
histological type (Table 2.2.18a; Calle et al., 2003;
Primary tumours of the brain are a relatively Oh et al., 2005; Samanic et al., 2006; Reeves et al.,
uncommon group of heterogeneous neoplastic 2007; Bhaskaran et al., 2014). There is consistently
diseases with variable natural histories from no evidence of associations between BMI and the
benign to malignant. There are about 130 histo- development of all brain tumours. [This observa-
logical types arising from the many cell types tion was robust when restricting the analyses to
that support and line the brain tissue and the non-smokers only (Reeves et al., 2007; Bhaskaran
central nervous system. Primary brain tumours et al., 2014).]
occur across the age spectrum, from childhood
through adulthood. The most common type, (b) Cohort studies of glioma
which arises from the glial cells, is called glioma Five cohort studies (all in European and
and accounts for approximately 30% of all brain North American populations) (Benson et al.,
tumours in adults (Wiedmann et al., 2013). In 2008; Moore et al., 2009; Michaud et al., 2011;
turn, gliomas are of at least three types – astro- Edlinger et al., 2012; Wiedmann et al., 2013)
cytoma, oligodendroglioma, and ependymoma reported on associations between baseline BMI
– and are graded into four grades (1 and 2 are and the development of glioma (Table 2.2.18a).
low-grade; 3 and 4, also known as glioblastoma There is consistently no evidence of associations
multiforme, are high-grade) (Ricard et al., 2012). between BMI and the development of glioma.
The next most common group is menin- One study stratified by low- and high-grade
gioma, which accounts for approximately 20% glioma and reported no difference.
of brain tumours. Many of these are benign and The NIH-AARP cohort study (Moore
slow-growing, but – as occurs with other brain et al., 2009) reported on the associations between
tumour types – benign tumours can undergo recalled BMI at age 18 years and the development
malignant transformation. of glioma later in life and noted a positive asso-
Established risk factors for brain tumours ciation (P = 0.003) [the numbers of cases in the
include hereditary conditions, such as neuro- upper BMI categories were small; n = 11 for BMI
fibromatosis, and ionizing radiation. of 30–34.9 kg/m2, and no cases in the highest
In 2001, the Working Group of the IARC category of BMI ≥ 35 kg/m2].
Handbook on weight control and physical activity
(IARC, 2002) concluded that the evidence of an (c) Cohort studies of meningioma
association between avoidance of weight gain
Five cohort studies (all in European and North
and brain cancers, including meningioma, was
American populations) (Jhawar et al., 2003;
inadequate.
Benson et al., 2008; Johnson et al., 2011; Michaud
(a) Cohort studies of tumours of the brain and et al., 2011; Wiedmann et al., 2013) reported
central nervous system combined on associations between baseline BMI and the
development of meningioma (Table 2.2.18a). All
Essentially all of the evidence of associations reported statistically significant or borderline
between measures of body fatness and primary significant positive associations, with increased
brain tumours applies to tumours in adulthood. risks ranging from 1.4 to 2.13.
Five large prospective cohort studies reported
associations between BMI and cancers of the
401
Two cohort studies (Johnson et al., 2011;

Michaud et al., 2011) reported on associations
between baseline waist circumference and
meningioma incidence. In both, significant posi-
tive associations were noted.
(d) Case–control studies

See Table 2.2.18b.
Two case–control studies (Cabaniols et al.,
2011; Little et al., 2013) examined relationships
between BMI and risk of glioma, and no asso-
ciations were found. Two further case–control
studies, one in women only (Claus et al., 2013)
and the other in men only (Schildkraut et al.,
2014), examined relationships between BMI and
meningioma and found positive associations in
both studies, similar to those found in the cohort
studies.
A meta-analysis (Niedermaier et al., 2015),
including 2982 meningioma cases from 12 cohort
and case–control studies reported positive asso-
ciations with meningioma: with normal weight
as the reference group, the relative risk was 1.21
(95% CI, 1.01–1.43) for overweight and 1.54 (95%
CI, 1.32–1.79) for obesity.
402
Table 2.2.18a Cohort studies of measures of body fatness and cancers of the brain and central nervous system
Reference Total number of Organ site Exposure categories Exposed Relative risk Covariates Comments
Cohort subjects or subtype cases (95% CI)
mortality
Brain and central nervous system combined
Calle et al. (2003) 404 576 Brain BMI Age, education level,
Cancer Prevention Men 18.5–24.9 370 1.00 smoking, physical
Study II population- Mortality 25–29.9 461 0.98 (0.85–1.13) activity, alcohol
based cohort 30–34.9 68 0.79 (0.61–1.03) consumption,
USA 35–39.9 – – marital status, race,
1982–1998 [Ptrend] [0.14] aspirin use, fat
495 477 Brain BMI intake, vegetable
Women 18.5–24.9 467 1.00 intake; in women,
Mortality 25–29.9 213 1.02 (0.87–1.21) also adjusted for
30–34.9 64 1.10 (0.84–1.44) HRT use
35–39.9 12 0.74 (0.42–1.32)
≥ 40 – –
[Ptrend] [0.96]
Oh et al. (2005) 781 283 Brain BMI Age, smoking
Korean civil servants Men < 18.5 4 1.07 (0.39–2.93) status, alcohol
and teachers from the Incidence 18.5–22.9 105 1.00 consumption,
Korea National Health 23.0–24.9 69 1.09 (0.79–1.50) frequency of regular
Insurance Corporation 25.0–26.9 32 0.84 (0.55–1.28) exercise, family
Republic of Korea 27.0–29.9 21 1.47 (0.90–2.38) history of cancer,
1992–2001 ≥ 30 3 1.79 (0.57–2.66) area of residence
[Ptrend] [0.241]

Samanic et al. (2006) 362 552 Brain BMI Age, year, smoking
Swedish Construction Men ICD-7: 193.0 18.5–24.9 519 1.00
Sweden ≥ 30 46 0.86 (0.63–1.16)
1971–1999 [Ptrend] [> 0.5]
403
404

mortality
Reeves et al. (2007) 1.2 million Brain BMI Age, region, SES, Similar results
Million Women Study Women ICD-10: C71 < 22.5 113 1.14 (0.95–1.38) reproductive when restricting to
United Kingdom Incidence 22.5–24.9 133 1.00 (0.84–1.19) history, never-smokers or
1995–2005 25.0–27.4 143 1.27 (1.08–1.50) smoking, alcohol excluding the first
27.5–29.9 83 1.19 (0.96–1.47) consumption, 2 yr of follow-up
≥ 30 99 1.08 (0.88–1.32) physical activity
per 10 kg/m2 1.01 (0.81–1.26) Where appropriate:
1.2 million BMI time since
Women < 22.5 123 1.17 (0.98–1.40) menopause, HRT
Mortality 22.5–24.9 143 1.00 (0.85–1.18) use
25.0–27.4 158 1.29 (1.10–1.51)
27.5–29.9 90 1.18 (0.96–1.45)
≥ 30 131 1.31 (1.10–1.56)
per 10 kg/m2 1.17 (0.95–1.43)
Bhaskaran et al. (2014) 5.24 million Brain and BMI 2974 Age, diabetes status, Very similar risk
United Kingdom Men and women central per 5 kg/m2 1.04 (0.99–1.10) smoking, alcohol estimates for never-
Clinical Practice Incidence nervous consumption, smokers (n = 1359
Research Database system calendar year, SES incident cases)
United Kingdom
1987–2012
Glioma
Benson et al. (2008) 1 184 225 Glioma BMI Height, SES,
Million Women Study Women ICD-O: < 25 259 1.00 smoking, alcohol
United Kingdom Incidence 9380–9481 25–29.9 241 1.20 (1.01–1.44) intake, parity, age
1996–2001 ≥ 30 106 1.07 (0.84–1.34) (yr) at first birth,
[Ptrend] [0.10] duration of OC use,
physical activity,
study region
Moore et al. (2009) 270 395 Glioma BMI Age at baseline, age
NIH-AARP cohort Men and women ICD-O-3: < 18.5 4 1.66 (0.59–4.64) squared, sex, race,
USA Incidence 9380–9460 18.5–24.9 82 1.00 highest level of
(8.2 years) 25–29.9 95 0.90 (0.67–1.22) education, marital
30–34.9 46 1.29 (0.89–1.86) status
≥ 35 9 0.74 (0.37–1.48)
[Ptrend] [0.95]
mortality
Moore et al. (2009) BMI at age 18 yr No significant
(cont.) < 18.5 26 0.69 (0.45–1.05) associations
18.5–24.9 175 1.00 observed with BMI
25–29.9 24 1.04 (0.67–1.59) at age 35 yr or at age
30–34.9 11 3.74 (2.03–6.90) 50 yr
≥ 35 – –
[Ptrend] [0.003]
Michaud et al. (2011) 380 775 Glioma BMI Age, country, sex,
EPIC cohort Men and women ICD-O-2: < 20 13 1.08 (0.60–1.92) education level
From 1999 (8.4 years) Incidence 9380–9460, 20–24.9 125 1.00
9505 25–29.9 147 1.04 (0.81–1.34)
≥ 30 55 1.06 (0.76–1.48)
[Ptrend] [0.80]
WC, quartiles Age, country, sex,
Q1 73 1.00 education level,
Q2 82 0.90 (0.65–1.24) height
Q3 73 0.82 (0.59–1.16)
Q4 90 0.97 (0.69–1.35)
[Ptrend] [0.81]
Edlinger et al. (2012) 578 462 Low-grade BMI, quintiles Year of birth (in
Metabolic Syndrome Men and women glioma Q1 21 1.00 decades), cohort,
and Cancer Project Incidence ICD-7: 193 Q2 16 0.69 (0.33–1.42) smoking status

(Me-Can) Q3 21 0.90 (0.46–1.77)
Austria, Norway, Q4 24 1.00 (0.51–1.95)
Sweden Q5 16 0.66 (0.31−1.38)
1972–2005 High-grade BMI, quintiles
glioma Q1 65 1.00
ICD-7: 193 Q2 72 0.98 (0.68–1.43)
Q3 82 1.06 (0.73–1.52)
Q4 99 1.23 (0.87–1.75)
Q5 92 1.14 (0.80–1.64)
405
406

mortality
Wiedmann et al. 74 242 Glioma BMI Age, sex
(2013) Men and women ICD-O-3: < 20 6 0.67 (0.29–1.56)
Nord-Trøndelag Incidence 9380–9480 20–24.9 79 1.00
Health Study (HUNT 25–29.9 49 0.88 (0.61–1.27)
1 Study) ≥ 30 14 1.04 (0.58–1.85)
Norway [Ptrend] [0.87]
From 1991 (23.5 yr)
Meningioma
Jhawar et al. (2003) 121 700 Meningioma BMI Age, menopausal
Nurses’ Health Study Women (self- < 22 22 1.00 status,
USA Incidence reported) 22–24.9 31 1.10 (0.61–1.97) postmenopausal
1.2 million person- ≥ 25 58 1.61 (0.96–2.70) HRT use
years [Ptrend] [0.06]
Benson et al. (2008) 1 184 225 Meningioma BMI Height, SES,
Million Women Study Women < 25 154 1.00 smoking, alcohol
United Kingdom Incidence 25–29.9 120 1.01 (0.79–1.29) intake, parity, age
1996–2001 ≥ 30 84 1.40 (1.08–1.87) (yr) at first birth,
[Ptrend] [0.03] duration of OC use,
physical activity,
study region
Johnson et al. (2011) 27 791 Meningioma BMI Age BMI at age 18 yr
Iowa Women’s Health Women ICD-9: 18.5–24.9 41 1.00 and at age 30 yr not
Study Incidence 192.1, 192.3, 25–29.9 36 0.92 (0.59–1.44) associated with risk
USA 225.2, 225.4, 30–34.0 35 2.14 (1.36–3.36)
291 021 person-years 237.6 ≥ 35 13 1.99 (1.06–3.71)
[Ptrend] [0.0007]
WC (in)
< 30.25 22 1.00
30.26–33.50 20 0.92 (0.50–1.69)
33.51–37.75 35 1.56 (0.92–2.67)
> 37.75 44 2.13 (1.28–3.56)
[Ptrend] [0.0006]
mortality
Michaud et al. (2011) 380 775 Meningioma BMI Age, country, sex,
EPIC cohort Men and women < 20 7 1.00 (0.46–2.19) education level
From 1999 (8.4 yr) Incidence 20–24.9 70 1.00
25–29.9 87 1.34 (0.97–1.86)
≥ 30 39 1.48 (0.98–2.23)
[Ptrend] [0.05]
Meningioma WC, quartiles Age, country, sex,
Q1 32 1.00 education level,
Q2 45 1.18 (0.73–1.88) height
Q3 41 1.06 (0.65–1.72)
Q4 66 1.71 (1.08–2.73)
[Ptrend] [0.01]
Wiedmann et al. 74 242 Meningioma BMI Age, sex When stratifying
(2013) Men and women ICD-O-3: < 20 6 0.82 (0.35–1.92) by sex, positive
Nord-Trøndelag Incidence 9530–9539 20–24.9 59 1.00 associations
Health Study (HUNT 25–29.9 51 1.22 (0.83–1.80) (borderline
1 Study) ≥ 30 22 1.48 (0.89–2.45) significant)
Norway [Ptrend] [0.08] observed in women
23.5 yr only
BMI, body mass index (in kg/m 2); CI, confidence interval; EPIC, European Prospective Investigation into Cancer and Nutrition; HRT, hormone replacement therapy; ICD, International
Classification of Diseases; NIH-AARP, National Institutes of Health–AARP Diet and Health Study; OC, oral contraceptive; SES: socioeconomic status; WC, waist circumference; yr,
year or years

407
408

Table 2.2.18b Case–control studies of measures of body fatness and cancers of the brain and central nervous system
Reference Total number of cases Exposure categories Exposed cases Relative risk Adjustment for confounding
Study location Sex (95% CI)
Glioma
Cabaniols et al. (2011) 122 BMI in recent past Age, sex
France Men and women < 25 1.00
2005 Hospital ≥ 25 49 0.70 (0.41–1.18)
Little et al. (2013) 643 BMI in adulthood, recent past Age, race, education level, state
USA Men < 18.5 8 2.47 (0.63–9.70) of residence
2004–2012 Population 18.5–24.9 133 1.00
25–29.9 311 1.26 (0.94–1.69)
≥ 30 191 1.26 (0.91–1.75)
[Ptrend] [0.67]
460 BMI in adulthood, recent past
Women < 18.5 10 0.80 (0.34–1.87)
Population 18.5–24.9 203 1.00
25–29.9 136 0.95 (0.70–1.29)
≥ 30 111 1.11 (0.98–1.03)
[Ptrend] [0.63]
643 BMI at age 21 yr
Men < 18.5 34 0.67 (0.41–1.09)
Population 18.5–24.9 391 1.00
25–29.9 182 1.16 (0.89–1.52)
≥ 30 29 0.77 (0.45–1.31)
[Ptrend] [0.054]
460 BMI at age 21 yr
Women < 18.5 69 0.68 (0.48–0.96)
Population 18.5–24.9 324 1.00
25–29.9 39 1.39 (0.85–2.27)
≥ 30 23 1.66 (0.85–3.23)
[Ptrend] [0.004]
Study location Sex (95% CI)
Meningioma
Claus et al. (2013) 1127 BMI Race, education level,
USA Women < 23.4 303 1.00 menopausal status, age at
2006–2011 Population 23.4–26.6 237 1.06 (0.83–1.35) menopause, age at menarche,
26.6–30.9 269 1.13 (0.89–1.45) number of full-term
≥ 30.9 308 1.29 (1.01–1.65) pregnancies, age at first live
[Ptrend] [0.04] birth, ever use of OC, ever use
of HRT, ever use of fertility
medications, smoking, alcohol
consumption, breastfeeding,
geographical location
Schildkraut et al. (2014) 456 BMI Age, race
USA Men < 25 84 1.00
2006–2012 Population 25–29.9 206 1.66 (1.17–2.34)
30–34.9 102 1.92 (1.28–2.90)
≥ 35 58 1.64 (1.02–2.64)
BMI, body mass index (in kg/m 2); CI, confidence interval; HRT, hormone replacement therapy; OC, oral contraceptive; yr, year or years

409
References (Phila), 4(9):1385–92. doi:10.1158/1940-6207.CAPR-11-

0014 PMID:21685234
Moore SC, Rajaraman P, Dubrow R, Darefsky AS,
Benson VS, Pirie K, Green J, Casabonne D, Beral V; Million Koebnick C, Hollenbeck A, et al. (2009). Height, body
Women Study Collaborators (2008). Lifestyle factors mass index, and physical activity in relation to glioma
and primary glioma and meningioma tumours in the risk. Cancer Res, 69(21):8349–55. doi:10.1158/0008-
Million Women Study cohort. Br J Cancer, 99(1):185– 5472.CAN-09-1669 PMID:19808953
90. doi:10.1038/sj.bjc.6604445 PMID:18560401 Niedermaier T, Behrens G, Schmid D, Schlecht I, Fischer
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, B, Leitzmann MF (2015). Body mass index, physical
Leon DA, Smeeth L (2014). Body-mass index and risk activity, and risk of adult meningioma and glioma: a
of 22 specific cancers: a population-based cohort study meta-analysis. Neurology, 85(15):1342–50. doi:10.1212/
of 5.24 million UK adults. Lancet, 384(9945):755–65. WNL.0000000000002020 PMID:26377253
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 Oh SW, Yoon YS, Shin SA (2005). Effects of excess
Cabaniols C, Giorgi R, Chinot O, Ferahta N, Spinelli weight on cancer incidences depending on cancer
V, Alla P, et al. (2011). Links between private habits, sites and histologic findings among men: Korea
psychological stress and brain cancer: a case-control National Health Insurance Corporation Study. J Clin
pilot study in France. J Neurooncol, 103(2):307–16. Oncol, 23(21):4742–54. doi:10.1200/JCO.2005.11.726
doi:10.1007/s11060-010-0388-1 PMID:20835749 PMID:16034050
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D;
(2003). Overweight, obesity, and mortality from cancer Million Women Study Collaboration (2007). Cancer
in a prospectively studied cohort of U.S. adults. N Engl incidence and mortality in relation to body mass
J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423 index in the Million Women Study: cohort study. BMJ,
PMID:12711737 335(7630):1134. doi:10.1136/bmj.39367.495995.AE
Claus EB, Calvocoressi L, Bondy ML, Wrensch M, Wiemels PMID:17986716
JL, Schildkraut JM (2013). Exogenous hormone Ricard D, Idbaih A, Ducray F, Lahutte M, Hoang-Xuan
use, reproductive factors, and risk of intracranial K, Delattre JY (2012). Primary brain tumours in
meningioma in females. J Neurosurg, 118(3):649–56. adults. Lancet, 379(9830):1984–96. doi:10.1016/S0140-
doi:10.3171/2012.9.JNS12811 PMID:23101448 6736(11)61346-9 PMID:22510398
Edlinger M, Strohmaier S, Jonsson H, Bjørge T, Manjer Samanic C, Chow WH, Gridley G, Jarvholm B, Fraumeni
J, Borena WT, et al. (2012). Blood pressure and JF Jr (2006). Relation of body mass index to cancer
other metabolic syndrome factors and risk of brain risk in 362,552 Swedish men. Cancer Causes Control,
tumour in the large population-based Me-Can 17(7):901–9. PMID:16841257
cohort study. J Hypertens, 30(2):290–6. doi:10.1097/ Schildkraut JM, Calvocoressi L, Wang F, Wrensch M,
HJH.0b013e32834e9176 PMID:22179083 Bondy ML, Wiemels JL, et al. (2014). Endogenous
IARC (2002). Weight control and physical activity. Lyon, and exogenous hormone exposure and the risk of
France: IARC Press (IARC Handbooks of Cancer meningioma in men. J Neurosurg, 120(4):820–6.
Prevention, Vol. 6). Available from: http://publications. doi:10.3171/2013.12.JNS131170 PMID:24484233
iarc.fr/376. Wiedmann M, Brunborg C, Lindemann K, Johannesen
Jhawar BS, Fuchs CS, Colditz GA, Stampfer MJ (2003). TB, Vatten L, Helseth E, et al. (2013). Body mass index
Sex steroid hormone exposures and risk for menin- and the risk of meningioma, glioma and schwannoma
gioma. J Neurosurg, 99(5):848–53. doi:10.3171/ in a large prospective cohort study (the HUNT Study).
jns.2003.99.5.0848 PMID:14609164 Br J Cancer, 109(1):289–94. doi:10.1038/bjc.2013.304
Johnson DR, Olson JE, Vierkant RA, Hammack JE, Wang PMID:23778522
AH, Folsom AR, et al. (2011). Risk factors for menin-
gioma in postmenopausal women: results from the
Iowa Women’s Health Study. Neuro Oncol, 13(9):1011–9.
doi:10.1093/neuonc/nor081 PMID:21750006
Little RB, Madden MH, Thompson RC, Olson JJ, Larocca
RV, Pan E, et al. (2013). Anthropometric factors in
relation to risk of glioma. Cancer Causes Control,
24(5):1025–31. PMID:23456313
Michaud DS, Bové G, Gallo V, Schlehofer B, Tjønneland
A, Olsen A, et al. (2011). Anthropometric measures,
physical activity, and risk of glioma and meningioma
in a large prospective cohort study. Cancer Prev Res
410
2.2.19 Cancer of the thyroid either sex, but a positive trend across BMI quin-
tiles in women only (Ptrend = 0.02). In a Norwegian
Cancer of the thyroid includes a variety population-based cohort, Engeland et al. (2006)
of histological types, ranging from the most observed no association in men, but a positive
common group of differentiated cancers (papil- association in women; the estimated relative risk
lary carcinoma and follicular carcinoma) to for BMI ≥ 30 kg/m2 compared with the refer-
medullary carcinoma and anaplastic (undiffer- ence BMI of 18–24.9 kg/m2 was 1.29 (95% CI,
entiated) carcinoma. Globally, thyroid cancer 1.13–1.46). In the Radiologic Technologists Study
incidence has been increasing during the past in the USA (Meinhold et al. (2010), no associa-
three decades; incidence rates in women are tion was observed in men, whereas the associa-
generally 2–3 times those in men. Known risk tion in women was also positive (RR, 1.74; 95%
factors include exposure to radiation for all CI, 1.03–2.94). A systematic review, including
thyroid cancers, and iodine deficiency for folli- 11 studies, estimated the relative risk of thyroid
cular carcinoma. cancer for obese compared with normal-weight
In 2001, the Working Group of the IARC individuals to be 1.53 (95% CI, 0.89–2.64) in
Handbook on weight control and physical activity men and 1.57 (95% CI, 1.13–2.19) in women
(IARC, 2002) concluded that the evidence of an (Schmid et al., 2015). Another systematic review
association between avoidance of weight gain and (Zhang et al., 2014), including 16 cohort studies,
thyroid cancer was inadequate. The 2007 WCRF estimated an overall relative risk of thyroid
review did not draw any conclusions about body cancer of 1.29 (95% CI, 1.20–1.37) in relation to
fatness and thyroid cancer risk (WCRF/AICR, obesity, with similar risk estimates in men and
2007). in women [the Working Group noted that this
(a) Cohort studies study provided limited information]. A pooled
analysis by Kitahara et al. (2016) of 22 cohorts
The evidence from cohort studies since 2000 including 2296 incident cases found a modest
includes 15 publications (excluding analyses that positive association between baseline BMI and
were later updated and analyses based on fewer thyroid cancer risk overall, and the association
than 100 incident cases), including a large pooled was stronger in men (RR per 5 kg/m2, 1.17; 95%
analysis of 22 cohorts (Kitahara et al., 2016). CI, 1.06–1.28) than in women (RR per 5 kg/m2,
Table 2.2.19a presents results from these studies 1.04; 95% CI, 1.00–1.09).
for BMI at baseline, with comments on findings A total of four studies assessed the association
according to other measures of body fatness, between body fatness and thyroid cancer risk by
such as weight changes over the life-course, waist histological subtype (Engeland et al., 2006; Kabat
circumference, or waist-to-hip ratio. et al., 2012; Rinaldi et al., 2012; Kitahara et al.,
In general, the evidence from cohort studies 2016). The association with BMI was similar for
supports a positive association between BMI and the papillary and follicular histological subtypes.
thyroid cancer, with most studies reporting a In the only study that assessed BMI at younger
significantly increased risk at the highest versus ages (Kitahara et al., 2016), thyroid cancer risk
lowest category of BMI and/or a significant dose– was similar for BMI in young adulthood (RR per
response relationship. However, in those studies 5 kg/m2, 1.13; 95% CI, 1.02–1.25) and BMI later in
that provided estimates for women and men adult life (RR per 5 kg/m2, 1.06; 95% CI, 1.02–1.10);
separately, inconsistent findings were observed a positive association was also reported with BMI
across studies. Almquist et al. (2011) found no gain in adult life (RR per 5 kg/m2, 1.07; 95% CI,
association between BMI and thyroid cancer in 1.00–1.15), after adjustment for BMI.
411
Two studies assessed anthropometric meas-

ures of body fatness other than BMI. In the pooled
analysis by Kitahara et al. (2016), a weaker positive
association was found with waist circumference
(RR per 5 cm, 1.03; 95% CI, 1.01–1.05) than with
BMI (RR per 5 kg/m2, 1.06; 95% CI, 1.02–1.10).
In the EPIC cohort (Rinaldi et al., 2012), associa-
tions with waist circumference and waist-to-hip
ratio were similar to those observed with BMI.

Six informative case–control studies were
identified that evaluated the association between
BMI and thyroid cancer, including two larger
pooled analyses (Table 2.2.19b). One study (Cléro
et al., 2010) that combined two of the five studies
but that did not offer additional information was
excluded. In two studies, the total number of
cases in men was less than 50 (Guignard et al.,
2007; Suzuki et al., 2008); therefore, only data
for women are reported. Two studies (Guignard
et al., 2007; Xu et al., 2014) were restricted to
papillary carcinomas.
Overall, there appeared to be an association
between elevated current BMI (in adulthood)
and the occurrence of thyroid cancer. There
was some indication that this relationship was
stronger in women than in men. [However, this
may reflect small case numbers in the studies,
especially in men, related to the low prevalence
of the disease.] Two of the studies evaluated
the associations between BMI at age 18 years
(Brindel et al., 2009) and at age 20 years (Suzuki
et al., 2008) and thyroid cancer, and noted some
evidence for an association with thyroid cancer
occurrence.
412
Table 2.2.19a Cohort studies of measures of body fatness and cancer of the thyroid
Location Sex
Samanic et al. (2004) 4 500 700 Obesity Age, calendar year Obesity defined as discharge
United States Veterans Men White men: diagnosis of obesity: ICD-8:
cohort Incidence or Non-obese 811 1.00 277; ICD-9: 278.0
USA mortality Obese 64 1.40 (1.09–1.81) Cancers diagnosed within
1969–1996 Black men: 1 yr of obesity diagnoses
Non-obese 156 1.00 were excluded from the
Obese 13 1.92 (1.09–3.40) study
In White men only, higher
risk of adrenal thyroid
cancer
Oh et al. (2005) 781 283 BMI Age, smoking,
Korea National Health Men < 18.5 3 0.82 (0.20–3.34) alcohol
Insurance Corporation 18.5–22.9 72 1.00 consumption,
cohort 23–24.9 70 1.52 (1.07–2.14) exercise, family
Republic of Korea 25–26.9 53 2.00 (1.38–2.89) history of cancer,
1992–2002 27–29.9 28 2.23 (1.40–3.55) area of residence
≥ 30 – –
[Ptrend] [< 0.001]
Engeland et al. (2006) 963 523 BMI Age Association was similar for
Norwegian population- Men < 18.5 2 0.47 (0.12–1.87) age 50–74 yr
based cohort Incidence 18.5–24.9 412 1.00
Norway 25–29.9 322 1.12 (0.97–1.30)
1972–2003 ≥ 30 42 1.14 (0.82–1.56)
[Ptrend] [0.005]

1 037 424 BMI Age Association was similar for
Women < 18.5 30 0.68 (0.47–0.98) age 20–49 yr and stronger
Incidence 18.5–24.9 1187 1.00 for age 50–74 yr (57%
25–29.9 710 1.08 (0.98–1.20) increased risk). Somewhat
≥ 30 341 1.29 (1.13–1.46) stronger associations for
[Ptrend] [0.001] follicular carcinoma vs
papillary carcinoma
Samanic et al. (2006) 362 552 BMI Age, calendar year,
Swedish Construction Men 18.5–24.9 89 1.00 smoking
Sweden ≥ 30 9 0.98 (0.49–1.96)
1971–1999 [Ptrend] [0.48]
413
414

Location Sex
Song et al. (2008) 170 481 BMI Age, height,
Female public servants Women < 18.5 3 0.35 (0.11–1.10) smoking, alcohol
Republic of Korea 18.5–20.9 40 0.78 (0.52–1.16) consumption,
1994–2003 21–22.9 89 1.00 exercise, SES
23–24.9 115 1.05 (0.79–1.41)
25–26.9 93 1.08 (0.80–1.47)
27–29.9 59 1.02 (0.72–1.45)
≥ 30 11 0.70 (0.35–1.40)
per 1 kg/m2 1.02 (0.98–1.04)
Clavel-Chapelon et al. 91 909 BMI Age, year of birth, Large body shape
(2010) Women < 18.5 3 0.35 (0.11–1.12) history of benign (Sörensen’s silhouette) at
E3N cohort (female 18.5–22 99 1.00 thyroid conditions, baseline and at age 35–40 yr,
teachers) 22–25 129 1.39 (1.07–1.81) smoking, iodine but not at age 20–25 yr,
France 25–30 62 1.18 (0.86–1.63) associated with increased
1990–2005 ≥ 30 24 1.76 (1.12–2.76) risk
[Ptrend] [0.005]
Leitzmann et al. (2010) 484 326 BMI Age, sex, physical For WC, positive association
NIH-AARP cohort Men and women 18.5–24.9 107 1.00 activity, race/ in men but not in women.
USA Incidence 25–29.9 153 1.27 (0.99–1.64) ethnicity, For waist-to-hip ratio, null
1995–2003 ≥ 30 92 1.39 (1.05–1.85) education level, association in either sex
[Ptrend] [0.007] smoking, alcohol
consumption, OC
use
Meinhold et al. (2010) 21 207 BMI Year of birth,
Radiologic Men 18.5–24.9 13 1.00 smoking, radiation
Technologists Study Incidence 25–29.9 15 0.89 (0.42–1.90) exposure, history
USA 30–34.5 9 1.91 (0.80–4.56) of benign thyroid
1983–2006 ≥ 35 2 2.14 (0.60–7.67) conditions
[Ptrend] [0.11]
69 506 BMI
Women < 18.5 6 0.96 (0.42–2.18)
Incidence 18.5–24.9 144 1.00
25–29.9 44 0.90 (0.64–1.27)
30–34.5 26 1.41 (0.92–2.16)
≥ 35 16 1.74 (1.03–2.94)
[Ptrend] [0.04]
Location Sex
Almquist et al. (2011) 289 866 BMI, quintiles Age, smoking
7 population-based Men Q1 23 1.00
cohorts Incidence Q2 35 1.41 (0.83–2.39)
Austria, Norway, and Q3 20 0.77 (0.42–1.41)
Sweden Q4 29 1.09 (0.63–1.89)
2006–2016 Q5 26 1.00 (0.57–1.77)
[Ptrend] [0.61]
288 834 BMI, quintiles Age, smoking
Women Q1 41 1.00
Incidence Q2 37 0.84 (0.54–1.31)
Q3 51 1.10 (0.73–1.68)
Q4 59 1.22 (0.81–1.84)
Q5 67 1.40 (0.93–2.09)
[Ptrend] [0.02]
Kabat et al. (2012) 144 319 BMI Age, age at first Waist-to-hip ratio and
Women’s Health Women < 25 92 1.00 pregnancy, weight change, null
Initiative Incidence 25– < 30 99 1.06 (0.79–1.42) education level, association. Similar
USA 30–35 71 1.40 (1.00–1.94) smoking, alcohol associations for papillary
1993–2011 ≥ 35 32 0.97 (0.62–1.50) consumption, carcinoma and follicular
[Ptrend] [0.39] exercise, history carcinoma
of benign thyroid
conditions
Rinaldi et al. (2012) 343 765 BMI Age, centre, Only 58 incident cases

EPIC cohort Women < 18.5 3 0.27 (0.09–0.84) smoking identified in men (results not
Europe Incidence 18.5–24.9 290 1.00 presented). Similar findings
1992–2009 25–29.9 145 1.12 (0.91–1.38) for papillary carcinoma only.
≥ 30 66 1.19 (0.89–1.59) WC and waist-to-hip ratio
[Ptrend] 4 [0.042] both positively associated
with risk
Bhaskaran et al. (2014) 5.24 million BMI 941 Age, sex, year, Similar association in never-
Clinical Practice Men and women per 5 kg/m2 1.09 (1.00–1.19) diabetes, alcohol smokers
Research Datalink Incidence consumption,
United Kingdom smoking, SES
1987–2012
415
416

Location Sex
Kitahara et al. (2016) 578 922 BMI Age, alcohol WC less associated. Similar
Pooled analysis of 22 Men 15–18.4 2 0.66 (0.16–2.67) consumption, associations for papillary
cohorts Incidence 18.5–24.9 191 1.0 physical activity, carcinoma and follicular
Asia, Australia, 25–29.9 327 1.23 (1.02–1.47) race/ethnicity, carcinoma
Europe, and North ≥ 30 129 1.35 (1.07–1.71) marital status,
America per 5 kg/m2 1.17 (1.06–1.28) education level,
1979–2009 smoking
774 373 BMI Similar associations for
Women 15–18.4 29 0.86 (0.59–1.24) WC. Similar associations
Incidence 18.5–24.9 995 1.0 for papillary carcinoma and
25–29.9 615 1.02 (0.93–1.14) follicular carcinoma
≥ 30 356 1.05 (0.92–1.19)
per 5 kg/m2 1.04 (1.00–1.09)
BMI at baseline
per 5 kg/m2 1.06 (1.02–1.10)
BMI in young adulthood
per 5 kg/m2 1.13 (1.02–1.25)
BMI gain in adult life
per 5 kg/m2 1.07 (1.00–1.15)
BMI, body mass index (in kg/m ); CI, confidence interval; EPIC, European Prospective Investigation into Cancer and Nutrition; NIH-AARP, National Institutes of Health–AARP Diet
2
and Health Study; OC, oral contraceptive; SES, socioeconomic status; WC, waist circumference; yr, year or years
Table 2.2.19b Case–control studies of measures of body fatness and cancer of the thyroid
Reference Total number of Exposure categories Exposed Odds ratio Adjustment for Comments
Period Sex
Source of controls
Dal Maso et al. (2000) Men: 417 BMI, tertiles NR Men: Age, history of
Pooled analysis of 12 Women: 2056 T1 1.0 radiation exposure
case–control studies Population and T2 0.8 (0.6–1.1)
China, Greece, Italy, hospital T3 1.0 (0.8–1.4)
Japan, Norway, [Ptrend] [0.71]
Sweden, Switzerland, BMI, tertiles NR Women:
and USA
T1 1.0
T2 1.0 (0.9–1.2)
T3 1.2 (1.0–1.4)
[Ptrend] [0.04]
Guignard et al. (2007) Women: 279 BMI Age, reference year, Papillary and follicular
New Caledonia Population <18.5 7 0.99 (0.35–2.80) ethnicity, smoking, carcinomas only
1993–1999 18.5–24.99 80 1.00 number of full- The risk was greater in
25.0–29.9 87 1.18 (0.75–1.86) term pregnancies, women aged > 50 yr
30.0–34.9 61 1.92 (1.14–3.22) miscarriages, Data for men NR because of
≥ 35.0 41 1.85 (1.02–3.35) and irregular the low number of cases
[Ptrend] [0.01] menstruations
Suzuki et al. (2008) Women: 131 Current BMI, tertiles Age, smoking habits, Papillary and follicular
Japan Hospital 15.4–20.4 31 1.00 drinking habits, carcinomas only
2001–2005 20.4–22.9 51 1.01 (0.59–1.74) regular exercise, family Null associations with BMI
22.9–37.0 49 1.48 (0.86–2.57) history of thyroid or weight change since age
[Ptrend] [0.141] cancer, past history 20 yr
BMI at age 20 yr, tertiles of thyroid diseases, Data for men NR because of

total non-alcohol the low number of cases
14.9–19.2 31 1.00
energy intake, referral
19.3–21.1 45 1.42 (0.85–2.38)
pattern to the hospital,
21.2–33.4 50 1.18 (0.69–2.01)
menopausal status, age
[Ptrend] [0.526]
at menarche, parity,
HRT use
417
418

Period Sex
Source of controls
Brindel et al. (2009) Men: 23 BMI before diagnosis Height, ethnicity,
French Polynesia Women: 177 Men: education level,
1979–2004 Population; < 18.5 0 – smoking, interviewer,
matched by date of 18.5–24.9 7 1.0 radiation to head or
birth and sex 25.0–29.9 11 5.9 (0.8–40.8) neck before age 15 yr
30.0–34.9 2 3.1 (0.2–42.1) In women, also
≥ 35.0 3 3.2 (0.3–39.2) adjusted for number of
Women: full-term pregnancies,
< 18.5 7 0.8 (0.3–2.4) menopausal status
18.5–24.9 74 1.0
25.0–29.9 44 3.5 (1.7–7.4)
30.0–34.9 25 1.2 (0.6–2.6)
≥ 35.0 27 3.0 (1.3–7.1)
BMI at age 18 yr
Men:
< 18.5 1 0.05 (0.0–1.0)
18.5–24.9 16 1.0
25.0–29.9 3 0.8 (0.1–6.3)
≥ 30.0 2 4.8 (0.2–113)
Women:
< 18.5 26 0.6 (0.3–1.2)
18.5–24.9 117 1.0
25.0–29.9 32 3.7 (1.6–8.4)
≥ 30.0 5 1.2 (0.3–5.2)
Xu et al. (2014) Men: 557 BMI Age, sex, race/ Papillary carcinoma
Pooled analysis Women: 1360 < 18.5 35 0.82 (0.52–1.30) ethnicity, study centre only. Body fat percentage
Germany, Italy, USA Hospital 18.5–24.9 581 1.00 (calculated by the formula of
1993–2013 25–29.9 422 1.67 (1.38–2.03) Deurenberg) also associated
≥ 30 319 3.91 (3.02–5.05) with increased risk, overall
[Ptrend] [0.001] and by sex
Period Sex
Source of controls
Xhaard et al. (2015) Men and women: BMI All: Stratified by sex, No differences in risk were
France 761 < 18.5 52 1.00 region, and age and observed when restricting
2005–2010 Population 18.5–24.9 496 1.15 (0.77–1.71) adjusted for education to papillary carcinomas
25–29.9 138 1.23 (0.77–1.96) level, ethnicity, (n = 676 cases)
≥ 30 72 1.56 (0.92–2.66) smoking status, family
[Ptrend] [0.09] history of thyroid
BMI Women: cancer, and number of
< 18.5 45 1.00 pregnancies (in women
18.5–24.9 384 1.25 (0.82–1.90) only)
25–29.9 102 1.50 (0.89–2.51)
≥ 30 65 1.78 (1.01–3.14)
[Ptrend] [0.03]
BMI, body mass index (in kg/m 2); CI, confidence interval; HRT, hormone replacement therapy; NR, not reported; yr, year or years

419
References analysis of 22 prospective studies. Thyroid, 26(2):306–

18. doi:10.1089/thy.2015.0319 PMID:26756356
Leitzmann MF, Brenner A, Moore SC, Koebnick C, Park
Almquist M, Johansen D, Björge T, Ulmer H, Lindkvist Y, Hollenbeck A, et al. (2010). Prospective study of
B, Stocks T, et al. (2011). Metabolic factors and risk of body mass index, physical activity and thyroid cancer.
thyroid cancer in the Metabolic Syndrome and Cancer Int J Cancer, 126(12):2947–56. PMID:19795465
Project (Me-Can). Cancer Causes Control, 22(5):743– Meinhold CL, Ron E, Schonfeld SJ, Alexander BH,
51. doi:10.1007/s10552-011-9747-2 PMID:21380729 Freedman DM, Linet MS, et al. (2010). Nonradiation
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, risk factors for thyroid cancer in the US Radiologic
Leon DA, Smeeth L (2014). Body-mass index and risk Technologists Study. Am J Epidemiol, 171(2):242–52.
of 22 specific cancers: a population-based cohort study doi:10.1093/aje/kwp354 PMID:19951937
of 5.24 million UK adults. Lancet, 384(9945):755–65. Oh SW, Yoon YS, Shin SA (2005). Effects of excess
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 weight on cancer incidences depending on cancer
Brindel P, Doyon F, Rachédi F, Boissin JL, Sebbag J, Shan sites and histologic findings among men: Korea
L, et al. (2009). Anthropometric factors in differenti- National Health Insurance Corporation Study. J Clin
ated thyroid cancer in French Polynesia: a case-con- Oncol, 23(21):4742–54. doi:10.1200/JCO.2005.11.726
trol study. Cancer Causes Control, 20(5):581–90. PMID:16034050
doi:10.1007/s10552-008-9266-y PMID:19043789 Rinaldi S, Lise M, Clavel-Chapelon F, Boutron-Ruault MC,
Clavel-Chapelon F, Guillas G, Tondeur L, Kernaleguen Guillas G, Overvad K, et al. (2012). Body size and risk
C, Boutron-Ruault MC (2010). Risk of differentiated of differentiated thyroid carcinomas: findings from the
thyroid cancer in relation to adult weight, height and EPIC study. Int J Cancer, 131(6):E1004–14. doi:10.1002/
body shape over life: the French E3N cohort. Int J ijc.27601 PMID:22511178
Cancer, 126(12):2984–90. PMID:19950225 Samanic C, Chow WH, Gridley G, Jarvholm B,
Cléro E, Leux C, Brindel P, Truong T, Anger A, Teinturier Fraumeni JF Jr (2006). Relation of body mass index
C, et al. (2010). Pooled analysis of two case-control to cancer risk in 362,552 Swedish men. Cancer Causes
studies in New Caledonia and French Polynesia of Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9
body mass index and differentiated thyroid cancer: the PMID:16841257
importance of body surface area. Thyroid, 20(11):1285– Samanic C, Gridley G, Chow WH, Lubin J, Hoover
93. doi:10.1089/thy.2009.0456 PMID:20932181 RN, Fraumeni JF Jr (2004). Obesity and cancer
Dal Maso L, La Vecchia C, Franceschi S, Preston- risk among white and black United States veterans.
Martin S, Ron E, Levi F, et al. (2000). A pooled Cancer Causes Control, 15(1):35–43. doi:10.1023/B:-
analysis of thyroid cancer studies. V. Anthropometric CACO.0000016573.79453.ba PMID:14970733
factors. Cancer Causes Control, 11(2):137–44. Schmid D, Ricci C, Behrens G, Leitzmann MF (2015).
doi:10.1023/A:1008938520101 PMID:10710197 Adiposity and risk of thyroid cancer: a systematic
Engeland A, Tretli S, Akslen LA, Bjørge T (2006). Body review and meta-analysis. Obes Rev, 16(12):1042–54.
size and thyroid cancer in two million Norwegian men doi:10.1111/obr.12321 PMID:26365757
and women. Br J Cancer, 95(3):366–70. doi:10.1038/ Song YM, Sung J, Ha M (2008). Obesity and risk of
sj.bjc.6603249 PMID:16832414 cancer in postmenopausal Korean women. J Clin
Guignard R, Truong T, Rougier Y, Baron-Dubourdieu D, Oncol, 26(20):3395–402. doi:10.1200/JCO.2007.15.7867
Guénel P (2007). Alcohol drinking, tobacco smoking, PMID:18612154
and anthropometric characteristics as risk factors Suzuki T, Matsuo K, Hasegawa Y, Hiraki A, Kawase T,
for thyroid cancer: a countrywide case-control study Tanaka H, et al. (2008). Anthropometric factors at age
in New Caledonia. Am J Epidemiol, 166(10):1140–9. 20 years and risk of thyroid cancer. Cancer Causes
doi:10.1093/aje/kwm204 PMID:17855390 Control, 19(10):1233–42. doi:10.1007/s10552-008-
IARC (2002). Weight control and physical activity. Lyon, 9194-x PMID:18618280
France: IARC Press (IARC Handbooks of Cancer WCRF/AICR (2007). Food, nutrition, physical activity,
Prevention, Vol. 6). Available from: http://publications. and the prevention of cancer: a global perspective.
iarc.fr/376. Washington (DC), USA: American Institute for Cancer
Kabat GC, Kim MY, Thomson CA, Luo J, Wactawski- Research. Available from: http://www.aicr.org/assets/
Wende J, Rohan TE (2012). Anthropometric factors docs/pdf/reports/Second_Expert_Report.pdf.
and physical activity and risk of thyroid cancer in post- Xhaard C, de Vathaire F, Cléro E, Maillard S, Ren Y,
menopausal women. Cancer Causes Control, 23(3):421– Borson-Chazot F, et al. (2015). Anthropometric risk
30. doi:10.1007/s10552-011-9890-9 PMID:22212611 factors for differentiated thyroid cancer in young men
Kitahara CM, McCullough ML, Franceschi S, Rinaldi S, and women from eastern France: a case-control study.
Wolk A, Neta G, et al. (2016). Anthropometric factors Am J Epidemiol, 182(3):202–14. doi:10.1093/aje/kwv048
and thyroid cancer risk by histological subtype: pooled PMID:26133374
420
Xu L, Port M, Landi S, Gemignani F, Cipollini M, Elisei

R, et al. (2014). Obesity and the risk of papillary thyroid
cancer: a pooled analysis of three case-control studies.
Thyroid, 24(6):966–74. doi:10.1089/thy.2013.0566
PMID:24555500
Zhang W, Bai X, Ge H, Cui H, Wei Z, Han G (2014). Meta-
analysis in the association between obesity and risk
of thyroid cancer. Int J Clin Exp Med, 7(12):5268–74.
PMID:25664030
421
2.2.20 Haematopoietic malignancies of leukaemia (CML). Also included are findings

lymphoid origin from prospective studies with at least 50 cases
for any specific subtype, meta-analyses or pooled
Haematopoietic malignancies are a heter- analyses from prospective and case–control
ogeneous group of cancers that arise from the studies, and findings from case–control studies
blood, the bone marrow, and lymphoid tissue, with at least 50 cases that were not included in a
and the cells of origin are either lymphoid or meta-analysis or pooled analysis. For malignan-
myeloid. Historically, haematopoietic cancers cies for which there was evidence suggesting a
were grouped into five major categories: Hodgkin relationship between BMI or weight at baseline
lymphoma, non-Hodgkin lymphoma (NHL), and risk, the Working Group also weighed into
multiple myeloma (also referred to as plasma their evaluation studies that assessed weight
cell myeloma), acute leukaemia, and chronic change and weight during young adulthood in
leukaemia. However, these historical group- relation to risk. Not all studies separated the
ings do not reflect the current understanding haematopoietic cancers according to the current
of etiology and pathogenesis or current clinical WHO classification system. Therefore, findings
practice. In 2001, the World Health Organization are presented about relationships between BMI
(WHO) introduced a new classification system and both individual haematological cancers and
(Jaffe et al., 2001), which was subsequently groups of cancers.
updated (Swerdlow et al., 2008) and is consid- Table 2.2.20a and Table 2.2.20b (both
ered the reference standard for classification of web only, available at: http://publications.iarc.
these malignancies. The WHO classification has fr/570) present data for cohort and case–control
been adopted worldwide, and the terminology studies, respectively, for subsites with inadequate
has been incorporated into the third edition of evidence; Table 2.2.20c and Table 2.2.20d present
the International Classification of Diseases for the corresponding studies for subsites with suffi-
Oncology (ICD-O-3) (WHO, 2013). cient or limited evidence.
Handbook on weight control and physical activity (a) Hodgkin lymphoma
(IARC, 2002) concluded that the evidence of an There are at least four individual prospec-
association between avoidance of weight gain and tive studies and one meta-analysis of BMI at
cancers of the haematopoietic system (e.g. NHL, baseline in relation to incidence of Hodgkin
multiple myeloma) was inadequate. In the current lymphoma (Table 2.2.20a, web only, available
review are included epidemiological studies of at: http://publications.iarc.fr/570). Results from
BMI, weight, or waist circumference at baseline three prospective studies in men (i.e. the United
in relation to risk of the more common types of States Veterans cohort, the NIH-AARP cohort,
haematopoietic malignancies for which the body and the Swedish Construction Worker cohort)
of evidence was substantial enough to review. showed increased risks (Samanic et al., 2004,
These include Hodgkin lymphoma, NHL, 2006; Lim et al., 2007), none of which were statis-
B-cell lymphoma overall, chronic lymphocytic tically significant, whereas a Norwegian cohort
leukaemia (CLL)/small lymphocytic lymphoma study found a positive association in women but
(SLL), diffuse large B-cell lymphoma (DLBCL), not in men (Engeland et al., 2007). [Most studies
follicular lymphoma, multiple myeloma, T-cell had limited statistical power, particularly at the
lymphoma, leukaemia overall, acute myeloid high end of the BMI categories.] However, in
leukaemia (AML), and chronic myeloid the meta-analysis of five studies, obesity was
422
associated with a statistically significant 41% Three meta-analyses showed a positive associ-
higher risk of Hodgkin lymphoma compared ation between BMI and NHL incidence and/or
with normal BMI (Larsson & Wolk, 2011). mortality (Larsson & Wolk, 2007a, 2011; Renehan
Only three case–control studies with at et al., 2008), whereas one pooled analysis found
least 50 cases have evaluated the relationship no association (Whitlock, et al., 2009). [The
between BMI and risk of Hodgkin lymphoma inconsistent evidence from individual prospec-
(Table 2.2.20b, web only, available at: http:// tive studies and meta-analyses may be due to the
publications.iarc.fr/570). The largest study, variation in histological subtypes included in a
including 618 cases from the Scandinavian classification of NHL.]
Lymphoma Etiology Study and 3187 population No association between waist circumference
controls, did not find a relationship between and NHL incidence was seen in the Women’s
BMI and risk of Hodgkin lymphoma in indi- Health Initiative in the USA (Kabat et al., 2012).
viduals younger or older than 45 years, assessed However, in a cohort in Taiwan, China, high
separately because of the bimodal distribution of abdominal obesity (waist circumference ≥ 90 cm
the disease (Chang et al., 2005). A second large in men and ≥ 80 cm in women) was associated
study, including 567 cases and 697 controls, also with an 86% higher risk of fatal NHL compared
did not find a relationship between BMI and risk with lower waist circumference (Chu et al., 2011).
of Hodgkin lymphoma in subgroups defined by A total of 11 hospital-based or popula-
sex and age (Li et al., 2013). However, a smaller tion-based case–control studies have evaluated
study of 216 cases and 216 matched controls, the relationship between BMI and risk of any
which considered BMI 5 years before cancer NHL (Table 2.2.20b, web only, available at:
diagnosis, found an increased risk of Hodgkin http://publications.iarc.fr/570). A meta-anal-
lymphoma with BMI ≥ 30 kg/m2 compared with ysis of six of these reports published in 2004
normal BMI in men, but not in women (Willett and 2005 (Pan et al., 2004; Skibola et al., 2004;
& Roman, 2006). Bosetti et al., 2005; Cerhan et al., 2005; Chang
et al., 2005; Willett et al., 2005) reported a rela-
(b) Non-Hodgkin lymphoma tive risk of NHL of 1.22 (95% CI, 1.00–1.50) in
There are at least 21 individual prospective individuals with BMI ≥ 30 kg/m2 (Larsson &
studies and 4 meta-analyses or pooled analyses Wolk, 2007a). A subsequent pooled analysis from
of BMI and/or weight at baseline in relation the InterLymph Consortium included data from
to NHL (Table 2.2.20a, web only, available at: 10 000 cases of NHL and 16 000 controls drawn
http://publications.iarc.fr/570). There were no from 18 case–control studies identified through
associations of either BMI or weight with NHL the International Lymphoma Epidemiology
incidence or mortality in 12 individual prospec- Consortium (Willett et al., 2008). That study
tive studies (Samanic et al., 2004, 2006; Fujino did not find a relationship between BMI and
et al., 2007; Maskarinec et al., 2008; Song et al., risk of NHL, with a relative risk of NHL of 0.84
2008; Andreotti et al., 2010; De Roos et al., 2010; (95% CI, 0.72–0.99) in individuals with BMI of
Kanda et al., 2010; Hemminki et al., 2011; Kabat 30–39.9 kg/m2 and a relative risk of 0.63 (95% CI,
et al., 2012; Bertrand et al., 2013; Bhaskaran et al., 0.40–0.99) in those with BMI ≥ 40 kg/m2.
2014). The other nine studies found positive asso-
ciations in men and/or women (Calle et al., 2003; (c) B-cell lymphoma
Oh et al., 2005; Rapp et al., 2005; Chiu et al., 2006; The association between excess body
Engeland et al., 2007; Lim et al., 2007; Reeves fatness and the incidence of B-cell lymphoma
et al., 2007; Troy et al., 2010; Chu et al., 2011). was examined in three individual prospective
423
studies (Table 2.2.20a, web only, available at: No association between waist circumference
http://publications.iarc.fr/570) [notably, under and CLL/SLL incidence was found in the three
the current classification, this includes all B-cell studies that examined this relationship (Ross
malignancies previously included under NHL]. et al., 2004; Kabat et al., 2012; Saberi Hosnijeh
Although no association was found with BMI et al., 2013).
and/or weight in men or in women in the EPIC Five case–control studies with at least 50
cohort (Britton et al., 2008), there were statisti- cases assessed the relationship between BMI and
cally significant positive trends with weight in risk of CLL/SLL (Table 2.2.20b, web only, avail-
the California Teachers Study (Lu et al., 2009) able at: http://publications.iarc.fr/570) and found
and with BMI in the Cancer Prevention Study II no association between BMI and CLL/SLL risk
Nutrition Cohort (Patel et al., 2013). (Chang et al., 2005; Pan et al., 2005; Morton et al.,
In the one study that assessed waist circum- 2008; Chen et al., 2011; Kelly et al., 2012).
ference, there was no association with the inci- (ii) Diffuse large B-cell lymphoma
dence of B-cell lymphoma in either men or
women (Britton et al., 2008). Associations of baseline BMI and/or weight
with risk of DLBCL have been examined in at
(d) Subtypes of B-cell lymphoma least nine individual prospective studies and two
meta-analyses (Table 2.2.20c). Most individual
(i) Chronic lymphocytic leukaemia/small
prospective studies found no evidence of an
lymphocytic lymphoma
association (Lim et al., 2007; Britton et al., 2008;
Most of the individual prospective studies Maskarinec et al., 2008; Lu et al., 2009; Pylypchuk
(Table 2.2.20a, web only, available at: http:// et al., 2009; Kabat et al., 2012; Bertrand et al.,
publications.iarc.fr/570) found no associations of 2013). However, two large studies in the USA did
BMI and/or weight at baseline with the incidence report statistically significant trends between
of CLL or CLL/SLL (Ross et al., 2004; Samanic baseline weight (Troy et al., 2010) or BMI
et al., 2006; Engeland et al., 2007; Lim et al., 2007; (Patel et al., 2013) and DLBCL incidence. Both
Lu et al., 2009; Pylypchuk et al., 2009; Kabat meta-analyses also showed statistically signif-
et al., 2012; Bertrand et al., 2013; Patel et al., icant positive associations. One meta-analysis
2013; Saberi Hosnijeh et al., 2013). However, in reported a relative risk per 5 kg/m2 increase of
the United States Veterans study, the largest indi- 1.13 (95% CI, 1.02–1.26) (Larsson & Wolk, 2011).
vidual prospective study, the risk of CLL was In the other meta-analysis, both overweight and
30% higher in obese White men and 72% higher obesity in men and women were associated with
in obese Black men compared with non-obese increased risk (Castillo et al., 2014).
men (Samanic et al., 2004). In the Prostate, Lung, Six individual studies assessed the associ-
Colorectal, and Ovarian Cancer Screening Trial ation between BMI or weight in early adult-
in the USA, baseline weight, but not BMI, was hood and incidence of DLBCL. In the two large
positively associated with risk (Ptrend < 0.215) (Troy studies in the USA, there were statistically
et al., 2010). Although an earlier meta-analysis of significant positive associations of weight at age
three cohort studies suggested a 25% higher risk 20 years (Ptrend = 0.013) (Troy et al., 2010) and of
of CLL for obesity versus normal weight (Larsson young adult BMI in men and women combined
& Wolk, 2008), an updated meta-analysis of six (Ptrend = 0.02) (Bertrand et al., 2013) with risk
prospective studies found no association between of DLBCL. However, in none of the four other
BMI as a continuous measure and incidence of studies was BMI and/or body weight at age
CLL/SLL (Larsson & Wolk, 2011). 18 years (Lu et al., 2009; Patel et al., 2013), at
424
age 20 years (Pylypchuk et al., 2009), or at age the two studies that examined this relationship
21 years (Maskarinec et al., 2008) associated (Britton et al., 2008; Kabat et al., 2012).
with DLBCL incidence. Similarly, adult weight The largest study evaluating the association
gain was not associated with risk of DLBCL in between BMI and follicular lymphoma (Table
any of the studies that examined this association 2.2.20b, web only, available at: http://publications.
(Maskarinec et al., 2008; Troy et al., 2010; Patel iarc.fr/570) was a pooled analysis of 3530 cases
et al., 2013). and 22 639 population controls from 19 case–
In the EPIC cohort, there was a 2-fold (RR, control studies in the InterLymph Consortium,
2.03; 95% CI, 0.96–4.28) higher incidence of which found no relationship between BMI and
DLBCL for waist circumference ≥ 102 cm versus risk of follicular lymphoma (Linet et al., 2014).
< 102 cm in men (based on only 21 cases in Two additional case–control studies not included
the group with high waist circumference), and in the pooled analysis also found no associa-
no association in women (Britton et al., 2008). tion between adult BMI and risk of follicular
Similarly, there was no association between waist lymphoma (Pan et al., 2005; Chen et al., 2011).
circumference and risk of DLBCL in the Women’s Therefore, the relationship between BMI and
Health Initiative in the USA (Kabat et al., 2012). risk of NHL varies by subtype, with a positive
A pooled analysis of 19 case–control studies association seen in some studies limited to the
from the InterLymph Consortium of 4667 cases risk of DLBCL, but not in studies assessing the
of DLBCL and 22 639 controls found a signifi- risk of any NHL or of follicular lymphoma.
cant positive association between risk of DLBCL
and young adult BMI, but not usual adult BMI (e) Multiple myeloma
(Cerhan et al., 2014). A case–control study from In the individual prospective studies that
the National Enhanced Cancer Surveillance examined the association of baseline BMI and/
System in Canada, including 419 cases of DLBCL, or weight with multiple myeloma incidence or
found an odds ratio for individuals with BMI mortality (Table 2.2.20c), most found positive
≥ 30 kg/m2 of 1.35 (95% CI, 0.99–1.83) (Pan et al., associations for at least one measure of excess
2005). Another case–control study, by Chen body fatness at baseline (Calle et al., 2003;
et al., (2011), including 245 cases of DLBCL, did Samanic et al., 2004; Blair et al., 2005; Birmann
not find a relationship between BMI and risk of et al., 2007; Engeland et al., 2007; Fujino et
DLBCL (Table 2.2.20d). al., 2007; Reeves et al., 2007; Troy et al., 2010;
(iii) Follicular lymphoma Hofmann et al., 2013). In particular, a positive
association was observed in the largest studies. In
None of the nine individual prospective the United States Veterans cohort of more than
studies (Lim et al., 2007; Britton et al., 2008; 4 million men, the risk of multiple myeloma was
Maskarinec et al., 2008; Lu et al., 2009; Pylypchuk 22% higher in obese White men and 26% higher
et al., 2009; Troy et al., 2010; Kabat et al., 2012; in obese Black men compared with non-obese
Bertrand et al., 2013; Patel et al., 2013) or the one men (Samanic et al., 2004). Similarly, the Million
meta-analysis (Larsson & Wolk, 2011) showed Women Study in the United Kingdom found
any evidence of an association between BMI positive associations between BMI and multiple
and/or weight and the incidence of follicular myeloma incidence and mortality (31% and 56%
lymphoma (Table 2.2.20a, web only, available at: increase, respectively, per 10 kg/m2) (Reeves et al.,
http://publications.iarc.fr/570). 2007). In a Norwegian cohort study of more than
Waist circumference was also not associated 2 million men and women whose height and
with the incidence of follicular lymphoma in weight were measured at baseline in 1963, there
425
were statistically significant dose-related positive Pylypchuk et al., 2009; De Roos et al., 2010; Lu
associations between BMI and risk of multiple et al., 2010; Patel et al., 2013), whereas in two
myeloma in men (RR, 1.14 for overweight and large studies young adult BMI was positively
1.28 for obesity vs normal BMI; Ptrend < 0.001) associated with risk (Troy et al., 2010; Hofmann
and in women (RR, 1.12 for overweight, 1.23 for et al., 2013). In the large pooled analysis by Teras
grade I, 1.42 for grade II, and 1.57 for grade III et al. (2014) there was a statistically significant
obesity vs normal BMI; Ptrend < 0.001) (Engeland positive association between increasing levels of
et al., 2007). One study found an inverse associ- young adult BMI (beginning in the overweight
ation (Samanic et al., 2006), and several studies category) and multiple myeloma mortality,
found no association (Oh et al., 2005; Fernberg although there was no association for change in
et al., 2007; Pylypchuk et al., 2009; De Roos et al., BMI during adulthood.
2010; Lu et al., 2010; Kanda et al., 2010; Patel et al., High waist circumference was associated
2013; Bhaskaran et al., 2014). with increased multiple myeloma incidence
Several meta-analyses or pooled analyses in one prospective study of postmenopausal
of excess body fatness in relation to multiple women (Blair et al., 2005), but not in two other
myeloma incidence and/or mortality have been studies (Britton et al., 2008; Lu et al., 2010). In
conducted (Larsson & Wolk, 2007b; Renehan the large pooled analysis by Teras et al. (2014),
et al., 2008; Parr et al., 2010; Wallin & Larsson, there was a statistically significant positive
2011; Teras et al., 2014). No association between association between waist circumference and
BMI and multiple myeloma mortality was found multiple myeloma mortality in men and women
in the Asia-Pacific Cohort Study Collaboration combined (RR per 5 cm increase, 1.06).
(Parr et al., 2010). However, in the meta-anal- Five case–control studies have evaluated
ysis by Wallin & Larsson (2011), which included the relationship between BMI and the risk of
studies worldwide, overweight and obesity multiple myeloma, four of which were included
were associated with a statistically significantly in a meta-analysis (Larsson & Wolk, 2007b; Table
increased risk of multiple myeloma incidence 2.2.20d). An increased risk of multiple myeloma
(RR, 1.12 for overweight and 1.21 for obesity, was reported in individuals who were overweight
based on 15 studies) and mortality (RR, 1.15 (RR, 1.43; 95% CI, 1.23–1.68) and those who were
for overweight and 1.54 for obesity, based on 5 obese (RR, 1.82; 95% CI, 1.47–2.26). One addi-
studies). The two earlier meta-analyses (Larsson tional study reported no significant association
& Wolk, 2007b; Renehan et al., 2008) found (Wang et al., 2013).
statistically significant positive associations
of a similar magnitude. Consistent with these (f) T-cell lymphoma
findings, in a pooled analysis of data from 20 In the Cancer Prevention Study II Nutrition
prospective studies (Teras et al., 2014), there was Cohort (Table 2.2.20a, web only, available at:
a statistically significant positive association http://publications.iarc.fr/570), there was a posi-
between BMI and multiple myeloma mortality tive association between BMI and the incidence
(RR per 5 kg/m2 increase in BMI, 1.09). of T-cell lymphoma (Ptrend = 0.013) (Patel et al.,
Given the observed associations between 2013). However, in two European cohort studies
baseline BMI and risk of multiple myeloma, there was no association (Lukanova et al., 2006;
associations with young adult BMI and with Lim et al., 2007). In the Cancer Prevention Study
BMI change were also examined. Several studies II Nutrition Cohort, BMI at age 18 years was not
found no association between young adult BMI associated with the incidence of T-cell lymphoma
and risk of multiple myeloma (Fujino et al., 2007; (Patel et al., 2013).
426
Table 2.2.20c Cohort studies of measures of body fatness and haematopoietic malignancies of lymphoid origin with
sufficient or limited evidence
Reference Total number of Organ site Exposure categories Exposed cases Relative risk Covariates Comments
Cohort subjects (ICD code) (95% CI)
Location Sex
mortality
Diffuse large B-cell lymphoma
Lim et al. (2007) 473 984 DLBCL BMI Age, ethnicity,
NIH-AARP Men and women ICD-O-2: 18.5–24.9 119 1.00 education
cohort Incidence 9680–9684, 25–29.9 141 0.92 (0.72–1.18) level, alcohol
USA 9688, 30–34.9 61 1.10 (0.81–1.51) intake, cigarette
1995–2003 9710–9712, ≥ 35 21 1.17 (0.73–1.88) smoking, height,
9715 [Ptrend] [0.42] physical activity
Britton et al. 141 425 DLBCL BMI Age, study centre Also examined
(2008) Men < 25 24 1.00 height, hip
EPIC cohort Incidence 25–29.9 37 0.83 (0.39–1.76) circumference, and
10 European ≥ 30 10 0.94 (0.56–1.59) waist-to-hip ratio
countries [Ptrend] [0.63]
1993–1998 Weight (kg)
< 72.7 19 1.00
72.7–79.8 13 0.59 (0.29–1.20)
79.9–87.7 20 0.90 (0.46–1.74)
≥ 87.8 19 0.86 (0.42–1.77)
[Ptrend] [1.00]
WC (cm)
< 102 44 1.00
≥ 102 21 2.03 (0.96–4.28)

230 558 DLBCL BMI Age, study centre Also examined
Women < 25 30 1.00 height, hip
Incidence 25–29.9 31 1.27 (0.63–2.55) circumference, and
≥ 30 12 1.54 (0.92–2.57) waist-to-hip ratio
[Ptrend] [0.28]
427
428

Location Sex
mortality
Britton et al. Weight (kg)
(2008) < 72.7 14 1.00
(cont.) 72.7–79.8 12 0.78 (0.35–1.69)
79.9–87.7 21 1.32 (0.66–2.68)
≥ 87.8 26 1.62 (0.81–3.25)
[Ptrend] [0.06]
WC (cm)
< 88 47 1.00
≥ 88 21 0.88 (0.42–1.85)
Maskarinec et al. 87 079 DLBCL BMI at baseline Age, ethnicity,
(2008) Men ICD-O-3: < 22.5 23 0.65 (0.35–1.21) education
Multiethnic Incidence 9675, 9680, 22.5–24.9 44 1.00 level, alcohol
Cohort 9684 25.0–29.9 60 0.90 (0.56–1.43) consumption
1993–2002 ≥ 30.0 23 0.78 (0.40–1.52)
[Ptrend] [0.69]
BMI at age 21 yr
< 18.5 14 0.56 (0.27–1.15)
18.5–24.9 105 1.00
25.0–29.9 17 0.78 (0.41–1.48)
≥ 30.0 5 1.03 (0.36–2.91)
[Ptrend] [0.51]
< 152.0 47 1.00
152.0–170.0 37 1.97 (1.16–3.36)
170.1–192.0 32 1.36 (0.75–2.49)
> 192.0 35 1.87 (0.95–3.68)
[Ptrend] [0.12]
< 130.0 43 1.00
130.0–145.0 34 0.87 (0.50–1.53)
145.1–165.0 38 1.24 (0.64–2.41)
> 165.0 27 1.26 (0.63–2.50)
[Ptrend] [0.33]
Location Sex
mortality
Maskarinec et al. Annual weight change (lb)
(2008) 0 or loss 32 1.00
(cont.) ≤ 1 89 1.07 (0.62–1.86)
> 1 31 1.14 (0.56–2.34)
[Ptrend] [0.69]
105 972 DLBCL BMI at baseline Age, ethnicity,
Women ICD-O-3: < 22.5 27 1.41 (0.66–3.00) education
Incidence 9675, 9680, 22.5–24.9 30 1.00 level, alcohol
9684 25.0–29.9 43 1.06 (0.58–1.96) consumption, age
≥ 30.0 28 1.45 (0.75–2.82) at first birth
[Ptrend] [0.80]
BMI at age 21 yr
< 18.5 16 1.02 (0.50–2.10)
18.5–24.9 91 1.00
25.0–29.9 11 1.08 (0.50–2.33)
≥ 30.0 4 0.94 (0.25–3.55)
[Ptrend] [1.00]
< 125.0 26 1.00
125.0–143.0 38 0.74 (0.40–1.38)
143.1–167.0 35 1.35 (0.67–2.75)
> 167.0 30 1.20 (0.57–2.52)

[Ptrend] [0.40]
< 105.0 22 1.00
105.0–118.0 34 0.70 (0.35–1.41)
118.1–127.0 34 0.97 (0.48–1.96)
> 127.0 33 1.10 (0.53–2.29)
[Ptrend] [0.44]
Annual weight change (lb)
0 or loss 19 1.00
≤ 1 85 0.56 (0.21–1.55)
> 1 28 0.93 (0.34–2.54)
[Ptrend] [0.85]
429
430

Location Sex
mortality
Lu et al. (2009) 121 216 DLBCL BMI at baseline Weight, height, Also included
California Women < 20 17 1.42 (0.83–2.42) age at menarche, results for height
Teachers Study Incidence 20–24.9 64 1.00 and physical and physical
USA 25–29.9 41 1.07 (0.72–1.59) activity activity
1995–2007 ≥ 30 26 1.37 (0.86–2.16)
[Ptrend] [0.50]
BMI at age 18 yr
< 19.5 37 0.98 (0.62–1.56)
19.5–20.7 36 1.00
20.8–22.4 32 0.90 (0.56–1.45)
> 22.4 43 1.23 (0.79–1.92)
[Ptrend] [0.30]
Weight (kg) at baseline
< 56.7 33 1.24 (0.76–2.03)
56.7– < 63.5 33 1.00
63.5– < 73.0 40 0.90 (0.57–1.43)
≥ 73.0 42 1.08 (0.68–1.72)
[Ptrend] [0.81]
< 52.6 33 0.88 (0.54–1.41)
52.6– < 57.2 45 1.00
57.2– < 61.7 30 1.16 (0.72–1.84)
≥ 61.7 40 1.23 (0.79–1.92)
[Ptrend] [0.19]
Pylypchuk et al. 5000 DLBCL BMI Age, sex Case–cohort
(2009) Men and women ICD-O-3: < 18.5 3 1.91 (0.58–6.30) design
Netherlands Incidence 9675, 9680, 18.5–24.9 112 1.00
Cohort Study on 9684 25–29.9 101 1.16 (0.88–1.53)
Diet and Cancer ≥ 30 8 0.62 (0.30–1.30)
The Netherlands [Ptrend] [0.77]
1986–1999 per 4 kg/m2 0.92 (0.77–1.10)
Location Sex
mortality
Pylypchuk et al. BMI at age 20 yr
(2009) < 20 39 0.96 (0.61–1.50)
(cont.) 20–21.4 43 1.00
21.5–22.9 41 0.99 (0.64–1.54)
23–24.9 43 1.35 (0.88–2.10)
≥ 25 16 1.29 (0.71–2.35)
[Ptrend] [0.12]
per 4 kg/m2 1.20 (0.98–1.47)
Troy et al. (2010) 142 982 DLBCL BMI at baseline Age, race/
PLCO Trial Men and women < 18.5 4 – ethnicity,
USA Incidence 18.5–24.9 58 1.00 education level
1993–2006 25–29.9 87 1.07 (0.76–1.50)
≥ 30 63 1.58 (1.10–2.27)
[Ptrend] [0.056]
BMI at age 20 yr
< 18.5 17 1.22 (0.74–2.02)
18.5–24.9 157 1.00
25–29.9 35 1.19 (0.82–1.73)
≥ 30 1 –
[Ptrend] [0.230]
Weight change (kg) per 10 yr
Loss 10 0.70 (0.35–1.39)

Gain 0–2 53 1.00
Gain 2.1–4 66 1.13 (0.78–1.63)
Gain 4.1–6 46 1.32 (0.88–1.97)
Gain ≥ 6 37 1.41 (0.91–2.18)
[Ptrend] [0.114]
Weight (kg) at baseline, quartiles (sex-specific)
Men: Women:
< 77.4 < 61.5 51 1.00
77.4–85.5 61.5–70.0 46 1.05 (0.71–1.57)
85.6–95.5 70.1–80.0 54 1.18 (0.81–1.74)
> 95.5 > 80.0 63 1.63 (1.12–2.37)
[Ptrend] [< 0.01]
431
432

Location Sex
mortality
Troy et al. (2010) Weight (kg) at age 20 yr, quartiles (sex-specific)
(cont.) Men: Women:
< 64.2 < 51.9 40 1.00
64.2–72.7 51.9–54.5 41 1.26 (0.81–1.95)
72.8–79.5 54.6–59.1 66 1.46 (0.98–2.17)
> 79.5 > 59.1 65 1.67 (1.12–2.50)
[Ptrend] [0.013]
Larsson & Wolk 6 studies DLBCL BMI NR
(2011) Men and women per 5 kg/m2 1.13 (1.02–1.26)
Meta-analysis Incidence
Multiple locations
1999–2010
Kabat et al. (2012) 158 975 DLBCL BMI at baseline Age, smoking, Also included
Women’s Health Women ICD-O-3: < 25 99 1.00 alcohol estimates for
Initiative Incidence 9678–9680, 25– < 30 115 1.23 (0.93–1.62) consumption, height, hip
USA 9684 30– < 35 55 1.11 (0.78–1.58) education circumference,
1993–2009 ≥ 35 33 1.30 (0.85–1.99) level, ethnicity, waist-to-hip ratio,
[Ptrend] [0.25] physical activity, and weight/BMI
Weight (kg) at baseline energy intake, at ages 18 yr, 35 yr,
< 62.0 73 1.00 substudy and 50 yr
62.0– < 70.4 79 1.09 (0.78–1.51)
70.4– < 81.6 80 1.11 (0.79–1.56)
≥ 81.6 70 1.05 (0.72–1.52)
[Ptrend] [0.77]
WC (cm) at baseline
< 76.1 70 1.00
76.1– < 84.6 80 1.13 (0.82–1.58)
84.6– < 95.0 68 1.02 (0.72–1.44)
> 95.0 84 1.28 (0.91–1.81)
[Ptrend] [0.25]
Location Sex
mortality
Bertrand et al. 163 184 DLBCL Adult BMI [261] Age, height,
(2013) Men and women per 5 kg/m2 1.10 (0.91–1.33) smoking, physical
Nurses’ Health Incidence [Ptrend] [0.31] activity, race
Study and Health Young adult BMI [241]
Professionals per 5 kg/m2 1.29 (1.05–1.57)
Follow-up Study [Ptrend] [0.02]
USA 46 390 DLBCL Adult BMI Age, height,
1976–2008 Men 15–22.9 11 1.00 smoking, physical
Incidence 23–24.9 25 1.57 (0.75–3.28) activity, race
25–26.9 23 1.58 (0.75–3.34)
27–29.9 17 1.65 (0.75–3.64)
30–45 10 2.18 (0.88–5.40)
per 5 kg/m2 1.30 (0.92–1.82)
[Ptrend] [0.14]
Young adult BMI
15–18.4 4 1.36 (0.46–4.02)
18.5–22.9 40 1.00
23–24.9 19 0.94 (0.54–1.64)
25–29.9 17 1.16 (0.65–2.08)
30–45 4 2.70 (0.93–7.86)
per 5 kg/m2 1.29 (0.89–1.88)
[Ptrend] [0.18]

116 794 DLBCL Adult BMI Age, height,
Women 15–22.9 60 1.00 smoking, physical
Incidence 23–24.9 38 0.97 (0.64–1.46) activity, race
25–26.9 31 1.06 (0.69–1.65)
27–29.9 23 0.85 (0.52–1.38)
30–45 33 1.36 (0.88–2.10)
per 5 kg/m2 1.04 (0.88–1.23)
[Ptrend] [0.65]
433
434

Location Sex
mortality
Bertrand et al. Young adult BMI
(2013) 15–18.4 14 0.71 (0.40–1.24)
(cont.) 18.5–22.9 104 1.00
23–24.9 18 0.99 (0.60–1.64)
25–29.9 17 1.26 (0.75–2.11)
30–45 4 1.39 (0.51–3.81)
per 5 kg/m2 1.28 (1.01–1.63)
[Ptrend] [0.04]
Patel et al. (2013) 152 423 DLBCL BMI at baseline Age, sex, family
Cancer Men and women < 18.5 1 0.28 (0.04–1.97) history of
Prevention Study Incidence 18.5– < 25 159 1.00 haematopoietic
II Nutrition 25– < 30 199 1.30 (1.05–1.61) cancer, education
Cohort ≥ 30 85 1.62 (1.23–2.12) level, smoking
USA [Ptrend] [0.0001] status, physical
1992–2007 BMI at age 18 yr activity, alcohol
< 18.5 52 0.86 (0.64–1.17) consumption
18.5– < 22.5 245 1.00
22.5– < 25 88 1.07 (0.83–1.38)
25– < 30 44 1.01 (0.72–1.42)
≥ 30 7 1.30 (0.60–2.80)
[Ptrend] [0.32]
Adult weight change (lb)
Loss > 5 11 0.60 (0.32–1.10)
Loss 5 to gain 20 147 1.00
Gain 21–40 142 0.97 (0.77–1.22)
Gain 41–60 83 0.97 (0.74–1.28)
Gain > 60 52 1.11 (0.80–1.54)
[Ptrend] [0.25]
Location Sex
mortality
Castillo et al. NR DLBCL BMI
(2014) Men and women Overweight 1.14 (1.04–1.24)
Meta-analysis Incidence Obese 1.29 (1.16–1.43)
of 10 cohorts, 6 NR DLBCL BMI
case–control Men Overweight 1.27 (1.09–1.47)
Incidence Obese 1.40 (1.00–1.95)
NR DLBCL BMI
Women Overweight 1.22 (1.07–1.38)
Incidence Obese 1.34 (1.16–1.54)
Multiple myeloma
Calle et al. (2003) 495 477 Multiple BMI Age, race,
Cancer Women myeloma 18.5–24.9 341 1.00 education level,
Prevention Study Mortality 25–29.9 187 1.12 (0.93–1.34) smoking, physical
II 30–34.9 72 1.47 (1.13–1.91) activity, alcohol
USA ≥ 35 20 1.44 (0.91–2.28) consumption,
1982–1998 [Ptrend] [0.004] marital status,
404 576 BMI aspirin use, fat
Men 18.5–24.9 259 1.00 and vegetable
Mortality 25–29.9 368 1.18 (1.01–1.39) consumption
30–34.9 70 1.44 (1.10–1.89)
≥ 35 11 1.71 (0.93–3.14)
[Ptrend] [0.002]

Samanic et al. 4 500 700 Multiple Obesity Age, calendar Obesity defined as
(2004) Men myeloma White men: year discharge diagnosis
United States Incidence ICD-9: 203 Non-obese 2817 1.00 of obesity: ICD-8:
USA Black men:
1969–1996 Non-obese 1509 1.00
Obese 89 1.26 (1.02–1.56)
Blair et al. (2005) 37 083 Multiple BMI Age Also included
Iowa Women’s Women myeloma 18.5–24.9 30 1.0 analyses of height,
Health Study Incidence 25–29.9 37 1.3 (0.78–2.0) waist-to-hip
USA ≥ 30 28 1.5 (0.92–2.6) ratio, and hip
1986–2001 [Ptrend] [0.10] circumference
435
436

Location Sex
mortality
Blair et al. (2005) Weight (lb)
(cont.) < 138 19 1.0
138–160 40 2.0 (1.1–3.4)
≥ 161 36 1.9 (1.1–3.4)
[Ptrend] [0.04]
WC (in)
< 31.75 19 1.0
31.76–36.25 37 1.9 (1.1–3.2)
≥ 36.26 39 2.0 (1.1–3.5)
[Ptrend] [0.02]
Oh et al. (2005) 781 283 Multiple BMI Age, smoking,
Korea National Men myeloma < 18.5 2 1.19 (0.29–4.96) alcohol intake,
Health Insurance Incidence 18.5–22.9 36 1.00 physical activity,
Corporation 23–24.9 45 1.72 (1.11–2.68) family history of
Republic of Korea 25–26.9 14 0.96 (0.51–1.77) cancer, urban/
1992–2001 27–29.9 6 0.98 (0.30–3.32) rural residence
≥ 30 0 –
[Ptrend] [0.61]
Samanic et al. 362 552 Multiple BMI Attained age,
(2006) Men myeloma 18.5–24.9 231 1.00 calendar year,
Swedish Incidence ICD-7: 203 25–29.9 201 0.96 (0.79–1.16) smoking
Construction ≥ 30 20 0.58 (0.37–0.93)
Worker Cohort [Ptrend] [0.06]
Sweden
1958–1999
Birmann et al. 136 623 Multiple BMI Age, sex, physical
(2007) Men and women myeloma < 22 28 1.0 activity, cohort
Nurses’ Health Incidence 22–24.9 64 1.2 (0.8–1.9)
Study and Health 25–29.9 84 1.3 (0.9–2.0)
Professionals ≥ 30 39 1.5 (0.9–2.5)
Follow-up Study [Ptrend] [0.11]
combined
USA
1980–2002
Location Sex
mortality
Birmann et al. 89 663 BMI Age, physical
(2007) Women < 22 21 1.0 activity
(cont.) Incidence 22–24.9 32 1.1 (0.7–2.0)
Nurses’ Health 25–29.9 53 1.6 (1.0–2.7)
Study ≥ 30 23 1.2 (0.7–2.2)
[Ptrend] [0.43]
Health 46 960 BMI Age, physical
Professionals Men < 22 7 1.0 activity
Follow-up Study Incidence 22–24.9 32 1.3 (0.5–2.9)
25–29.9 31 1.0 (0.4–2.2)
≥ 30 16 2.4 (1.0–6.0)
[Ptrend] [0.07]
Engeland et al. 963 709 Plasma cell BMI Age, birth cohort
(2007) Men myeloma < 18.5 11 0.69 (0.38–1.25)
Norwegian Incidence 18.5–24.9 1596 1.00
cohort 25–29.9 1417 1.14 (1.06–1.22)
Norway ≥ 30 209 1.28 (1.10–1.47)
1963–2001 [Ptrend] [< 0.001]
1 038 010 BMI
Women < 18.5 24 0.85 (057–1.27)
Incidence 18.5–24.9 1161 1.00
25–29.9 1125 1.12 (1.03–1.22)

30–34.9 436 1.23 (1.10–1.38)
35–39.9 110 1.42 (1.17–1.74)
≥ 40 26 1.57 (1.06–2.31)
[Ptrend] [< 0.001]
Fernberg et al. 336 381 Multiple BMI Attained age,
(2007) Men myeloma 18.5–25 256 1.00 snuff use, daily
Swedish Incidence 25.1–30 236 1.04 (0.86–1.24) tobacco smoking
construction > 30 27 0.70 (0.46–1.06)
workers
Sweden
1971–2004
437
438

Location Sex
mortality
Fujino et al. NR Multiple BMI Age, area of study [No information
(2007) Men myeloma < 18.5 3 0.96 (0.29–3.16) provided on
Japan Mortality 18.5–24 36 1.00 follow-up or
Collaborative 25–29 5 0.70 (0.27–1.80) number of people
Cohort Study ≥ 30 0 N.A in study]
Japan Weight (kg)
NR < 55 12 1.00
55–62 20 1.51 (0.73–3.11)
≥ 63 15 1.41 (0.64–3.12)
< 55 25 1.00
55–60 12 0.91 (0.38–2.14)
≥ 61 10 0.98 (0.40–2.42)
NR Multiple BMI Age, area of study [No information
Women myeloma < 18.5 2 0.59 (0.14–2.48) provided on
Mortality 18.5–24 31 1.00 follow-up or
25–29 7 0.77 (0.34–1.77) number of people
≥ 30 4 4.34 (1.51–12.5) in study]
Weight (kg)
< 49 18 1.00
49–54 12 0.93 (0.44–1.96)
≥ 55 17 1.17 (0.59–2.33)
< 47 24 1.00
47–52 9 0.76 (0.32–1.81)
≥ 53 11 0.87 (0.38–1.97)
Larsson & Wolk 9 cohort studies Multiple BMI 6987 total
(2007b) Men and women myeloma per 5 kg/m2 1.11 (1.03–1.19)
Multiple locations
1994–2007
Location Sex
mortality
Larsson & Wolk 9 cohort studies BMI 1492 total
(2007b) Men and women per 5 kg/m2 1.19 (1.12–1.28)
(cont.) Mortality
9 cohort studies BMI 8479 total Similar values
Men and women Normal 1.00 for BMI per
Incidence and Overweight 1.12 (1.07–1.18) 5 kg/m2 for men
mortality Obese 1.27 (1.15–1.41) and women
per 5 kg/m2 1.14 (1.09–1.20) separately
Reeves et al. 1 222 630 Multiple BMI Age, geographical
(2007) Women myeloma < 22.5 76 0.80 (0.64–1.00) region, SES,
Million Women Incidence ICD-10: C90 22.5–24.9 127 1.00 (0.84–1.19) reproductive
Study 25–27.4 118 1.11 (0.92–1.32) history, smoking
United Kingdom 27.5–29.9 73 1.11 (0.88–1.40) status, alcohol
1996–2005 ≥ 30 97 1.16 (0.95–1.42) intake, physical
per 10 kg/m2 1.31 (1.04–1.65) activity
1 222 630 BMI
Women < 22.5 46 0.99 (0.74–1.32)
Mortality 22.5–24.9 63 1.00 (0.78–1.28)
25–27.4 68 1.26 (0.99–1.59)
27.5–29.9 38 1.13 (0.82–1.55)
≥ 30 69 1.63 (1.28–2.08)
per 10 kg/m2 1.56 (1.15–2.10)

Britton et al. 141 425 Multiple BMI Age, study centre Also examined
(2008) Men myeloma < 25 43 1.00 height, hip
EPIC cohort Incidence 25–29.9 72 1.33 (0.79–2.23) circumference, and
10 European ≥ 30 24 1.17 (0.80–1.72) waist-to-hip ratio;
countries [Ptrend] [0.26] analyses by weight
1993–1998 and WC gave
similar results
230 558 Multiple BMI Age, study centre Also examined
Women myeloma < 25 59 1.00 height, hip
Incidence 25–29.9 49 0.93 (0.55–1.56) circumference, and
≥ 30 21 1.06 (0.72–1.58) waist-to-hip ratio;
[Ptrend] [0.89] analyses by weight
and WC gave
439
similar results
440

Location Sex
mortality
Renehan et al. 7 studies Multiple BMI
(2008) Men myeloma per 5 kg/m2 1.11 (1.05–1.18)
Multiple locations 6 studies BMI
1966–2007 Women per 5 kg/m2 1.11 (1.07–1.15)
Incidence
Pylypchuk et al. 5000 Multiple BMI Age, sex Case–cohort
(2009) Men and women myeloma < 25 135 1.00 design
Netherlands Incidence ICD-O-3: 25–29.9 126 1.23 (0.95–1.58) Similar results for
Cohort Study on 9731, 9732, ≥ 30 18 1.13 (0.68–1.88) BMI at age 20 yr
Diet and Cancer 9734 [Ptrend] [0.17]
The Netherlands per 4 kg/m2 1.13 (0.97–1.31)
1986–1999
De Roos et al. 81 219 Multiple BMI at enrolment Age, minority Similar results for
(2010) Women myeloma < 25 39 1.00 race, education BMI at age 18 yr,
Women’s Health Incidence 25–29.9 35 1.03 (0.65–1.63) level, region age 35 yr, and age
Initiative 30–34.9 10 0.66 (0.33–1.33) of the USA, 50 yr
USA ≥ 35 7 0.83 (0.37–1.87) smoking
1994–2008 [Ptrend] [0.37]
Kanda et al. 94 547 Plasma cell BMI Age, sex, Also included
(2010) Men and women myeloma < 18.5 2 0.56 (0.13–2.36) study area, estimates for
Japanese men and Incidence ICD-O-3: 18.5–22.9 33 0.70 (0.42–1.15) pack-years of height
women 9731, 9732 23.0–24.9 29 1.00 smoking, alcohol Similar results for
Japan 25–29.9 22 0.79 (0.45–1.38) consumption weight at baseline
1992–2006 ≥ 30 2 0.76 (0.18–3.20) and at age 20 yr
per 1 kg/m2 1.01 (0.95–1.09)
Weight (kg), quartiles (sex-specific)
Men: Women:
30–57 27–49 22 1.00
58–63 50–53 21 1.05 (0.57–1.93)
64–69 54–59 25 1.35 (0.74–2.46)
70–115 60–98 20 1.14 (0.59–2.21)
per 5 kg 1.06 (0.93–1.22)
Location Sex
mortality
Lu et al. (2010) 121 216 Multiple BMI at baseline Height, race Also included
California Women myeloma < 20 9 0.92 (0.45–1.86) estimates for hip
Teachers Study Incidence 20–24.9 55 1.00 circumference,
USA 25–29.9 28 0.83 (0.53–1.31) waist-to-hip ratio,
1995–2007 ≥ 30 14 0.86 (0.48–1.55) waist-to-height
[Ptrend] [0.55] ratio, and height
Weight (lb) at baseline Similar results for
< 131 38 1.00 BMI at age 18 yr,
131–154 36 0.85 (0.54–1.36) for weight at age
≥ 155 32 0.71 (0.43–1.16) 18 yr, and for WC
[Ptrend] [0.18]
Parr et al. (2010) 326 387 Myeloma BMI Age, sex, smoking
Asia-Pacific Men and women ICD-9: 203 < 18.5 3 1.94 (0.57–6.68)
Cohort Studies Mortality ICD-10: C90 18.5–24.9 12 1.00 (0.70–1.43)
Collaboration 25–29.9 19 0.87 (0.54–1.41)
1961–1999 ≥ 30 25 1.20 (0.59–2.43)
Average follow- per 5 kg/m2 10 1.05 (0.73–1.50)
up 4 yr [Ptrend] [0.78]
Troy et al. (2010) 142 982 Plasma cell BMI at baseline Age, race/
PLCO Trial Men and women myeloma < 18.5 2 – ethnicity,
USA Incidence 18.5–24.9 57 1.00 education level
1993–2006 25–29.9 112 1.45 (1.05–2.01)

≥ 30 66 1.69 (1.18–2.41)
[Ptrend] [< 0.01]
BMI at age 20 yr
< 18.5 12 0.71 (0.40–1.29)
18.5–24.9 173 1.00
25–29.9 41 1.33 (0.94–1.88)
≥ 30 12 3.08 (1.71–5.54)
[Ptrend] [< 0.001]
441
442

Location Sex
mortality
Troy et al. (2010) Weight change (kg) per 10 yr
(cont.) Loss 16 1.00 (0.56–1.78)
Gain 0–2 52 1.00
Gain 2.1–4 78 1.40 (0.98–2.00)
Gain 4.1–6 51 1.48 (1.00–2.20)
Gain > 6 42 1.55 (1.02–2.36)
[Ptrend] [0.216]
Wallin & Larsson 15 studies Multiple BMI NR
(2011) Men and women myeloma Overweight 1.12 (1.07–1.18)
Meta-analysis Incidence ICD-O-3: Obesity 1.21 (1.08–1.35)
Multiple locations 9732/3 per 5 kg/m2 1.12 (1.08–1.16)
5 studies Multiple BMI NR
Men and women myeloma Overweight 1.15 (1.05–1.27)
Mortality ICD-O-3: Obesity 1.54 (1.35–1.76)
9732/3 per 5 kg/m2 1.21 (1.13–1.30)
Hofmann et al. 305 618 Multiple BMI at baseline Age, sex, race Analyses also for
(2013) Men and women myeloma < 18.5 1 0.30 (0.04–2.17) women and men
NIH-AARP Incidence ICD-O-3: 18.5–22.49 53 1.0 separately
cohort 9732 22.5–24.9 99 1.02 (0.73–1.43)
USA 25–29.9 207 1.09 (0.80–1.48)
1995–1996 30–34.9 82 1.26 (0.89–1.78)
≥ 35 34 1.55 (1.01–2.39)
[Ptrend] [0.008]
BMI at age 50 yr
< 18.5 3 0.78 (0.25–2.49)
18.5–22.49 73 1.00
22.5–24.9 129 1.14 (0.85–1.52)
25–29.9 193 1.16 (0.88–1.54)
30–34.9 45 1.23 (0.84–1.80)
≥ 35 18 1.77 (1.05–2.99)
[Ptrend] [0.04]
Location Sex
mortality
Hofmann et al. BMI at age 35 yr
(2013) < 18.5 7 0.77 (0.36–1.66)
(cont.) 18.5–22.49 136 1.00
22.5–24.9 159 1.42 (1.12–1.79)
25–29.9 131 1.27 (0.99–1.63)
30–34.9 22 1.41 (0.89–2.22)
≥ 35 8 2.53 (1.24–5.18)
[Ptrend] [0.004]
BMI at age 18 yr
< 18.5 55 0.93 (0.69–1.25)
18.5–22.49 237 1.00
22.5–24.9 86 1.12 (0.88–1.44)
≥ 25 64 1.38 (1.04–1.82)
[Ptrend] [0.015]
Patel et al. (2013) 152 423 Multiple BMI at baseline Age, sex, family
Cancer Men and women myeloma < 18.5 1 0.32 (0.04–2.30) history of
Prevention Study Incidence 18.5– < 25 144 1.00 haematopoietic
II Nutrition 25– < 30 149 1.00 (0.79–1.26) cancer, education
Cohort ≥ 30 58 1.17 (0.86–1.60) level, smoking
USA [Ptrend] [0.25] status, physical
1992–2007 activity, alcohol
consumption

BMI at age 18 yr
< 18.5 44 0.89 (0.64–1.24)
18.5–< 22.5 197 1.00
22.5–< 25 66 1.01 (0.75–1.34)
25– < 30 31 0.92 (0.61–1.37)
≥ 30 7 1.77 (0.82–3.84)
[Ptrend] [0.37]
443
444

Location Sex
mortality
Patel et al. (2013) Adult weight change (lb)
(cont.) Loss > 5 10 0.77 (0.40–1.47)
Loss 5 to gain 20 105 1.00
Gain 21–40 133 1.25 (0.96–1.61)
Gain 41–60 68 1.08 (0.79–1.47)
Gain > 60 28 0.81 (0.53–1.24)
[Ptrend] [0.85]
Bhaskaran et al. 5 243 978 Multiple BMI 2969 Age, sex,
(2014) Men and women myeloma per 5 kg/m2 1.03 (0.98–1.09) diabetes,
Clinical Practice Incidence ICD-10: C90 [Ptrend] [0.15] smoking, alcohol
Research consumption,
Datalink SES, calendar
United Kingdom year
1987–2012
Teras et al. (2014) 1 564 218 Multiple BMI at baseline Race, sex,
Pooled analysis of Men and women myeloma 15.0–18.4 15 1.21 (0.71–2.06) education
20 cohorts Mortality ICD-9: 203; 18.5−20.9 85 1.02 (0.79–1.32) level, marital
Multiple locations ICD-10: C90 21.0−22.9 171 1.00 status, alcohol
1970–2002 23.0–24.9 302 1.22 (1.01–1.47) consumption,
27.5–29.9 215 1.24 (1.01–1.52) smoking
30.0–34.9 178 1.23 (0.99–1.52)
≥ 35 71 1.52 (1.15–2.02)
per 5 kg/m2 1.09 (1.03–1.16)
Young adult BMI
15.0–18.4 121 0.99 (0.80–1.23)
18.5−20.9 319 0.91 (0.78–1.07)
21.0−22.9 275 1.00
23.0–24.9 160 1.04 (0.85–1.26)
25.0–27.4 92 1.11 (0.87–1.40)
27.5–29.9 31 1.49 (1.03–2.16)
≥ 30.0 26 1.82 (1.22–2.73)
per 5 kg/m2 1.22 (1.09–1.35)
Location Sex
mortality
Teras et al. (2014) BMI gain
(cont.) ≤ −2.5 34 1.12 (0.77–1.64)
−2.5 to < 0 67 0.84 (0.64–1.10)
0–2.5 221 1.00
2.5–4.9 266 1.04 (0.87–1.24)
5.0–7.4 220 1.17 (0.96–1.41)
7.5–9.9 113 1.10 (0.87–1.38)
≥ 10 103 1.17 (0.92–1.50)
per 1 kg/m2 1.06 (0.98–1.14)
647 478 WC (cm), quartiles (sex-specific) Race, sex, Also provided
Men and women Men: Women: education estimates for waist-
Mortality < 90 < 70 112 1.00 level, marital to-hip ratio and
90–99 70–79 216 1.28 (1.01–1.62) status, alcohol height
100–109 80–89 153 1.32 (1.02–1.71) consumption,
≥ 110 ≥ 90 108 1.47 (1.10–1.96) physical activity,
per 5 cm 1.06 (1.02–1.10) smoking
656 771 BMI at baseline Race, education Also provided
Men 15.0–18.4 1 – level, marital estimates for waist-
Mortality 18.5−20.9 17 0.97 (0.57–1.67) status, alcohol to-hip ratio and
21.0−22.9 63 1.00 consumption, height
25.0–27.4 219 1.20 (0.90–1.59) smoking

27.5–29.0 130 1.29 (0.95–1.75)
30.0–34.9 93 1.28 (0.93–1.78)
≥ 35 24 1.48 (0.91–2.38)
per 5 kg/m2 1.11 (1.00–1.22)
Young adult BMI
15.0–18.4 40 0.85 (0.60–1.21)
18.5−20.9 136 0.91 (0.73–1.15)
21.0−22.9 155 1.00
23.0–24.9 92 0.88 (0.68–1.14)
25.0–27.4 62 1.00 (0.74–1.34)
27.5–29.0 21 1.47 (0.93–2.32)
≥ 30.0 10 1.36 (0.72–2.59)
per 5 kg/m2 1.15 (0.98–1.35)
445
446

Location Sex
mortality
(cont.) ≤ −2.5 11 1.04 (0.55–1.97)
−2.5 to < 0 33 0.96 (0.65–1.42)
0–2.5 117 1.00
2.5–4.9 147 1.04 (0.81–1.33)
5 0–7.4 108 1.05 (0.80–1.38)
7.5–9.9 60 1.18 (0.85–1.64)
≥ 10 40 1.20 (0.82–1.76)
per 1 kg/m2 1.07 (0.94–1.21)
WC (cm)
< 90 62 1.00
90–99 144 1.25 (0.93–1.69)
100–109 83 1.26 (0.90–1.77)
≥ 110 38 1.38 (0.91–2.08)
per 5 cm 1.06 (1.01–1.12)
907 447 BMI at baseline Race, education Also provided
Women 15.0–18.4 14 1.39 (0.79–2.43) level, marital estimates for waist-
Mortality 18.5−20.9 68 1.01 (0.75–1.38) status, alcohol to-hip ratio and
21.0−22.9 108 1.00 consumption, height
25.0–27.4 132 1.11 (0.86–1.44) smoking
27.5–29.0 85 1.20 (0.90–1.60)
30.0–34.9 85 1.18 (0.89–1.58)
≥ 35 47 1.51 (1.06–2.15)
per 5 kg/m2 1.07 (0.99–1.16)
Young adult BMI
15.0–18.4 81 1.11 (0.84–1.47)
18.5−20.9 183 0.94 (0.75–1.19)
21.0−22.9 120 1.00
23.0–24.9 68 1.31 (0.97–1.76)
25.0–27.4 30 1.28 (0.86–1.91)
27.5–29.0 10 1.42 (0.75–2.71)
≥ 30.0 16 2.32 (1.37–3.92)
per 5 kg/m2 1.27 (1.10–1.47)
Location Sex
mortality
(cont.) ≤ −2.5 23 1.16 (0.72–1.89)
−2.5 to < 0 34 0.75 (0.51–1.10)
0–2.5 104 1.00
2.5–4.9 119 1.02 (0.78–1.33)
5.0–7.4 112 1.28 (0.98–1.68)
7.5–9.9 53 1.00 (0.71–1.40)
≥ 10 63 1.12 (0.81–1.56)
per 1 kg/m2 1.04 (0.95–1.15)
WC (cm)
< 70 50 1.00
70–79 72 1.32 (0.90–1.94)
80–89 70 1.42 (0.94–2.13)
≥ 90 70 1.54 (1.00–2.36)
per 5 cm 1.05 (1.00–1.11)
BMI, body mass index (in kg/m 2); CI, confidence interval; DLBCL, diffuse large B-cell lymphoma; EPIC, European Prospective Investigation into Cancer and Nutrition; ICD,
International Classification of Diseases; ICD-O, International Classification of Diseases for Oncology; NIH-AARP, National Institutes of Health–AARP Diet and Health Study; NR, not
reported; PLCO Trial, Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial; RR, relative risk; SES, socioeconomic status; WC, waist circumference; yr, year or years

447
448

Table 2.2.20d Case–control studies of measures of body fatness and haematopoietic malignancies of lymphoid origin with
sufficient or limited evidence
Study location Total number of controls (95% CI)
Diffuse large B-cell lymphoma
Pan et al. (2005) 419 from National Adult BMI 2 yr before interview/diagnosis Age, province, sex, education level,
Canada Enhanced Cancer 18.5– < 25 162 1.00 pack-years of smoking, alcohol
1994–1997 Surveillance System 25– < 30 184 1.37 (1.09–1.73) consumption, exposure to some
3106 ≥ 30 69 1.35 (0.99–1.83) chemicals, occupational exposures,
Population [Ptrend] [0.015] physical activity, energy intake
Chen et al. (2011) 245 Usual adult BMI assessed via interview Age, race, total energy intake
USA 868 < 25 77 1.0
1996–2000 Population 25–30 56 1.5 (1.0–2.2)
> 30 28 1.1 (0.7–1.8)
Cerhan et al. (2014) 4667 Young adult BMI
Pooled analysis 22 639 15– < 18.5 64 0.93 (0.69–1.24)
from InterLymph 18.5– < 22.5 517 1.00
Consortium of 19 case– 22.5– < 25 276 1.11 (0.93–1.31)
control studies 25– < 30 226 1.47 (1.22–1.77)
Europe, Japan, North 30–50 54 1.58 (1.12–2.23)
America [Ptrend] [0.002]
Usual adult BMI
15– < 18.5 33 0.58 (0.39–0.85)
18.5– < 22.5 722 1.00
22.5– < 25 850 0.91 (0.81–1.03)
25– < 30 1310 0.93 (0.83–1.04)
30– < 35 419 0.95 (0.82–1.10)
35–50 175 1.06 (0.86–1.30)
[Ptrend] [0.042]
Multiple myeloma
Larsson & Wolk (2007b) 1166 total BMI Note: the reference category was ≤ 25
Meta-analysis of 4 case– 8247 total ≤ 25 1.00 in all but 3 studies
control studies 25–29.9 1.43 (1.23–1.68)
Studies published in ≥ 30 1.82 (1.47–2.26)
1994–2007
Table 2.2.20d (continued)
Study location Total number of controls (95% CI)
Wang et al. (2013) 278 from Los Angeles Self-reported BMI 1 yr before cancer diagnosis or at time of Sex, age ± 5 yr, race
USA County Multiple Myeloma interview
1985–1992 Case–Control Study All:
278 < 25 116 1.00
Population 25–29.9 98 0.75 (0.51–1.10)
30–34.9 43 0.98 (0.59–1.62)
≥ 35 21 1.86 (0.84–4.14)
Men:
< 25 58 1.00
25–29.9 65 0.85 (0.52–1.39)
30–34.9 19 0.96 (0.46–2.01)
≥ 35 8 1.80 (0.51–6.30)
Women:
< 25 58 1.00
25–29.9 33 0.62 (0.34–1.17)
30–34.9 24 0.92 (0.45–1.88)
≥ 35 11 1.56 (0.55–4.40)
BMI, body mass index (in kg/m 2); CI, confidence interval; yr, year or years

449
References 16(10):1203–14. doi:10.10 07/s10552- 0 05- 0358-7

PMID:16215871
Cerhan JR, Kricker A, Paltiel O, Flowers CR, Wang SS,
Andreotti G, Hou L, Beane Freeman LE, Mahajan R, Monnereau A, et al. (2014). Medical history, lifestyle,
Koutros S, Coble J, et al. (2010). Body mass index, family history, and occupational risk factors for diffuse
agricultural pesticide use, and cancer incidence in large B-cell lymphoma: the InterLymph Non-Hodgkin
the Agricultural Health Study cohort. Cancer Causes Lymphoma Subtypes Project. J Natl Cancer Inst
Control, 21(11):1759–75. doi:10.1007/s10552-010-9603-9 Monogr, 2014(48):15–25. doi:10.1093/jncimonographs/
PMID:20730623 lgu010 PMID:25174023
Bertrand KA, Giovannucci E, Zhang SM, Laden F, Rosner Chang ET, Hjalgrim H, Smedby KE, Akerman M, Tani
B, Birmann BM (2013). A prospective analysis of body E, Johnsen HE, et al. (2005). Body mass index and risk
size during childhood, adolescence, and adulthood of malignant lymphoma in Scandinavian men and
and risk of non-Hodgkin lymphoma. Cancer Prev Res women. J Natl Cancer Inst, 97(3):210–8. doi:10.1093/
(Phila), 6(8):864–73. doi:10.1158/1940-6207.CAPR-13- jnci/dji012 PMID:15687364
0132 PMID:23803416 Chen Y, Zheng T, Lan Q, Foss F, Kim C, Chen X, et al.
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, (2011). Cytokine polymorphisms in Th1/Th2 pathway
Leon DA, Smeeth L (2014). Body-mass index and risk genes, body mass index, and risk of non-Hodgkin
of 22 specific cancers: a population-based cohort study lymphoma. Blood, 117(2):585–90. doi:10.1182/blood-
of 5.24 million UK adults. Lancet, 384(9945):755–65. 2010-07-295097 PMID:20952689
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 Chiu BC, Gapstur SM, Greenland P, Wang R, Dyer A
Birmann BM, Giovannucci E, Rosner B, Anderson KC, (2006). Body mass index, abnormal glucose metab-
Colditz GA (2007). Body mass index, physical activity, olism, and mortality from hematopoietic cancer.
and risk of multiple myeloma. Cancer Epidemiol Cancer Epidemiol Biomarkers Prev, 15(12):2348–54.
Biomarkers Prev, 16(7):1474–8. doi:10.1158/1055-9965. doi:10.1158/1055-9965.EPI-06-0007 PMID:17164355
EPI-07-0143 PMID:17627013 Chu DM, Wahlqvist ML, Lee MS, Chang HY (2011).
Blair CK, Cerhan JR, Folsom AR, Ross JA (2005). Central obesity predicts non-Hodgkin’s lymphoma
Anthropometric characteristics and risk of multiple mortality and overall obesity predicts leukemia
myeloma. Epidemiology, 16(5):691–4. doi:10.1097/01. mortality in adult Taiwanese. J Am Coll Nutr,
ede.0000172135.61188.2d PMID:16135948 30(5):310–9. doi:10.1080/07315724.2011.10719974
Bosetti C, Dal Maso L, Negri E, Talamini R, Montella M, PMID:22081617
Franceschi S, et al. (2005). Re: Body mass index and De Roos AJ, Ulrich CM, Ray RM, Mossavar-Rahmani Y,
risk of malignant lymphoma in Scandinavian men and Rosenberg CA, Caan BJ, et al. (2010). Intentional weight
women. J Natl Cancer Inst, 97(11):860–1. doi:10.1093/ loss and risk of lymphohematopoietic cancers. Cancer
jnci/dji150 PMID:15928310 Causes Control, 21(2):223–36. doi:10.1007/s10552-009-
Britton JA, Khan AE, Rohrmann S, Becker N, Linseisen J, 9453-5 PMID:19851877
Nieters A, et al. (2008). Anthropometric characteristics Engeland A, Tretli S, Hansen S, Bjørge T (2007). Height
and non-Hodgkin’s lymphoma and multiple myeloma and body mass index and risk of lymphohematopoi-
risk in the European Prospective Investigation etic malignancies in two million Norwegian men and
into Cancer and Nutrition (EPIC). Haematologica, women. Am J Epidemiol, 165(1):44–52. doi:10.1093/aje/
93(11):1666–77. doi :10 . 332 4 / h a em atol .13 0 78 kwj353 PMID:17041129
PMID:18835833 Fernberg P, Odenbro A, Bellocco R, Boffetta P, Pawitan
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ Y, Zendehdel K, et al. (2007). Tobacco use, body mass
(2003). Overweight, obesity, and mortality from cancer index, and the risk of leukemia and multiple myeloma:
in a prospectively studied cohort of U.S. adults. N Engl a nationwide cohort study in Sweden. Cancer Res,
J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423 67(12):5983–6. doi:10.1158/0008-5472.CAN-07-0274
PMID:12711737 PMID:17575169
Castillo JJ, Ingham RR, Reagan JL, Furman M, Dalia S, Fujino Y; Japan Collaborative Cohort Study for Evaluation
Mitri J (2014). Obesity is associated with increased rela- of Cancer (2007). Anthropometry, development history
tive risk of diffuse large B-cell lymphoma: a meta-ana- and mortality in the Japan Collaborative Cohort Study
lysis of observational studies. Clin Lymphoma Myeloma for Evaluation of Cancer (JACC). Asian Pac J Cancer
Leuk, 14(2):122–30. doi:10.1016/j.clml.2013.10.005 Prev, 8(Suppl):105–12. PMID:18260709
PMID:24360912 Hemminki K, Li X, Sundquist J, Sundquist K (2011).
Cerhan JR, Bernstein L, Severson RK, Davis S, Colt JS, Obesity and familial obesity and risk of cancer.
Blair A, et al. (2005). Anthropometrics, physical Eur J Cancer Prev, 20(5):438–43. doi:10.1097/
activity, related medical conditions, and the risk of CEJ.0b013e32834761c0 PMID:21606843
non-Hodgkin lymphoma. Cancer Causes Control,
450
Hofmann JN, Moore SC, Lim U, Park Y, Baris D, Lim U, Morton LM, Subar AF, Baris D, Stolzenberg-
Hollenbeck AR, et al. (2013). Body mass index and Solomon R, Leitzmann M, et al. (2007). Alcohol,
physical activity at different ages and risk of multiple smoking, and body size in relation to incident
myeloma in the NIH-AARP diet and health study. Am Hodgkin’s and non-Hodgkin’s lymphoma risk. Am J
J Epidemiol, 177(8):776–86. doi:10.1093/aje/kws295 Epidemiol, 166(6):697–708. doi:10.1093/aje/kwm122
PMID:23543160 PMID:17596266
IARC (2002). Weight control and physical activity. Lyon, Linet MS, Vajdic CM, Morton LM, de Roos AJ, Skibola
France: IARC Press (IARC Handbooks of Cancer CF, Boffetta P, et al. (2014). Medical history, lifestyle,
Prevention, Vol. 6). Available from: http://publications. family history, and occupational risk factors for folli-
iarc.fr/376. cular lymphoma: the InterLymph Non-Hodgkin
Jaffe ES, Harris NL, Stein H, Vardiman JW, editors Lymphoma Subtypes Project. J Natl Cancer Inst
(2001). Pathology and genetics of tumours of haemato- Monogr, 2014(48):26–40. doi:10.1093/jncimonographs/
poietic and lymphoid tissues. 3rd ed. Lyon, France: lgu006 PMID:25174024
International Agency for Research on Cancer. Lu Y, Prescott J, Sullivan-Halley J, Henderson KD, Ma H,
Kabat GC, Kim MY, Jean-Wactawski-Wende, Bea JW, Chang ET, et al. (2009). Body size, recreational phys-
Edlefsen KL, Adams-Campbell LL, et al. (2012). ical activity, and B-cell non-Hodgkin lymphoma risk
Anthropometric factors, physical activity, and risk among women in the California Teachers Study. Am
of non-Hodgkin’s lymphoma in the Women’s Health J Epidemiol, 170(10):1231–40. doi:10.1093/aje/kwp268
Initiative. Cancer Epidemiol, 36(1):52–9. doi:10.1016/j. PMID:19822569
canep.2011.05.014 PMID:21816698 Lu Y, Sullivan-Halley J, Henderson KD, Ma H, Horn-Ross
Kanda J, Matsuo K, Inoue M, Iwasaki M, Sawada PL, Reynolds P, et al. (2010). Anthropometric charac-
N, Shimazu T, et al.; Japan Public Health Center- teristics and multiple myeloma risk. Epidemiology,
based Prospective Study Group (2010). Association 21(2):272–3. doi:10.1097/EDE.0b013e3181cc9241
of anthropometric characteristics with the risk of PMID:20160567
malignant lymphoma and plasma cell myeloma in a Lukanova A, Björ O, Kaaks R, Lenner P, Lindahl B,
Japanese population: a population-based cohort study. Hallmans G, et al. (2006). Body mass index and cancer:
Cancer Epidemiol Biomarkers Prev, 19(6):1623–31. results from the Northern Sweden Health and Disease
doi:10.1158/1055-9965.EPI-10-0171 PMID:20501768 Cohort. Int J Cancer, 118(2):458–66. doi:10.1002/
Kelly JL, Fredericksen ZS, Liebow M, Shanafelt TD, ijc.21354 PMID:16049963
Thompson CA, Call TG, et al. (2012). The association Maskarinec G, Erber E, Gill J, Cozen W, Kolonel LN
between early life and adult body mass index and (2008). Overweight and obesity at different times in
physical activity with risk of non-Hodgkin lymphoma: life as risk factors for non-Hodgkin’s lymphoma: the
impact of gender. Ann Epidemiol, 22(12):855–62. Multiethnic Cohort. Cancer Epidemiol Biomarkers
doi:10.1016/j.annepidem.2012.10.002 PMID:23146413 Prev, 17(1):196–203. doi:10.1158/1055-9965.EPI-07-
Larsson SC, Wolk A (2007a). Obesity and risk of non-Hodg- 0716 PMID:18187389
kin’s lymphoma: a meta-analysis. Int J Cancer, Morton LM, Wang SS, Cozen W, Linet MS, Chatterjee
121(7):1564–70. doi:10.1002/ijc.22762 PMID:17443495 N, Davis S, et al. (2008). Etiologic heterogeneity
Larsson SC, Wolk A (2007b). Body mass index and risk among non-Hodgkin lymphoma subtypes. Blood,
of multiple myeloma: a meta-analysis. Int J Cancer, 112(13):5150–60. doi:10.1182/blood-2008-01-133587
121(11):2512–6. doi:10.1002/ijc.22968 PMID:17680557 PMID:18796628
Larsson SC, Wolk A (2008). Overweight and obesity Oh SW, Yoon YS, Shin SA (2005). Effects of excess
and incidence of leukemia: a meta-analysis of cohort weight on cancer incidences depending on cancer
studies. Int J Cancer, 122(6):1418–21. doi:10.1002/ sites and histologic findings among men: Korea
ijc.23176 PMID:18027857 National Health Insurance Corporation Study. J Clin
Larsson SC, Wolk A (2011). Body mass index and risk of Oncol, 23(21):4742–54. doi:10.1200/JCO.2005.11.726
non-Hodgkin’s and Hodgkin’s lymphoma: a meta-ana- PMID:16034050
lysis of prospective studies. Eur J Cancer, 47(16):2422– Pan SY, Johnson KC, Ugnat AM, Wen SW, Mao Y;
30. doi:10.1016/j.ejca.2011.06.029 PMID:21733676 Canadian Cancer Registries Epidemiology Research
Li Q, Chang ET, Bassig BA, Dai M, Qin Q, Gao Y, et al. Group (2004). Association of obesity and cancer risk
(2013). Body size and risk of Hodgkin’s lymphoma by in Canada. Am J Epidemiol, 159(3):259–68. doi:10.1093/
age and gender: a population-based case-control study aje/kwh041 PMID:14742286
in Connecticut and Massachusetts. Cancer Causes Pan SY, Mao Y, Ugnat AM; Canadian Cancer Registries
Control, 24(2):287–95. doi:10.1007/s10552-012-0100-1 Epidemiology Research Group (2005). Physical activity,
PMID:23208661 obesity, energy intake, and the risk of non-Hodgkin’s
lymphoma: a population-based case-control study.
451
Am J Epidemiol, 162(12):1162–73. doi:10.1093/aje/ and leptin receptor polymorphisms, and non-Hodgkin

kwi342 PMID:16269580 lymphoma. Cancer Epidemiol Biomarkers Prev,
Parr CL, Batty GD, Lam TH, Barzi F, Fang X, Ho SC 13(5):779–86. PMID:15159310
et al.; Asia-Pacific Cohort Studies Collaboration (2010). Song Y-M, Sung J, Ha M (2008). Obesity and risk of
Body-mass index and cancer mortality in the Asia- cancer in postmenopausal Korean women. J Clin
Pacific Cohort Studies Collaboration: pooled analyses Oncol, 26(20):3395–402. doi:10.1200/JCO.2007.15.7867
of 424,519 participants. Lancet Oncol, 11(8):741–52. PMID:18612154
doi:10.1016/S1470-2045(10)70141-8 PMID:20594911 Swerdlow SH, Campo E, Harris NL, Jaffe ES, Pileri SA,
Patel AV, Diver WR, Teras LR, Birmann BM, Gapstur SM Stein H, et al., editors (2008). WHO classification of
(2013). Body mass index, height and risk of lymphoid tumours of haematopoietic and lymphoid tissues. 4th
neoplasms in a large United States cohort. Leuk ed. Lyon, France: International Agency for Research on
Lymphoma, 54(6):1221–7. doi:10.3109/10428194.2012.7 Cancer.
42523 PMID:23098244 Teras LR, Kitahara CM, Birmann BM, Hartge PA, Wang
Pylypchuk RD, Schouten LJ, Goldbohm RA, Schouten SS, Robien K, et al. (2014). Body size and multiple
HC, van den Brandt PA (2009). Body mass index, myeloma mortality: a pooled analysis of 20 prospec-
height, and risk of lymphatic malignancies: a prospective studies. Br J Haematol, 166(5):667–76. doi:10.1111/
tive cohort study. Am J Epidemiol, 170(3):297–307. bjh.12935 PMID:24861847
doi:10.1093/aje/kwp123 PMID:19478235 Troy JD, Hartge P, Weissfeld JL, Oken MM, Colditz
Rapp K, Schroeder J, Klenk J, Stoehr S, Ulmer H, Concin GA, Mechanic LE, et al. (2010). Associations between
H, et al. (2005). Obesity and incidence of cancer: a anthropometry, cigarette smoking, alcohol consump-
large cohort study of over 145,000 adults in Austria. tion, and non-Hodgkin lymphoma in the Prostate,
Br J Cancer, 93(9):1062–7. doi:10.1038/sj.bjc.6602819 Lung, Colorectal, and Ovarian Cancer Screening
PMID:16234822 Trial. Am J Epidemiol, 171(12):1270–81. doi:10.1093/aje/
Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D; kwq085 PMID:20494998
Million Women Study Collaboration (2007). Cancer Wallin A, Larsson SC (2011). Body mass index and risk
incidence and mortality in relation to body mass of multiple myeloma: a meta-analysis of prospective
index in the Million Women Study: cohort study. BMJ, studies. Eur J Cancer, 47(11):1606–15. doi:10.1016/j.
335(7630):1134. doi:10.1136/bmj.39367.495995.AE ejca.2011.01.020 PMID:21354783
PMID:17986716 Wang SS, Voutsinas J, Chang ET, Clarke CA, Lu Y, Ma H,
Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen et al. (2013). Anthropometric, behavioral, and female
M (2008). Body-mass index and incidence of cancer: reproductive factors and risk of multiple myeloma: a
a systematic review and meta-analysis of prospec- pooled analysis. Cancer Causes Control, 24(7):1279–89.
tive observational studies. Lancet, 371(9612):569–78. doi:10.1007/s10552-013-0206-0 PMID:23568533
doi:10.1016/S0140-6736(08)60269-X PMID:18280327 Whitlock G, Lewington S, Sherliker P, Clarke R, Emberson
Ross JA, Parker E, Blair CK, Cerhan JR, Folsom AR (2004). J, Halsey J, et al.; Prospective Studies Collaboration
Body mass index and risk of leukemia in older women. (2009). Body-mass index and cause-specific mortality
Cancer Epidemiol Biomarkers Prev, 13(11 Pt 1):1810–3. in 900 000 adults: collaborative analyses of 57 prospec-
PMID:15533912 tive studies. Lancet, 373(9669):1083–96. doi:10.1016/
Saberi Hosnijeh F, Romieu I, Gallo V, Riboli E, Tjønneland S0140-6736(09)60318-4 PMID:19299006
A, Halkjær J, et al. (2013). Anthropometric characteris- WHO (2013). International Classification of Diseases for
tics and risk of lymphoid and myeloid leukemia in the Oncology. 3rd ed., 1st revision. Geneva, Switzerland:
European Prospective Investigation into Cancer and World Health Organization.
Nutrition (EPIC). Cancer Causes Control, 24(3):427–38. Willett EV, Morton LM, Hartge P, Becker N, Bernstein
doi:10.1007/s10552-012-0128-2 PMID:23288400 L, Boffetta P, et al.; InterLymph Consortium (2008).
Samanic C, Chow WH, Gridley G, Jarvholm B, Non-Hodgkin lymphoma and obesity: a pooled
Fraumeni JF Jr (2006). Relation of body mass index analysis from the InterLymph Consortium. Int J Cancer,
to cancer risk in 362,552 Swedish men. Cancer Causes 122(9):2062–70. doi:10.1002/ijc.23344 PMID:18167059
Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9 Willett EV, Roman E (2006). Obesity and the risk of
PMID:16841257 Hodgkin lymphoma (United Kingdom). Cancer Causes
Samanic C, Gridley G, Chow WH, Lubin J, Hoover Control, 17(8):1103–6. doi:10.1007/s10552-006-0042-6
RN, Fraumeni JF Jr (2004). Obesity and cancer PMID:16933061
risk among white and black United States veterans. Willett EV, Skibola CF, Adamson P, Skibola DR, Morgan
Cancer Causes Control, 15(1):35–43. doi:10.1023/B:- GJ, Smith MT, et al. (2005). Non-Hodgkin’s lymphoma,
CACO.0000016573.79453.ba PMID:14970733 obesity and energy homeostasis polymorphisms.
Skibola CF, Holly EA, Forrest MS, Hubbard A, Bracci Br J Cancer, 93(7):811–6. doi:10.1038/sj.bjc.6602762
PM, Skibola DR, et al. (2004). Body mass index, leptin PMID:16160698
452
2.2.21 Other haematopoietic malignancies circumference with risk of AML in women

(Ptrend = 0.06), but not in men (Saberi Hosnijeh
(a) Myeloid leukaemia et al., 2013). In that study, there were also no asso-
(i) Cohort studies ciations of waist circumference with CML inci-
There have been only two prospective studies dence in either men or women (Saberi Hosnijeh
of BMI and/or weight in relation to total myeloid et al., 2013).
leukaemia incidence (Table 2.2.20a, web only, (ii) Case–control studies
available at: http://publications.iarc.fr/570). In
Three case–control studies have evaluated
the Japan Collaborative Cohort Study, compared
the relationship between BMI and the risk of
with BMI 18.5–24 kg/m2, BMI ≥ 30 kg/m2 was
developing various subtypes of leukaemia (Table
associated with a statistically significantly higher
2.2.20b, web only, available at: http://publications.
risk (Fujino et al., 2007). In the EPIC cohort, BMI
iarc.fr/570). In a study of 420 cases of AML from
was positively associated with risk in women
the Minnesota Cancer Surveillance System,
(Ptrend = 0.04), but no association was found in
Poynter et al. (2016) found a non-significant
men (Saberi Hosnijeh et al., 2013).
increase in risk of AML with high BMI in women
Statistically significant positive associations
only. Kasim et al. (2005) found an increased risk
between BMI and risk of AML were observed
of all leukaemia, AML, and CML in obese versus
in postmenopausal women in the USA (Ross
normal-weight individuals in a case–control
et al., 2004), in the United States Veterans cohort
study of 1068 people with leukaemia from the
(Samanic et al., 2004), and in a Norwegian cohort
Canadian Enhanced Survival Surveillance
(Engeland et al., 2007). However, there were no
System. Finally, Strom et al. (2009) found a trend
associations of BMI or weight with risk in other
towards an increased risk of CML with BMI at
studies in European men (Samanic et al., 2006;
age 25 years, at age 40 years, and at diagnosis in
Fernberg et al., 2007; Saberi Hosnijeh et al., 2013)
a case–control study of 253 cases of CML from
or women (Saberi Hosnijeh et al., 2013). Of six
MD Anderson Cancer Center in the USA.
individual prospective studies of BMI and/or
weight in relation to CML incidence (Samanic (b) Leukaemia not otherwise specified
et al., 2004, 2006; Engeland et al., 2007; Fernberg
At least six individual cohort studies found
et al., 2007; Kabat et al., 2013; Saberi Hosnijeh
no association between BMI and total leukaemia
et al., 2013), only one (Engeland et al., 2007)
incidence or mortality (Table 2.2.20a, web only,
found clear evidence of a positive association. In
available at: http://publications.iarc.fr/570; Oh
a meta-analysis of prospective studies, obesity
et al., 2005; Samanic et al., 2006; Andreotti et al.,
was associated with a statistically significant
2010; De Roos et al., 2010; Saberi Hosnijeh et al.,
52% higher risk of AML and a 26% higher risk
2013; Batty et al., 2015). Conversely, positive asso-
of CML compared with normal weight (Larsson
ciations were found in at least eight other studies,
& Wolk, 2008).
conducted in the Republic of Korea, Taiwan,
Only two studies have examined associa-
China, the United Kingdom, and the USA (Calle
tions of abdominal obesity with risk of myeloid
et al., 2003, in men only; Ross et al., 2004; Samanic
leukaemia. In the Iowa Women’s Health Study,
et al., 2004; Chiu et al., 2006; Reeves et al., 2007;
waist circumference was positively associated
Song et al., 2008; Chu et al., 2011; Bhaskaran
with risk of AML (Ptrend = 0.04) (Ross et al.,
et al., 2014). Positive associations in men and in
2004). Similarly, in the EPIC cohort, there was
women were found in a meta-analysis of seven
suggestive evidence for an association of waist
453
prospective studies (Renehan et al., 2008). In a Chiu BC, Gapstur SM, Greenland P, Wang R, Dyer A
meta-analysis of 10 studies in men and women (2006). Body mass index, abnormal glucose metab-
olism, and mortality from hematopoietic cancer.
combined, there was a 39% increased risk of Cancer Epidemiol Biomarkers Prev, 15(12):2348–54.
leukaemia incidence for obese versus normal doi:10.1158/1055-9965.EPI-06-0007 PMID:17164355
BMI (Larsson & Wolk, 2008). Similarly, in the Chu DM, Wahlqvist ML, Lee MS, Chang HY (2011).
Central obesity predicts non-Hodgkin’s lymphoma
Asia-Pacific Cohort Studies Collaboration, mortality and overall obesity predicts leukemia
there was a positive association between BMI mortality in adult Taiwanese. J Am Coll Nutr,
and leukaemia mortality (Parr et al., 2010). 30(5):310–9. doi:10.1080/07315724.2011.10719974
However, a pooled analysis of almost 1 million PMID:22081617
De Roos AJ, Ulrich CM, Ray RM, Mossavar-Rahmani Y,
people found no association between BMI and Rosenberg CA, Caan BJ, et al. (2010). Intentional weight
leukaemia mortality (Whitlock et al., 2009). loss and risk of lymphohematopoietic cancers. Cancer
Although waist circumference was not asso- Causes Control, 21(2):223–36. doi:10.1007/s10552-009-
ciated with total leukaemia incidence in the Iowa 9453-5 PMID:19851877
Engeland A, Tretli S, Hansen S, Bjørge T (2007). Height
Women’s Health Study (Ross et al., 2004) or in and body mass index and risk of lymphohematopoi-
the EPIC cohort (Saberi Hosnijeh et al., 2013), in etic malignancies in two million Norwegian men and
the MJ Health Screening Center study, in Taiwan, women. Am J Epidemiol, 165(1):44–52. doi:10.1093/aje/
kwj353 PMID:17041129
China, abdominal obesity (waist circumference Fernberg P, Odenbro A, Bellocco R, Boffetta P, Pawitan
of ≥ 90 cm in men and ≥ 80 cm in women) was Y, Zendehdel K, et al. (2007). Tobacco use, body mass
associated with an 87% higher risk of death from index, and the risk of leukemia and multiple myeloma:
leukaemia compared with lower waist circum- a nationwide cohort study in Sweden. Cancer Res,
67(12):5983–6. doi:10.1158/0008-5472.CAN-07-0274
ference (Chu et al., 2011). PMID:17575169
Fujino Y; Japan Collaborative Cohort Study for Evaluation
of Cancer (2007). Anthropometry, development history
and mortality in the Japan Collaborative Cohort Study
References for Evaluation of Cancer (JACC). Asian Pac J Cancer
Prev, 8(Suppl):105–12. PMID:18260709
Andreotti G, Hou L, Beane Freeman LE, Mahajan R, Kabat GC, Wu JW, Moore SC, Morton LM, Park Y,
Koutros S, Coble J, et al. (2010). Body mass index, Hollenbeck AR, et a l. (2013). Lifestyle and dietary factors
agricultural pesticide use, and cancer incidence in in relation to risk of chronic myeloid leukemia in the
the Agricultural Health Study cohort. Cancer Causes NIH-AARP Diet and Health Study. Cancer Epidemiol
Control, 21(11):1759–75. doi:10.1007/s10552-010-9603-9 Biomarkers Prev, 22(5):848–54. doi:10.1158/1055-9965.
PMID:20730623 EPI-13-0093 PMID:23625904
Batty GD, Calvin CM, Brett CE, Čukić I, Deary IJ (2015). Kasim K, Levallois P, Abdous B, Auger P, Johnson KC
Childhood body weight in relation to morbidity from (2005). Lifestyle factors and the risk of adult leukemia
cardiovascular disease and cancer in older adulthood: in Canada. Cancer Causes Control, 16(5):489–500.
67-year follow-up of participants in the 1947 Scottish doi:10.1007/s10552-004-7115-1 PMID:15986104
Mental Survey. Am J Epidemiol, 182(9):775–80. Larsson SC, Wolk A (2008). Overweight and obesity
doi:10.1093/aje/kwv154 PMID:26443418 and incidence of leukemia: a meta-analysis of cohort
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, studies. Int J Cancer, 122(6):1418–21. doi:10.1002/
Leon DA, Smeeth L (2014). Body-mass index and risk ijc.23176 PMID:18027857
of 22 specific cancers: a population-based cohort study Oh SW, Yoon YS, Shin SA (2005). Effects of excess
of 5.24 million UK adults. Lancet, 384(9945):755–65. weight on cancer incidences depending on cancer
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 sites and histologic findings among men: Korea
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ National Health Insurance Corporation Study. J Clin
(2003). Overweight, obesity, and mortality from cancer Oncol, 23(21):4742–54. doi:10.1200/JCO.2005.11.726
in a prospectively studied cohort of U.S. adults. N Engl PMID:16034050
J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423 Parr CL, Batty GD, Lam TH, Barzi F, Fang X, Ho SC
PMID:12711737 et al.; Asia-Pacific Cohort Studies Collaboration (2010).
Body-mass index and cancer mortality in the Asia-
454
of 424,519 participants. Lancet Oncol, 11(8):741–52.

doi:10.1016/S1470-2045(10)70141-8 PMID:20594911
Poynter JN, Richardson M, Blair CK, Roesler MA, Hirsch
BA, Nguyen P, et al. (2016). Obesity over the life
course and risk of acute myeloid leukemia and myel-
odysplastic syndromes. Cancer Epidemiol, 40:134–40.
doi:10.1016/j.canep.2015.12.005 PMID:26720913
335(7630):1134. doi:10.1136/bmj.39367.495995.AE
PMID:17986716
Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen
M (2008). Body-mass index and incidence of cancer:
a systematic review and meta-analysis of prospec-
tive observational studies. Lancet, 371(9612):569–78.
doi:10.1016/S0140-6736(08)60269-X PMID:18280327
Ross JA, Parker E, Blair CK, Cerhan JR, Folsom AR (2004).
Body mass index and risk of leukemia in older women.
Cancer Epidemiol Biomarkers Prev, 13(11 Pt 1):1810–3.
PMID:15533912
Saberi Hosnijeh F, Romieu I, Gallo V, Riboli E, Tjønneland
A, Halkjær J, et al. (2013). Anthropometric characteris-
tics and risk of lymphoid and myeloid leukemia in the
European Prospective Investigation into Cancer and
Nutrition (EPIC). Cancer Causes Control, 24(3):427–38.
doi:10.1007/s10552-012-0128-2 PMID:23288400
Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9
PMID:16841257
CACO.0000016573.79453.ba PMID:14970733
Song Y-M, Sung J, Ha M (2008). Obesity and risk of
Oncol, 26(20):3395–402. doi:10.1200/JCO.2007.15.7867
PMID:18612154
Strom SS, Yamamura Y, Kantarijian HM, Cortes-Franco
JE (2009). Obesity, weight gain, and risk of chronic
myeloid leukemia. Cancer Epidemiol Biomarkers Prev,
18(5):1501–6. doi:10.1158/1055-9965.EPI-09-0028
PMID:19423527
S0140-6736(09)60318-4 PMID:19299006
455
2.2.22 Cancers of the head and neck et al., 2014) or in the large pooled analysis (Gaudet
et al., 2015).
Head and neck cancer refers to a group In contrast, quartiles of waist circumfer-
of cancers that develop in (i) the oral cavity; ence were positively associated with risk (RR for
(ii) the pharynx, including the nasopharynx, highest vs lowest quartile, 2.00; 95% CI. 1.24–3.23;
the oropharynx, and the hypopharynx; (iii) the Ptrend < 0.001) in the NIH-AARP study (Etemadi
larynx; (iv) the paranasal sinuses and the nasal et al., 2014). Similarly, in the large pooled analysis
cavity; and (v) the salivary glands. of 20 prospective studies, there was a 9% increase
Most head and neck cancers are squamous in risk (95% CI, 1.03–1.16) per 5 cm increase in
cell carcinomas. Because of the established asso- waist circumference (Ptrend = 0.006) (Gaudet et al.,
ciations of head and neck cancer with tobacco 2015).
use, and because BMI is inversely associated
with tobacco use, it is important that associa- (ii) Cancers of the pharynx (nasopharynx,
tions of BMI with risk of head and neck cancers oropharynx, and/or hypopharynx)
carefully consider potential confounding and/or There was no association between BMI and
effect modification by tobacco use. Notably, only risk of nasopharyngeal cancer in the only study
prospective studies with at least 50 cases for any that assessed this relationship (Samanic et al.,
specific site were included in this review. 2004). Similarly, there is no evidence that BMI
In 2001, the Working Group of the IARC is associated with risk of oropharyngeal cancer
Handbook on weight control and physical activity incidence (Gaudet et al., 2012, 2015) or mortality
(IARC, 2002) concluded that the evidence of an (Gaudet et al., 2012), or with hypopharyngeal
association between avoidance of weight gain cancer incidence (Gaudet et al., 2015). In the
and cancers of the head and neck was inadequate. NIH-AARP cohort, BMI < 18.5 kg/m2 was asso-
ciated with a higher risk of oropharyngeal and
(a) Cohort studies hypopharyngeal cancer incidence compared
See Table 2.2.22a (web only, available at: with BMI 18.5– < 25 kg/m2 (Etemadi et al. 2014).
http://publications.iarc.fr/570). [There were only three cases in the exposed
group.]
(i) Cancer of the oral cavity
Waist circumference was not associated with
The association between BMI and risk of oropharyngeal or hypopharyngeal cancer inci-
cancer of the oral cavity has been examined in two dence in the NIH-AARP cohort study (Etemadi
individual prospective studies (Bhaskaran et al., et al., 2014) or in the large pooled analysis (Gaudet
2014; Etemadi et al., 2014) and in a large pooled et al., 2015).
analysis of data from 20 prospective studies
(Gaudet et al., 2015). All of these studies adjusted (iii) Cancer of the larynx
for both tobacco use and alcohol consumption. Since 2000, there have been two individual
In the United Kingdom data linkage study of prospective studies (Samanic et al., 2004; Etemadi
more than 5 million men and women, there et al., 2014) and one large pooled analysis of 20
was a statistically significant inverse association prospective studies (Gaudet et al., 2015) of the
(RR per 5 kg/m2 increase in BMI, 0.81; 95% CI, association between BMI and risk of cancer of
0.74–0.89; Ptrend < 0.0001) (Bhaskaran et al., 2014). the larynx (Table 2.2.22a, web only, available at:
No significant association was observed in the http://publications.iarc.fr/570). In the large study
NIH-AARP cohort study in the USA (Etemadi of more than 4.5 million United States Veterans,
there was a statistically significantly lower risk
456
of laryngeal cancer for obese compared with obese was not associated with a higher incidence
non-obese White and Black men (Samanic et al., compared with being non-obese in either White
2004). [Neither tobacco use nor alcohol consump- men or Black men.
tion was included in the statistical model; there- (vi) Cancer of the head and neck or upper
fore, confounding by these factors is likely.] In the aerodigestive tract
NIH-AARP cohort study in the USA, in which
both tobacco use and alcohol consumption were For head and neck cancer incidence overall,
adjusted for in the model, BMI was not associ- in the United States Veterans study there was a
ated with risk of laryngeal cancer (Etemadi et al., significantly lower risk for obese compared with
2014). Conversely, in the pooled analysis, there non-obese Black men and White men, without
was a statistically significant positive association adjustment for tobacco use or alcohol consump-
between BMI and risk (RR per 5 kg/m2 increase, tion (Samanic et al., 2004). Most other prospec-
1.42; 95% CI, 1.19–1.70) (Gaudet et al., 2015). tive studies found a weak inverse association
In the NIH-AARP study (Etemadi et al., or no association between BMI at baseline and
2014), there was no evidence of an association incidence of head and neck cancer (Wolk et al.,
between waist circumference and risk of laryn- 2001; Gaudet et al., 2012; Hashibe et al., 2013;
geal cancer, whereas a weak positive association Etemadi et al., 2014). When the pooled analysis
was reported in the pooled analysis (RR per 5 cm of data from 20 prospective studies was stratified
increase, 1.10; 95% CI, 0.99–1.22; Ptrend = 0.08) by smoking status, BMI was positively associated
(Gaudet et al., 2015). with risk in never-smokers but was inversely
associated with risk in current smokers (Gaudet
(iv) Cancer of the oral cavity, pharynx, and et al., 2015).
larynx combined BMI was inversely associated with head and
In two studies, the Asia-Pacific Cohort neck cancer mortality (Ptrend = 3 × 10−10) in the
Studies Collaboration (Parr et al., 2010) and the Cancer Prevention Study II in the USA (Gaudet
Cancer Prevention Study II (Gaudet et al., 2012), et al., 2012), and in a smaller cohort study in
BMI was inversely associated with death from Switzerland a weaker inverse association was
cancer of the oral cavity, pharynx, and larynx found between BMI and death from cancer of
combined. In contrast, in the pooled analysis, the upper aerodigestive tract (Meyer et al., 2015).
an incremental increase in BMI of 5 kg/m2 was In the only study that examined the associ-
associated with a 36% increase in risk (Gaudet ation between BMI at younger ages and risk of
et al., 2015). Results from the Agricultural Health head and neck cancer, no association was found
Study (Andreotti et al., 2010) were inconclusive. with increased BMI at age 20 years or at age
The association between waist circumference 50 years, or with percentage change in BMI from
and the risk of cancer of the oral cavity, pharynx, age 20 years or age 50 years to baseline (Hashibe
and larynx combined was examined in the large et al., 2013).
pooled analysis of 20 prospective studies, and no Waist circumference was positively associ-
evidence of association was observed (Gaudet ated with risk of head and neck cancer incidence
et al., 2015). both in the NIH-AARP cohort study (Etemadi
et al., 2014) and in the pooled analysis of 20
(v) Cancer of the salivary glands
prospective studies, in which a 5 cm increase in
There has been only one study of the associa- waist circumference was associated with a 4%
tion between BMI and incidence of salivary gland increase in risk (95% CI, 1.03–1.05) (Gaudet et al.,
cancer (Samanic et al., 2004). In that study, being 2015).
457
(b) Case–control studies Four case–control studies stratified the

Since 2000, a total of seven independent case– analyses by smoking status. In one early study
control studies, conducted in Australia, China, in the USA, the inverse association was more
Cuba, India, Europe, Sudan, and the USA, and pronounced in current or ever-smokers than in
one large multicentre case–control study (nine never-smokers (Kabat et al., 1994). In two more
countries) have reported on the association of recent studies in the USA, a similar pattern was
BMI with various combinations of cancers of the observed in African Americans but not in Whites
head and neck (Table 2.2.22b, web only, available (Petrick et al., 2014) and in both HPV-positive
at: http://publications.iarc.fr/570). In addition, and HPV-negative individuals (Tan et al., 2015).
Gaudet at al. (2010) and Lubin et al. (2010, 2011) In contrast, the IARC Multicenter Oral Cancer
performed pooled reanalyses of 15–17 case– Study, which included a total of 1670 cases and
control studies with stratification by smoking 1732 controls from nine countries worldwide,
status, by alcohol consumption status, and by found statistically significant inverse associa-
subsite (Table 2.2.22c, web only, available at: tions of BMI (country-specific tertiles) with risk
http://publications.iarc.fr/570). of oral and oropharyngeal squamous cell carci-
In most studies, BMI was assessed on the nomas in both tobacco users and never-users, as
basis of self-reported height and body weight, well as in alcohol consumers and never-drinkers
referring to either a recent period (mostly 1 or (Kreimer et al., 2006). Similarly, a pooled reanal-
2 years) before disease diagnosis or to a period ysis of the data from 17 case–control studies,
in the more distant past (e.g. at age 30 years). which included a total of 12 716 cases and 17 438
All original studies adjusted for potential controls (INHANCE consortium; Gaudet et al.,
confounding by smoking or alcohol consump- 2010) (see Table 2.2.22c, web only, available at:
tion, in addition to variable adjustments for other http://publications.iarc.fr/570), found inverse
potential confounding factors. relationships of BMI with the risk of cancers
Most of the studies found an inverse asso- of the oral cavity, pharynx, and larynx, in men
ciation of BMI with cancer risk. In several and women combined, in ever-smokers (for
studies, compared with normal-weight individ- BMI ≥ 30 kg/m2 vs 18.5– < 25 kg/m2: OR, 0.38;
uals (18.5 kg/m2 ≤ BMI < 25 kg/m2), those who 95% CI, 0.30–0.49) but not in never-smokers.
were overweight or obese had reduced risks of Furthermore, the increase in risk in underweight
head and neck cancer (Rajkumar et al., 2003; (BMI < 18.5 kg/m2) compared with normal-
Rodriguez et al., 2004; Kreimer et al., 2006; weight (18.5– < 25 kg/m2) individuals was signif-
Peters et al., 2008; Radoï et al., 2013; Petrick icant only in the smokers (OR, 2.13; 95% CI,
et al., 2014 in African Americans only); being 1.75–2.58) (Gaudet et al., 2010).
underweight (BMI < 18.5 kg/m2) was associated
with an approximately 2-fold increase in risk in
two large-scale studies (French ICARE study,
689 cases and 3481 controls, Radoï et al., 2013;
United States CHANCE study, 1289 cases and
1361 controls, Petrick et al., 2014). In the one
study that additionally reported recalled body
weight at age 30 years (Radoï et al., 2013), this
inverse association was also observed for past
BMI.
458
References analysis from the IARC multinational case-control

study. Int J Cancer, 118(9):2293–7. doi:10.1002/ijc.21577
PMID:16331628
Andreotti G, Hou L, Beane Freeman LE, Mahajan R, Lubin JH, Gaudet MM, Olshan AF, Kelsey K, Boffetta
Koutros S, Coble J, et al. (2010). Body mass index, P, Brennan P, et al. (2010). Body mass index, ciga-
agricultural pesticide use, and cancer incidence in rette smoking, and alcohol consumption and cancers
the Agricultural Health Study cohort. Cancer Causes of the oral cavity, pharynx, and larynx: mode-
Control, 21(11):1759–75. doi:10.1007/s10552-010-9603-9 ling odds ratios in pooled case-control data. Am J
PMID:20730623 Epidemiol, 171(12):1250–61. doi:10.1093/aje/kwq088
Bhaskaran K, Douglas I, Forbes H, dos-Santos-Silva I, PMID:20494999
Leon DA, Smeeth L (2014). Body-mass index and risk Lubin JH, Muscat J, Gaudet MM, Olshan AF, Curado
of 22 specific cancers: a population-based cohort study MP, Dal Maso L, et al. (2011). An examination of male
of 5.24 million UK adults. Lancet, 384(9945):755–65. and female odds ratios by BMI, cigarette smoking, and
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328 alcohol consumption for cancers of the oral cavity,
Etemadi A, O’Doherty MG, Freedman ND, Hollenbeck pharynx, and larynx in pooled data from 15 case-con-
AR, Dawsey SM, Abnet CC (2014). A prospective trol studies. Cancer Causes Control, 22(9):1217–31.
cohort study of body size and risk of head and neck doi:10.1007/s10552-011-9792-x PMID:21744095
cancers in the NIH-AARP Diet and Health Study. Meyer J, Rohrmann S, Bopp M, Faeh D; Swiss National
Cancer Epidemiol Biomarkers Prev, 23(11):2422–9. Cohort Study Group (2015). Impact of smoking and
doi:10.1158/1055-9965.EPI-14-0709-T PMID:25172872 excess body weight on overall and site-specific cancer
Gaudet MM, Kitahara CM, Newton CC, Bernstein L, mortality risk. Cancer Epidemiol Biomarkers Prev,
Reynolds P, Weiderpass E, et al. (2015). Anthropometry 24(10):1516–22. doi:10.1158/1055-9965.EPI-15-0415
and head and neck cancer: a pooled analysis of cohort PMID:26215293
data. Int J Epidemiol, 44(2):673–81. doi:10.1093/ije/ Parr CL, Batty GD, Lam TH, Barzi F, Fang X, Ho SC,
dyv059 PMID:26050257 et al.; Asia-Pacific Cohort Studies Collaboration (2010).
Gaudet MM, Olshan AF, Chuang SC, Berthiller J, Zhang Body-mass index and cancer mortality in the Asia-
ZF, Lissowska J, et al. (2010). Body mass index and Pacific Cohort Studies Collaboration: pooled analyses
risk of head and neck cancer in a pooled analysis of of 424,519 participants. Lancet Oncol, 11(8):741–52.
case-control studies in the International Head and doi:10.1016/S1470-2045(10)70141-8 PMID:20594911
Neck Cancer Epidemiology (INHANCE) Consortium. Peters ES, Luckett BG, Applebaum KM, Marsit CJ,
Int J Epidemiol, 39(4):1091–102. doi:10.1093/ije/dyp380 McClean MD, Kelsey KT (2008). Dairy products, lean-
PMID:20123951 ness, and head and neck squamous cell carcinoma.
Gaudet MM, Patel AV, Sun J, Hildebrand JS, McCullough Head Neck, 30(9):1193–205. doi:10.1002/hed.20846
ML, Chen AY, et al. (2012). Prospective studies of PMID:18642285
body mass index with head and neck cancer incidence Petrick JL, Gaudet MM, Weissler MC, Funkhouser
and mortality. Cancer Epidemiol Biomarkers Prev, WK, Olshan AF (2014). Body mass index and risk of
21(3):497–503. doi:10.1158/1055-9965.EPI-11-0935 head and neck cancer by race: the Carolina Head and
PMID:22219317 Neck Cancer Epidemiology Study. Ann Epidemiol,
Hashibe M, Hunt J, Wei M, Buys S, Gren L, Lee YC (2013). 24(2):160–164.e1. doi:10.1016/j.annepidem.2013.11.004
Tobacco, alcohol, body mass index, physical activity, PMID:24342030
and the risk of head and neck cancer in the Prostate, Radoï L, Paget-Bailly S, Cyr D, Papadopoulos A, Guida F,
Lung, Colorectal, and Ovarian (PLCO) cohort. Tarnaud C, et al. (2013). Body mass index, body mass
Head Neck, 35(7):914–22. doi:10.1002/hed.23052 change, and risk of oral cavity cancer: results of a large
PMID:22711227 population-based case-control study, the ICARE study.
IARC (2002). Weight control and physical activity. Lyon, Cancer Causes Control, 24(7):1437–48. doi:10.1007/
France: IARC Press (IARC Handbooks of Cancer s10552-013-0223-z PMID:23677332
Prevention, Vol. 6). Available from: http://publications. Rajkumar T, Sridhar H, Balaram P, Vaccarella S,
iarc.fr/376. Gajalakshmi V, Nandakumar A, et al. (2003). Oral
Kabat GC, Chang CJ, Wynder EL (1994). The role of cancer in Southern India: the influence of body size,
tobacco, alcohol use, and body mass index in oral and diet, infections and sexual practices. Eur J Cancer Prev,
pharyngeal cancer. Int J Epidemiol, 23(6):1137–44. 12(2):135–43. doi:10.1097/00008469-200304000-00007
doi:10.1093/ije/23.6.1137 PMID:7721514 PMID:12671537
Kreimer AR, Randi G, Herrero R, Castellsagué X, La Rodriguez T, Altieri A, Chatenoud L, Gallus S, Bosetti
Vecchia C, Franceschi S; IARC Multicenter Oral C, Negri E, et al. (2004). Risk factors for oral and
Cancer Study Group (2006). Diet and body mass, and pharyngeal cancer in young adults. Oral Oncol,
oral and oropharyngeal squamous cell carcinomas:
459
40(2):207–13. doi:10.1016/j.oraloncology.2003.08.014
PMID:14693246
CACO.0000016573.79453.ba PMID:14970733
Tan X, Nelson HH, Langevin SM, McClean M, Marsit
CJ, Waterboer T, et al. (2015). Obesity and head and
neck cancer risk and survival by human papilloma-
virus serology. Cancer Causes Control, 26(1):111–9.
doi:10.1007/s10552-014-0490-3 PMID:25398682
Control, 12(1):13–21. doi:10.1023/A:1008995217664
PMID:11227921
460
2.2.23 Malignant melanoma and 0.87 (95% CI, 0.70–1.08) in women (based
on 6 studies).
Malignant melanoma is the most lethal of the Three cohorts have examined weight at
cancers of the skin. The incidence of melanoma earlier ages in relation to risk of melanoma. In
varies between countries and is related to skin both the Nurses’ Health Study and the Male
colour, with a higher risk for populations with Health Professionals Follow-Up Study, BMI at
lighter skin. Melanoma is known to be caused by 10 years before baseline was not related to risk
exposure to ultraviolet radiation in people who (Pothiawala et al., 2012); in the study of agricul-
are susceptible because of family history and/or tural workers in the USA (Dennis et al., 2008),
who have a tendency to burn easily as a result of recalled BMI at age 20 years was positively asso-
exposure to sunlight. ciated, with an estimated relative risk for BMI
In 2001, the Working Group of the IARC ≥ 25 kg/m2 of 2.55 (95% CI, 1.52–4.30).
Handbook on weight control and physical activity
(IARC, 2002) concluded that the evidence of an (b) Case–control studies
association between avoidance of weight gain
The meta-analysis by Sergentanis et al. (2013)
and malignant melanoma was inadequate.
included 10 published case–control studies that
(a) Cohort studies evaluated the association between BMI and risk
of melanoma (Table 2.2.23b, web only, available
The evidence published since 2000 includes at: http://publications.iarc.fr/570). The associ-
eight cohort studies (excluding analyses that ation between BMI and melanoma was signifi-
were later updated and analyses based on fewer cant both in overweight men and in obese men,
than 100 incident cases) (Table 2.2.23a, web although there was considerable between-study
only, available at: http://publications.iarc.fr/570) heterogeneity. No such association was observed
and one meta-analysis (Table 2.2.23b, web only, in women. When the cohort and case–control
available at: http://publications.iarc.fr/570). studies were combined, the pooled effect esti-
In most studies, there was no association mate was 1.31 (95% CI, 1.18–1.45) in overweight
between BMI and risk of melanoma (Calle men and 1.31 (95% CI, 1.19–1.44) in obese men.
et al., 2003; Rapp et al., 2005; Dennis et al., In women, no association was observed in either
2008; Pothiawala et al., 2012; Bhaskaran et al., category (Sergentanis et al., 2013). [There was
2014). However, findings by sex have not been evidence for confounding by exposure to sunlight
consistent. In two studies in men only, the in women.]
estimated relative risk for BMI ≥ 30 kg/m2 was The pooled analysis of case–control studies
1.35 (95% CI, 1.06–1.73) in Swedish construc- (Olsen et al., 2008) assessed BMI in early adult-
tion workers (Samanic et al., 2006) and 1.29 hood and weight change in relation to risk of
(95% CI, 1.14–1.46) in White men in the United melanoma in women. There was no association
States Veterans cohort (Samanic et al., 2004). In between BMI in early adulthood and mela-
the Million Women Study (Reeves et al., 2007), noma risk, but an elevated risk was associ-
the risk was also significantly increased (RR per ated with a weight gain of 2 kg or more during
10 kg/m2, 1.24; 95% CI, 1.03–1.48). In a meta-ana- adult life (pooled OR, 1.5, 95% CI, 1.1–2.0) (see
lysis of cohort studies (Sergentanis et al., 2013), Table 2.2.23b, web only, available at: http://
the estimated relative risk of obesity was 1.30 publications.iarc.fr/570).
(95% CI, 1.17–1.45) in men (based on 7 studies)
461
References
Leon DA, Smeeth L (2014). Body-mass index and risk
of 5.24 million UK adults. Lancet, 384(9945):755–65.
doi:10.1016/S0140-6736(14)60892-8 PMID:25129328
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ
(2003). Overweight, obesity, and mortality from cancer
in a prospectively studied cohort of U.S. adults. N Engl
J Med, 348(17):1625–38. doi:10.1056/NEJMoa021423
PMID:12711737
Dennis LK, Lowe JB, Lynch CF, Alavanja MC (2008).
Cutaneous melanoma and obesity in the Agricultural
Health Study. Ann Epidemiol, 18(3):214–21.
doi:10.1016/j.annepidem.2007.09.003 PMID:18280921
iarc.fr/376.
Olsen CM, Green AC, Zens MS, Stukel TA, Bataille V,
Berwick M, et al. (2008). Anthropometric factors
and risk of melanoma in women: a pooled analysis.
Int J Cancer, 122(5):1100–8. doi:10.1002/ijc.23214
PMID:17990316
Pothiawala S, Qureshi AA, Li Y, Han J (2012). Obesity and
the incidence of skin cancer in US Caucasians. Cancer
Causes Control, 23(5):717–26. doi:10.1007/s10552-012-
9941-x PMID:22450736
PMID:16234822
335(7630):1134. doi:10.1136/bmj.39367.495995.AE
PMID:17986716
Control, 17(7):901–9. doi:10.1007/s10552-006-0023-9
PMID:16841257
CACO.0000016573.79453.ba PMID:14970733
Sergentanis TN, Antoniadis AG, Gogas HJ, Antonopoulos
CN, Adami HO, Ekbom A, et al. (2013). Obesity and
risk of malignant melanoma: a meta-analysis of cohort
and case-control studies. Eur J Cancer, 49(3):642–57.
doi:10.1016/j.ejca.2012.08.028 PMID:23200191
462
2.3. Excess body fatness in early life 2.3.1 Weight and height measured in
and subsequent cancer risk childhood
WHO defines children as individuals younger The few prospective studies that directly
than 19 years (WHO, 2016). The scope of this measured weight and height in childhood
section includes children and young adults up to and related these parameters with subse-
age 25 years, the age range collectively referred quent cancer occurrence have been reviewed
to as early life. recently (Simmonds et al., 2015). These include
It is generally held that childhood obesity data from the Helsinki Birth Cohort Study
is strongly associated with obesity in adult- (ages 7 years and 15 years; Hilakivi-Clarke
hood. According to a recent systematic review et al., 2001), the 1946 United Kingdom Medical
(Simmonds et al., 2015), obese children are more Research Council National Survey of Health and
than 5 times as likely as non-obese children to Development (ages 2–15 years; De Stavola et al.,
be obese as adults. However, childhood BMI is 2004), the Copenhagen School Health Records
not a good predictor of the occurrence of obesity Register (ages 7–15 years; Ahlgren et al., 2006;
in adulthood; 80% of people older than 30 years Aarestrup et al., 2014; Berentzen et al., 2014;
who are obese were not obese in adolescence. Kitahara et al., 2014a, 2014b; Cook et al., 2015),
Similarly, many obesity-related diseases occur in the Norwegian health surveys (ages 14–19 years;
adults who had a healthy weight in childhood. Engeland et al., 2003; Bjørge et al., 2004, 2008),
Few comprehensive reviews or meta-ana- the Israeli army (ages 16–19 years; Levi et al.,
lyses are available on the topic of body shape 2011; Leiba et al., 2013), and the Harvard Alumni
and weight in early life and subsequent cancer Health Study (ages 18–21 years; Gray et al., 2012).
risk. The literature review for this section iden- [These cohorts have the advantage that height
tified three categories of studies: (i) prospective and weight were directly measured, but they
studies that directly measured weight and height have relatively small sample sizes. Because the
in childhood and related these parameters with baseline data were collected more than half a
subsequent cancer occurrence; (ii) prospective century ago, extrapolation to the current child-
cohort studies that determined body shape hood and adolescent population may not apply,
in early adulthood by recall and related these and it is not always clear whether these cohorts
parameters with subsequent cancer occurrence; were representative of the general population.]
and (iii) studies that determined trajectories of The relationship between weight and height in
body shape (from repeated determinations) from childhood and subsequent cancer occurrence
childhood to late adulthood and related these is presented separately for cancer of the breast
parameters with subsequent cancer occurrence. (Table 2.3a) and for other cancers (Table 2.3b).
[The Working Group considered that the Table 2.3a lists study characteristics and
relationship between weight at birth and subse- breast cancer risk estimates from three studies
quent cancer risk was beyond the scope of this (Hilakivi-Clarke et al., 2001; De Stavola et al.,
Handbook.] 2004; Ahlgren et al., 2006), which included a
total of 3576 breast cancer cases. There was no
evidence that excess weight directly measured in
childhood is associated with subsequent breast
cancer risk. Indeed, there is some evidence of an
inverse association.
463
Table 2.3b lists, for boys and/or girls, study is a risk of recall bias, but distributions of recalled
characteristics and risk estimates of mortality BMI have been tested against BMI distributions
and incidence for the following types of cancer: from population data contemporaneous with the
colon cancer (Bjørge et al., 2008; Levi et al., respective age strata and were found to be similar
2011), rectal cancer (Levi et al., 2011), oesopha- (Renehan et al., 2012). It is worth remembering
geal adenocarcinoma (Cook et al., 2015), gastric that the mean values of BMI distributions of a
non-cardia cancer (Levi et al., 2013), hepato- cohort at ages 18–25 years are considerably lower
cellular carcinoma (Berentzen et al., 2014), than those in later adulthood. For example, in
pancreatic cancer (Levi et al., 2012), ovarian the NIH-AARP cohort, the mean BMI at age
cancer (Engeland et al., 2003), prostate cancer 18 years was 21.5 kg/m2 in men and 20.8 kg/m2
(Gray et al., 2012; Aarestrup et al., 2014; Batty in women (Renehan et al., 2012). In addition,
et al., 2015), renal cancer (Bjørge et al., 2004; there is a survival bias, in that individuals have
Leiba et al., 2013), urothelial cancer (Leiba et al., had to survive to baseline age (typically > 50)
2012), glioma (Kitahara et al., 2014a), and thyroid to participate in the cohort study. Finally, in
cancer (Farfel et al., 2014; Kitahara et al., 2014b). these studies, risk estimates from multivariate
Although the number of studies per cancer type analyses are commonly expressed as those from
is small, for boys, excess weight in childhood and separate models adjusted for several potential
adolescence (generally expressed per increase of confounders and as those from models adjusted
1 or 2 standard deviations in BMI) was generally for several potential confounders plus base-
associated with increased risk of colon cancer line (current-age) BMI. The latter models are of
(but not rectal cancer), oesophageal adenocar- mechanistic relevance; for the purpose of a public
cinoma, hepatocellular carcinoma, pancreatic health message in this Handbook, risk estimates
cancer, renal cancer, or urothelial cancer. There from the former models are reported.]
was no association with subsequent prostate These studies are dealt with in the individual
cancer occurrence. For girls, there was evidence cancer site-specific sections. Here, specific note is
that excess weight in childhood and adoles- made in relation to breast cancer.
cence (generally expressed per increase of 1 or Prospective cohort studies of recalled BMI
2 standard deviations in BMI) was associated at ages 18–25 years and subsequent postmeno-
with increased risk of colon cancer (but not pausal or premenopausal breast cancer risk
rectal cancer), oesophageal adenocarcinoma, are presented in Table 2.3c and Table 2.3d,
hepatocellular carcinoma, and ovarian cancer. respectively.
The association with renal cancer was uncertain For BMI at ages 18–25 years determined by
[because of a large confidence interval]. No asso- recall, several cohort studies showed no associ-
ciations were seen for glioma or thyroid cancer ation (van den Brandt et al., 1997; Suzuki et al.,
in either sex. 2011; Fagherazzi et al., 2013; Krishnan et al.,
2013; Catsburg et al., 2014) or inverse associations
2.3.2 Body shape in early adulthood (Ahn et al., 2007; Palmer et al., 2007; Baer et al.,
determined by recall 2010; Kawai et al., 2010; White et al., 2012) with
subsequent breast cancer risk. The same level of
A larger number of prospective cohort association was observed for postmenopausal
studies have determined body shape in early (Table 2.3c) and premenopausal (Table 2.3d)
adulthood (ages 18–25 years) by recall, typically women.
using the Sørensen scale (silhouette drawings), Some studies additionally evaluated BMI or
and converting the results to BMI values. [There weight at ages younger than 18 years determined
464
by recall: age at menarche in the French cohort associated with a higher risk of colorectal cancer
(Fagherazzi et al., 2013), at age 12 years in the Iowa and oesophageal adenocarcinoma; in addition,
Women’s Health Study (Bardia et al., 2008), and the heavy-stable/increase trajectory was associ-
at ages 5 years and 10 years in the Nurses’ Health ated with a higher risk of pancreatic cancer and
Study (1988–2004) and the Nurses’ Health Study a lower risk of advanced prostate cancer.
II (1989–2005) cohorts (Baer et al., 2010) (data In the French E3N cohort, Fagherazzi et al.
not shown in tables). These studies are consistent (2013) evaluated the risk of breast cancer associ-
in showing that body fatness at ages 5–12 years or ated with body shape (using the Sørensen scale)
age at menarche is independently and inversely at ages 8 years, age at menarche, 20–25 years,
associated with subsequent premenopausal (Baer and 35–40 years. Six lifetime trajectories of body
et al., 2010) and postmenopausal breast cancer shape were derived, using a finite mixture model-
(Bardia et al., 2008; Baer et al., 2010; Fagherazzi ling approach (Jones & Nagin, 2007). In this
et al., 2013). analysis, from age 8 years and/or at menarche,
a constantly elevated body size was associated
2.3.3 Trajectories of body shape determined with a significantly decreased risk of ER-positive
from early life and PR-positive postmenopausal breast cancer
(approximately 80% of breast cancers). No signif-
Additional information may be gained icant association with other body shape trajecto-
by exploring weight changes with time and ries was found.
cancer risk. Recently, Song et al. (2016) reported
combined analyses from the Nurses’ Health Study
(73 581 women) and the Health Professionals
Follow-up Study (32 632 men) for several cancer
sites (Table 2.3e). Using a data-driven latent class
approach, they identified five distinct trajectories
of body shape from age 5 years to age 60 years:
maintained a lean body shape (lean-stable),
started lean and experienced a moderate increase
in body shape (lean-moderate increase), started
lean and gained a substantial amount of weight
(lean-marked increase), maintained a medium
body shape (medium-stable), and started heavy
and maintained or gained weight (heavy-stable/
increase). Compared with women with the lean-
stable trajectory, women with the lean-marked
increase and the heavy-stable/increase trajecto-
ries had higher risks of colorectal, oesophageal,
pancreatic, renal, and endometrial cancers. For
postmenopausal breast cancer risk, early-life
adiposity with no loss in later life (heavy-stable/
increase trajectory) showed no association,
whereas late-life adiposity (lean-marked increase
trajectory) was positively associated. In men,
excess body fatness during any life period was
465
466

Table 2.3a Prospective studies of childhood cohorts where weight and height were directly measured and subsequent risk
of cancer of the breast
Reference Number at Number at Number of Adult age at final Childhood Relative risk
Cohort baseline follow-up breast cancers follow-up (years) age at (95% CI) per SD or unit
Period of study (Birth cohort) measurement increase in BMI
(years)
Hilakivi-Clarke et al. (2001) 3447 3447 177 Minimum, 38 7 0.91 (0.73–1.05)
Helsinki Birth Cohort (1924–1933) (76% > 50) 15 0.85 (0.70–1.00)
1971–1995
De Stavola et al. (2004) 2547 2187 59 47–53 2 1.02 (0.78–1.33)
United Kingdom Medical Research (March 1946) 4 0.88 (0.67–1.14)
Council National Survey of Health 7 0.87 (0.66–1.15)
and Development 11 0.89 (0.68–1.18)
1946–1999 15 0.86 (0.65–1.14)
Ahlgren et al. (2006) 161 063 117 415 3340 NR 14 0.97 (0.96–0.98)
Girls in Copenhagen, Denmark (1930–1975)
(Copenhagen School Health Records
Register)
Until 2001
BMI, body mass index (in kg/m 2); CI, confidence interval; NR, not reported; SD, standard deviation
Table 2.3b Prospective studies of childhood cohorts where weight and height were directly measured and subsequent risk
of other cancers, by sex and by organ site
Reference Number at Number at Number of Adult age at final Childhood age Relative risk
Cohort baseline follow-up cancers follow-up (years) at measurement (95% CI) per SD or unit increase in BMI
Period of (years)
recruitment
Boys
Colon cancer: mortality
Bjørge et al. (2008) 114 977 NR 97 Mean, 40 14–19 ≥ 85th percentile vs 25th–75th percentile:
Norwegian Cancer (1963–1975) 2.1 (1.1–4.1)
Registry
Colon cancer: incidence
Levi et al. (2011) 1 109 864 NR 445 19–57 16–19 1.21 (1.07–1.38)b
Israeli military cohort (1947–1966)
Rectal cancer: incidence
Levi et al. (2011) 1 109 864 NR 193 19–57 16–19 0.96 (0.88–1.10)b
Oesophageal adenocarcinoma: incidence
Cook et al. (2015) 188 360 128 330 216 > 40 7 1.11 (0.95–1.30)
Boys in Copenhagen, (1930–1989) 8 1.10 (0.94–1.29)
Denmark (Copenhagen 9 1.15 (0.98–1.35)
School Health Records 10 1.18 (1.00–1.38)
Register) 11 1.21 (1.03–1.42)
12 1.25 (1.07–1.47)
13 1.25 (1.06–1.46)
Gastric non-cardia: incidence
Levi et al. (2013) 1 088 530 NR 130 19–57 16–19 vs BMI 18.5–24.9:

Israeli military cohort (1967/2005–2006) BMI 25–29.9: 0.98 (0.51–1.89)
BMI ≥ 30: 2.62 (0.96–7.15)
Hepatocellular carcinoma: incidence
Berentzen et al. (2014) 188 360 144 417 229 Median, 59 7 1.18 (1.01–1.37)
Boys in Copenhagen, (1930–1980) 8 1.17 (1.00–1.37)
Register) 11 1.31 (1.12–1.53)
12 1.36 (1.16–1.59)
13 1.36 (1.17–1.60)
Pancreatic cancer: incidence
Levi et al. (2012) 720 927 NR 98 29–56 16–19 1.17 (0.96–1.52)b
467
468

Table 2.3b (continued)
Period of (years)
recruitment
Prostate cancer: mortality
Gray et al. (2012) 19 593 NR NR NR Mean, 18.4 1.04 (0.93–1.16)
Harvard Alumni Health (1914–1952)
Study
Prostate cancer: incidence
Aarestrup et al. (2014) 188 360 133 647 3355 Median, 66.5 7 1.04 (0.98–1.10)
Boys in Copenhagen, (1930–1969) (range, 40–81) 8 1.04 (0.98–1.11)
Register) 11 1.02 (0.96–1.08)
12 1.02 (0.96–1.08)
13 1.02 (0.96–1.09)
Batty et al. (2015) 2332 2332 109 Maximum, 77 11 0.97 (0.80–1.18)
Scottish Mental Health 1947–2014
Survey
Scotland, United Kingdom
Renal cancer: incidence
Registry
Leiba et al. (2013) 1 110 835 NR 274 Mean, 44 16–19 1.19 (1.04–1.37)b
Urothelial cancer:a
incidence
Leiba et al. (2012) 1 110 835 NR 661 Mean, 35 16–19 1.21 (1.06–1.38)b
Glioma: incidence
Kitahara et al. (2014a) 188 360 162 295 355 > 40 7 1.01 (0.86–1.17)
Boys in Copenhagen, 8 1.04 (0.89–1.22)
Register) 11 1.02 (0.87–1.19)
12 1.00 (0.86–1.17)
13 1.04 (0.89–1.21)
Period of (years)
recruitment
Thyroid cancer: incidence
Farfel et al. (2014) 1 145 865 NR 425 19–57 16–19 BMI, Q5 vs Q1:
Israeli military cohort (1967–2005) 1.19 (0.87–1.63)
Kitahara et al. (2014b) 165 978 162 632 64 > 40 7 1.22 (0.93–1.60)
Boys in Copenhagen, 8 1.24 (0.94–1.63)
Register) 11 1.24 (0.94–1.65)
12 1.25 (0.94–1.66)
13 1.25 (0.93–1.66)
Girls
Colon cancer: mortality
Registry
Oesophageal adenocarcinoma: incidence
Cook et al. (2015) 184 276 126 723 38 > 40 7 1.30 (0.90–1.87)
Girls in Copenhagen, (1931–1971) 8 1.41 (0.97–2.06)
Register) 11 1.63 (1.12–2.36)
12 1.55 (1.07–2.26)
13 1.68 (1.15–2.44)

Hepatocellular carcinoma: incidence
Berentzen et al. (2014) 184 276 141 467 62 Median, 60.2 7 1.20 (0.90–1.60)
Girls in Copenhagen, (1930–1980) 8 1.12 (0.84–1.50)
Register) 11 1.05 (0.78–1.40)
12 1.15 (0.85–1.54)
13 1.23 (0.93–1.65)
Ovarian cancer
Engeland et al. (2003) NR 111 883 7882 Mean, 41 14–19 1.22 (1.01–1.49)b
Norwegian Cancer (1963–1999)
Registry
469
470

Period of (years)
recruitment
Renal cancer: incidence
Registry
Glioma: incidence
Kitahara et al. (2014a) 184 276 158 130 253 > 40 7 0.96 (0.79–1.16)
Girls in Copenhagen, 8 0.95 (0.79–1.16)
Register) 11 0.93 (0.76–1.13)
12 0.91 (0.75–1.10)
13 1.01 (0.83–1.22)
Thyroid cancer: incidence
Farfel et al. (2014) 478 445 NR 323 19–57 16–19 BMI, Q5 vs Q1:
Israeli military cohort (1989–2005) 1.14 (0.81–1.60)
Kitahara et al. (2014b) 161 262 158 453 171 > 40 7 1.13 (0.96–1.33)
Girls in Copenhagen, 8 1.12 (0.95–1.32)
Register) 11 1.11 (0.94–1.31)
12 1.09 (0.92–1.29)
13 1.13 (0.96–1.34)
BMI, body mass index (in kg/m 2); CI, confidence interval; NR, not reported; SD, standard deviation.
a Bladder, ureter, and renal pelvis.
b Taken from the systematic review and meta-analysis by Simmonds et al. (2015).
Table 2.3c Prospective cohort studies of BMI at ages 18–25 years determined by recall and subsequent risk of cancer of the
breast in postmenopausal women
Reference Total number in Follow-up Baseline age (years) Recall age Number of cases Relative risk
Cohort cohort period (years) (years) (95% CI)
Country
van den Brandt et al. (1997) 62 573 4.3 55–69 20 626 Per 8 kg/m2:
Netherlands Cohort Study 0.79 (0.58–1.08)
The Netherlands
Ahn et al. (2007) 99 039 3.9 50–71 18 2111 BMI ≥ 30.0 vs 18.5–22.4:
NIH-AARP Diet and All postmenopausal HRT non-users
Health Study 0.48 (0.27–0.86)
USA HRT current users
0.65 (0.35–1.23)
Palmer et al. (2007) 9542 10 21–69 18 442 BMI ≥ 25.0 vs < 20.0:
Black Women’s Health 0.55 (0.37–0.82)
Study
USA
Baer et al. (2010) 188 860 16 NHS, 30–55 20 4974 Per 1 kg/m2:
Nurses’ Health Study (NHS) NHS II, 25–42 0.93 (0.90–0.95)
and NHS II
USA
Kawai et al. (2010) 10 106 12.8 40–64 20 108 BMI ≥ 23.8 vs < 20.5:
Miyagi Cohort Study 0.44 (0.24–0.81)
Japan
Suzuki et al. (2011) 41 594 10 40–59 20 232 Per 5 kg/m2:
Japan Public Health Cohort 0.77 (0.59–1.02)
Study
Japan

White et al. (2012) 82 971 NR 45–75 21 3030 BMI ≥ 30.0 vs < 20.0–24.9:
Multiethnic Cohort 0.63 (0.43–0.91)
USA
Fagherazzi et al. (2013) 81 089 NR 40–64 20–25 2828 Level ≥ 4 vs level 1:a
French E3N cohort 0.86 (0.74–1.00)
France
Krishnan et al. (2013) 14 441 16.5 27–76 (99% 40–69) 18–21 668 Per 5 kg/m2:
Melbourne Collaborative 0.90 (0.79–1.04)
Cohort Study
Australia
471
472

Table 2.3c (continued)
Reference Total number in Follow-up Baseline age (years) Recall age Number of cases Relative risk
Cohort cohort period (years) (years) (95% CI)
Country
Catsburg et al. (2014) 2210 12 67 20 541 BMI ≥ 30.0 vs 18.5–24.9:
Canadian Study of Diet, 0.21 (0.03–1.59)
Lifestyle and Health
Canada
BMI, body mass index (in kg/m 2); CI, confidence interval; HRT, hormone replacement therapy; NR, not reported
a Participants were asked to recall their body fatness by using a 9-level figure drawing, where level 1 represents the most lean and level 9 represents the most overweight.
Table 2.3d Prospective cohort studies of BMI at ages 18–25 years determined by recall and subsequent risk of cancer of the
breast in premenopausal women
Reference Total number in Follow-up period Baseline age Recall age Number of cancers Relative risk
Cohort cohort (years) (years) (years) (95% CI)
Country
Palmer et al. (2007) 42 538 10 21–69 18 491 BMI ≥ 25.0 vs < 20.0:
Black Women’s Health Study 0.63 (0.46–0.87)
USA
Baer et al. (2010); Michels et 188 860 16 NHS, 30–55 20 2188 Per 1 kg/m2:
al. (2012) NHS II, 25–42 0.89 (0.86–0.93)
Nurses’ Health Study (NHS)
and NHS II
USA
Suzuki et al. (2011) 41 594 10 40–59 20 220 Per 5 kg/m2:
Japan Public Health Cohort 0.78 (0.57–1.06)
Study
Japan
Fagherazzi et al. (2013) 81 089 NR 40–64 20–25 745 Level ≥ 4 vs level 1:a
French E3N cohort 1.22 (0.88–1.69)
France
Catsburg et al. (2014) 1110 14 45 20 556 BMI ≥ 30.0 vs 18.5–24.9:
Canadian Study of Diet, 0.96 (0.33–2.81)
Lifestyle and Health
Canada
BMI, body mass index (in kg/m 2); CI, confidence interval; NR, not reported
a Participants were asked to recall their body fatness by using a 9-level figure drawing, where level 1 represents the most lean and level 9 represents the most overweight.

473
474

Table 2.3e Relative risk of selected cancers according to trajectories of body shape from age 5 years to age 60 years in
women and in men
Cancer type Category of body shape trajectorya

Lean-stable Lean-moderate increase Lean-marked increase Medium-stable Heavy-stable/increase
Women
Number of participants 13 183 18 405 18 217 23 288 11 699
Colorectal cancer 1.00 0.97 (0.80–1.17) 1.22 (1.00–1.49) 1.02 (0.85–1.22) 1.40 (1.13–1.74)
Oesophageal adenocarcinoma 1.00 1.02 (0.29–3.63) 2.56 (0.82–8.03) 1.04 (0.30–3.57) 2.19 (0.63–7.70)
Pancreatic cancer 1.00 1.18 (0.82–1.69) 1.36 (0.93–1.98) 1.15 (0.81–1.63) 1.39 (0.91–2.12)
Kidney cancer 1.00 1.26 (0.78–2.04) 1.89 (1.19–3.03) 1.05 (0.65–1.69) 1.92 (1.15–3.21)
Postmenopausal breast cancer 1.00 1.30 (1.17–1.45) 1.41 (1.26–1.58) 1.05 (0.94–1.17) 1.11 (0.97–1.28)
Endometrial cancer 1.00 0.99 (0.75–1.29) 1.57 (1.21–2.03) 0.94 (0.73–1.22) 2.08 (1.59–2.73)
Ovarian cancer 1.00 0.88 (0.66–1.16) 0.93 (0.70–1.25) 0.88 (0.67–1.15) 0.84 (0.59–1.19)
Men
Number of participants 5946 6881 14 225 5725 4929
Colorectal cancer 1.00 1.36 (1.03–1.80) 1.23 (0.95–1.60) 1.26 (0.92–1.72) 1.47 (1.05–2.05)
Oesophageal adenocarcinoma 1.00 1.90 (0.67–5.34) 2.09 (0.80–5.48) 1.53 (0.48–4.84) 3.01 (1.04–9.13)
Pancreatic cancer 1.00 0.85 (0.54–1.35) 1.20 (0.81–1.78) 1.12 (0.70–1.80) 1.50 (0.92–2.46)
Kidney cancer 1.00 1.05 (0.67–1.64) 0.94 (0.63–1.43) 1.07 (0.66–1.74) 0.93 (0.53–1.64)
Advanced prostate cancer 1.00 1.16 (0.91–1.47) 0.97 (0.78–1.21) 1.00 (0.76–1.32) 0.67 (0.47–0.95)
a Trajectories of body shape: maintained a lean body shape (lean-stable); started lean and experienced a moderate increase in body shape (lean-moderate increase); started lean and
gained a substantial amount of weight (lean-marked increase); maintained a medium body shape (medium-stable); started heavy and maintained or gained weight (heavy-stable/
increase).
Source: Song et al. (2016). Data for women are from the Nurses’ Health Study, and data for men are from the Health Professionals Follow-up Study.
References De Stavola BL, dos Santos Silva I, McCormack V, Hardy

RJ, Kuh DJ, Wadsworth ME (2004). Childhood growth
and breast cancer. Am J Epidemiol, 159(7):671–82.
Aarestrup J, Gamborg M, Cook MB, Sørensen TI, Baker doi:10.1093/aje/kwh097 PMID:15033645
JL (2014). Childhood body mass index and the risk of Engeland A, Tretli S, Bjørge T (2003). Height, body mass
prostate cancer in adult men. Br J Cancer, 111(1):207– index, and ovarian cancer: a follow-up of 1.1 million
12. doi:10.1038/bjc.2014.266 PMID:24867696 Norwegian women. J Natl Cancer Inst, 95(16):1244–8.
Ahlgren M, Melbye M, Wohlfahrt J, Sørensen TI (2006). doi:10.1093/jnci/djg010 PMID:12928351
Growth patterns and the risk of breast cancer in women. Fagherazzi G, Guillas G, Boutron-Ruault MC, Clavel-
Int J Gynecol Cancer, 16(Suppl 2):569–75. doi:10.1111/ Chapelon F, Mesrine S (2013). Body shape throughout
j.1525-1438.2006.00698.x PMID:17010075 life and the risk for breast cancer at adulthood in the
Ahn J, Schatzkin A, Lacey JV Jr, Albanes D, Ballard- French E3N cohort. Eur J Cancer Prev, 22(1):29–37.
Barbash R, Adams KF, et al. (2007). Adiposity, adult doi:10.1097/CEJ.0b013e328355ec04 PMID:22694827
weight change, and postmenopausal breast cancer Farfel A, Kark JD, Derazne E, Tzur D, Barchana M, Lazar L,
risk. Arch Intern Med, 167(19):2091–102. doi:10.1001/ et al. (2014). Predictors for thyroid carcinoma in Israel:
archinte.167.19.2091 PMID:17954804 a national cohort of 1,624,310 adolescents followed
Baer HJ, Tworoger SS, Hankinson SE, Willett WC (2010). for up to 40 years. Thyroid, 24(6):987–93. doi:10.1089/
Body fatness at young ages and risk of breast cancer thy.2013.0173 PMID:24483833
throughout life. Am J Epidemiol, 171(11):1183–94. Gray L, Lee IM, Sesso HD, Batty GD (2012). Association
doi:10.1093/aje/kwq045 PMID:20460303 of body mass index in early adulthood and middle age
Bardia A, Vachon CM, Olson JE, Vierkant RA, Wang with future site-specific cancer mortality: the Harvard
AH, Hartmann LC, et al. (2008). Relative weight at Alumni Health Study. Ann Oncol, 23(3):754–9.
age 12 and risk of postmenopausal breast cancer. doi:10.1093/annonc/mdr270 PMID:21677311
Cancer Epidemiol Biomarkers Prev, 17(2):374–8. Hilakivi-Clarke L, Forsén T, Eriksson JG, Luoto R,
doi:10.1158/1055-9965.EPI-07-0389 PMID:18250344 Tuomilehto J, Osmond C, et al. (2001). Tallness and
Batty GD, Calvin CM, Brett CE, Čukić I, Deary IJ (2015). overweight during childhood have opposing effects
Childhood body weight in relation to morbidity from on breast cancer risk. Br J Cancer, 85(11):1680–4.
cardiovascular disease and cancer in older adulthood: doi:10.1054/bjoc.2001.2109 PMID:11742488
67-year follow-up of participants in the 1947 Scottish Jones N, Nagin D (2007). Advances in group-based
Mental Survey. Am J Epidemiol, 182(9):775–80. trajectory modeling and a SAS procedure for esti-
doi:10.1093/aje/kwv154 PMID:26443418 mating them. Sociol Methods Res, 35(4):542–71.
Berentzen TL, Gamborg M, Holst C, Sørensen TI, Baker doi:10.1177/0049124106292364
JL (2014). Body mass index in childhood and adult Kawai M, Minami Y, Kuriyama S, Kakizaki M, Kakugawa
risk of primary liver cancer. J Hepatol, 60(2):325–30. Y, Nishino Y, et al. (2010). Adiposity, adult weight
doi:10.1016/j.jhep.2013.09.015 PMID:24076363 change and breast cancer risk in postmenopausal
Bjørge T, Engeland A, Tverdal A, Smith GD (2008). Body Japanese women: the Miyagi Cohort Study. Br J
mass index in adolescence in relation to cause-specific Cancer, 103(9):1443–7. doi:10.1038/sj.bjc.6605885
mortality: a follow-up of 230,000 Norwegian adoles- PMID:20842123
cents. Am J Epidemiol, 168(1):30–7. doi:10.1093/aje/ Kitahara CM, Gamborg M, Berrington de González A,
kwn096 PMID:18477652 Sørensen TI, Baker JL (2014b). Childhood height and
Bjørge T, Tretli S, Engeland A (2004). Relation of body mass index were associated with risk of adult
height and body mass index to renal cell carcinoma thyroid cancer in a large cohort study. Cancer Res,
in two million Norwegian men and women. Am J 74(1):235–42. doi:10.1158/0008-5472.CAN-13-2228
Epidemiol, 160(12):1168–76. doi:10.1093/aje/kwh345 PMID:24247722
PMID:15583369 Kitahara CM, Gamborg M, Rajaraman P, Sørensen TI,
Catsburg C, Kirsh VA, Soskolne CL, Kreiger N, Bruce Baker JL (2014a). A prospective study of height and
E, Ho T, et al. (2014). Associations between anthro- body mass index in childhood, birth weight, and risk of
pometric characteristics, physical activity, and breast adult glioma over 40 years of follow-up. Am J Epidemiol,
cancer risk in a Canadian cohort. Breast Cancer Res 180(8):821–9. doi:10.1093/aje/kwu203 PMID:25205831
Treat, 145(2):545–52. doi:10.1007/s10549-014-2973-z Krishnan K, Bassett JK, MacInnis RJ, English DR, Hopper
PMID:24781974 JL, McLean C, et al. (2013). Associations between weight
Cook MB, Freedman ND, Gamborg M, Sørensen TI, in early adulthood, change in weight, and breast cancer
Baker JL (2015). Childhood body mass index in rela- risk in postmenopausal women. Cancer Epidemiol
tion to future risk of oesophageal adenocarcinoma. Biomarkers Prev, 22(8):1409–16. doi:10.1158/1055-
Br J Cancer, 112(3):601–7. doi:10.1038/bjc.2014.646 9965.EPI-13-0136 PMID:23720403
PMID:25562436
475
Leiba A, Kark JD, Afek A, Derazne E, Barchana M, Tzur D, 20 years, subsequent weight change and breast cancer
et al. (2013). Adolescent obesity and paternal country risk defined by estrogen and progesterone receptor
of origin predict renal cell carcinoma: a cohort study of status – the Japan Public Health Center-based prospec-
1.1 million 16 to 19-year-old males. J Urol, 189(1):25–9. tive study. Int J Cancer, 129(5):1214–24. doi:10.1002/
doi:10.1016/j.juro.2012.08.184 PMID:23164387 ijc.25744 PMID:21064092
Leiba A, Kark JD, Afek A, Levi Z, Barchana M, Tzur van den Brandt PA, Dirx MJ, Ronckers CM, van den
D, et al. (2012). Overweight in adolescence is related Hoogen P, Goldbohm RA (1997). Height, weight,
to increased risk of future urothelial cancer. Obesity weight change, and postmenopausal breast cancer
(Silver Spring), 20(12):2445–50. doi:10.1038/oby.2012.83 risk: the Netherlands Cohort Study. Cancer Causes
PMID:22510956 Control, 8(1):39–47. doi:10.1023/A:1018479020716
Levi Z, Kark JD, Afek A, Derazne E, Tzur D, Furman PMID:9051321
M, et al. (2012). Measured body mass index in White KK, Park SY, Kolonel LN, Henderson BE, Wilkens
adolescence and the incidence of pancreatic cancer LR (2012). Body size and breast cancer risk: the
in a cohort of 720,000 Jewish men. Cancer Causes Multiethnic Cohort. Int J Cancer, 131(5):E705–16.
Control, 23(2):371–8. doi:10.1007/s10552-011-9886-5 doi:10.1002/ijc.27373 PMID:22120517
PMID:22212610 WHO (2016). Report of the Commission on Ending
Levi Z, Kark JD, Barchana M, Liphshitz I, Zavdi O, Tzur D, Childhood Obesity. Geneva, Switzerland: World
et al. (2011). Measured body mass index in adolescence Health Organization. Available from: http://www.who.
and the incidence of colorectal cancer in a cohort of int/end-childhood-obesity/news/launch-final-report/
1.1 million males. Cancer Epidemiol Biomarkers Prev, en/.
20(12):2524–31. doi:10.1158/1055-9965.EPI-11-0531
PMID:22056504
Levi Z, Kark JD, Shamiss A, Derazne E, Tzur D, Keinan-
Boker L et a l. (2013). Body mass index and socioeconomic
status measured in adolescence, country of origin, and
the incidence of gastroesophageal adenocarcinoma in
a cohort of 1 million men. Cancer, 119(23):4086–93.
doi:10.1002/cncr.28241 PMID:24129941
Michels KB, Terry KL, Eliassen AH, Hankinson SE,
Willett WC (2012). Adult weight change and inci-
dence of premenopausal breast cancer. Int J Cancer,
130(4):902–9. doi:10.1002/ijc.26069 PMID:21413008
Palmer JR, Adams-Campbell LL, Boggs DA, Wise LA,
Rosenberg L (2007). A prospective study of body size
and breast cancer in black women. Cancer Epidemiol
Biomarkers Prev, 16(9):1795–802. doi:10.1158/1055-
9965.EPI-07-0336 PMID:17855697
Renehan AG, Flood A, Adams KF, Olden M, Hollenbeck
AR, Cross AJ, et al. (2012). Body mass index at different
adult ages, weight change, and colorectal cancer risk in
the National Institutes of Health-AARP Cohort. Am
J Epidemiol, 176(12):1130–40. doi:10.1093/aje/kws192
PMID:23186750
Simmonds M, Burch J, Llewellyn A, Griffiths C, Yang H,
Owen C, et al. (2015). The use of measures of obesity in
childhood for predicting obesity and the development
of obesity-related diseases in adulthood: a system-
atic review and meta-analysis. Health Technol Assess,
19(43):1–336. doi:10.3310/hta19430 PMID:26108433
Song M, Willett WC, Hu FB, Spiegelman D, Must A, Wu
K, et al. (2016). Trajectory of body shape across the
lifespan and cancer risk. Int J Cancer, 138(10):2383–95.
doi:10.1002/ijc.29981 PMID:26704725
Suzuki R, Iwasaki M, Inoue M, Sasazuki S, Sawada N,
Yamaji T, et al.; Japan Public Health Center-based
Prospective Study Group (2011). Body weight at age
476
2.4 Excess body fatness in cancer diagnosis. Associations were observed between
survivors measures of adiposity and prognosis, but there
were many pitfalls to interpretations, biases,
2.4.1 Studies of weight at diagnosis and and confounding. The evidence linking obesity
cancer outcomes to cancer survival was rated as “limited-sugges-
tive”, primarily because of concerns about the
An increasing number of observational timing of baseline BMI analysis in relation to
studies are focusing on the association between cancer diagnosis in some studies.
excess body fatness and prognosis in cancer Fewer studies have evaluated the associa-
survivors. Specifically, more than 100 indi- tion between body fatness and cancer prognosis
vidual reports have evaluated the relationship in other malignancies. A meta-analysis that
between BMI or body weight at the time of diag- evaluated the relationship between obesity and
nosis of early-stage breast cancer and the risk of colorectal cancer outcomes included 16 reports
breast cancer recurrence, breast cancer-related that encompassed 58 917 individuals followed
mortality, and all-cause mortality. up for a median of 9.9 years (Lee et al., 2015).
A meta-analysis of 82 reports on this topic (all Obesity before diagnosis of colorectal cancer was
but 8 of which had a median follow-up of at least associated with an increased risk of colorectal
5 years) incorporated data from 213 075 women cancer-specific mortality (RR, 1.22; 95% CI,
(Chan et al., 2014). Women who were obese (BMI 1.00–1.35) and all-cause mortality (RR, 1.25;
> 30.0 kg/m2) at the time of diagnosis of breast 95% CI, 1.14–1.36). Obesity after diagnosis of
cancer had a 35% increased risk (RR, 1.35; 95% colorectal cancer was also associated with an
CI, 1.24–1.47) of breast cancer-related mortality increased risk of all-cause mortality (RR, 1.08;
and a 41% increased risk of all-cause mortality 95% CI, 1.03–1.13).
compared with women who were of normal Excess body fatness has also been linked with
weight at the time of breast cancer diagnosis. The biochemical recurrence of cancer (rising levels
association between obesity and poor outcomes of prostate-specific antigen [PSA]) in men with
was seen in both postmenopausal and premeno- early-stage prostate cancer treated with radical
pausal breast cancer survivors, with summary prostatectomy or external beam radiation. A
relative risks for all-cause mortality in obese meta-analysis of 26 studies, including 36 927
versus normal-weight women of 1.75 (95% CI, men, estimated a 16% increase in the risk of
1.26–2.41) in women with premenopausal breast elevated PSA levels with each 5 kg/m2 increase
cancer and 1.34 (95% CI, 1.18–1.53) in women in BMI (RR, 1.16; 95% CI, 1.08–1.24) (Hu et al.,
with postmenopausal breast cancer. 2014).
In another study, the WCRF Continuous A meta-analysis of 14 studies that assessed
Update Project reviewed data on the associa- BMI before or shortly after diagnosis in women
tion in female breast cancer survivors between with ovarian cancer estimated a hazard ratio for
weight and the risk of dying of breast cancer, all-cause mortality of 1.17 (95% CI, 1.03–1.34)
second cancers, or any cause (WCRF/AICR, for obese versus non-obese patients (Protani
2014). The report stressed the importance of et al., 2012). Another meta-analysis of 13 cohort
taking into account the timing of weight meas- studies of individuals with pancreatic cancer
urement, focusing on three main time points: reported an adjusted hazard ratio for pancre-
(i) before diagnosis; (ii) less than 12 months after atic cancer-related mortality of 1.06 (95% CI,
diagnosis; and (iii) more than 12 months after 1.02–1.11) in overweight patients and of 1.31 (95%
477
CI, 1.20–1.42) in obese patients versus normal- A recent meta-analysis of 12 studies exam-
weight patients (Majumder et al., 2015). ined the association between weight gain
Meta-analyses and/or systematic reviews after diagnosis of breast cancer and prognosis
on obesity and cancer survival have also been (Playdon et al., 2015). High weight gain after
conducted in patients with endometrial cancer breast cancer diagnosis (> 10% of body weight
(Arem & Irwin, 2013; Nakao et al., 2014) and with at diagnosis) increased the risk of both all-cause
childhood leukaemia (Amankwah et al., 2015). mortality and breast cancer-specific mortality,
[It is unclear whether the relationship between whereas moderate weight gain (5–10%) did not
obesity and increased risk of cancer-related (HR, 0.98; 95% CI, 0.83–1.15). The increased
mortality stems from differences in the biological risk was observed among women with a BMI at
aggressiveness or subtypes of cancers that develop diagnosis of less than 25 kg/m2 and of 25 kg/m2
in obese versus non-obese patients. Some studies or more. In an earlier analysis of a prospective
have suggested that obese individuals are more cohort study of 5204 non-smoking women with
likely to develop biologically aggressive cancers early-stage breast cancer, those who gained more
with poorer outcomes, or to have more advanced than 2 kg/m2 had a significantly increased risk of
disease at the time of diagnosis. For example, death from breast cancer compared with women
studies have shown that obese individuals are at who maintained a stable weight; the relative risk
increased risk of developing biologically aggres- of death from breast cancer was 1.35 (95% CI,
sive prostate cancers, but not of developing lower- 0.93–1.95) for weight gain of 0.5–2 kg/m2 and
grade prostate cancers (see Section 2.2.14). Some 1.64 (95% CI, 1.07–2.51) for weight gain of more
reports suggest that obese women are more likely than 2 kg/m2 (Kroenke et al., 2005). In contrast,
to develop poorly differentiated and hormone in another study of 1692 women with early-
receptor-negative breast cancers (Stark et al., stage breast cancer, no association was observed
2010; Abdel-Maksoud et al., 2012), although between weight gain and breast cancer recur-
other reports suggest that obese women are more rence or all-cause mortality, even among women
likely to develop slower-growing hormone recep- who gained more than 10% of their baseline body
tor-positive breast cancers (Borgquist et al., 2009; weight (Caan et al., 2006).
Canchola et al., 2012; Biglia et al., 2013). A few
recent studies that have used genomic profiling 2.4.3 Intervention trials of weight-loss
techniques have suggested that obese women intervention and dietary modification
who develop hormone receptor-positive cancers
are more likely to have luminal B cancers, which No data were available to the Working Group
have been shown to have a worse prognosis, about the impact of a weight-loss intervention on
compared with luminal A cancers (Kwan et al., cancer recurrence, cancer-related mortality, or
2015; Ligibel et al., 2015). See Section 2.2.9 for all-cause mortality in cancer survivors.
more detailed data on risk estimates by subtype Two randomized trials assessed the impact of
of breast cancer.] dietary modification on disease-free and overall
survival in women with early-stage breast cancer.
The Women’s Intervention Nutrition Study
2.4.2 Studies of weight change after cancer
randomized 2400 women to a low-fat dietary
diagnosis and cancer outcomes
intervention or usual care (control group)
Fewer studies have investigated the associ- (Chlebowski et al., 2008). Patients assigned to
ation between weight change after cancer diag- the intervention group reduced their dietary fat
nosis and recurrence-free or overall survival. intake for the duration of the 5-year intervention.
478
Intervention participants experienced an average Several ongoing studies are testing the
weight loss of 6 lb (2.7 kg). An initial analysis hypothesis that weight loss after cancer diag-
of study results demonstrated a 24% reduction nosis reduces the risk of cancer recurrence or
in breast cancer recurrence compared with the progression in individuals with early-stage
control group (HR, 0.76; 95% CI, 0.60–0.98) cancer (Courneya et al., 2008; Rack et al., 2010;
(Chlebowski et al., 2006), although the differ- Villarini et al., 2012; Crane et al., 2014; Parsons
ence lost statistical significance with further et al., 2014).
follow-up (Chlebowski et al., 2008). Unplanned
subset analysis suggested that the impact of the
intervention differed in women with ER-positive References
cancers versus those with ER-negative cancers,
with a hazard ratio for recurrence in the inter- Abdel-Maksoud MF, Risendal BC, Slattery ML, Giuliano
vention group versus controls of 0.58 (95% CI, AR, Baumgartner KB, Byers TE (2012). Behavioral risk
factors and their relationship to tumor characteristics
0.37–0.91) in women with ER-negative cancers in Hispanic and non-Hispanic white long-term breast
and 0.85 (95% CI, 0.63–1.14) in women with cancer survivors. Breast Cancer Res Treat, 131(1):169–
ER-positive cancers (Pinteraction = 0.15). [The weight 76. doi:10.1007/s10549-011-1705-x PMID:21822637
loss experienced by participants in the Women’s Amankwah EK, Saenz AM, Hale GA, Brown PA (2015).
Association between body mass index at diagnosis and
Intervention Nutrition Study may have contrib- pediatric leukemia mortality and relapse: a system-
uted to the reduced risk of cancer recurrence in atic review and meta-analysis. Leuk Lymphoma,
intervention participants in that study.] 57(5):1140–8. doi:10.3109/10428194.2015.1076815
PMID:26453440
In contrast, the Women’s Healthy Eating and Arem H, Irwin ML (2013). Obesity and endometrial
Living study randomized 3088 women to a coun- cancer survival: a systematic review. Int J Obes (Lond),
selling programme for a diet very high in fruits 37(5):634–9. doi:10.1038/ijo.2012.94 PMID:22710929
and vegetables and low in fat or printed guidelines Biglia N, Peano E, Sgandurra P, Moggio G, Pecchio S,
Maggiorotto F, et al. (2013). Body mass index (BMI)
(Pierce et al., 2007). Adherence to the dietary and breast cancer: impact on tumor histopathologic
intervention was good, with intervention partic- features, cancer subtypes and recurrence rate in pre
ipants increasing their daily intake of vegetables and postmenopausal women. Gynecol Endocrinol,
by 65% and of fruits by 25%, and reducing their 29(3):263–7. doi:10.3109/09513590.2012 .736559
PMID:23174088
daily intake of fat by 13%. [Of note, participants Borgquist S, Jirström K, Anagnostaki L, Manjer J,
consumed on average seven servings of fruits Landberg G (2009). Anthropometric factors in rela-
and vegetables per day at baseline.] Participants tion to different tumor biological subgroups of post-
menopausal breast cancer. Int J Cancer, 124(2):402–11.
randomized to the dietary intervention group doi:10.1002/ijc.23850 PMID:18798278
did not lose weight compared with controls. The Caan BJ, Emond JA, Natarajan L, Castillo A, Gunderson
dietary intervention had no impact on rates of EP, Habel L, et al. (2006). Post-diagnosis weight gain
recurrence (HR for recurrence in intervention and breast cancer recurrence in women with early stage
breast cancer. Breast Cancer Res Treat, 99(1):47–57.
group vs controls, 0.96; 95% CI, 0.80–1.14). doi:10.1007/s10549-006-9179-y PMID:16541317
[There were several differences between the Canchola AJ, Anton-Culver H, Bernstein L, Clarke CA,
trials, including in the degree of reduction in Henderson K, Ma H, et al. (2012). Body size and the
dietary fat intake achieved by intervention partic- risk of postmenopausal breast cancer subtypes in
the California Teachers Study cohort. Cancer Causes
ipants, the baseline diets, the delivery method of Control, 23(3):473–85. doi:10.1007/s10552-012-9897-x
the dietary intervention, the timing of enrolment PMID:22286371
relative to breast cancer diagnosis, and the study Chan DS, Vieira AR, Aune D, Bandera EV, Greenwood
DC, McTiernan A, et al. (2014). Body mass index and
population.] survival in women with breast cancer – systematic
479
literature review and meta-analysis of 82 follow-up on outcomes of hematopoietic SCT: a meta-analysis.

studies. Ann Oncol, 25(10):1901–14. doi:10.1093/ Bone Marrow Transplant, 49(1):66–72. doi:10.1038/
annonc/mdu042 PMID:24769692 bmt.2013.128 PMID:23955636
Chlebowski R, Blackburn G, Elashoff R (2006). Mature Parsons JK, Pierce JP, Mohler J, Paskett E, Jung SH,
analysis from the Women’s Intervention Nutrition Humphrey P, et al. (2014). A randomized trial of diet
Study (WINS) evaluating dietary fat reduction and in men with early stage prostate cancer on active
breast cancer outcome. Breast Cancer Res Treat, surveillance: rationale and design of the Men’s Eating
100:S16. and Living (MEAL) Study (CALGB 70807 [Alliance]).
Chlebowski RT, Blackburn GL, Hoy MK, Thomson CA, Contemp Clin Trials, 38(2):198–203. doi:10.1016/j.
Giuliano AE, McAndrew P, et al.; Women’s Interven cct.2014.05.002 PMID:24837543
tion Nutrition Study (2008). Survival analyses from the Pierce JP, Natarajan L, Caan BJ, Parker BA, Greenberg ER,
Women’s Intervention Nutrition Study (WINS) evalu- Flatt SW, et al. (2007). Influence of a diet very high in
ating dietary fat reduction and breast cancer outcome. vegetables, fruit, and fiber and low in fat on prognosis
[Abstract] J Clin Oncol, 26(15 Suppl):522. doi:10.1200/ following treatment for breast cancer: the Women’s
jco.2008.26.15_suppl.522 Healthy Eating and Living (WHEL) randomized trial.
Courneya KS, Booth CM, Gill S, O’Brien P, Vardy J, JAMA, 298(3):289–98. doi:10.1001/jama.298.3.289
Friedenreich CM, et al. (2008). The Colon Health and PMID:17635889
Life-Long Exercise Change trial: a randomized trial of Playdon MC, Bracken MB, Sanft TB, Ligibel JA, Harrigan
the National Cancer Institute of Canada Clinical Trials M, Irwin ML (2015). Weight gain after breast cancer
Group. Curr Oncol, 15(6):279–85. PMID:19079628 diagnosis and all-cause mortality: systematic review
Crane T, Basen-Engquist K, Alberts D, et al. (2014). and meta-analysis. J Natl Cancer Inst, 107(12):djv275.
Baseline characteristics of ovarian cancer survivors doi:10.1093/jnci/djv275 PMID:26424778
enrolled in a lifestyle intervention trial: LIVES study. Protani MM, Nagle CM, Webb PM (2012). Obesity and
Cancer Survivorship Research Conference, Atlanta, ovarian cancer survival: a systematic review and
GA, June 18–20, 2014. meta-analysis. Cancer Prev Res (Phila), 5(7):901–10.
Hu MB, Xu H, Bai PD, Jiang HW, Ding Q (2014). Obesity doi:10.1158/1940-6207.CAPR-12-0048 PMID:22609763
has multifaceted impact on biochemical recurrence Rack B, Andergassen U, Neugebauer J, Salmen J, Hepp
of prostate cancer: a dose-response meta-analysis of P, Sommer H, et al. (2010). The German SUCCESS
36,927 patients. Med Oncol, 31(2):829. doi:10.1007/ C study – the first European lifestyle study on
s12032-013-0829-8 PMID:24390417 breast cancer. Breast Care (Basel), 5(6):395–400.
Kroenke CH, Chen WY, Rosner B, Holmes MD (2005). doi:10.1159/000322677 PMID:21494405
Weight, weight gain, and survival after breast cancer Stark A, Stahl MS, Kirchner HL, Krum S, Prichard J, Evans
diagnosis. J Clin Oncol, 23(7):1370–8. doi:10.1200/ J (2010). Body mass index at the time of diagnosis and
JCO.2005.01.079 PMID:15684320 the risk of advanced stages and poorly differentiated
Kwan ML, Kroenke CH, Sweeney C, Bernard PS, Weltzien cancers of the breast: findings from a case-series study.
EK, Castillo A, et al. (2015). Association of high obesity Int J Obes (Lond), 34(9):1381–6. doi:10.1038/ijo.2010.69
with PAM50 breast cancer intrinsic subtypes and gene PMID:20351736
expression. BMC Cancer, 15(1):278. doi:10.1186/s12885- Villarini A, Pasanisi P, Traina A, Mano MP, Bonanni
015-1263-4 PMID:25884832 B, Panico S, et al. (2012). Lifestyle and breast cancer
Lee J, Meyerhardt JA, Giovannucci E, Jeon JY (2015). recurrences: the DIANA-5 trial. Tumori, 98(1):1–18.
Association between body mass index and prognosis PMID:22495696
of colorectal cancer: a meta-analysis of prospective WCRF/AICR (2014). Continuous Update Project Report.
cohort studies. PLoS ONE, 10(3):e0120706. doi:10.1371/ Diet, nutrition, physical activity and breast cancer
journal.pone.0120706 PMID:25811460 survivors. Washington (DC), USA: American Institute
Ligibel JA, Cirrincione CT, Liu M, Citron M, Ingle JN, for Cancer Research. Available from: https://www.wcrf.
Gradishar W, et al. (2015). Body mass index, PAM50 org/sites/default/files/Breast-Cancer-Survivors-2014-
subtype, and outcomes in node-positive breast cancer: Report.pdf.
CALGB 9741 (Alliance). J Natl Cancer Inst, 107(9):
djv179. PMID:26113580
Majumder K, Gupta A, Arora N, Singh PP, Singh S
(2015). Premorbid obesity and mortality in patients
with pancreatic cancer: a systematic review and
meta-analysis. Clin Gastroenterol Hepatol, S1542-
3565(15):01341–5. PMID:26460214
Nakao M, Chihara D, Niimi A, Ueda R, Tanaka H,
Morishima Y, et al. (2014). Impact of being overweight
480
2.5 Sustained weight loss and [It is important to note that many of the
cancer risk: illustrative examples published trials and observational studies were
not designed to document weight loss, and
Studies investigating whether weight loss weight change may reflect both intentional
protects against cancer occurrence are limited to weight loss (with uncertainty about what exactly
a few observational studies on weight reduction the intervention was) and unintentional weight
in relation to breast cancer incidence and on the loss (which is potentially illness-induced).]
impact of intentional weight loss after bariatric
surgery on cancer risk in morbidly obese patients. 2.5.2 Studies of bariatric surgery and cancer
risk
2.5.1 Studies of weight loss and cancer risk
Several prospective intervention trials or
Few observational studies have been able retrospective cohort studies (Christou et al.,
to assess the impact of weight loss on cancer 2008; Adams et al., 2009; Sjöström et al., 2009;
risk. Women from the Nurses’ Health Study Ward et al., 2014) and reviews (Tee et al., 2013;
who had never used postmenopausal HRT and Maestro et al., 2015) have evaluated the effect
had lost 10 kg or more sustainably since meno- of bariatric surgery on cancer risk, comparing
pause [duration not reported] had a lower risk the risk of cancer in patients who underwent
of postmenopausal breast cancer than those who bariatric surgery with that in an obese control
maintained their weight since menopause (RR, group who did not undergo surgery (Table 2.5).
0.43; 95% CI, 0.21–0.86) (Eliassen et al., 2006). Overall, in most studies the risk of cancer at all
However, no association was found between sites in obese patients was significantly reduced
short-term (4-year) weight loss and subsequent after bariatric surgery. A 45% decrease in risk of
cancer risk in the same cohort (Rosner et al., all cancers combined was estimated in a recent
2015). In contrast, regardless of use of postmeno- meta-analysis (RR, 0.55; 95% CI, 0.41–0.73)
pausal HRT, adult weight loss was unrelated to (Tee et al., 2013). The extent of the cancer-pro-
postmenopausal breast cancer risk compared tective effect of bariatric surgery seems to be
with stable weight in the NIH-AARP study (Ahn more pronounced in women than in men: in the
et al., 2007), the EPIC-PANACEA study (Emaus Swedish Obese Subjects study, after a median
et al., 2014), and the Cancer Prevention Study follow-up of more than 10 years, the relative
II for the first 5 years of follow-up (Teras et al., risk was 0.58 (95% CI, 0.44–0.77) in women
2011); however, in the Cancer Prevention Study and 0.97 (95% CI, 0.62–1.52) in men (Sjöström
II an inverse association was suggested in women et al., 2009). Also, there are broadly consistent
who maintained a weight loss of 10 lb [4.5 kg] inverse associations with the subsequent risk of
or more for the next 4 years. Similarly, in the female sex hormone-sensitive cancers, notably
Women’s Health Initiative Dietary Intervention endometrial cancer and breast cancer (Adams
Trial, no effect of weight loss on postmenopausal et al., 2009; Tee et al., 2013; Ward et al., 2014).
breast cancer risk was found in overweight or [However, there were methodological problems
obese women (Neuhouser et al., 2015). in the study designs because of confounding by
indication, and failure to adequately capture the
extent of body weight reduction after bariatric
surgery.]
481
Studies using population-level registry

data (i.e. standardized population cohorts) for
comparison purposes have reported an increased
incidence of colorectal cancer in obese men who
underwent bariatric surgery compared with the
expected risk in the general population (Östlund
et al., 2010; Derogar et al., 2013). [Because the
general population was used as comparator, the
median BMI (not reported) would have been
considerably less than that for the treatment
group, and the observed increase in incidence
might reflect the effect of the premature morbidly
obese status rather than of the surgery itself.
Therefore, any comparison with the general
population may be misleading in the evaluation
of the effects of bariatric surgery on subsequent
cancer risk in obese patients.]
482
Table 2.5 Studies of obese patients who underwent bariatric surgery and subsequent cancer risk
Reference Study design Surgery group Control group Cancer site Surgery cases Relative risk Adjustments
Location Mean follow- (cohort) (95% CI) Comments
up (years) Control cases
(cohort)
Men and women
Christou et al. Retrospective Bariatric patients Diagnosis of “morbid All sitesa 21 (1035) 0.22 (0.14–0.35) Age, sex, BMI
(2008) hospital- in regional obesity” from 487 (5746)
Canada based database hospital records or
Maximum, BMI not available prescription
5.0 BMI unknown
Adams et al. Retrospective Roux-en-Y gastric State document All sitesb 254 (6596) 0.76 (0.65–0.89) Age, sex, BMI
(2009) registry bypass applicants with a self- 477 (9442) Data also reported for
Utah, USA 12.5 Mean BMI, 44.9 reported BMI > 35 “Obesity-related 104 (6596) 0.62 (0.49–0.78) the 31 individual cancer
Mean BMI, 47.4 sites”c 253 (9442) sites
Colorectum 25 (6596) 0.70 (0.43–1.15)
52 (9442)
Women
Adams et al. Retrospective Roux-en-Y gastric State document All sitesb 215 (5654) 0.73 (0.62–0.87) Age, BMI
(2009) registry bypass applicants with a self- 412 (7872)
Utah, USA Median, 12.5 Mean BMI, 44.9 reported BMI > 35 Breast 25 (5654) 0.91 (0.67–1.24)
Mean BMI, 47.4 52 (7872)
Premenopausal 49 (5654) 0.93 (0.63–1.37)
breast 65 (7872)
Postmenopausal 24 (5654) 0.96 (0.57–1.63)
breast 40 (7872)
Corpus uteri 14 (5654) 0.22 (0.13–0.40)

98 (7872)
Sjöström et al. Prospective Mean BMI, 42.2 Matched using 18 All sitesd 79 (1420) 0.58 (0.44–0.77) Age, smoking, weight
(2009) intervention anthropometric, 130 (1447) change, energy intake,
Sweden trial cardiovascular, and and matching
10.9 biochemical indices Also significantly
Mean BMI, 41.6 reduced for melanoma
and haematopoietic
cancers
Ward et al. Retrospective All female All female admissions Corpus uteri 424 (103 797) 0.29 (0.26–0.32) None
(2014) clinical data patients with with an associated 43 921 (7 328 061)
USA repository a history of diagnosis of obesity
Unknown bariatric surgery BMI unknown
BMI unknown
483
484

Table 2.5 (continued)
Reference Study design Surgery group Control group Cancer site Surgery cases Relative risk Adjustments
Location Mean follow- (cohort) (95% CI) Comments
up (years) Control cases
(cohort)
Men
Adams et al. Retrospective Roux-en-Y gastric State document All sitesb 39 (942) 1.02 (0.69–1.51) Age, BMI
(2009) registry bypass applicants with a self- 65 (1570)
Utah, USA 12.5 Mean BMI, 44.9 reported BMI > 35
Mean BMI, 47.4
Sjöström et al. Prospective Mean BMI, 40.6 Matched using 18 All sitesd 39 (590) 0.97 (0.62–1.52) Age, smoking, weight
(2009) intervention anthropometric, 39 (590) change, energy intake,
Sweden trial cardiovascular, and and matching
10.9 biochemical indices Results were not
Mean BMI, 39.2 statistically significant
for any of the
individual cancer sites
a Includes colorectum, pancreas, breast, endometrium, kidney, melanoma, myeloma, and non-Hodgkin lymphoma.
b Includes 31 cancer sites and “other”.
c Includes colorectum, oesophagus (adenocarcinoma), liver, gall bladder, pancreas, postmenopausal breast, corpus and uterus, kidney, non-Hodgkin lymphoma, leukaemia, and
multiple myeloma.
d Includes colorectum, stomach, liver, pancreas, kidney, bladder, lung and bronchia, haematopoietic system, and melanoma for both sexes, and breast, cervix, and endometrium in
women and prostate in men.

BMI, body mass index (in kg/m 2); CI, confidence interval
References Subjects Study): a prospective, controlled intervention

trial. Lancet Oncol, 10(7):653–62. doi:10.1016/S1470-
2045(09)70159-7 PMID:19556163
Adams TD, Stroup AM, Gress RE, Adams KF, Calle Tee MC, Cao Y, Warnock GL, Hu FB, Chavarro JE (2013).
EE, Smith SC, et al. (2009). Cancer incidence and Effect of bariatric surgery on oncologic outcomes: a
mortality after gastric bypass surgery. Obesity (Silver systematic review and meta-analysis. Surg Endosc,
Spring), 17(4):796–802. doi:10.1038/oby.2008.610 27(12):4449–56. doi:10.1007/s00464-013-3127-9
PMID:19148123 PMID:23949484
Ahn J, Schatzkin A, Lacey JV Jr, Albanes D, Ballard- Teras LR, Goodman M, Patel AV, Diver WR, Flanders WD,
Barbash R, Adams KF, et al. (2007). Adiposity, adult Feigelson HS (2011). Weight loss and postmenopausal
weight change, and postmenopausal breast cancer breast cancer in a prospective cohort of overweight and
risk. Arch Intern Med, 167(19):2091–102. doi:10.1001/ obese US women. Cancer Causes Control, 22(4):573–9.
archinte.167.19.2091 PMID:17954804 doi:10.1007/s10552-011-9730-y PMID:21327461
Christou NV, Lieberman M, Sampalis F, Sampalis Ward KK, Roncancio AM, Shah NR, Davis MA, Saenz CC,
JS (2008). Bariatric surgery reduces cancer risk McHale MT, et al. (2014). Bariatric surgery decreases the
in morbidly obese patients. Surg Obes Relat Dis, risk of uterine malignancy. Gynecol Oncol, 133(1):63–6.
4(6):691–5. doi:10.1016/j.soard.2008.08.025 PMID: doi:10.1016/j.ygyno.2013.11.012 PMID:24680593
19026373
Derogar M, Hull MA, Kant P, Östlund M, Lu Y, Lagergren
J (2013). Increased risk of colorectal cancer after
obesity surgery. Ann Surg, 258(6):983–8. doi:10.1097/
SLA.0b013e318288463a PMID:23470581
Eliassen AH, Colditz GA, Rosner B, Willett WC,
Hankinson SE (2006). Adult weight change and risk of
postmenopausal breast cancer. JAMA, 296(2):193–201.
doi:10.1001/jama.296.2.193 PMID:16835425
Emaus MJ, van Gils CH, Bakker MF, Bisschop CN,
Monninkhof EM, Bueno-de-Mesquita HB, et al.
(2014). Weight change in middle adulthood and breast
cancer risk in the EPIC-PANACEA study. Int J Cancer,
135(12):2887–99. doi:10.1002/ijc.28926 PMID:24771551
Maestro A, Rigla M, Caixàs A (2015). Does bariatric
surgery reduce cancer risk? A review of the litera-
ture. Endocrinol Nutr, 62(3):138–43. doi:10.1016/j.
endonu.2014.12.005 PMID:25637364
Neuhouser ML, Aragaki AK, Prentice RL, Manson JE,
Chlebowski R, Carty CL, et al. (2015). Overweight,
obesity, and postmenopausal invasive breast
cancer risk: a secondary analysis of the Women’s
Health Initiative randomized clinical trials. JAMA
Oncol, 1(5):611–21. doi:10.1001/jamaoncol.2015.1546
PMID:26182172
Östlund MP, Lu Y, Lagergren J (2010). Risk of obesity-
related cancer after obesity surgery in a popula-
tion-based cohort study. Ann Surg, 252(6):972–6.
doi:10.1097/SLA.0b013e3181e33778 PMID:20571362
Rosner B, Eliassen AH, Toriola AT, Hankinson SE, Willett
WC, Natarajan L, et al. (2015). Short-term weight gain
and breast cancer risk by hormone receptor classifica-
tion among pre- and postmenopausal women. Breast
Cancer Res Treat, 150(3):643–53. doi:10.1007/s10549-
015-3344-0 PMID:25796612
Sjöström L, Gummesson A, Sjöström CD, Narbro K,
Peltonen M, Wedel H, et al.; Swedish Obese Subjects
Study (2009). Effects of bariatric surgery on cancer
incidence in obese patients in Sweden (Swedish Obese
485
3. CANCER-PREVENTIVE EFFECTS IN
EXPERIMENTAL ANIMALS
3.1 Methodological considerations In energy restriction protocols, diets are

formulated so that animals fed different amounts
3.1.1 Definition of dietary/calorie restriction of calories still receive the same levels of other
Dietary/calorie restriction involves altering nutrients, such that the only variable is energy
food intake qualitatively or quantitatively to intake (Thompson et al., 2002, 2003; Thompson,
control body weight gain and presumably main- 2015). In other words, there is selective reduc-
tain body composition. Several terms have been tion in energy intake (nutrient density) while
used to describe the different methodological feeding the same level of micronutrients as those
approaches in which amount, nutrient, calorie, of the control group. Nutrient density is altered
or energy intake is restricted or modulated to in a manner that facilitates investigation of the
control body weight (Thompson et al., 2002, effects of energy restriction without changing
2003; Thompson, 2015). These terms are not the amounts of nutrients and other dietary
synonyms, and it is important to be aware of factors. These studies may provide information
the methodological distinctions between the most relevant to understanding the mechanisms
approaches when designing experiments. involved. Terms used to describe this approach
Dietary restriction protocols refer to feeding a include calorie restriction (CR), caloric restric-
reduced amount of a complete diet, such that less tion, dietary energy restriction, and energy
total nutrients and dietary factors are ingested. If restriction. In this section, the term dietary
excessive, this approach can lead to an intake of restriction (DR) will be used to include both
micronutrients and/or macrocomponents that is dietary and calorie restriction.
incompatible with optimal health and survival.
In early studies, this was the most commonly 3.1.2 Design issues in studies of dietary
used approach to study cancer prevention. These restriction
studies have provided valuable insights, but DR has been one of the most widely used
the results should be interpreted with caution, methods for studying cancer prevention in exper-
because this approach does not allow the detec- imental models. In DR protocols, control and
tion of effects that are specifically due to dietary carcinogen-treated animals are diet-restricted
modulation. Terms used to describe this approach by amounts that result in lower body weights
include dietary restriction, food restriction, and relative to those of controls fed ad libitum (AL),
total dietary restriction. without weight loss. Also, diet-restricted animals
487
generally survive longer (Maeda et al., 1985; Yu (b) Selection of model

et al., 1985). Characteristics that should be considered
However, inherent within DR protocols are when choosing a model to study cancer in humans
factors that could confound the assessment of include the similarity of tumour morphology
tumorigenesis. An important factor in carcino- and biological traits (e.g. hormonal responsive-
genesis studies is that the sensitivity of bioassays ness) to those observed in humans. Such models
to detect xenobiotics-induced carcinogenic are available for many organ sites. For example,
responses may be altered by DR. A chemical that in rat models for breast cancer, not only are the
is observed to be toxic or carcinogenic in animals tumours in the rat morphologically similar to
fed AL might not produce the same effects in tumours in humans, but the majority are ovarian
diet-restricted or otherwise leaner animals, i.e. steroid-responsive. N-nitrosobis(2-oxopropyl)
carcinogenic activity might be underestimated amine (BOP)-induced ductular pancreatic cancer
in the diet-restricted animals. In addition, DR in the Syrian hamster is a model that has been
may induce a variety of pleiotropic responses useful in studying aspects of pancreatic cancer in
that affect metabolism, distribution, and dispo- humans (Pour et al., 1993), whereas other carcin-
sition of xenobiotics. Furthermore, because ogen-induced pancreatic tumours are acinar cell
diet-restricted animals live longer than those fed carcinomas, which are not a common lesion in
AL, evaluations performed at age 2 years result humans. There is a variety of animal models
in comparisons at disproportionate times in the for cancers of the mammary gland, colon, liver,
respective lifespans (NTP, 1997). prostate, pancreas, skin, and pituitary gland,
(a) Experimental systems and cancers of the haematopoietic system,
including lymphoma and leukaemia. In earlier
Animal models have been used to study how studies, spontaneous tumour models were used.
DR modulates the development and/or progres- Increasingly in recent decades, researchers have
sion of cancer in humans. Typically, animals are used transgenic mice to model cancer in humans
fed diets that result in 20–40% reduction in daily and to determine whether dietary interventions
food/calorie intake relative to controls fed AL can prevent cancer development driven by genes
while maintaining adequate nutrition (Klurfeld known to be mutated in human cancer (Hursting
et al., 1989a; Kritchevsky, 1999; Hursting et al., et al., 2011). These models are described in the
2010, 2013). The animals have lower body weight, subsequent sections.
proportionately smaller skeletal size, lower
percentage of total body fat, and decreased weight (c) Selection of intervention protocol
of most internal organs (with the exception of When a chemical carcinogen is used in DR
the brain and testes) compared with animals fed models, the timing of carcinogen administration
AL, indicating that malnutrition is not involved and dietary intervention needs to be taken into
(Keenan et al., 2013). DR models are designed consideration. The intervention should gener-
to directly parallel and model conditions in ally be separate from the administration of the
humans. The range of DR controls body weight chemical carcinogen. However, it is important
and is intended to resemble the range of energy to assess the impact of the intervention both on
balances that occur in healthy people. the cancer induction phase (before and at the
time of treatment with the carcinogen) and on
the promotion/progression (development) of the
cancer (after treatment with the carcinogen).
488
Models involving a short induction phase provide humans. In the first regimen, animals are fed a
the ability to assess the impact of diet on early or restricted amount of energy (i.e. energy restriction
late stages of promotion. If tumours are allowed to various extents) and continue to gain weight,
to develop before the intervention, it is possible but at a slower rate than animals fed AL. This
to assess the impact of the intervention on the DR does not usually cause weight loss, because
regression or progression of the lesions. When restriction is started shortly after weaning. In
spontaneous or genetically modified models are the second regimen, when DR is started in older
used, the dietary intervention can be applied animals, restriction initially causes weight loss,
at different ages, depending on the individual but then diet-restricted animals maintain a body
model (Thompson et al., 2002; Everitt & Alder, weight that is lower than that of animals that
2013). consume food AL. The third regimen mimics a
cyclic pattern of dieting. Animals are subjected
(d) Selection of diet to intermittent and repetitive periods of DR or
Laboratory rodents can be fed three types of total fasting that result in alternating patterns
diets: (i) non-purified, (ii) purified or semi-puri- of weight loss and weight gain (Thompson et al.,
fied, and (iii) chemically defined (Everitt & Alder, 2002; Cleary & Grossmann, 2011).
2013; Lipman & Leary, 2015). Purified diets are
the most widely used in cancer-related studies 3.2 Overview of the effects of excess
in rodents. These diets should meet and label
the minimum requirements for protein and fat body weight
and maximum levels of fibre and ash; however, 3.2.1 Obesity models
the percentages of the various macronutrients
can vary. Purified diets (previously known as The use of rodent models to study excess
semi-synthetic or semi-purified diets) are formu- body weight and associated diseases in humans
lated using refined ingredients, including sugars, has steadily increased (Kanasaki & Koya, 2011).
proteins, carbohydrates, and fats, with added These models have several modulating factors
mineral and vitamin mixtures. that include species, strain, age of animals, type
Non-purified and purified diets have been of diet, level of fat, and type of control diet;
used in studies of dietary impact on tumorigen- inflammation, metabolic status, and endocrine
esis. In these types of studies, it is very impor- status may be associated confounding factors.
tant that the diet of both the intervention and Ray & Cleary (2013) and Cleary (2013) have
the control group be adequate in all nutrients. published comprehensive reviews on the use of
Non-purified diets have the advantage that they such animal models.
are formulated using whole food ingredients. Most mouse models used to study obesity
Purified diets have the advantage that each and cancer are genetically manipulated (trans-
component can be changed independently of genic) animals: animals are either genetically
other constituents in a highly controlled fashion. modified to induce carcinogenicity and fed a
In most rodent models, DR administered modified diet to induce obesity, or genetically
throughout life appears to be more effective in modified to induce obesity and administered
controlling body weight than regimens started in chemicals to induce cancer. There are several
adult animals (Ross & Bras, 1971; Hursting et al., genetic mutations that result in obesity. One
2010, 2013). There are three commonly used common disturbance is in the function of leptin,
animal DR regimens that model approaches to a critical anorexigenic adipokine that conveys
and patterns of body weight regulation in adult information about adipose status; leptin levels
489
in serum/plasma are elevated in proportion Other methods for inducing obesity in

to adipose tissue mass. Genetically obese mice rodents include surgical removal of the ovaries
include Avy yellow obese mice, leptin-deficient to induce weight gain and a postmenopausal-like
C57BL/6L-LepobLepob (originally termed ob/ob) state (Nkhata et al., 2009), or damaging the
mice, and leptin receptor-deficient LeprdbLeprdb hypothalamus by injection of gold thioglucose
(originally termed db/db) mice. LepobLepob mice (GTG), which results in overeating, rapid weight
are homozygous recessive and do not produce gain, and subsequent obesity (Bergen et al., 1998).
leptin (Zhang et al., 1994), whereas LeprdbLeprdb These models have been used to study mammary
mice have a defect in the leptin receptor (OB-R) tumorigenesis (see Section 3.2.2).
and manifest high circulating levels of leptin
(Frederich et al., 1995). Both strains are obese at 3.2.2 Cancer of the mammary gland
a young age, and develop hyperinsulinaemia and
insulin resistance. The Avy yellow obese mouse See Table 3.1.
has mutations in the agouti gene that cause ubiq- Various rodent models have been used to
uitous expression of the agouti protein, which assess the association between obesity and
results in appetite stimulation, leading to hyper- cancer of the mammary gland, including genetic,
insulinaemia (Wolff et al., 1999). diet-induced, and GTG-induced obesity models.
Genetically obese rat strains, such as the The Avy yellow obese mouse model was one of the
Zucker and Corpulent strains that have leptin first genetic obesity models used to study both
receptor defects, have also been used in cancer spontaneous and chemically induced cancer of
studies. The Zucker “fatty” rat carries a mutation the mammary gland. In an early study, the time
on the Lepr gene for obesity, which is inherited as to tumour detection (latency period) for spon-
a Mendelian recessive trait and leads to extreme, taneous mammary tumours initiated by the
early-onset obesity by age 3 weeks (Zucker & mouse mammary tumour virus (MMTV) was
Zucker, 1961). determined in breeding and virgin Avy mice. The
The diet-induced obesity (DIO) model is also incidence of spontaneous mammary tumours
used to study the interplay between cancer and was 96–100% by 8 months in virgin obese mice,
increased body weight in mice and rats. Although whereas it reached 100% at 15 months in virgin
this model is generally considered to be closest to lean mice (Heston & Vlahakis, 1961, 1962). The
the development of obesity in humans, the utility difference disappeared in the breeding mice, with
of the model is limited by the fact that humans approximately 100% incidence at 8 months in
do not normally consume extremely high quan- both lean and obese mice. In another study using
tities of fat in their diets. In DIO studies, the fat Avy mice, the time course of appearance of hyper-
content of the diets is increased from 10% of total plastic alveolar nodules and mammary tumours
calories to 30–60% of total calories. High-fat diets was determined in virgin “viable yellow” (Avy/A)
(HFDs) also contain higher calorie density than [obese] and non-yellow (A/a) (C3H/HeNIcrWf ×
control or standard rodent chow, and therefore VY/Wf)F1 [lean] female mice. Hyperplastic alve-
are in fact high-fat, high-calorie diets. Although olar mammary nodules occurred by age 16 weeks
the response is species- and strain-dependent, the in virgin yellow obese mice compared with age
animals gain weight rapidly and develop other 19 weeks in their non-yellow lean counterparts.
health complications, such as glucose intolerance In addition, the incidence of hyperplastic alve-
and insulin resistance or diabetes. This response olar nodules was increased among yellow obese
may be species- and strain-specific, because not females compared with non-yellow lean females
all species gain weight when placed on a HFD. by age 36 weeks. Mammary adenocarcinomas
490
Table 3.1 Effect of obesity on the development of mammary tumours in mice and rats
Species Obesity model Cancer etiology Results (obese vs lean animals) Reference
Mouse A yellow obese
vy MMTV Shortened latency until 100% Heston &
incidence of tumours Vlahakis (1961,
1962)
Mouse Avy yellow obese Spontaneous Shortened latency for Wolff et al.
hyperplastic alveolar nodules and (1979)
adenocarcinomas
Mouse Avy yellow obese DMBA Shortened latency and increased Wolff et al.
incidence of tumours (1982)
Mouse LepobLepob Spontaneous Shortened latency but decreased Heston &
incidence of tumours Vlahakis (1962)
Mouse LepobLepob Transgenic MMTV- No tumours; increased incidence Cleary et al.
TGF-α of tumours in wild-type and (2004c)
heterozygous mice (see text)
Rat Zucker rat MNU No effect on latency of tumours; Lee et al. (2001)
decreased incidence of carcinomas
Rat Zucker fa/fa rat; LA/Ncp DMBA Shortened latency and increased Klurfeld et al.
corpulent rat incidence of tumours (1991); Hakkak
et al. (2005)
Rat Zucker rat, ovariectomized DMBA Shortened latency and increased Hakkak et al.
incidence of tumours (no tumours in (2007)
lean animals)
Mouse GTG-induced Spontaneous (C3H) Shortened latency until 50% incidence Waxler et al.
of tumours (1953)
Mouse GTG-induced Implantation of T47-D Increased incidence of tumours Nkhata et al.
human breast cancer cells (2009)
Mouse Diet-induced (33% fat diet; Transgenic MMTV- Shortened latency and increased Cleary et al.
mice divided into groups TGF-α (C57BL/6) incidence of palpable tumours; some (2004a); Dogan
based on weight gain) high-grade adenocarcinomas et al. (2007)
Mouse Diet-induced (33% fat diet; Transgenic MMTV-neu No effect on latency or incidence Cleary et al.
mice divided into groups (FVB/N) of tumours; earlier onset of second (2004b); Khalid
based on weight gain) tumours, increased multiplicity et al. (2010)
Mouse Diet-induced (5.2 kcal/g or Implantation of Significantly increased tumour Nuñez et al.
3.8 kcal/g) mammary tumour cells volume and growth rate (2008)
from Wnt-1 transgenic
mice
Mouse Diet-induced, Implantation of Increased tumour volume Rossi et al.
ovariectomized mammary tumour cells (2016)
from Wnt-1 transgenic
mice
Rat Obesity-prone Sprague- MNU Shortened latency; increased tumour Matthews et al.
Dawley incidence and tumour weight (2014)
DMBA, 7,12-dimethylbenz[a]anthracene; GTG, gold thioglucose; MMTV, mouse mammary tumour virus; MNU, N-methyl-N-nitrosourea;
TGF-α, transforming growth factor alpha.
Adapted from Ray & Cleary (2013) by permission from Springer Nature, © 2013.
491
were first observed at an earlier age in yellow tumours when treated with N-methyl-N-
obese mice than in non-yellow lean mice (Wolff nitrosourea (MNU) compared with lean rats (Lee
et al., 1979). et al., 2001). However, two strains of genetically
In a study of chemically induced mammary obese rats, the LA/Ncp Corpulent rat and the
tumours, obese Avy (BALB/c) mice aged 8 weeks Zucker fa/fa rat, administered DMBA had higher
treated with 1.5 mg/kg of 7,12-dimethylbenz[a] incidences and significantly shorter latency of
anthracene (DMBA) weekly for 2 weeks or mammary tumours by approximately 112 days
with 6.0 mg/kg of DMBA weekly for 6 weeks and 150 days after administration, respectively,
had higher incidences of mammary tumours compared with lean rats. Whereas multiplicity
compared with the lean mice at both doses. In and tumour size were increased in Corpulent
addition, the latency period was shorter for the rats, such effects were not observed in Zucker fa/
Avy mice than for the lean mice at both doses of fa rats (Klurfeld et al., 1991; Hakkak et al., 2005).
DMBA (Wolff et al., 1982). Lean and obese ovariectomized Zucker rats
The LepobLepob mouse is another model that were treated with DMBA at 50 days and killed
has been used to study obesity and cancer. In an at 135 days after DMBA treatment. Obese rats
early study, obese LepobLepob female mice had had higher incidences and shorter latency of
decreased incidences of spontaneous mammary mammary tumours compared with lean rats,
tumours compared with lean mice; however, which did not develop any mammary tumours
tumours were first detected at an earlier age, (Hakkak et al., 2007).
10.7 months for obese mice versus 17.6 months Chemically induced obesity has been another
for lean mice (Heston & Vlahakis, 1962). approach to study increased body weight and
In another study, obese LepobLepob female cancer of the mammary gland. Injection of GTG
mice crossed with transgenic mice overex- to destroy the hypothalamus results in over-
pressing human transforming growth factor eating and obesity (Bergen et al., 1998). Obese
alpha (MMTV-TGF-α) were used as a model C3H mice injected with 10 mg of GTG at age
for postmenopausal mammary tumorigenesis. 2–3 months to induce obesity developed 50%
The MMTV-TGF-α strain develops 30% inci- incidence of spontaneous mammary tumours by
dence of mammary tumours by age 16 weeks 295 days, whereas the incidence was only 19% in
and is useful in assessing tumour incidence lean control mice (Waxler et al., 1953). In another
and latency. MMTV-TGF-α/LepobLepob mice did study, ovariectomized mice aged 6 weeks were
not develop mammary tumours by age 2 years. treated with GTG, followed 4 weeks later by
However, the incidence was 50% for wild-type implantation of the T47-D human breast cancer
mice and 67% for heterozygous mice (Cleary estrogen-positive cell line with and without
et al., 2004c). Similar results were obtained with estrogen implants. When assessed at 30 weeks,
the leptin receptor-deficient (LeprdbLeprdb) model GTG-obese mice without estrogen implants had
(Cleary et al., 2004d). [The lack of development 100% tumour incidence, compared with 50% for
of mammary tumours in these two genetically GTG-lean controls, 20% for lean vehicle controls,
obese mouse strains has been attributed to prob- and 0% for GTG-obese mice with estrogen
lems with basic mammary gland development implants (Nkhata et al., 2009).
as well as the now-known involvement of leptin HFDs have also been used to investigate
signalling in tumorigenesis.] mammary tumour development in rats and
The Zucker rat is another genetic obesity mice (Cleary, 2013; Ray & Cleary, 2013). In one
model that has been used to study mammary approach, tumour-prone MMTV-TGF-α trans-
tumours. Zucker rats developed fewer mammary genic C57BL/6 mice aged 10 weeks were fed the
492
same HFDs and then divided into groups based 3.2.3 Cancer of the colon
on whether they gained weight (obesity-prone) or
did not gain weight (obesity-resistant). All groups See Table 3.2.
had incidences of mammary tumours between Elevated body weight and obesity in associa-
72% and 82%; however, obesity-prone mice tion with cancers of the colon and intestine have
developed mammary tumours at an earlier age been investigated in several transgenic and DIO
than obesity-resistant or control (low-fat) mice. rodent models. Genetically obese LepobLepob and
In addition, some obesity-prone mice developed LeprdbLeprdb mice treated with azoxymethane
a more malignant variant of mammary adeno- (AOM) or MNU to induce cancer of the colon
carcinoma (Cleary et al., 2004a; Dogan et al., had increases in the multiplicity of pre-neoplastic
2007). In MMTV-neu mice, tumour latency aberrant crypt foci in the colon compared with
was similar in mice fed a HFD compared with control lean mice (Hirose et al., 2004; Hayashi
a low-fat diet (LFD) (Cleary et al., 2004b; Khalid et al., 2007; Bobe et al., 2008; Ealey et al., 2008).
et al., 2010); however, twice as many HFD mice Similar effects were observed in studies with
developed a second tumour, compared with obese gastrin gene knockout (GAS-KO) mice
LFD mice (Khalid et al., 2010). In another study, (Cowey et al., 2005), KK-Ay mice (derived from
C57BL/6 mice were fed either a high-calorie diet Avy yellow obese mice) (Teraoka et al., 2011), and
with 5.2 kcal/g (obese) or 3.8 kcal/g (overweight) Zucker rats (Raju & Bird, 2003) administered
or a 30% CR diet (lean). The mice were inocu- AOM. In studies with Zucker obese (fa/fa) and
lated with mammary tumour cells from Wnt-1 lean (Fa/Fa) rats, tumours of the colon induced by
transgenic mice. Tumour volume and growth administration of AOM or MNU were observed
rate were higher in obese mice and overweight in the obese rats, whereas none occurred in the
mice than in lean animals (Nuñez et al., 2008). lean rats (Weber et al., 2000; Lee et al., 2001; Ray
Ovariectomized female C57BL/6 mice were & Cleary, 2013).
fed a control diet or a DIO regimen, resulting Mice harbouring mutations in the adenoma-
in a normal weight or an obese phenotype, tous polyposis coli (Apc) gene develop tumours of
respectively. At week 24, mice were injected the intestine and colorectum and have also been
with MMTV-Wnt-1 mouse mammary tumour used to study the association between obesity
cells. At 36 months, mean tumour volumes were and colorectal cancer. When Apc1638N/+ mice
higher in DIO mice than in control animals were crossed with genetically obese Leprdb mice,
(Rossi et al., 2016). the resulting obese mice developed increased
Matthews et al. (2014) used a somewhat numbers of colon adenomas by age 6 months,
different approach, with administration of compared with non-obese Apc mice, which did
MNU. Sprague-Dawley rats that had been bred not develop tumours (Gravaghi et al., 2008).
to be obesity-resistant or obesity-prone were fed In male and female C57BL/6 mice fed a
a moderately HFD from age 20 days and followed high-calorie diet followed by subcutaneous
up for development of mammary tumours. injection of the MC38 colon carcinoma cell line,
Tumour incidence was significantly higher in obese mice had increased numbers of palpable
the obesity-prone rats (91.1%) than in the obesi- tumours and significantly higher average tumour
ty-resistant rats (65.1%). In addition, tumour size compared with non-obese mice (Yakar et al.,
weight was increased and latency was shortened 2006; Algire et al., 2010).
in obesity-prone rats, compared with obesity-
resistant rats.
493
Table 3.2 Effect of obesity on the development of colon tumours and pre-neoplastic lesions in
mice and rats
Species Obesity model Cancer etiology Results (obese vs lean animals) Reference
Mouse Lep Lep
ob ob AOM Increased multiplicity of ACF Hirose et al. (2004); Hayashi et al. (2007);
Bobe et al. (2008); Ealey et al. (2008)
Mouse LeprdbLeprdb AOM Increased multiplicity of ACF Hirose et al. (2004); Hayashi et al. (2007);
Ealey et al. (2008)
Mouse LeprdbLeprdb MNU Increased multiplicity of ACF Ealey et al. (2008)
Mouse Gastrin gene AOM Increased multiplicity of ACF Cowey et al. (2005)
knockout
(GAS-KO)
Mouse KK-Ay AOM Increased multiplicity of ACF Teraoka et al. (2011)
at age 13 wk; increased tumour
incidence at age 19 wk
Mouse LeprdbLeprdb Apc1638N/– Increased incidence of colon Gravaghi et al. (2008)
tumours at age 6 mo (no tumours
in lean animals)
Rat Zucker AOM Increased multiplicity of ACF Raju & Bird (2003)
Rat Zucker (fa/fa) AOM Increased multiplicity of ACF Weber et al. (2000)
and incidence of colon tumours
Rat Zucker (fa/fa) MNU Increased incidence of colon Lee et al. (2001)
tumours
Mouse Diet-induced s.c. injection Increased number and size of Yakar et al. (2006); Algire et al. (2010)
of MC38 colon colon tumours
carcinoma cell line
ACF, aberrant crypt foci; AOM, azoxymethane; mo, month or months; MNU, N-methyl-N-nitrosourea; s.c., subcutaneous; wk, week or weeks.
3.2.4 Cancer of the liver Lepob mouse is a congenic obese strain that
develops spontaneous hepatocellular adenomas
See Table 3.3. and carcinomas at a younger age and a higher
Several genetically modified animal models incidence than either parental strain (FLS and
have been used to study the link between obesity Lepob/Lepob) (Soga et al., 2010).
and cancer of the liver. For example, genetically The outbred obese, diabetic, male Swiss-
obese yellow agouti (Avy) mice and LepobLepob Webster mouse is another model for studying
mice had increased incidences of liver tumours, the link between obesity and cancer of the liver.
which also developed at a younger age in obese These mice are polyuric, polydipsic, glucosuric,
mice compared with lean mice (Heston & and hyperglycaemic. Compared with their lean
Vlahakis, 1961, 1962). counterparts, they develop a high incidence of
The genetically obese LepobLepob mouse is late-onset HCC (Lemke et al., 2008). Strain–diet
often used as an animal model for non-alcoholic interactions are another important consideration
fatty liver disease in humans. These mice have for genetically controlled, diet-induced HCC.
increased incidences of hepatocellular carci- For example, male C57BL/6J mice made obese by
noma (HCC) and of focal hepatocyte hyperplasia feeding a HFD developed HCC, compared with
(considered to be a pre-neoplastic lesion) at an none in mice fed a LFD; in contrast, a HFD had
earlier age compared with lean littermates (Yang little effect on A/J mice similarly treated (Hill-
et al., 2001). The fatty liver Shionogi (FLS)-Lepob/ Baskin et al., 2009).
494
Table 3.3 Effect of obesity on the development of liver tumours and pre-neoplastic lesions in
mice and rats
Species Obesity model Cancer Results (obese vs lean animals) Reference

etiology
Mouse Avy Spontaneous Shortened latency and increased incidence of Heston & Vlahakis
tumours (1961)
Mouse LepobLepob Spontaneous Shortened latency and increased incidence of Heston & Vlahakis
tumours (1962)
Mouse LepobLepob Spontaneous Shortened latency and increased incidence of Yang et al. (2001)
tumours and of focal hepatocyte hyperplasia
Mouse LepobLepob crossed with Spontaneous Increased incidence of hepatocellular adenoma Soga et al. (2010)
fatty liver Shionogi and carcinoma at age 12 mo
(FLS)
Rat Obese, diabetic Swiss- Spontaneous Increased incidence of late-onset hepatocellular Lemke et al. (2008)
Webster carcinoma (male mice only)
Mouse Diet-induced C57BL/6 Spontaneous Increased incidence of hepatocellular Hill-Baskin et al.
(58% fat) carcinoma (none in lean mice) (2009)
Mouse Diet-induced A/J Spontaneous No effect Hill-Baskin et al.
(58% fat) (2009)
Mouse LepobLepob DEN Increased incidence and number and size of Park et al. (2010)
tumours
Mouse Diet-induced C57BL/6 DEN Increased incidence and number and size of Park et al. (2010)
(58% fat) tumours
Mouse Diet-induced IL6−/−; DEN + Induction of hepatocellular carcinoma without Park et al. (2010)
TNFR1−/− phenobarbital phenobarbital promotion in mice fed high-fat
(59% fat) promotion diet but not in mice fed low-fat diet
DEN, diethylnitrosamine; mo, month or months.
Carcinogen-induced protocols involve induc- 3.2.5 Cancer of the prostate

tion of hepatocellular tumours by administration
of diethylnitrosamine (DEN). After administra- See Table 3.4.
tion of DEN, DIO C57BL6 mice and genetically There has been a lack of suitable experimental
obese LepobLepob mice had increased incidence, animal models that mimic the development and
numbers, and size of hepatocellular tumours, progression of cancer of the prostate in humans.
compared with lean mice fed a LFD. When IL-6- Therefore, few published studies have reported
deficient (IL6−/−) and tumour necrosis factor the long-term effects of a HFD and/or obesity
receptor-deficient (TNFR1−/−) mice were fed on the development of prostate cancer. In recent
either a LFD or a HFD and treated with DEN, years, several genetically engineered mouse
the majority of the mice fed the HFD devel- models of prostate cancer have been developed
oped HCC without phenobarbital promotion. In (Parisotto & Metzger, 2013). The model that is
contrast, DEN-treated mice fed the LFD did not most commonly used is the transgenic adeno-
develop HCC unless treated with phenobarbital carcinoma of the mouse prostate (TRAMP)
(Park et al., 2010). mouse model. When maintained on a C57BL/6
genetic background, male TRAMP mice develop
prostatic intraepithelial neoplasia (PIN) (pre-
neoplastic lesions) in all lobes between age
495
Table 3.4 Effect of obesity on the development of prostate tumours and pre-neoplastic lesions in
mice
Obesity model Cancer etiology Results (obese vs lean animals) Reference

Diet-induced, “Western TRAMP mice More advanced disease; increased Llaverias et al. (2010)
diet” (40% fat vs chow) percentage of metastasis
Diet-induced (33% fat) TRAMP mice More advanced disease; higher Bonorden et al. (2012)
incidence of metastasis (see text for
comment)
GTG-induced TRAMP mice Less advanced disease; lower percentage Bonorden et al. (2012)
of metastasis
Diet-induced in C57BL/6 Implantation of TRAMP-C2 Faster tumour growth Bonorden et al. (2012)
cells
Diet-induced (60% fat) Transgenic Hi-Myc mice More advanced disease Blando et al. (2011)
Diet-induced s.c. injection of RM1 prostate Larger tumours Ribeiro et al. (2010)
carcinoma cell line
LepobLepob s.c. injection of RM1 prostate Larger tumours Ribeiro et al. (2010)
carcinoma cell line
LeprdbLeprdb s.c. injection of RM1 prostate Smaller tumours Ribeiro et al. (2010)
carcinoma cell line
GTG, gold thioglucose; s.c., subcutaneous; TRAMP, transgenic adenocarcinoma of the mouse prostate.
2 months and age 3 months, and progression to poorly differentiated tumours, than mice that
poorly differentiated neuroendocrine carcinoma weighed less, and a higher incidence of metas-
occurs by age 4–7 months. tasis. [It should be noted that these results were
When fed a “Western-type” diet, enriched not statistically significant and also that the
in both fat and cholesterol, TRAMP-C57BL6 aggressiveness of the development of prostate
mice had accelerated tumour incidence and cancer in the TRAMP mouse affects its useful-
burden compared with mice fed a control chow ness in this type of study.]
diet. Mice fed the Western-type diet had more Implantation of TRAMP-C2 cells into
advanced disease, characterized by highly inva- diet-induced obese mice showed similar effects.
sive and less well differentiated tumours and In contrast, GTG-induced obesity in TRAMP
fewer high-grade PIN, in contrast to the chow-fed mice led to a reduction in disease progression
mice, which had only high-grade PIN. Increased and metastasis (Bonorden et al., 2012).
incidences of metastasis to the lung (67% vs 43%) The Hi-Myc transgenic mouse model of pros-
were also observed in mice fed the Western-type tate cancer was used to study the effect of modu-
diet (Llaverias et al., 2010). lating dietary energy balance on the development
In another study, TRAMP mice were fed and progression of prostate cancer. The mice
a moderately HFD (33% of calories from fat) were placed on one of three diets: 30% CR; a
from age 7 weeks, and at age 18 weeks they were modified AIN-76A diet with 10% of calories from
divided into obesity-resistant, overweight, and fat (overweight); or a DIO diet with 60% of calo-
obesity-prone groups and were then followed ries from fat (obese). All three groups had similar
up until age 50 weeks (Bonorden et al., 2012). incidences of hyperplasia and low-grade PIN at
An LFD group was also included. Obesity- age 3 months and 6 months. The CR group had
prone mice tended to have more severe lesions, significantly reduced incidence of in situ adeno-
including a higher incidence of moderate and carcinomas at 3 months compared with the DIO
496
Table 3.5 Effect of obesity on the development of skin tumours in mice

Lep Lep
ob ob s.c. injection of Increased number of metastatic tumour foci in Mori et al. (2006)
B16BL6 mouse the lung
melanoma cells
LeprdbLeprdb s.c. injection of Increased number of metastatic tumour foci in Mori et al. (2006)
B16BL6 mouse the lung
melanoma cells
LepobLepob s.c. injection of Significantly larger tumours Brandon et al. (2009)
B16F10 melanoma
cells
MC4R−/− s.c. injection of Significantly larger tumours Brandon et al. (2009)
B16F10 melanoma
cells
Diet-induced: HFD s.c. injection of Time to tumour formation similar, but tumours Pandey et al. (2012)
B16F10 mouse progressed more rapidly; increased tumour
melanoma cells weight and volume
Diet-induced: pelleted + SKH-1 hairless mice + Shortened latency and increased multiplicity: Dinkova-Kostova et al.
powdered (overweight) or UV radiation obese > overweight > lean (2008)
powdered (obese)
HFD, high-fat diet; s.c., subcutaneous; UV, ultraviolet.
group, and at 6 months compared with both the 3.2.6 Cancer of the skin (melanoma)
overweight group and the DIO group. The DIO
regimen also significantly increased (P = 0.02) See Table 3.5.
the incidence of invasive adenocarcinoma (95%), Several studies have been performed to assess
compared with the overweight group (65%) and the effect of obesity on the progression and metas-
with the 30% CR group (no invasive adenocarci- tasis of skin tumours. In one study, B16BL6 mela-
nomas) (Blando et al., 2011). noma cells were injected into the tail vein of male
Cancer cell lines have also been used to study genetically obese LepobLepob and LeprdbLeprdb
the induction and progression of prostate cancer. mice to assess the effects of obesity on metastasis.
Obese LepobLepob and LeprdbLeprdb mice, DIO At 14 days after injection, the number of meta-
mice, and control C57BL/6J mice were injected static tumour foci was significantly increased in
subcutaneously with RM1, a murine andro- the lungs of both obese strains compared with
gen-insensitive prostate carcinoma cell line, and control C57BL/6 mice (Mori et al., 2006).
evaluated 14 days after inoculation. The tumours In another study, obese LepobLepob mice, obese
induced in the obese LepobLepob mice and the melanocortin receptor 4 knockout MC4R−/−
DIO mice were significantly larger (P < 0.001), mice, lean wild-type mice, and pair-fed lean
and those induced in the LeprdbLeprdb mice were Lepob−/− mice were injected subcutaneously with
significantly smaller (P = 0.047) than those in the B16F10 melanoma cells. The resulting tumours
controls (Ribeiro et al., 2010). were significantly larger in the obese LepobLepob
(5.1 ± 0.9 g) and MC4R−/− mice (5.1 ± 0.7 g) than
in the lean wild-type mice (1.9 ± 0.3 g) or the
pair-fed Lepob–/– mice (0.95 ± 0.2 g) (Brandon
et al., 2009).
497
Table 3.6 Effect of obesity on the development of pancreatic tumours and pre-neoplastic lesions
in mice
Obesity Cancer etiology Results (obese vs lean animals) Reference

model
LepobLepob s.c. injection of PAN02 pancreatic Larger tumours; increased incidence of Zyromski et al. (2009)
and tumour cells metastases
LeprdbLeprdb
Diet-induced s.c. injection of PAN02 pancreatic Larger tumours and increased tumour White et al. (2010)
(60% fat) tumour cells weight
Diet-induced KrasG12D More advanced PanIN lesions Dawson et al. (2013)
(HFHC diet)
Diet-induced KrasG12D + Ink4a deficiency: LSL-Kras/ Increased number of PanIN, more Lashinger et al. (2013)
(HFHC diet) Pdx-1-Cre/Ink4a/Arf advanced PanIN, increased number of
PDAC, and increased number of pancreatic
desmoplastic (fibrotic) stroma
Diet-induced KrasG12D with or without COX Increased numbers of PanIN and PDAC in Philip et al. (2013)
(HFHC diet) conditional knockout: LSL-Kras/ all 3 models
Ela-CreERT; COXKO/LSL-Kras/Ela-
CreERT and LSL-Kras/Pdx-1-Cre
HFHC, high-fat, high-calorie; PanIN, pancreatic intraepithelial neoplasia; PDAC, pancreatic ductal adenocarcinoma; s.c., subcutaneous.
In another study, male C57BL/6J mice made tumour latency and an increased multiplicity of
obese by feeding a HFD for 6 months and subse- squamous cell carcinoma/papilloma compared
quently injected subcutaneously with B16F10 with the control group; the overweight group had
murine melanoma cells developed significantly intermediate values.
larger tumours compared with LFD control
mice (Pandey et al., 2012). Although there was 3.2.7 Cancer of the pancreas
no noticeable difference in the time of initiation
of tumour formation between the two groups, See Table 3.6.
tumours in HFD mice progressed more rapidly After BOP treatment, increased incidence
than those in controls. The average tumour and multiplicity of cancer of the pancreas was
weight was 3.52 g in HFD mice and 0.92 g in reported in an early study in Syrian hamsters
control mice, and the average tumour volume fed a HFD compared with those fed a LFD (Birt
was 1920 mm3 in HFD mice and 924 mm3 in et al., 1981). More recent studies assessing the
control mice. link between obesity and pancreatic cancer have
One study involved the induction of skin used inoculation of cell lines and have examined
cancer by whole-body exposure to ultraviolet progression rather than the induction and devel-
radiation. SKH-1 hairless mice made obese by opment of pancreatic tumours. Obese LepobLepob
feeding a powdered AIN-76A diet exclusively and LeprdbLeprdb mice injected subcutaneously
from age 5 weeks to age 30 weeks, or made over- with PAN02 murine pancreatic adenocarci-
weight by feeding a pelleted diet followed by a noma cells developed larger tumours, and a
powdered diet, were exposed to ultraviolet radia- significantly greater number of them developed
tion twice a week for 17 weeks (Dinkova-Kostova metastases compared with lean mice. Tumour
et al., 2008). A control group received the pelleted weights at 5 weeks after inoculation were highest
diet only. The obese group had a shortened in the LepobLepob mice, intermediate in the
498
Table 3.7 Effect of obesity on the development of endometrial tumours in mice

Diet-induced (AIN-93G- Transgenic Pten +/− Increased incidence of glandular hyperplasia with atypia; Yu et al. (2010)
based; 58% fat) 1 adenocarcinoma (0 in lean animals)
Table 3.8 Effect of obesity on the development of acute lymphoblastic leukaemia in mice

Diet-induced (60% fat) Transgenic BCR/ABL Shortened latency for development of B-cell-derived and Yun et al. (2010)
T-cell-derived ALL
Diet-induced (60% fat) AKR/J Shortened latency for development of B-cell-derived and Yun et al. (2010)
T-cell-derived ALL
ALL, acute lymphoblastic leukaemia.
LeprdbLeprdb mice, and lowest in the lean mice. and shorter survival time than control mice
Tumour weights were also positively correlated (Dawson et al., 2013; Lashinger et al., 2013; Philip
with body weights, and tumours from both obese et al., 2013).
groups exhibited higher proliferation rates than
those from lean mice (Zyromski et al., 2009). 3.2.8 Cancer of the endometrium of
DIO C57BL/6 mice injected with PAN02 murine the uterus
pancreatic tumour cells also had significantly
larger tumours compared with control mice, and See Table 3.7.
tumour weights were positively correlated with Heterozygous phosphatase and tensin homo-
body weights (White et al., 2010). logue deleted on chromosome 10 Pten+/− mice
In mice, the Kras mutation combined with develop spontaneous multifocal glandular
Ink4a/Arf deficiency induces development hyperplasia and endometrial cancer between age
of pre-neoplastic pancreatic intraepithelial 28 weeks and age 52 weeks. Feeding Pten+/− mice
neoplasia (PanIN) lesions and their progression a HFD increased the incidence of focal atypical
to invasive pancreatic ductal adenocarcinoma glandular hyperplasia and malignant lesions
(PDAC). The association between obesity and from 58% in the Pten+/− mice fed a control diet
pancreatic cancer was tested in three DIO studies to 78% in the obese Pten+/− mice (Yu et al., 2010).
using mice with acinar cell-specific expression of
KrasG12D alone, or mice crossed with the COX2 3.2.9 Leukaemia
conditional knockout or with Ink4a deficiency. See Table 3.8.
The mice were fed a control diet (12% of calories Genetically modified models to study the
from fat), or a high-fat, high-calorie diet (40–60% link between obesity and haematological malig-
of calories from fat) for 4–14 weeks. In all three nancies such as leukaemia and lymphoma are
studies, mice fed the high-fat, high-calorie diet limited. The progression of acute lymphoblastic
had increased numbers of PanIN lesions, more leukaemia (ALL) was tested in two animal
advanced PanIN, increased numbers of PDAC, models (transgenic BCR/ABL and AKR/J mice)
499
fed a HFD (60% of calories from fat). In both [The basic diet had only 2% fat; no statistics were
models, the obese mice developed both B-cell- presented.]
derived and T-cell-derived acute lymphoblastic In another study using the same mouse strain,
leukaemia significantly earlier than control mice a 27% DR or 36% DR compared with the AL mice
(Yun et al., 2010). was used (Tannenbaum, 1945b). No mice in the
36% DR group developed spontaneous mammary
tumours, compared with 54% in the AL group
3.3 Preventive effects of dietary/ and 12% in the 27% DR group. [No statistics were
calorie restriction provided, and the diets fed were very low in fat
(2–3%).] In a second study, parous DBA female
3.3.1 Cancer of the mammary gland
mice were subjected to DR at various degrees
The effect of dietary/calorie restriction on (12–31% DR) beginning at age 21–25 weeks. A
mammary tumorigenesis has been studied exten- 50% reduction in incidence of spontaneous
sively over many decades. In the IARC Handbook mammary tumours was observed for the 31% DR
on weight control and physical activity (IARC, group. [No statistics were presented.]
2002), the Working Group concluded that there Engelman et al. (1990) studied the effects
was “sufficient evidence in experimental animals of DR and/or increased fat levels in the diet on
for a cancer-preventive effect of avoidance of the development of spontaneous mammary
weight gain by restriction of dietary energy tumours associated with MMTV in C3H/HeOu
intake” on tumours of the mammary gland. female mice. The mice were assigned to one of
The models used for these initial investigations five experimental groups between age 6 weeks
were primarily rats with carcinogen-induced and age 8 weeks. This included four groups fed
mammary tumours or mice with spontaneous purified diets (AL LFD, 40% DR LFD, AL HFD,
tumour development. Studies on the prevention and 40% DR HFD) and an AL group fed regular
of mammary tumours by dietary/calorie restric- laboratory chow. Mice were followed up until age
tion are presented in Table 3.9 for mice and in 60 weeks. The chow-fed mice reached 100% inci-
Table 3.10 for rats, in chronological order. The dence of spontaneous mammary tumours by age
text presents the key studies, by type of model 46 weeks, whereas the AL HFD mice and the AL
and/or intervention. LFD mice did so by age 58 weeks and 64 weeks,
respectively. At termination of the study, the
(a) Mice incidence was 15% for the 40% DR HFD mice
Tannenbaum (1945a) studied the effect of and 0% for the 40% DR LFD mice. Body weights
overall moderate CR, carbohydrate restriction, were significantly reduced in the 40% DR groups
and DR with increased fat intake – all in the compared with the AL groups (P < 0.001).
range of about 17% DR – in virgin DBA female Engelman et al. (1994) also tried to identify
mice from age 10 weeks until age 136 weeks. The potential critical periods for the impact of DR
mammary tumour incidence was lowest (47%) on the development of spontaneous mammary
in the carbohydrate-restricted group, compared tumours in C3H/HeOu mice. Mice were sepa-
with 74% in the group fed AL, whereas the inci- rated into three groups: (i) fed AL, (ii) contin-
dence was 65% in mice with overall CR and 87% uously 40% DR, or (iii) 40% DR only from age
in the high-fat group. In addition, tumour latency 4 weeks to age 12 weeks (40% DR4–12), after which
was extended for the carbohydrate-restricted they were fed a HFD AL [this was not described].
group compared with the other three groups. Mice were followed up until age 60 weeks; the
incidence of mammary tumours was 83% in the
500
Table 3.9 Studies on the prevention of mammary tumours by dietary/calorie restriction in female mice
Strain Dose and duration Type of diet, dosing Type of tumours: Statistical Comments
Age at start of carcinogen regimena, and duration of Tumour incidence (number of significance vs AL,
Number of administration intervention (if not until tumours/effective number of unless otherwise
animals termination of study) animals, and/or percentage), specified
Duration of study multiplicity, or other outcomes
Reference as specified
DBA Spontaneous Cereal-based (fox chow), Mammary tumours: NR Diets low in fat; bw reduced in
9–10 wk tumours 2% fat AL, 74% all DR groups
n = 200 16% DR-CHO restricted 16% DR-CHO restricted, 47% Latency extended for the 16%
126 wk 16% DR-all components 16% DR-all components, 65% DR-CHO restricted group
Tannenbaum 23% DR-18% fat 23% DR-18% fat, 87%
(1945a)
DBA Spontaneous Cereal-based (fox chow) Mammary tumours [no NR Diets low in fat; bw of DR
4 wk tumours diluted with cornstarch, indication of pathology]: groups reduced
n = 150 2–3% fat AL, 54% Latency extended
96 wk DR: 27% or 36%, cornstarch 27% DR, 12%
Tannenbaum removed 36% DR, 0%
(1945b)
DBA parous Spontaneous Cereal-based (fox chow) Mammary tumours [no NR Diets low in fat; bw reduced to
21–25 wk tumours diluted with cornstarch, indication of pathology]: 76% and 69% of AL
n = 140 2–3% fat AL, 73% DR started in adult animals
134–138 wk DR: 12%, 18%, 24%, or 31%, 12% DR, 57%
Tannenbaum cornstarch removed 18% DR, 63%
(1945b) 24% DR, 68%
31% DR, 36%
C3H/HeOu Spontaneous Purified diets, either high- Mammary adenocarcinoma: Weibull Tumour latency: chow < AL
6–8 wk tumours (with CHO (sucrose), low-fat Chow, 100% distribution by high-fat < AL low-fat << 40%

n = 215 MMTV) (4.5%) or high-fat (60%), no- AL low-fat, 100% survival analysis DR high-fat ≈ 40% DR low-fat
60 wk CHO, AL or 40% DR 40% DR low-fat, 0%* (P values NR) Bw significantly reduced
Engelman et al. Chow-fed control AL high-fat, 100% [*Significant vs in DR groups vs AL groups
(1990) 40% DR high-fat, 15%** AL low-fat] (P < 0.001)
[**Significant vs
AL high-fat]
C3H/HeOu Spontaneous Purified diets Mammary tumours: *P < 0.000 001
4 wk tumours 40% DR AL, 83% **P = 0.009 vs 40%
n = 144 40% DR age 4–12 wk 40% DR, 13%*,** DR 4–12 wk
age 60 wk 40% DR 4–12 wk, 50%*** ***P = 0.004
Engelman et al.
(1994)
501
502

B6C3F1 Spontaneous NIH-31 diet Mammary tumours (benign and NR
4 wk tumours 40% DR from age 14 wk malignant combined):
n = 102 AL, 27%
Lifetime 40% DR, 4%
Sheldon et al.
(1996)
C57BL6 Spontaneous NIH-31 diet Mammary tumours (benign and NR
n = 102 AL, 14%
Lifetime 40% DR, 0%
Sheldon et al.
(1996)
B6D2F1 Spontaneous NIH-31 diet Mammary tumours (benign and NR
n = 102 AL, 27%
Lifetime 40% DR, 0%
Sheldon et al.
(1996)
MMTV-TGF-α Human TGF-α AIN-93M Mammary adenocarcinoma: [*Significant, χ2 Final bw similar in all 3
C57BL6 CDR, mice pair-fed to IDR AL, 77% test] (P values NR) groups, although IDR mice
10 wk (21% DR) CDR, 44%* lost bw during the 50% DR
n = 93 IDR, mice fed 3 wk of 50% IDR, 3%* periods
age 79–80 wk DR + 3 wk of AL for 11 cycles Latency shorter in AL vs DR
Cleary et al. (2002) (21% DR) Only 1 tumour detected at
necropsy for IDR (n = 1, hence
no statistics)
HER2/neu Transgene- AIN-93M with 10% fat Mammary adenocarcinoma: NS Bw reduced in CDR and IDR
9 wk heterozygous for CDR, 25% CHO restriction AL, 37.5% vs AL
n = 96 HER2/neu IDR, mice fed 3 wk of 50% CDR, 33.3%
age 80 wk DR + 3 wk of AL for 11 cycles IDR, 22.5%
Pape-Ansorge et al. (~25% DR)
(2002)
MMTV-TGF-α Human TGF-α Chow, 33% fat; at age 34 wk, Histologically confirmed NS Latency shorter for obesity-
C57BL6 divided (based on weight mammary adenocarcinoma: prone mice (P < 0.0001);
10 wk status) into obesity-prone, AL, 72% higher multiplicity per
n = 76 overweight, and obesity- Obesity-resistant, 82% tumour-bearing animal
age 85 wk resistant groups; 1 group Obesity-prone, 76% (P < 0.015)
Cleary et al. chow-fed, lean
(2004a)
HER2/neu Transgene Diet NR Mammary adenocarcinoma: Metformin used as a CR
8 wk AL 100% incidence in both groups NS mimetic
n = 106 AL-metformin (100 mg/kg Mean latency: P < 0.05 No effect on bw
10 mo bw in drinking-water 5 d/ 187 d vs 178 d
Anisimov et al. wk) Mean tumour size: P < 0.05
(2005) 1.71 cm vs 1.59 cm
MMTV-TGF-α Human TGF-α AIN-93M Mammary adenocarcinoma: *P < 0.001, χ2 test Final bw similar for AL
C57BL6 CDR, mice pair-fed to IDR AL, 84% and IDR (within 1 wk of
10 wk (14% DR) CDR, 27%* refeeding); significantly
n = 100 IDR, mice fed 3 wk of 50% IDR, 15%* higher than for CDR mice
age 79–80 wk DR + 3 wk of AL for 11 cycles
Cleary et al. (2007) (11% DR)
MMTV-TGF-α Human TGF-α AIN-93M Mammary adenocarcinoma: *Significant, χ2 test Final bw for IDR and CDR
C57BL6 CDR, mice pair-fed to IDR AL, 71% (P values NR) significantly lower (P < 0.001)

10 wk (27% DR) CDR, 35.4%* than for AL
n = 200 IDR, mice fed 3 wk of 50% IDR, 9.1%* IDR mice limited to intake
age 79–82 wk DR + 3 wk pair-fed to AL for of AL mice during refeeding,
Rogozina et al. 11 cycles (25% DR) thus higher DR; latency
(2009) extended for both CDR and
IDR vs AL
HER2/neu Transgene Diet NR Mammary adenocarcinoma: No effect on bw
8 wk AL > 90% incidence in both groups NS
n = 75 AL-metformin (100 mg/kg Slower kinetic of tumour P < 0.001
Lifetime bw in drinking-water 5 d/ incidence with metformin
Anisimov et al. wk) Latency: 223 d vs 197 d P < 0.05
(2010)
503
504

C57BL6 (OVX) M-Wnt or E-Wnt AIN-76A, AL Allograft: P < 0.02 Tumour growth enhanced by
6–8 wk cells implanted 30% DR Tumour growth (area in mm2) HFD for M-Wnt cells but not
n = 120 8 wk after diets HFD (60% fat) reduced by 30% DR for both cell for E-Wnt cells
14 wk started, and mice lines
Dunlap et al. followed up for
(2012) 6 wk
C57BL6 (OVX) Wnt-1 cells AIN-76A Allograft: Bw reduced in 30% DR vs AL
6 wk implanted and mice 30% DR Tumour weight: P = 0.0002
n = 45 followed up for 0.045 g vs 0.39 g
4 wk 4 wk Latency: NS
Nogueira et al. 16.3 wk vs 15.6 wk
(2012)
HER2/neu Transgene- AL-CON, AIN-93M with Mammary adenocarcinoma: *P < 0.05 vs AL- Bw reduced in CDR and IDR
10 wk homozygous for 10% fat calories from soy oil AL-CON, 87% CON vs AL
n = 178 HER2/neu CDR-CON, 25% CHO CDR-CON, 47%*
age 60 wk restriction IDR-CON, 59%*
Mizuno et al. IDR-CON, 3 wk of 50% DR
(2013) + 3 wk pair-fed to AL for 11
cycles (25% DR)
Same groups with EPA AL-EPA, 63% **P < 0.05 vs CDR-
substituted for some of the CDR-EPA, 40% EPA and AL-EPA
soy oil IDR-EPA, 15%**
MMTV-TGF-α Human TGF-α AIN-93M based diet with Mammary adenocarcinoma: *P < 0.0001, χ2 test Bw of AL mice greater than
C57BL6 22.7% fat AL, 66.7% CDR and IDR. IDR mice lost
10 wk CDR, mice matched to IDR CDR, 52.3%* and regained weight with
n = 135 (22% DR) IDR, 4.4%* each cycle to values similar to
age 79–82 wk IDR, mice fed 3 wk of 50% Latency: AL < CDR* < IDR* *P < 0.0001, those of CDR
Rogozina et al. DR + 3 wk pair-fed to AL for Kaplan–Meier
(2013) 11 cycles (22% DR)
See Rogozina et al.
(2009)
C3(1)-TAg FVB/N Transgenic C3(1)- LFD (10% fat) Basal-like mammary tumours: Bw of HFD-7 decreased and
3 wk TAg HFD (60% fat) Tumour volume: remained at level of LFD
n = 40 HFD for 7 wk, switched to ~3-fold higher in HFD vs LFD P = 0.0024 mice; both significantly lower
27 wk LFD for ~17 wk HFD-7 vs LFD NS than in HFD mice (P = 0.019);
Sundaram et al. latency and multiplicity not
(2014) affected
[Not clear whether effect due
to weight loss or to change in
diet composition]
HER2/neu/p53KO Transgene Teklad 4% mash Mammary adenocarcinoma: Metformin used as a CR
60 d AL Tumour multiplicity NS mimetic
NR AL-metformin (150 mg/kg) Tumour weight NS [No bw data]
11 mo Survival not affected
Thompson et al.
(2015)
AL, ad libitum; bw, body weight; CDR, chronic dietary restriction; CHO, carbohydrate; d, day or days; CON, control; CR, calorie restriction; DR, dietary restriction; EPA,
eicosapentaenoic acid; HFD, high-fat diet; IDR, intermittent dietary restriction; LFD, low-fat diet; mo, month or months; MMTV, mouse mammary tumour virus; NR, not reported; NS,
not significant; OVX, ovariectomized; TGF-α, transforming growth factor alpha; vs, versus; wk, week or weeks.
a The (%) indicates the percentage of calories from fat.

505
506

Table 3.10 Studies on the prevention of mammary tumours by dietary/calorie restriction in female rats
Strain Dose and Type of diet, dosing regimen, Type of tumours: Statistical Comments
Age at start duration of and duration of intervention as Tumour incidence (number of significance
Number of carcinogen appropriate tumours/effective number of
animals administration animals, and/or percentage),
Duration of multiplicity, or other outcomes
study as specified
Reference
Wistar Spontaneous AL, 20% DR after 1 mo for 23 mo Mammary tumours: P < 0.001
NR tumours 17%, 3%
~50/group
24 mo
Tucker (1979)
Sprague-Dawley DMBA, 5 mg, Rat chow, 50% DR for 7 d before and Palpable mammary tumours: *P < 0.05 vs AL Latency extended in 50%
57 d once for 30 d after injection of DMBA. AL, 76% or 50% DR+E2 DR rats
n = 104 One 50% DR group received various 50% DR, 29%*
26 wk hormone treatments (e.g. E2) 50% DR+E2, 71%
Sylvester et al.
(1981)
Sprague-Dawley DMBA, 5 mg, Rat chow: Palpable mammary tumours: *P < 0.05 No significant effect
57 d once AL; 50% DR for 1 wk before and for 80.9%, 27.8%*, 76.2%, 75.0%, on latency or tumour
n = 101 1 wk after DMBA; 50% DR for 2 wk 75.0% multiplicity
22 wk starting 1 wk after DMBA; 50% DR for
Sylvester et al. 2 wk starting 3 wk after DMBA; 50%
(1982) DR for 4 wk starting 5 wk after DMBA
F344 DMBA, 65 mg/kg Purified diet, dextrose reduced in 30% Mammary tumours: NR
52 d bw, once fat diets, 1 d after DMBA: AL 30% fat, 73%, 43%, 7%
n = 45 AL 5% fat, 14% DR 30% fat
24 wk
Boissonneault
et al. (1986)
Sprague-Dawley DMBA, 5 mg, AIN-76A, sucrose reduced in DR diets, Histologically verified mammary *P < 0.005 Tumour latency extended
50 d once 1 wk after DMBA: AL, 10% DR, 20% tumours: in 30% DR and 40%
n = 100 DR, 30% DR, 40% DR 60%, 60%, 40%*, 35%*, 5%* DR groups. Tumour
20 wk multiplicity reduced in
Klurfeld et al. 40% DR group
(1989a)
study as specified
Reference
Sprague-Dawley DMBA, 5 mg, AIN-76A, sucrose reduced in DR diets, Mammary tumours: *P < 0.01 vs AL Tumour multiplicity
50 d once 1 wk after DMBA: AL 5% fat, AL 15% 65%*, 85%, 80%, 60%*, 30%** 15% fat and AL reduced to similar levels
n = 100 fat, AL 20% fat, 25% DR 20% fat, 25% 20% fat in both 25% DR groups
20 wk DR 26.7% fat **P < 0.0001 vs
Klurfeld et al. AL groups
(1989b)
Sprague-Dawley DMBA, 5 mg, AIN-76A, sucrose reduced in DR diets, Palpable mammary tumours: *P < 0.001, Weight gain correlated
50 d once 1 wk after DMBA. 6 groups: 50%, 20%*, 60%, 40%, 45%, 30% group B vs A with tumour incidence
n = 120 (A) AL; (B) 25% DR, wk 1–16; (C) 25% and group B (r = 0.96) and with total
16 wk DR, wk 1–4; (D) 25% DR, wk 1–8; vs C calorie intake (r = 0.83)
Kritchevsky et al. (E) 25% DR, wk 5–12; (F) 25% DR,
(1989) wk 9–16
Sprague-Dawley DMBA, 5 mg, AIN-76A, sucrose reduced in DR diets, Mammary tumours: *P = 0.007, χ2 Multiplicity of palpable
50 d once 1 wk after DMBA: AL, 25% DR, 40% 90%, 61%*, 20%* test tumours reduced in 40%
n = 110 DR DR rats (P < 0.05)
20 wk
Ruggeri et al.
(1989a, b)
LA/N DMBA, 5 mg, AIN-76A, sucrose reduced in DR diets, Mammary tumours: NR Tumour multiplicity
65 d once 1 wk after DMBA: obese AL, obese 100%, 27%, 21% reduced in obese 25%

n = 49 25% DR, lean AL DR rats; bw in DR
17 wk significantly lower than in
Klurfeld et al. obese AL rats
(1991)
Sprague-Dawley MNU, 25 mg/kg Purified diet, 45% fat diet for ~10 wk Mammary tumours: *P < 0.05 vs 100% incidence of
50 d bw, at age 50 d, (until tumour of 1 cm3), then divided Multiplicity: corresponding mammary tumours;
n = 83 once into 4 groups: AL 45% fat, 30% DR 2.43, 1.74*, 2.35, 0.95* AL group bw reduced by 10% in DR
30 wk 45% fat, AL 25% fat, 30% DR 25% fat Tumour/bw (%): groups
Zhu et al. (1991) 3.8, 2.2*, 2.5, 1.3*
507
508

study as specified
Reference
F344 Spontaneous NIH-31; AL, 40% DR Mammary adenocarcinoma: Extended survival;
14 wk tumours Exp. 1: 8%, 2% [NS] reduced bw; reduced
n = 54/group Exp. 2: 5%, 0% [P = 0.048] tumour multiplicity
(exp. 1); n = 114– Mammary fibroadenoma: [Low incidence of
116/group (exp. 2) Exp. 1: 36%, 2% [P < 0.0001] tumours in AL animals]
Lifetime Exp. 2: 35%, 1% [P < 0.0001]
Thurman et al.
(1994)
F344 MNU, 50 mg/kg AIN-76A with cornstarch + cerelose; Mammary adenocarcinoma: *P < 0.05 Tumour multiplicity
50 d bw, at age 50 d AL, 20% DR 23.3%, 6.7%* reduced in AL vs 20%
n = 132 and 57 d, once DR; exercise by treadmill
20 wk running not effective in
Gillette et al. any group
(1997)
Sprague-Dawley MNU, 50 mg/kg AIN-93G with cornstarch + cerelose, Mammary carcinoma: Ptrend < 0.01, Dose-dependent
21 d bw, once fed AL, 10% DR, 20% DR, 40% DR 100%, 80%, 60%, 25% dose-dependent increased latency for DR
n = 75 reduction Additional results
35 d presented (Zhu et al.,
Zhu et al. (1997) 1999a, b)
ACI E2 treatment Purified diet, 5% fat; AL, 40% DR in Mammary tumours: *P < 0.001 vs Latency of palpable
49 d from age 59 d controls or mice treated with E2 AL, 0% AL+E2 tumour: 69 d after E2 for
n = 84 40% DR, 0% AL+E2 vs 104 d for 40%
220 d AL+E2, 100% DR+E2
Harvell et al. 40% DR+E2, 59%* Bw reduced in 40% DR
(2002) groups
Sprague-Dawley MNU, 50 mg, AIN-93G; AL, 40% DR (CHO reduced) Mammary adenocarcinoma: *P < 0.001, Bw similar; incidence
3 wk once for 6 wk, and then fed AL (DR-AL) Detectable tumour incidence at χ2 test of tumours in DR-AL
n = 66 day 42 after MNU: similar to AL by end of
90 d AL, 61%; DR, 11%* experiment
Zhu et al. (2002)
study as specified
Reference
Sprague-Dawley MNU, 50 mg, AIN-93G; 40% DR; 40% DR for 6 wk, Mammary adenocarcinoma: P < 0.001 Bw reduced in DR rats
3 wk once and then fed AL (DR-AL) Tumour volume of DR rats 2 meals daily
n = 108 significantly smaller than AL or
54 d DR-AL rats
Thompson et al.
(2004a)
Sprague-Dawley MNU, 50 mg, AIN-93G; 40% DR (CHO reduced) Mammary adenocarcinoma: *P < 0.01 2 meals daily
3 wk once from age ~4 wk 96%, 59%* **P < 0.001
n = 78 Multiplicity: 4.3, 1.0**
77–78 d
Thompson et al.
(2004b)
Wistar DMBA, 2 mg, AL, HFD (60% fat) Mammary tumours: NS Bw significantly lower in
7 wk once IDR-HFD (50% DR CHO reduction), HFD, 17.6% IDR-HFD rats
n = 90 first cycle at age 15 wk to loss of 20% – IDR-HFD, 8.8%
50 wk 4 cycles
Buison et al.
(2005)
Sprague-Dawley MNU, 50 mg/kg AIN-93G; AL, 40% DR (CHO) at age Mammary adenocarcinoma: *P < 0.0006
3 wk bw, at age 21 d 30 d for 6 wk, then divided into DR, AL, 96.6%

n = 99 DR-AL, DR+IGF-1 DR, 56.7%*
57–58 d DR-AL, 80%
Zhu et al. (2005) DR+IGF-1, 60%
Sprague-Dawley MNU, 50 mg/kg AIN-93G: AL, AL+0.03% Mammary adenocarcinoma: P < 0.005 2-deoxyglucose used as a
3 wk bw, at age 21 d 2-deoxyglucose 86.7%, 53.3% CR mimetic
n = 60 Multiplicity: P = 0.018 No difference in bw
7 wk 2.03, 1.37
Jiang et al.
(2008a)
509
510

study as specified
Reference
Sprague-Dawley MNU, 50 mg/kg AIN-93G: AL, 20% DR, 40% DR Mammary adenocarcinoma: *P < 0.001 Bw reduced in DR rats
21 d bw, at age 21 d 96%, 60%*, 23%*
n = 90 Multiplicity:
52 d 2.1, 1.1*, 0.3*
Jiang et al.
(2008b)
Sprague-Dawley MNU, 50 mg/kg SUMO32 diet (32% HFD corn oil-9.4% Mammary adenocarcinoma: *P < 0.01 OP rats weighed 15%
21 d bw, at age 21 d saturated fat) OP, 91%; OR, 65%* more than OR rats at
n = 103, 101 Obesity-prone (OP) study termination
60 d Obesity-resistant (OR) Tumour weight reduction
Matthews et al. of 80% for OR vs OP;
(2014) tumour latency extended
for OR vs OP
Sprague-Dawley MNU, 75 mg/kg Teklad standard diet (8% fat) ER+ mammary tumours: Metformin used as a CR
50 d bw, at age 50 d Control; metformin, 50 mg/kg/d; Tumour latency NS mimetic
NR metformin, 150 mg/kg/d Tumour weight NS 2 meals daily
126 d Bw NR
Thompson et al.
(2015)
Sprague-Dawley MNU, 50 mg/kg AIN-93G Mammary adenocarcinoma: *P < 0.003 Compounds studied as
21 d bw, once Control; metformin, buformin, or 83.3%, 93.3%, 43.3%*, 76.7% CR mimetics
n = 120 phenformin Bw reportedly measured,
(30/group) but no data presented
51 d
Zhu et al. (2015)
AL, ad libitum; bw, body weight; d, day or days; CHO, carbohydrate; CR, calorie restriction; DMBA, 7,12-dimethylbenz[a]anthracene; DR, dietary restriction; E2, 17β-estradiol; ER,
estrogen receptor; exp., experiment; F344, Fischer 344; HFD, high-fat diet; IDR, intermittent dietary restriction; IGF-1, insulin-like growth factor 1; MNU, N-methyl-N-nitrosourea; mo,
month or months; NR, not reported; NS, not significant; OP, obesity-prone; OR, obesity-resistant; vs, versus; wk, week or weeks.
AL group, 13% in the 40% DR group, and 50% in tumour incidence than in CDR, although not
the 40% DR4–12 group. significantly so (Pape-Ansorge et al., 2002;
The effect of 40% DR on ageing and longevity Mizuno et al., 2013).
in three different mouse strains was reported by In another transgenic mouse strain, C3(1)-
Sheldon et al. (1996). DR was initiated at age TAg mice were fed either a LFD (10% of calories
14 weeks and extended to 48 months in B6C3F1, from fat) or a HFD (60% of calories from fat) from
C57BL6, and B6D2F1 mice. AL mice in the age 3 weeks (Sundaram et al., 2014), and a group
B6C3F1, C57BL6, and B6D2F1 groups had inci- of the HFD mice was switched to the LFD after
dences of spontaneous mammary tumours of 7 weeks on the HFD. The switch to the LFD from
27%, 14%, and 27%, respectively, compared with the HFD resulted in weight loss to the level of the
4%, 0%, and 0% for the corresponding 40% DR LFD mice. The tumour volume in the HFD mice
groups. was 3 times that in the LFD mice, and switching
More recently, transgenic mice have been from the HFD to the LFD resulted in tumour
used to evaluate the effect of DR on develop- volumes similar to those in the LFD mice. [It is
ment of mammary tumours. Several studies not clear whether the findings are due to weight
were conducted using mice that overexpress loss or to the change in diet composition.]
human TGF-α, in which two modes of DR were Allograft models were also used to assess
compared (Cleary et al., 2002, 2007; Rogozina tumour progression in response to dietary
et al., 2009, 2013). In the initial experiments, intervention. For example, ovariectomized
mice received either intermittent DR (IDR) or C57BL6 mice were fed AL, 30% DR, or a HFD
chronic DR (CDR). IDR mice were subjected (60% of calories from fat) for 8 weeks, before
to 50% DR for 3-week intervals, followed by two types of Wnt cells (M-Wnt or E-Wnt cells)
3 weeks of refeeding AL. This resulted in an were implanted; the mice then continued on
overall DR of 21%, because of overconsumption their diets for 6 weeks while tumour growth was
during refeeding compared with what the AL monitored. DR reduced tumour growth for both
mice consumed. CDR mice were matched for cell lines compared with AL, whereas tumour
calorie intake for each 6-week cycle of 50% DR/ growth was enhanced by the HFD only for the
refeeding. The IDR mice had significantly lower M-Wnt cells (Dunlap et al., 2012). In another
mammary tumour incidence compared with the study, Wnt-1 cells were implanted in AL or 30%
CDR mice, i.e. mammary tumour incidences of DR mice. Tumour weight was lower in DR mice
3% and 15% in the two reports, compared with than in AL mice, but latency was not affected
77% and 84% for AL mice and 44% and 27% (Nogueira et al., 2012).
for the CDR mice (Cleary et al., 2002, 2007). In Another approach to assess the effects of
subsequent studies, IDR mice were pair-fed to the body weight independent of diet is to use mouse
AL group during the refeeding phases, resulting or rat strains that respond to HFD feeding with
in significantly lower body weights in both DR a range of body weights. Cleary et al. (2004a) fed
groups, with fluctuating body weights in the IDR MMTV-TGF-α mice on a C57BL6 background
group; both the CDR and IDR groups had lower a LFD or a moderately HFD (33% of calories
tumour incidence than the AL mice (Rogozina from fat) and then divided the mice into three
et al., 2009, 2013). This was also observed when groups (obesity-prone, overweight, and obesi-
the fat content of the diets was moderately ty-resistant), based on body weight status at age
increased (Rogozina et al., 2013). IDR compared 34 weeks. The heaviest group, obesity-prone,
with CDR was also examined in the transgenic had the shortest mammary tumour latency,
mouse strain HER2/neu. IDR resulted in lower compared with obesity-resistant mice fed the
511
same diet, i.e. with body weights similar to DR rats. Reduced body weight was associated
those of the LFD mice. Furthermore, the heavier with extended survival and reduced multiplicity.
mice had more palpable tumours than mice that Female Sprague-Dawley rats were admin-
weighed less, although this was not statistically istered 5 mg of DMBA dissolved in corn oil at age
significant. 50 days (Klurfeld et al., 1989a). The rats treated
A murine model of basal-like breast cancer with DMBA were then subjected to 10%, 20%,
was used to assess whether the obesity-induced 30%, or 40% DR from age 57 days and followed
pro-tumour effects are reversed by weight up for 20 weeks. The 10% DR had no effect on
normalization. Ovariectomized female C57BL/6 tumour incidence, but the 20%, 30%, and 40%
mice were fed a control diet or a DIO regimen DR resulted in incidences of 40%, 35%, and 5%,
for 17 weeks, resulting in a normal weight or an respectively. Tumour multiplicity was signifi-
obese phenotype, respectively. After 17 weeks, cantly reduced in the 40% DR group, and latency
mice on the DIO regimen were randomized to was extended in the 30% DR and 40% DR groups.
continue the DIO diet or switched to the control In another study also using DMBA (Ruggeri
diet. The resulting formerly obese mice had body et al., 1989a, b), only three groups of rats were
weights comparable to those of the controls. At included: AL, 25% DR, and 40% DR. The inci-
week 24, the mammary pads of all mice were dence of mammary tumours in these groups was
injected with MMTV-Wnt-1 mouse mammary 90%, 61%, and 20% (P = 0.007), respectively. In
tumour cells, and tumour growth was then the 40% DR group, the majority of the tumours
measured twice per week until 36 months. Mean were small and non-palpable (P < 0.05). The
tumour volumes in the DIO and formerly obese authors also used rats treated with DMBA and
mice were similar, and were higher than those in fed diets combining increased fat levels with
the controls (Rossi et al., 2016). DR and determined effects on development of
mammary tumours (Klurfeld et al., 1989b). The
(b) Rats experimental groups included AL rats fed diets
In a longevity study, the incidence of with 5%, 15%, and 20% of calories from fat, as
mammary tumours was significantly lower in well as 25% DR rats fed diets with 20% and 26.7%
20% DR Wistar rats (3%) than in the AL group of calories from fat. The incidence of mammary
(17%) (Tucker, 1979). tumours was significantly lower in the 25%
In one study (Thurman et al., 1994), groups DR, 26.7% fat group than in the other groups,
of female Fischer 344 (F344) rats were fed AL for which the incidences were in the range of
or subjected to 40% DR from age 14 weeks and 60–85%. Although there was only a slight reduc-
followed up for their lifetime in two experiments tion in incidence for the 25% DR, 20% fat group,
with a similar study design. In general, survival tumour weight and tumour multiplicity were
was extended by DR [the increase was modest, reduced to similar levels as in the other DR group
and no statistics were presented]. Incidence (i.e. 25% DR, 26.7% fat).
of spontaneous mammary adenocarcinoma A similar study was reported by Boissonneault
was 8% and 5% in the AL rats and 2% and 0% et al. (1986) using female F344 rats that were
[P = 0.048] in the 40% DR rats in the first and switched to experimental diets 1 day after DMBA
second experiment, respectively. Reductions in treatment (at age 52 days); these groups included
the incidence of mammary fibroadenoma were AL 30% fat, AL 5% fat, and 14% DR 30% fat. The
also reported, from 36% and 35% in the AL rats groups were then followed up for 24 weeks after
to 2% [P < 0.0001] and 1% [P < 0.0001] in the 40% DMBA treatment. The incidence of mammary
tumours was 73% for the AL 30% fat group
512
and 43% for the AL 5% fat group but only 7% and 29% in the 50% DR group. In the follow-up
for the 14% DR 30% fat group [no statistics were study, five groups of rats were used (Sylvester
reported]. et al., 1982): AL control rats, 50% DR 1 week
Klurfeld et al. (1991) also investigated the before and 1 week after DMBA treatment,
effects of DR on the development of DMBA- 50% DR for 2-week periods starting 1 week or
induced mammary tumours in genetically 3 weeks after DMBA treatment, and 50% DR for
obese rats. After administration of DMBA at age 4 weeks starting 5 weeks after DMBA treatment.
65 days, female LA/N Corpulent rats were fed All groups had similar incidences of mammary
purified diets either AL or 40% DR, and an AL tumours (75.0–80.9%), except for the group
lean group was also included. The body weight subjected to 50% DR for 1 week before and 1 week
of the obese 40% DR rats remained at a level after DMBA treatment, in which the incidence
substantially lower than that of the AL obese was only 27.8%. [The Working Group noted that
rats but higher than that of the AL lean rats. The DR started before administration of DMBA, and
incidence of mammary tumours was 100% in the hence the effect of DR on DMBA metabolism is
AL obese rats, compared with 27% in the obese unknown and might have been partly respon-
40% DR rats and 21% in the AL lean rats. [The sible for the observed effect.]
Working Group noted that this study assessed Zhu et al. (1991) used Sprague-Dawley rats
DR in obese rats.] administered MNU at age 50 days to induce
Several investigations have focused on timing mammary tumours. The rats were then fed a 45%
of DR interventions and development of mam- fat diet and followed up until the tumours reached
mary tumours. a volume of 1 cm3, which was 10 ± 2 weeks after
One study examined the impact of 25% DR administration of MNU. The rats were then
imposed at different times relative to the admin- divided into four groups: AL 45% fat (group 1),
istration of DMBA (Kritchevsky et al., 1989). 30% DR 45% fat (group 2), AL 25% fat (group
There were a total of six groups in the 16-week 3), and 30% DR 25% fat (group 4). The rats were
experiment: fed AL throughout (group A), fed then followed up for an additional 30 weeks, after
25% DR throughout (group B), fed 25% DR which tumour progression was assessed. DR
for the first 4 weeks (group C), fed 25% DR for reduced the number of tumours per animal, the
the first 8 weeks (group D), fed 25% DR for the tumour weight, and the tumour weight per body
8 weeks (weeks 5–12) in the middle of the exper- weight, compared with AL. Body weight was
iment (group E), and fed 25% DR for the last reduced by 10%. [No statistics were reported.]
8 weeks (group F). The incidence of mammary A rapidly developing carcinogen-induced
tumours was 50% in the AL rats and 20% in the mammary tumour model was developed in
rats fed 25% DR throughout the study. The other Sprague-Dawley rats to investigate the effect
groups had incidences of 30–60%; the incidence of DR on mammary tumorigenesis (Gillette
was 30% in the group fed 25% DR for the last et al., 1997; Zhu et al., 1997, 1999a, b, 2002, 2005;
8 weeks (group F). Thompson et al., 2004a, b). In this model, rats
Sylvester et al. (1981, 1982) also investigated are administered MNU at age 21 days and then
the effect of timing of DR in Sprague-Dawley rats followed up until age 100 days or more as the
treated with DMBA. In their first study, a 50% tumours develop; they are subjected to 40% DR
DR was imposed 1 week before and continued through carbohydrate restriction. This degree of
until 30 days after DMBA injection (Sylvester DR consistently and significantly reduced body
et al., 1981). After 26 weeks, the incidence of weight as well as mammary tumour development,
mammary tumours was 76% in the AL group as reflected by incidence and tumour volume.
513
In an additional study (Jiang et al., 2008b), 20% estrogen treatment, and at termination of the
DR led to an incidence of 60%, compared with study, mammary tumour incidence was 59%. As
96% in AL rats and 23% in the 40% DR rats. expected, body weight was reduced for the 40%
Multiplicity was also significantly reduced. [The DR rats. No tumours were detected in ACI rats
Working Group noted that in this study model, not treated with estrogen, whether they were fed
tumours develop in pre-pubertal animals.] AL or subjected to 40% DR.
Buison et al. (2005) used the DMBA
mammary tumour model with the carcinogen (c) Calorie restriction mimetics
administered at age 50 days and the rats followed An additional approach to study the effect of
up for 50 weeks. The intervention consisted of CR on mammary tumour development has been
feeding the rats a 60% HFD followed by 50% the use of CR mimetics. Metformin, the most
DR (with carbohydrate restriction) for 4 cycles common CR mimetic, did not have an effect on
of 20% weight loss, followed by refeeding; this mammary tumour development in the MNU
resulted in a 50% reduction in mammary tumour rat model, when MNU was administered at age
incidence, from 17.6% to 8.8% [not significant]. 50 days, or in transgenic HER2/neu/p53KO mice
The body weight of the IDR rats fluctuated and at (Thompson et al., 2015). Several earlier studies
termination of the study was significantly lower using the HER2 mouse model of breast cancer
than that of the HFD control rats. [The Working had reported some effects of metformin on
Group noted that a control group with chronic latency, but tumour incidence was not affected
DR is missing.] (Anisimov et al., 2005, 2010).
In another model (Matthews et al., 2014), In the rapidly emerging tumour model in
ovary-intact female Sprague-Dawley rats were rats, treatment with 2-deoxyglucose (Jiang et al.,
injected with 50 mg/kg of MNU at age 21 days. 2008a) but not with metformin (Thompson et al.,
Obesity-resistant or obesity-prone animals were 2015; Zhu et al., 2015) decreased the incidence of
fed a purified diet containing 32% of calories mammary tumours; buformin and phenformin
from fat. At termination of the study, obesi- both reduced tumour incidence.
ty-prone rats were approximately 15.5% heavier
than obesity-resistant rats. Obesity-resistant rats 3.3.2 Cancer of the colon
had lower incidence, multiplicity, and burden
of mammary carcinomas, with a concomitant See Table 3.11.
increase in cancer latency compared with obesi- The early studies of the effect of DR on cancer
ty-prone rats (P < 0.01 for all analyses). of the colon used carcinogen-induced models in
Another model for breast cancer is the ACI rats, whereas more recent studies used mouse
rat; when supplementary estrogen is given to models.
ovary-intact animals, this leads to develop- In one of the early studies, male Lobund
ment of mammary tumours (Shull et al., 1997). Sprague-Dawley rats were administered meth-
Harvell et al. (2002) determined the impact of ylazoxymethanol at 30 mg/kg at weaning, and
40% DR starting at age 7 weeks on mammary about 25% DR started either 10 days or 63 days
tumorigenesis in this model. By 216 days of after and continued until 140 days after admin-
estrogen treatment, 100% of the AL rats had at istration of methylazoxymethanol (Pollard et al.,
least one palpable mammary tumour, with the 1984). [It is not clear what the natural ingredient
first tumour detected at 69 days. In contrast, diet contained.] The long-term DR significantly
the first palpable mammary tumour in the 40% reduced tumour incidence and multiplicity,
DR group was not detected until 104 days of whereas there was no effect when DR was initiated
514
Table 3.11 Studies on the prevention of colon tumours by dietary/calorie restriction in rats and mice
Species, strain Route, dose, Type of diet, dosing regimen, Type of tumours: Statistical Comments
(sex) and duration and duration of intervention Tumour incidence (number of significance
Age at start of carcinogen tumours/effective number of
Number of animals administration animals, and/or percentage),
Rat
Lobund Sprague- MAM, s.c., 30 mg/kg Natural ingredient diet L-485; Tumours of colon and small Small group sizes may have
Dawley (M) bw AL or 25% DR, starting at intestine: affected findings, and possibly
Weanling either 10 d or 63 d after MAM, AL, 90%; 25% DR-10, 30% P < 0.0001 components of diet had
n = 76 or ADF from 8 d or 31 d after AL, 85%; 25% DR-63, 100% NS protective effect
20 wk MAM (each intervention own AL, 60%; ADF-8, 60% NS
Pollard et al. (1984) AL group) AL, 90%; ADF-31, 67% NS
F344 (M) AOM, 15 mg/kg bw, Semi-purified diet (23% fat); Colon adenoma or P < 0.05 Tumour multiplicity
5 wk once a wk from age AL or 30% DR from 4 d after adenocarcinoma (combined): significantly reduced by DR
n = 60 7 wk for 2 wk AOM, followed up for 32 wk AL, 83%; 30% DR, 33%
32 wk Colon adenocarcinoma: P < 0.05
Reddy et al. (1987) AL, 30%; 30% DR, 0%
Sprague-Dawley DMH, s.c., Two groups: pups raised 4/ Colon tumours: *P < 0.01 vs [The Working Group
(F, M) 10 mg/kg bw, once litter or 8/litter; weaned to 8/litter: M, 48%; F, 42% respective 8/ considered that using litter
Neonatal a wk for 10 wk from semi-purified diet (20% fat) 4/litter: M, 85%*; F, 60%* litter M or F size as an indicator of early-
n = 179 1 mo after weaning AL 4/litter pair-fed: M, 76%**; F, **P < 0.01 vs life access to nutrition made
32 wk Third group: 4/litter pair-fed 52% 8/litter M it difficult to evaluate the
Newberne et al. to 8/litter after weaning effect of DR on colon tumour
(1990) development]
Zucker lean and AOM, 10 mg/kg bw, Lean rats, AL Multiplicity of ACF: 100% incidence of ACF in all
obese (fa/fa) (F) at age 6 wk, once a wk Obese rats, AL All ACF NS groups; bw not affected by DR

6 wk for 2 wk Obese rats, 20–25% DR Advanced ACF (≥ 7 crypts) P < 0.001 for [Small n values; effect seen only
n = 32 obese, n = 16 DR vs obese when distinguishing between
lean AL advanced and early foci, which
15 wk is questionable]
Raju & Bird (2003)
Mouse
C57BL6 Apc Min (M) Transgenic AIN-76A Colon polyps: NS Bw reduced in DR vs AL (19.4 g
10 wk 25% DR (CHO restriction) Mean number/mouse: vs 25.9 g). No effect of feeding
28–30/group AL, 5; 25% DR, 7 a HFD (30% of calories from
9 wk corn oil)
Mai et al. (2003) [Values reported in graph]
515
516

Species, strain Route, dose, Type of diet, dosing regimen, Type of tumours: Statistical Comments
(sex) and duration and duration of intervention Tumour incidence (number of significance
Age at start of carcinogen tumours/effective number of
Number of animals administration animals, and/or percentage),
C57BL6 (M) MC38 murine HFD (60% fat) for 7 wk, then Allograft tumour: Bw of HFD and LCD mice
6 wk adenocarcinoma cells divided into 4 groups: Tumour latency: NS higher than other groups;
n = 80 (20/group) injected at wk 15 HFD (60% fat + 20% protein + HFD, 11.2 d; LCD, 11.4 d; 30% HCD mice maintained bw;
23 wk 20% CHO) DR, 20.1 d 30% DR mice lost bw and then
Wheatley et al. LCD (60% fat + 35% protein + Tumour size : *P < 0.001 vs maintained it
(2008) 5% CHO) HFD, 397.2 mm2; LCD, all the other
HCD (10% fat + 20% protein + 351.6 mm2; HCD, 474.6 mm2; groups
70% CHO) 30% DR, 162.4 mm2*
30% DR (10% fat + 20%
protein + 70% CHO)
BALB/c (M) CT26 murine colon LFD (10% fat) for 16 wk Allograft tumour: Metastasis to lung [number
4 wk carcinoma cells HFD (60% fat) for 16 wk Tumour volume increase P < 0.05 of tumour nodules] higher in
21/group inoculated at age from 3 wk after cell injection HFD vs LFD (P < 0.05)
24 wk 20 wk, and tumours (measured in situ) Bw slightly higher, by 5.9%, in
Park et al. (2012) harvested after 31 d Tumour weight (g): P < 0.05 HFD vs LFD
LFD, 1.2; HFD, 1.5 [BALB/c mice resistant to HFD]
C57BL6 (F) MC38 murine AIN-76A Allograft tumour: P < 0.001 Bw of DR mice remained stable
7 wk adenocarcinoma cells 30% DR (CHO restriction) for Tumour volume: over the 22 wk of restriction,
15/group inoculated at age 25 wk AL, 2286 mm3; 30% DR, whereas AL mice gained bw
25 wk 29 wk, and tumours 515 mm3 (P < 0.05)
Harvey et al. (2013) harvested after 24 d
FVB (M) AOM, 10 mg/kg bw, AIN-76A Colon tumours: *P < 0.05 vs 30% DR mice lost bw and then
10 wk once a wk for 5 wk 30% DR (CHO restriction) for Multiplicity after 20 wk: AL maintained it; increase in bw
12/group 5, 10, or 20 wk AL, 9.5; 30% DR, 7.2* and tumour number in HFD
up to 20 wk HFD (45% fat) group
Olivo-Marston [Values read from graph]
et al. (2014)
ACF, aberrant crypt foci; ADF, alternate-day fasting; AL, ad libitum; AOM, azoxymethane; bw, body weight; CHO, carbohydrate; d, day or days; DMH, dimethylhydrazine; DR, dietary
restriction; F, female; F344, Fischer 344; HCD, high-carbohydrate diet; HFD, high-fat diet; LCD, low-carbohydrate diet; LFD, low-fat diet; M, male; MAM, methylazoxymethanol; mo,
month or months; NS, not significant; s.c., subcutaneous; vs, versus; wk, week or weeks.
in the older rats. Alternate-day fasting was also models. Transgenic ApcMin mice that develop
initiated at either 8 days or 31 days after admin- “spontaneous” tumours were subjected to 25%
istration of methylazoxymethanol, but there was DR for 9 weeks and compared with AL mice (fed
no effect of this intervention. [The low number of AIN-76A-based diets). DR had no effect on the
animals per group in this experiment may have number of colon polyps. In addition, feeding
affected the study conclusions.] a HFD (30% of calories from fat) had no effect
Several other studies used the carcinogen (Mai et al., 2003). In another study, AOM was
AOM to induce cancer of the colon. For example, used to induce colon tumours in FVB male
Reddy et al. (1987) fed F344 male rats a HFD AL mice. AL mice were fed the AIN-76A diet and
or 30% DR from age 5 weeks. They were treated compared with mice subjected to 30% DR (with
with an AOM regimen beginning at age 7 weeks carbohydrate restriction). Mean numbers of
and followed up until age 32 weeks. No adeno- colon tumours were significantly reduced after
carcinomas were detected in the 30% DR group, 20 weeks of 30% DR (Olivo-Marston et al., 2014).
compared with a 30% incidence in the AL rats. In Allograft implants of colon cancer cell lines
a study with Zucker rats, AL lean and AL obese have also been used to assess the effects of DR
rats were used, as well as a 20–25% DR obese on tumour growth. MC38 cells were used in two
group, fed 75–80% of the consumption of the AL different studies (Wheatley et al., 2008; Harvey
lean rats. After 8 weeks of DR, there was no effect et al., 2013). In one study, female C57BL6 mice
on the multiplicity of total aberrant crypt foci were fed the AIN-76A diet or were subjected
in the colon (Raju & Bird, 2003). [The Working to 30% DR (with carbohydrate restriction)
Group noted the small n values; an effect was from age 7 weeks (Harvey et al., 2013); the cells
seen when distinguishing between advanced and were implanted at age 29 weeks, and tumours
early foci, which is questionable.] were harvested 24 days later. Tumour volume
Another approach by Newberne et al. (1990) was reduced significantly in the 30% DR mice
to study the effect of body weight on develop- compared with the AL mice (515 mm3 vs
ment of colon cancer was to use pups obtained 2286 mm3). In the other study (Wheatley et al.,
from litter sizes adjusted to either four or eight. 2008), male C57BL6 mice were fed a HFD (60%
Dimethylhydrazine was administered from of calories from fat) for 7 weeks and then divided
1 month after weaning for 10 weeks. In addition into four experimental groups, including a group
to rats raised in litter sizes of four or eight, some subjected to 30% DR (10% of calories from fat).
pups from the litters of four were pair-fed to the MC38 cells were injected at week 15. Tumour size
pups from the litters of eight. Tumour incidence was reduced in the 30% DR group compared with
was significantly higher in male and female all the other experimental groups. [The Working
rats raised in the smaller litters when fed AL Group noted that this study assessed DR in obese
compared with the corresponding groups raised mice.]
in the larger litters. Pair-fed male rats from litters The CT26 murine carcinoma cell line, which
of four had a higher incidence of colon tumours was developed from BALB/c mice, was used to
than the male rats from litters of eight. [The evaluate tumour growth in LFD versus HFD
Working Group considered that the use of litter mice (Park et al., 2012). The HFD group was fed
size as an indicator of food intake made interpre- from age 4–20 weeks; the cells were inoculated
tation difficult for the evaluation of the effect of at age 20 weeks, and the mice continued on their
DR on colon tumour development.] respective diets for an additional 31 days. Body
More recent studies on the effect of DR on weight was only slightly higher in the HFD mice
colon cancer development have focused on mice than in the LFD mice, whereas tumour volume
517
and weight were significantly higher in the HFD In a study by the United States National
group. Metastasis to the lungs, as determined Toxicology Program (NTP, 1997), administra-
by the number of tumour nodules, was signif- tion of salicylazosulfapyridine (SASP) in the feed
icantly higher in the HFD mice. [This study is decreased the body weights of male B6C3F1 mice
of interest because – although it did not use DR by 15% in the 2-year bioassay. To eliminate a
– the BALB/c mice were resistant to the HFD.] possible effect of the weight reduction by SASP on
the occurrence of neoplasms, a weight-matched
3.3.3 Cancer of the liver control group was included, in which the food
intake was restricted by 13–22% to reduce the
See Table 3.12. body weight to the level of AL mice treated with
Comprehensive studies of nutrition and SASP (15% less than for the untreated AL group).
ageing, conducted in collaboration with the The incidence of hepatocellular adenomas or of
United States Food and Drug Administration’s carcinomas was lower in the weight-matched
National Center for Toxicological Research and control group, although this difference was not
the United States National Institute on Aging statistically significant. However, the incidence
provided pathological data on mice subjected to of adenoma and carcinoma combined was signif-
40% DR (Blackwell et al., 1995; Sheldon et al., 1995, icantly lower in the weight-matched control
1996). In scheduled-sacrificed female B6C3F1 group compared with the untreated AL group.
mice, no liver tumours were found in DR mice Although the DR level in the 40% DR group was
until age 30 months, whereas in AL mice, 4.9% double (i.e. 40%), and therefore body weight was
(2 of 41 mice) had liver tumours at 24 months lower in the 40% DR group than in the weight-
and 13.3% (2 of 15 mice) had liver tumours at matched controls, the incidences of adenoma,
30 months (Sheldon et al., 1995). At 36 months, of carcinoma, and of adenoma and carcinoma
the incidences were similar between female AL combined were not significantly lower even
and DR mice, at 42.9% (6 of 14 mice) and 33.3% when compared with the untreated AL group.
(5 of 15 mice), respectively. In male B6C3F1 mice, In contrast, in the animals treated with SASP,
the incidence increased with advancing age in the 40% DR group had a significantly lower inci-
the AL group, and the incidence was significantly dence of hepatocellular adenoma (P < 0.001), of
lower in the DR group than in the AL group at carcinoma (P < 0.05), and of adenoma and carci-
24 months and 36 months. Necropsy data from noma combined (P < 0.001), compared with the
mice that died spontaneously or were killed SASP AL controls.
when moribund also showed that the incidence In the National Toxicology Program feeding
of liver tumours was lower in female and male study of scopolamine hydrobromide trihydrate
DR mice, compared with the respective control (SHT), the untreated weight-matched group
AL mice. In C57BL6 mice, the incidence of liver (20% lower body weight compared with the AL
tumours in scheduled-sacrificed male and female group) and the DR group had lower incidences
AL mice was less than 5%, and no tumours were of hepatocellular adenoma, and of adenoma and
found in the DR group (Blackwell et al., 1995). carcinoma combined, although for carcinoma
The incidence of liver tumours in male mice at the difference was not statistically significant
necropsy was also lower in the DR group than in (NTP, 1997). In animals treated with SHT, the
the AL group; there was no significant difference 40% DR group had a significantly lower inci-
in female mice between the AL group and the DR dence of hepatocellular adenoma (P < 0.05),
group at necropsy. of carcinoma (P < 0.05), and of adenoma and
carcinoma combined (P < 0.01), compared with
518
Table 3.12 Studies on the prevention of tumours of the liver by dietary/calorie restriction in mice
Strain (sex) Route, dose, and Type of diet, dosing Type of tumours: Statistical significance Comments
Age at start duration of carcinogen regimen, and duration of Tumour incidence (number of
Number of administration, intervention tumours/effective number of
animals duration of study animals, and/or percentage),
Duration of multiplicity, or other outcomes as
study specified
Reference
Swiss OF1 NDEA, 0.4 µmol/g bw, Nafag 857 diet, fed AL or Liver G6Pd foci; adenoma;
(M) i.p., at birth 30% DR, from age 12 wk, carcinoma; adenoma or carcinoma
Weanling killed at 24, 36, 48 wk (combined):
(21 d) Control (AL):
14–20/group 12 wk 17/20, 0/20, 0/20, 0/20
48 wk
24 wk 17/20, 16/20, 0/20, 16/20
Lagopoulos
et al. (1991) 36 wk 11/18, 18/18, 5/18, 18/18
48 wk 2/14, 9/14, 6/14, 14/14
30% DR:
12 wk 7/19*, 0/19, 0/19, 0/19 *P < 0.05 vs AL group,
χ2 test
24 wk 16/20, 0/20*, 0/20, 0/20* *P < 0.001 vs AL group,
Fisher exact test
36 wk 16/17*, 12/17, 1/17, 12/17* *P < 0.05 vs AL group,
χ2 test
48 wk 15/16*, 10/16, 0/16**, 10/16*** *P < 0.001, **P < 0.01,
***P < 0.05 vs AL
group, χ2 test or Fisher
exact test

C57BL6 Spontaneous neoplasms NIH-31 diet, fed AL or Hepatocellular tumours, mostly *P < 0.01 vs controls, Incidence of liver tumours
(F, M) 40% DR, from age 14 wk adenoma or carcinoma (%) Fisher exact test was small; therefore, data
4 wk (0–33 mo): of 0–27 mo and 28–33 mo
266/group Control (AL) F: 4/115 (3.5%), M: 18/156 (11.5%) were combined
SS at 40% DR F: 2/100 (2%), M: 2/106 (1.9%)*
12–36 mo, or
lifelong
Blackwell
et al. (1995)
519
520

study specified
Reference
B6C3F1 Spontaneous neoplasms NIH-31 diet, fed AL or Hepatocellular tumours, mostly *P < 0.05, **P < 0.001 Data of necropsied mice
(F, M) 40% DR; DR increased adenoma or carcinoma (%): vs controls, Fisher from the SS group and the
4 wk gradually: 10% at 14 wk, 0–27 mo, 28–40 mo exact test lifespan study group were
266/group 25% at 15 wk, and 40% at combined
SS at age 16 wk AL: 31 hepatocellular
12–42 mo, Control (AL) F: 7/56 (12.5%), 14/131 (10.7%) neoplasms, 13 of
or lifelong M: 13/39 (33.3%), 48/118 (40.7%) them carcinomas; 7
Sheldon 40% DR F: 0/10 (0%), 1/72 (1.4%)* metastasized to the lungs
et al. (1995) M: 0/7 (0%), 1/45 (2.2%)** 40% DR: 18 hepatocellular
neoplasms, 9 of
them carcinomas; 1
metastasized to the lungs
B6D2F1 Spontaneous neoplasms NIH-31 diet, fed AL or Benign and malignant liver tumours: *[NS]; **[P < 0.001]
(F, M) 40% DR
4 wk Control (AL) F: 13%, M: 24%
56/group 40% DR F: 4%*, M: 2%**
Lifelong
Sheldon
et al. (1996)
B6C3F1 (M) SASP, 2700 mg/kg bw in NIH-07 diet, fed AL or (on Hepatocellular adenoma; carcinoma;
6 wk corn oil by gavage, once average) 15% DR (weight- adenoma or carcinoma (combined):
50 or 52/ a day, 5 d/wk matched to SASP group),
group or 40% DR
104 wk Untreated, AL 13/50, 13/50, 24/50 *P < 0.05 vs untreated
NTP (1997) Untreated, 15% DR 8/50, 6/50, 14/50* AL group, χ2 test
Untreated, 40% DR 13/52, 7/52, 18/52 *P < 0.001, **P < 0.05
SASP, AL 42/50, 8/50, 44/50 vs SASP AL group, χ2
SASP, 40% DR 9/50*, 1/50**, 9/50* test or Fisher exact test
study specified
Reference
B6C3F1 (M) SHT, 25 mg/kg bw in NIH-07 diet, fed AL or (on Hepatocellular adenoma; carcinoma; Untreated weight-matched
6 wk corn oil by gavage, once average) 15% DR (weight- adenoma or carcinoma (combined): group had 20% lower bw
70/group a day, 5 d/wk matched to SHT group), or compared with AL group
104 wk 40% DR
NTP (1997) Untreated, AL 26/50, 6/50, 30/50 *P < 0.001 vs untreated
Untreated, 15% DR 5/50*, 5/50, 10/50* AL group, χ2 test
Untreated, DR 3/50*, 2/50, 5/50* *P < 0.001 vs untreated
AL group, χ2 test or
Fisher exact test
SHT, AL 8/50*, 7/50, 15/50** *P < 0.001, **P < 0.01
vs untreated AL group,
χ2 test
SHT, 40% DR 0/50*, 1/50*, 1/50** *P < 0.05, **P < 0.001
vs SHT AL group,
Fisher exact test
AL, ad libitum; bw, body weight; d, day or days; DR, dietary restriction; F, female; G6Pd, glucose-6-phosphatase-deficient; i.p., intraperitoneal; M, male; mo, month or months; NDEA,
N-nitrosodiethylamine; NS, not significant; NTP, National Toxicology Program; SASP, salicylazosulfapyridine; SHT, scopolamine hydrobromide trihydrate; SS, scheduled-sacrificed;
vs, versus; wk, week or weeks.

521
the SHT AL controls. [The protective effect on Another genetic model uses FVB-Tg (BK5.
adenoma development may have been due to COX-2) mice and calculates a composite score for
decreased food intake and was not a direct effect pancreatic dysplasia. In the study of Lashinger
of administration of SHT.] et al. (2011), formation of severe dysplasia in
In male Swiss OF1 mice treated with pancreatic ducts was lower in the 30% DR group
N-nitrosodiethylamine, the incidence of than in the AL group. When tumour cells were
glucose-6-phosphatase-deficient pre-neoplastic injected into wild-type mice, tumour weight at
foci was significantly lower in the 30% DR group 4 weeks after injection was significantly lower
compared with the AL group at age 12 weeks. in 30% DR mice. Lashinger et al. (2013) used
In the AL group, 80% of the mice had hepato- KrasG12D/Ink4a+/− male mice and observed longer
cellular adenoma or carcinoma at 24 weeks, and median survival in 30% DR mice than in AL
100% at 36 weeks and 48 weeks. The incidence mice, and no PDAC in 30% DR mice compared
of hepatocellular neoplasms (adenoma and carci- with 3 PDAC in AL mice.
noma combined) was significantly lower in the In Sprague-Dawley rats, 4–6% of the AL group
30% DR group than in the AL group at 24 weeks, rats spontaneously developed pancreatic islet
36 weeks, and 48 weeks (Lagopoulos et al., 1991). adenomas or carcinomas (Keenan et al., 1995).
The overall incidence of pancreatic islet neoplasms
3.3.4 Cancer of the pancreas over a 2-year study seemed to be lower in the 35%
DR group. [The baseline incidence was low, and
See Table 3.13. therefore the effect of 35% DR on spontaneous
The incidence of spontaneous pancreatic islet neoplasms could not be evaluated.] In another
tumours is very low in mice, on both C57BL6 study in Sprague-Dawley rats (Molon-Noblot et al.,
and B6C3F1 backgrounds (Blackwell et al., 1995; 2001), the incidence of adenomas was 24% and 18%
Sheldon et al., 1995). LSL-KrasG12D; Pdx-1/Cre in female and male AL groups, respectively. The
mice, a genetic model, develop pancreatic inhibition of adenomas by three different levels
precursor lesions such as PanIN, which progress of DR was significant in female rats, whereas the
to PDAC. In one study (Lanza-Jacoby et al., effect of DR was modest in male rats. In a third
2013), two regimens for 25% DR were used with study in male Sprague-Dawley rats (Duffy et al.,
this model. One was CDR at 25% less than the 2008), the incidence of islet adenomas was lower in
AL average intake; the other was IDR, i.e. 50% the 31% DR group than in the AL group, but this
restriction for 1 week after 100% provision of AL difference was not statistically significant.
intake for 1 week, and therefore IDR also reduced In male Lewis rats, the effect of DR on the
the calorie intake by 25% over the 2-week post-initiation phase of pancreatic carcinoma
interval. The body weight of the IDR mice fluc- induced by azaserine was assessed (Roebuck et al.,
tuated according to calorie intake. In the 100% 1993). Feeding AL for a limited time (5–6 hours per
feeding week, body weights were similar to those day), designated as a “meal-fed” regimen, reduced
in the AL group; in the 50% restriction week, the food intake to an equivalent of 10–15% DR
body weights were lower than those in the CDR relative to AL. The meal-fed regimen significantly
mice. The incidence of PanIN was lower in both reduced the incidence of adenomas and carci-
the CDR and IDR groups than in the AL group, nomas 14 months after azaserine initiation.
with a greater effect of DR in the IDR regimen In male Syrian golden hamsters, the inci-
than in the CDR regimen. PDAC was found in dence of pancreatic carcinoma, induced by BOP,
the AL group, whereas no PDAC was observed in did not differ among control AL, 20% CR, and
the CDR and IDR groups at age 44 weeks. 40% CR groups (Birt et al., 1997). However, the
522
Table 3.13 Studies on the prevention of tumours of the pancreas by dietary/calorie restriction in experimental animals
Species, strain (sex) Route, dose, and Type of diet, dosing Type of tumours: Statistical significance Comments
Age at start duration of carcinogen regimen, and duration Tumour incidence (number of
Number of animals administration, of intervention tumours/effective number of
Duration of study duration of study animals, and/or percentage),
Reference multiplicity, or other outcomes as
specified
Mouse
FVB-Tg (BK5.COX- Genetic model for Control diet (Research Composite score for pancreatic **P < 0.01
2) (F, M) dysplastic lesions Diets, Inc., New dysplasia:
6 wk Brunswick, NJ, AL: 10 ± 4
24/group #D12450B), fed AL or 30% DR: 5 ± 4**
20 wk 30% DR
Lashinger et al.
(2011)
FVB-WT (F, M) JC101 pancreatic Control diet (Research Tumour weight 4 wk after s.c. *P < 0.05
6 wk tumour cell injection at Diets, Inc., New transplantation:
24/group age 13–15 wk Brunswick, NJ, AL: 1.05 ± 0.38 g
20 wk #D12450B), fed AL or 30% DR: 0.47 ± 0.30 g*
Lashinger et al. 30% DR
(2011)
LSL-KrasG12D; Pdx-1/ Genetic model for Modified AIN-93 diets, PanIN (PanIN-2 or greater); PDAC: *P < 0.05 vs AL control IDR: 50% DR for 1
Cre (M) PDAC fed AL or 25% DR (CDR) Control (AL): 70%, 27% (n = 11) group, exact Poisson wk, then 100% of AL
6 wk or IDR, 6–44 wk CDR: 40%*, 0% (n = 15) regression analysis intake for 1 wk
31/group
IDR: 27%*/#, 0% (n = 16) *P < 0.05 vs AL control
44 wk
group, #P < 0.05 vs CDR
Lanza-Jacoby et al.
group, exact Poisson
(2013)
regression analysis

KrasG12D/Ink4a+/– Genetic model for Control diet (Research PanIN (PanIN-2 or greater); PDAC: NS
(M) PanIN-2 or PDAC Diets, Inc., New 10 wk study:
6 wk Brunswick, NJ, AL: 7/15 (46.7%), 3/15 (20%)
42–43/group #D12450B), fed AL 30% DR: 9/15 (60%), 0/15 (0%)
10 wk or 56 wk or 30% DR, from age 56 wk study, median survival: *P < 0.001, log-rank test
Lashinger et al. 6–9 wk AL: 20.0 wk
(2013) 30% DR: 30.0 wk*
523
524

specified
Rat
Lewis (M) Azaserine, 30 mg/kg bw, AIN-76A diet, fed AL or Pancreatic adenoma (%); carcinoma *P < 0.05 vs control
2 wk i.p., once, at age 2 wk 10–15% DR by meal- (%); adenoma multiplicity; group, χ2 test or Fisher
22–23/group feeding carcinoma multiplicity: exact test
14 mo #P < 0.05 vs AL
Control AL: 23/23 (100%), 15/23 (65%),
Roebuck et al. 8.04 ± 0.99, 1.60 ± 0.214 group, ANOVA with
(1993) 10–15% DR: 5/22 (23%)*, 0/22 (0%)*, Bonferroni test
1.00 ± 0#, 0
Sprague-Dawley Spontaneous neoplasms Purina Certified Rodent Pancreatic islet adenoma (%); islet Chow 5002: 21.4%
[Crl:CD® (SD) BR] Chow 5002 or 5002-9 at carcinoma (%): protein, 5.7% fat,
(F, M) various regimens 4.1% crude fibre,
36 d Chow 5002: energy value
70/group Control (AL) F: 3/70 (4.3%), 0/70 (0%) — 3.07 kcal/g;
106 wk M: 4/70 (6.2%), 4/70 (6.2%) Chow 5002-9:
Keenan et al. (1995) DR 6.5 h (AL for 6.5 h) F: 1/70 (1.4%), 1/70 (1.4%) NS 13.6% protein, 4.6%
M: 2/70 (2.9%), 7/70 (10%) NS fat, 15.7% crude
fibre, energy value
35% DR F: 1/70 (1.4%), 0/70 (0%) NS
2.36 kcal/g
M: 5/70 (7.1%), 1/70 (1.4%) NS
Baseline incidence
Chow 5002-9: was low, so that the
Control (AL) F: 3/70 (4.3%), 0/70 (0%) — effect of DR could
M: 3/70 (4.3%), 5/70 (7.1%) not be evaluated
DR (fed at the same F: 1/70 (1.4%), 2/70 (2.9%) NS
calorie intake as rats fed M: 3/70 (4.3%), 0/70 (0%) NS
35% DR chow 5002)
specified
Sprague-Dawley Spontaneous neoplasms Purina Certified Rodent Pancreatic islet adenoma (%); islet *P < 0.05 vs AL group,
[Crl:CD® (SD) IGS Diet, 21–28% DR, carcinoma (%): Fisher exact test
BR] (F, M) 28–32% DR, 52–53% DR
7 wk Control (AL) F: 12/50 (24%), 2/50 (4%)
15/group M: 9/50 (18%), 3/50 (6%)
100 wk 21–28% DR F: 3/52* (5.8%), 0/52 (0%)
Molon-Noblot et al. M: 16/50 (32%), 5/50 (10%)
(2001)
28–32% DR F: 3/52* (5.8%), 0/52 (0%)
M: 12/50 (24%), 0/50* (0%)
52–53% DR F: 1/51* (2%), 0/51 (0%)
M: 4/50 (8%), 3/50 (6%)
Sprague-Dawley Spontaneous neoplasms AIN-93M diet, fed AL Pancreatic islet adenoma: NS (poly-3 test)
[Crl:CD® (SD) BR] or 31% DR from age AL: 11.8% (n = 57)
(M) 6–114 wk 31% DR: 6.7% (n = 38)
6 wk
40 or 60/group
108 wk
Duffy et al. (2008)
Hamster
Syrian golden (M) BOP, 20 mg/kg bw, 3 Control diet, fed AL or Pancreatic carcinoma; multiplicity: Incidence: NS
8 wk weekly s.c. injections 20% CR or 40% CR for AL: 17/29, 0.9 ± 0.2 Multiplicity: *P < 0.02

25–35/group 42–44 wk 20% CR: 19/29, 1.2 ± 0.2 vs control
Up to 44 wk 40% CR: 17/26, 1.7 ± 0.3*
Birt et al. (1997)
AL, ad libitum; ANOVA, analysis of variance; BOP, N-nitrosobis(2-oxopropyl)amine; bw, body weight; CDR, chronic dietary restriction; CR, calorie restriction; d, day or days; DR,
dietary restriction; F, female; h, hour or hours; IDR, intermittent dietary restriction; i.p., intraperitoneal; M, male; mo, month or months; NS, not significant; PanIN, pancreatic
intraepithelial neoplasia; PDAC, pancreatic ductal adenocarcinoma; s.c., subcutaneous; vs, versus; wk, week or weeks; WT, wild-type.
525
multiplicity of pancreatic carcinoma was greater phase had significantly lower incidence of skin
in the 40% CR group than in the AL group. papilloma and skin carcinoma.
Birt et al. (1993) tested the effect of 35%
3.3.5 Cancer of the skin and subcutaneous CR from fat or carbohydrate in the chemically
tissue induced DMBA skin tumour model in female
SENCAR mice. The incidence of skin carcinoma
See Table 3.14. was significantly lower in the groups subjected to
Lifespan studies using mice and rats have CR from either fat or carbohydrate, and there was
indicated the low incidence of spontaneous skin no difference in incidence between the two CR
tumours (Blackwell et al., 1995; Sheldon et al., regimens, although papilloma multiplicity was
1995). In one study, less than 2% of scheduled-sac- greater in the carbohydrate-restricted (and thus
rificed male B6C3F1 mice had skin or subcuta- HFD) group, compared with the fat-restricted
neous tumours (Sheldon et al., 1995). In contrast, group [no statistics were reported]. A subsequent
female mice had relatively high incidences of skin study indicated that moderate (i.e. 20%) CR of
or subcutaneous tumours. However, there was a HFD or a LFD was ineffective, suggesting the
no statistical difference between the AL and DR importance of the level of CR (Birt et al., 1996).
groups for either sex (Sheldon et al., 1995). [The Birt et al. (1994) also tested the effect of
occurrence of skin tumours could be delayed in 40% DR in the two-stage promotion protocol,
the DR group.] which comprised early-stage treatment
Topical application of benzo[a]pyrene is one with DMBA at age 9 weeks, promotion by
of the models used to induce skin tumours in 12-O-tetradecanoylphorbol-13-acetate (TPA) for
rodents. Boutwell et al. (1949) demonstrated that 2 weeks, and subsequent late-stage promotion
50% DR reduced the incidence of skin carcinoma by mezerein for 15 weeks. The individual effect
induced by benzo[a]pyrene. of DR on the promotion phase by either TPA or
Birt et al. (1991) reported the effect of DR on mezerein was evaluated. Mice subjected to DR at
either the initiation or the promotion phase of age 10–27 weeks recovered their weight loss once
a chemically induced skin carcinoma model in they returned to AL feeding with a control diet at
female SENCAR mice. They used two regimens age 28 weeks. However, those mice subjected to
for 40% DR: one was a total DR of the control DR during the period of treatment with TPA and/
diet (TDR), and the other involved CR using or mezerein had a reduction of approximately
a diet that was low in fat and glucose but high 10% in body weight compared with the control
in protein and fibre. Mice that were subjected AL mice at age 71 weeks. DR during the entire
to 40% TDR or CR at the initiation phase had period of promotion, i.e. at age 10–27 weeks (DR/
reduced weight gain, and the subsequent AL DR group), and DR during the period of treat-
regimen led to recovery of the weight gain within ment with mezerein, at age 12–27 weeks (AL/
4 weeks. Mice that were subjected to 40% TDR DR group), significantly reduced the incidence
and CR starting at the promotion phase had and multiplicity of skin papilloma at 28 weeks;
significantly lower body weight according to the the cumulative incidence of skin carcinoma was
energy restriction levels. Neither 40% TDR nor also reduced (not significantly) in the DR/DR
CR at the initiation phase affected the incidence group and the AL/DR group. However, 2-week
or multiplicity of skin papilloma or the incidence DR feeding during TPA treatment was insuffi-
of skin carcinoma. Mice that were subjected to cient to produce an inhibitory effect of DR on
40% TDR and CR starting at the promotion skin tumorigenesis.
526
Table 3.14 Studies on the prevention of tumours of the skin and subcutaneous tissue by dietary/calorie restriction in mice
Strain (sex) Route, dose, and duration of Type of diet, dosing Type of tumours: Statistical Comments
Age at start carcinogen administration regimen, and duration of Tumour incidence (number of significance
Number of intervention (if not until tumours/effective number of
animals termination of study) animals, and/or percentage),
Duration of multiplicity, or other
study outcomes as specified
Reference
Rockland (F) 60 µg of benzo[a]pyrene (skin High-calorie, low-fat (AL), Skin carcinoma: **P < 0.01, χ2 test
~2–3 mo application), twice a wk for or low-calorie, low-fat (50%
48/group 19 wk from 26 d after beginning DR)
173 d of DR Control (AL) 32/39
Boutwell 50% DR 8/44**
et al. (1949)
C57BL/6 (M) 105 B16 melanoma cells injected Purina Laboratory Chow, Tumour volume (mm3)/1000 *P < 0.05,
4.5 mo s.c., 2 wk after beginning of DR fed AL or 40% DR (mean ± SE): Student t test
7–10/group Control (AL) 7.9 ± 1.5
5–6 wk 40% DR 2.3 ± 0.5*
Ershler et al.
(1986)
SENCAR (F) 10 nmol of DMBA (skin AIN; fed AL or 40% DR by Skin papilloma (%) at age Both TDR and CR from
6 wk or application), once at age TDR or CR, from 6–9 wk 30 wk; papilloma multiplicity 10–56 wk lowered bw by 30%
10 wk 9 wk; 3.2 nmol of TPA (skin or 10–56 wk at age 34 wk (mean ± SE); skin No significant difference
15–30/group application), twice a wk for carcinoma (%) until age 56 wk in tumour incidence and
Up to 56 wk 20 wk from age 10 wk Control (AL) 27/30 (90%), 5.7 ± 0.4, 15/24 — multiplicity between TDR
Birt et al. (71%) and CR groups
(1991) 40% TDR, 6–9 wk 24/30 (80%), 4.9 ± 0.9, 14/24 NS [Multiplicity at age 34 wk read
(58%) from figure]

40% TDR, 10–56 wk 11/20 (55%)***, 2.2 ± 0.4***, ***P < 0.001,
7/17 (41%)* *P < 0.05 vs
AL, χ2 test or
ANOVA
40% CR, 6–9 wk 25/29 (89%), 6.6 ± 0.9, 11/17 NS
(69%)
40% CR, 10–56 wk 13/23 (56%)***, 1.8 ± 0.4***, ***P < 0.001,
5/15 (28%)* *P < 0.05 vs
AL, χ2 test or
ANOVA
527
528

Reference
CD-1 (F) 51.2 µg of DMBA (skin Purina 5015 chow, fed Number of skin papillomas/ *P < 0.01,
~6–7 wk application), once at age 6 wk AL or ~35% DR from age mouse: Wilcoxon
42–59/group or 7 wk; 2 µg of TPA (skin 8–9 wk Mann–Whitney
100 d application), twice a wk for 82 d Control (AL) 5.5 sum test
Pashko & from 1 wk after beginning of DR 0.9*
Schwartz DR
(1992)
SENCAR (F) 10 nmol of DMBA (skin AIN diet fed AL, balanced Number of papillomas/mouse *P < 0.05 vs Bw reduced in HCD and HFD
6 wk application), once at age 9 wk; high fat (BHD) AL, 35% at 28 wk after DMBA; time to control and vs control AL and BHD AL
38–42/group 2.0 µg of TPA (skin application), CR with high carbohydrate carcinoma at 50% incidence BHD AL groups; no difference between
Up to 68 wk twice a wk for 20 wk from age (HCD), 35% CR with high #P < 0.01 vs HCD and HFD, or between
Birt et al. 10 wk fat (HFD), from age 10 wk control and control AL and BHD AL
(1993) Control AL 6.5, ~40 wk BHD AL, log-
BHD AL 6.2, ~44 wk rank test
35% CR HCD 1.5*, > 59 wk#
35% CR HFD 3.2*, > 59 wk#
SENCAR (F) 10 nmol of DMBA (skin Ingredient diet, fed AL Skin papilloma; papilloma Bw loss of ~10% in DR mice
7 wk application), once at age or 40% DR at various multiplicity at age 28 wk; skin vs AL at age 71 wk
30–52/group 9 wk; 3.2 nmol of TPA (skin regimens carcinoma (%) at age 71 wk: Multiplicity was tested by
Up to 62 wk application), twice a wk at Control (AL), age 7–71 wk 42%, 0.9, 14/48 (29%) — a Poisson random variable,
Birt et al. 10–11 wk, and 10 nmol of MEZ 40% DR, 10–27 wk; AL, 17%*, 0.3*, 3/31# (10%) *P < 0.05, #NS using the generalized
(1994) twice a wk at age 12–27 wk 28–71 wk vs AL, χ2 test or estimating equation approach
Fisher exact test of Liang and Zeger, and the
estimation of regression
40% DR, 10–11 wk; AL, 54%, 0.9, 10/30 (33%) NS
coefficients
12–71 wk
AL, 10–11 wk; DR, 17%*, 0.2*, 5/33 (15%) *P < 0.05 vs AL
12–27 wk; AL, 28–71 wk
Reference
B6C3F1 Spontaneous neoplasms NIH-31 diet, fed AL or Skin tumours (%): NS Incidence of spontaneous
(F, M) 40% DR; DR increased 0–27 mo, 28–33 mo, 34–39 mo, skin tumours was low even in
4 wk gradually: 10% at 14 wk, 40–51 mo: the AL groups
245–266/ 25% at 15 wk, and 40% at
group age 16 wk; SS at age 12, 18,
SS at 24, 30, 36, 42 mo
12−42 mo, or Control (AL) M: 1/72 (1.4%), 1/83 (1.2%),
lifelong 0/65 (0%), 0/15 (0%)
Sheldon et al. F: 4/97 (4.1%), 10/110 (9.1%),
(1995) 1/50 (2%), 0/1 (0%)
40% DR M: 0/49 (0%), 0/27 (0%), 0/48
(0%), 0/138 (0%)
F: 1/52 (1.9%), 2/38 (5.3%), 6/64
(9.4%), 3/111 (2.7%)
SENCAR (F) 10 nmol of DMBA (skin Fed AL, 20% or 40% CR Skin carcinoma (%):
9 wk application), once at age from fat or carbohydrate
35/group 9 wk; 3.2 nmol of TPA (skin using LFD (10% fat) or
45 wk application), twice a wk for HFD (42% fat), from age
Birt et al. 18 wk from age 10 wk 10 wk
(1996) Control: LFD AL 35% *P < 0.05 vs LFD

20% CR of LFD 20% AL, HFD AL,
40% CR of LFD 15%* and 20% CR of
HFD AL 35% HFD
20% CR of HFD 35%
40% CR of HFD 0%*
ICR (M), 25 µg of DMBA (skin Teklad 2018 diet, fed AL or Skin tumour: time at 25% and Kaplan–Meier survival
Nrf2 gene application), once at age 20% DR, 30% DR, or 40% 50% incidence: analysis was performed to
knockout 15 wk or 16 wk; 4 µg of TPA DR for 5–6 wk compare the 2 curves of
10 wk (skin application), twice a wk, Control (AL) 13 wk, 15 wk — papilloma occurrence
10–15/group from 2 wk after DMBA until 20% DR 18 wk, 25 wk (< 50% incidence) P < 0.05 Time of appearance of
up to 42 wk appearance of first papilloma 30% DR 18 wk, 20 wk tumour was recorded when
Pearson et al. 40% DR 42 wk, 42 wk at least 1 papilloma with a
(2008) radius > 1 mm was identified
529
530

Reference
129S1/SvImJ 25 µg of DMBA (skin AIN-93G diet, fed AL or Skin papilloma (%); *P < 0.0083 [Values read from graph]
(M) application), once; 4 µg of TPA 30% DR multiplicity ± SD vs WT AL,
3–5 mo (skin application), twice a wk, Bonferroni t
11–13/group from 2 wk after DMBA until tests
18 wk appearance of first papilloma Control (AL) 100%, 4.0 ± 2.7 —
Minor et al. 30% DR 58%, 1.0 ± 1.1*
(2011)
ICR (F) 25 nmol of DMBA (skin AIN-76A diet, fed AL,15% Skin papilloma (%) at age Bw lower in both DR groups
7 wk application), once at age 7 wk; DR, or 30% DR from age 30 wk; multiplicity at age 39
30/group 3.4 nmol of TPA, twice a wk 11 wk wk; skin carcinoma (%) at
58 wk (skin application), from age age 65 wk; multiplicity at age
Moore et al. 15 wk for 50 wk 65 wk:
(2012) Control (AL) 81%, 8.2, 92%, 1.6 —
15% DR 81%, 6.2*, 69%#, 1.6 *P < 0.05 vs AL,
Mann–Whitney
U test
#P < 0.05 vs AL,
χ2 test
30% DR 68%, 4.3*, 58%#, 1.0* *P < 0.05 vs AL,
Mann–Whitney
U test
#P < 0.05 vs AL,
χ2 test
AL, ad libitum; ANOVA, analysis of variance; BHD, balanced high-fat diet; bw, body weight; CR, calorie restriction; d, day or days; DMBA, 7,12-dimethylbenz[a]anthracene; DR, dietary
restriction; F, female; HCD, high-carbohydrate diet; HFD, high-fat diet; LFD, low-fat diet; M, male; MEZ, mezerein; mo, month or months; NS, not significant; s.c., subcutaneous; SD,
standard deviation; SE, standard error; SS, scheduled-sacrificed; TDR, total dietary restriction; TPA, 12-O-tetradecanoylphorbol-13-acetate; vs, versus; wk, week or weeks; WT, wild-
type.
Moore et al. (2012) used female ICR mice to 1995). In female B6C3F1 and B6D2F1 mice, the
assess the effect of 15% DR or 30% DR on the incidence of spontaneously occurring pituitary
promotion of skin tumours. At the end of the tumours was also lower in the DR group than in
experiment, the body weights were 36% lower in the AL group (Sheldon et al., 1995, 1996).
the 30% DR group and 15% lower in the 15% DR In a 2-year study in Sprague-Dawley rats fed
group, compared with the control AL group. The 35% DR with either a standard diet or a low-pro-
cumulative incidences of skin papilloma in the tein, high-fibre diet, incidences of spontaneous
experimental groups did not differ significantly pituitary adenomas were very high in both male
among these groups. However, the multiplicity and female AL rats (Keenan et al., 1995). In male
of skin papilloma was significantly lower in the rats, both 35% DR groups had lower incidences
15% DR and 30% DR groups compared with the of pituitary adenomas. In contrast, in female rats
AL group. The incidences of skin carcinoma were the preventive effect of 35% DR was observed
also significantly lower in the 15% DR and 30% only in the 35% DR group fed the low-protein,
DR groups; the multiplicity of skin carcinoma high-fibre diet and not in the 35% DR group fed
was significantly lower only in the 30% DR group the standard diet.
(Moore et al., 2012). In a lifespan study in F344 rats (Thurman
In the B16 melanoma cell injection model, et al., 1994), the incidence of pituitary tumours in
the tumour volume was significantly lower in the 40% DR group was significantly lower than
C57BL/6 mice subjected to 40% DR compared in the AL group in both female and male rats.
with AL mice (Ershler et al., 1986). The mean age at death of rats bearing pituitary
Using the two-stage skin tumorigenesis tumours was also higher in the 40% DR group
model in CD-1 mice treated with DMBA, Pashko compared with the AL group in both male and
& Schwartz (1992) reported that DR suppressed female rats.
TPA promotion of skin papillomas. Estrogen stimulates the proliferation of
The preventive effect of 40% DR on DMBA- prolactin-producing lactotrophs, and therefore
TPA-induced skin tumours was diminished in continuous administration of estrogen promotes
ICR mice (Pearson et al., 2008) and in 129S mice the development of prolactin-producing tumours
(Minor et al., 2011). in the rat. Pituitary weight can be measured as a
quantitative indicator of estrogen-induced pitu-
3.3.6 Cancer of the pituitary gland itary tumour development, because increased
weight correlates with increases in pituitary cell
See Table 3.15. number and DNA content. Several studies (Shull
In lifespan studies in mice, the incidence of et al., 1998; Spady et al., 1998, 1999; Harvell et al.,
pituitary tumours is very low in males (Blackwell 2001) have used this model to study pituitary
et al., 1995; Sheldon et al., 1995, 1996). tumour development, and some have shown a
The incidence in female scheduled-sacrificed reduction of tumour development with 40% DR.
C57BL6 control mice was reported to be 14% at [In this model, response to DR for the inhibition
24 months and 64% at 30 months. In DR mice, no of tumours depends on the strain of rat used.
pituitary tumours were found at 24 months and These results are confounded by the fact that body
30 months. Necropsies of mice that died spon- weight was reduced by estrogen administration
taneously or were killed when moribund also in the AL group, whereas it was not significantly
indicated a significant reduction in the incidence reduced in the DR group. In addition, there was
of pituitary tumours in the female DR group no indication of histopathology or of tumour
compared with the AL group (Blackwell et al., incidence. Therefore, these studies are regarded
531
532

Table 3.15 Studies on the prevention of tumours of the pituitary gland by dietary/calorie restriction in mice and rats
Species, Route, dose, and Type of diet, dosing Type of tumours: Statistical Comments
strain (sex) duration of carcinogen regimen, and duration of Tumour incidence (number of significance
Age at start administration intervention (if not until tumours/effective number of
Number of termination of study) animals, and/or percentage),
animals multiplicity, or other outcomes as
Duration of specified
study
Reference
Mouse
C57BL6 (F) Spontaneous neoplasms NIH-31 open formula diet, Pituitary tumours (%): [*P < 0.05,
Weanling fed AL or 40% DR from age Mice SS at 0–24 mo, 30 mo: **P < 0.0005 vs
266/group 4 wk; SS at 12, 18, 24, 30, controls, Fisher
SS at 12– 36 mo exact test]
36 mo, or Control (AL) 6/43 (14%), 9/14 (64%)
lifelong 40% DR 0/44 (0%)*, 0/15 (0%)**
Blackwell
Pituitary tumours (%): *P < 0.001 vs
et al. (1995)
Mice that died spontaneously or controls, Fisher
were killed when moribund in the exact test
SS and the lifespan study groups
(0–27 mo, 28–33 mo):
Control (AL) 17/75 (22.7%), 31/40 (77.5%)
40% DR 0/50 (0%)*, 2/50 (4%)*
B6C3F1 (F) Spontaneous neoplasms NIH-31 open formula diet, Pituitary tumours: NS
Weanling fed AL or 40% DR from age Mice SS at 24 mo, 30 mo, 36 mo:
266/group 4 wk; SS at 12, 18, 24, 30,
SS at 12– 36 mo
36 mo, or Control (AL) 0/41, 4/15, 3/14
lifelong 40% DR 0/42, 1/15, 0/15
Sheldon et al.
Pituitary tumours (%): *P < 0.001 vs
(1995)
Mice that died spontaneously or controls, Fisher
were killed when moribund in the exact test
(0–27 mo, 28–33 mo, 34–39 mo):
Control (AL) 3/56 (5.4%), 13/95 (13.7%), 15/36
(41.7%)
40% DR 0/10 (0%), 0/23 (0%)*, 1/49 (2.0%)*
study
Reference
B6D2F1 (F) Spontaneous neoplasms NIH-31 open formula diet, Pituitary tumours: NR
Weanling fed AL or 40% DR
4 wk Control (AL) 17%
56/group 40% DR 0%
Lifelong
Sheldon et al.
(1996)
Rat
F344 (F, M) Spontaneous neoplasms NIH-31 open formula diet, Pituitary tumours: *P < 0.01, Mean age at death of 40%
4 wk fed AL or 40% DR, from Mice SS at 18 mo, 24 mo, 30 mo: **P < 0.05 vs DR rats bearing pituitary
54/group age 14 wk controls, Fisher tumours was higher
SS at 18– Control (AL) F: 3/10, 11/12, 11/12 exact test compared with AL rats in
30 mo, or M: 6/10, 9/11, 9/9 both sexes
lifelong
40% DR F: 0/10, 4/12*, 4/12*
Thurman
et al. (1994) M: 2/12, 3/12**, 7/12
Pituitary tumours: *P < 0.01 vs
Mice that died spontaneously or controls, χ2 test
were killed when moribund:

Control (AL) F: 118/160, M: 92/128
40% DR F: 60/137*, M: 55/152*
533
534

study
Reference
Sprague- Spontaneous neoplasms Purina Certified Rodent Pituitary adenoma: No difference in incidence
Dawley Chow 5002 (energy value: of pituitary focal
[Crl:CD® (SD) 3.07 kcal/g) or 5002-9 hyperplasia between
BR] (F, M) (energy value: 2.36 kcal/g), groups
36 d fed at various regimens
70/group Chow 5002:
106 wk Control (AL) F: 50/70, M: 40/70 —
Keenan et al. DR 6.5 h (AL for 6.5 h) F: 49/70, M: 32/70 NS
(1995)
35% DR F: 47/70, M: 28/70* *P < 0.05 vs
controls, χ2 test
Chow 5002-9:
Control (AL) F: 55/70, M: 37/70 NS
DR (fed at the same calorie F: 31/70*, M: 19/70** *P < 0.001,
intake as rats fed 65% DR **P < 0.01 vs
chow 5002) controls, χ2 test
F344 (M) Silastic tubing implants Ingredient diet, fed AL or Prolactin-producing pituitary *P < 0.05 vs Pituitary weight was
32 d containing 5 mg of DES, s.c. 40% DR tumour: controls measured as a quantitative
6–8/group at age 39 d; animals killed Fold increase of the weight vs indicator of estrogen-
9 wk 8 wk after DES DES-untreated counterparts; induced pituitary
Shull et al. pituitary-to-bw ratio (bw ± SD): development
(1998) Control (AL) 11.2-fold, 55.8 × 10−5 (125 ± 7 g) DES treatment reduced
40% DR 3.5-fold*, 18.4 × 10−5 (80 ± 3 g) the food intake and thus
bw by ~50% in both AL
and DR groups
study
Reference
Holtzman (M) Silastic tubing implants Ingredient diet, fed AL or Prolactin-producing pituitary NS Pituitary weight was
32 d containing 5 mg of DES, s.c. 40% DR tumour: measured as a quantitative
6–8/group at age 39 d; animals killed Fold increase of the weight vs indicator of estrogen-
9 wk 8 wk after DES DES-untreated counterparts; induced pituitary
Shull et al. pituitary-to-bw ratio (bw ± SD): development
(1998) Control (AL) 5.3-fold, 31.6 × 10−5 (263 ± 10 g) DES treatment reduced
40% DR 4.1-fold, 30.2 × 10−5 (175 ± 5 g) the food intake and thus
the bw by ~40% in both
AL and DR groups
F344 (F-OVX) Silastic tubing implants Ingredient diet, fed AL, Prolactin-producing pituitary *P < 0.05 vs Pituitary weight was
57 d containing 27.5 mg of E2, 25% DR, or 40% DR tumour: controls measured as a quantitative
5–8/group by s.c. injection at age 63 d; Fold increase of the weight vs indicator of estrogen-
11 wk animals killed 10 wk after E2-untreated counterparts: induced pituitary
Spady et al. E2 Control (AL) 4.9-fold development
(1998) 25% DR 4.1-fold
40% DR 2.0-fold*
ACI (F-OVX) Silastic tubing implants Ingredient diet, fed AL or Prolactin-producing pituitary NS Pituitary weight was
35 d containing 27.5 mg of E2, 40% DR tumour: measured as a quantitative
12/group s.c. at age 45 d; animals Fold increase of the weight vs E2- indicator of estrogen-

22 wk killed 20 wk after E2 untreated counterparts; pituitary- induced pituitary
Spady et al. to-bw ratio (bw): development
(1999) Control (AL) 5.3-fold, 31.6 × 10−5 (160 g) E2 treatment reduced
40% DR 4.1-fold, 30.2 × 10−5 (85 g) the bw by 22% in AL
rats and 21% in DR rats,
vs respective untreated
groups
535
536

study
Reference
COP (F-OVX) Silastic tubing implants Ingredient diet, fed AL or Prolactin-producing pituitary NR Pituitary weight was
or ACI containing 27.5 mg of E2, 40% DR tumour: measured as a quantitative
(F-OVX) s.c. at age 45 d; animals Fold increase of the weight vs E2- indicator of estrogen-
35 d killed 12 wk after E2 untreated counterparts: induced pituitary
12/group COP development
13 wk Control (AL) 2.4-fold
Harvell et al. 40% DR 1.5-fold
(2001)
ACI
Control (AL) 3.4-fold
40% DR 4.5-fold
AL, ad libitum; bw, body weight; d, day or days; DES, diethylstilbestrol; DR, dietary restriction; E2, 17β-estradiol; F, female; h, hour or hours; M, male; mo, month or months; NR, not
reported; NS, not significant; OVX, ovariectomized; s.c.; subcutaneous; SD, standard deviation; SS, scheduled-sacrificed; vs, versus; wk, week or weeks.
as less informative and are considered to provide to DR. [The Working Group noted the difficulty
only supporting evidence.] in performing morphometric measures in the
prostate gland.]
3.3.7 Cancer of the prostate Another model has been the TRAMP mouse.
When TRAMP mice were subjected to 20%
See Table 3.16. DR from age 7 weeks for 4 weeks or 13 weeks,
Several rat models have been used to evaluate lesions of a lower grade were reported compared
the effects of DR on prostate cancer development. with AL mice [no body weight information was
Only one early study reported the effect of DR provided] (Suttie et al., 2003). In a second study
on spontaneous prostate cancer development in from the same group, 20% DR was not imple-
rats. Lobund-Wistar rats raised in conventional mented until age 20 weeks and had no effect on
or germ-free conditions were followed up for survival or lesion severity [body weight data were
up to 41 months. Rats subjected to 30% DR and not presented] (Suttie et al., 2005). In an addi-
raised in conventional conditions had reduced tional study using TRAMP mice, two different
incidence of prostate adenocarcinoma compared modes of DR were used, with the same overall
with the AL rats (6% vs 26%), but among rats degree of restriction, i.e. 25% DR (Bonorden
raised in germ-free conditions, the incidence et al., 2009a, b). Mice that received IDR – 2 weeks
was higher in the 30% DR rats than in the AL of 50% DR with 2 weeks of AL feeding, for
rats (10% vs 5%), although the overall incidence 11 cycles – had delayed time to prostate tumour
of prostate cancer was reduced for rats raised in detection, compared with both AL and 25% DR
germ-free conditions compared with those raised mice. Body weights were lower in the 25% DR
in conventional conditions (Pollard et al., 1989). mice than in the AL mice. [Although the find-
In a carcinogen-induced prostate cancer ings are of interest, the Working Group noted the
model, Wistar-Unilever rats were treated with strong influence of the transgene as the mice age,
the luteinizing hormone-releasing antagonist thus possibly limiting the model’s usefulness for
cyproterone, followed by treatment with testos- evaluating the effect of DR on prostate cancer
terone, followed by administration of MNU to development.]
induce prostate cancer. Rats subjected to 20% The Hi-Myc mouse model was also used to
DR had longer prostate cancer-free survival assess the effects of 30% DR or a HFD (60% fat)
compared with AL rats, and this was accompa- implemented at age 6 weeks compared with an
nied by reduced body weight (Boileau et al., 2003). AL group. At age 26 weeks, the incidence of pros-
However, in a similar study, no effect of 15% DR tate adenocarcinomas was 62% in the AL mice,
or 30% DR on incidence of prostate cancer was compared with no tumours observed in the 30%
reported, and DR did not reduce body weight DR group (Blando et al., 2011).
gain (McCormick et al., 2007).
Transgenic models have also been used to
3.3.8 Cancers of the haematopoietic system
determine the effect of DR on the development of
prostate cancer in rodents. In the probasin/SV40 (a) Lymphoma
T antigen transgenic rat model, 30% DR had no See Table 3.17.
effect on incidence of PIN or of adenocarcinoma Two lifespan studies in male B10C3F1
but significantly reduced the percentage ratio (Weindruch & Walford, 1982) and C57BL/6 mice
of the epithelial area to the whole prostate area (Volk et al., 1994) started DR at age 45–50 weeks
(which included PIN and tumour cells) (Kandori and assessed incidence of malignant lymphoma.
et al., 2005). There was reduced weight gain due DR (44% DR or 25% DR) significantly increased
537
538

Table 3.16 Studies on the prevention of tumours of the prostate by dietary/calorie restriction in rats and mice
Species, strain (sex) Route, dose, and duration of Type of diet, Type of tumours: Statistical Comments
Age at start carcinogen administration dosing regimen, Tumour incidence (number of significance
Number of animals and duration of tumours/effective number of
Duration of study intervention animals, and/or percentage),
specified
Rat
Lobund-Wistar (M) Spontaneous tumours in Cereal-based Prostate adenocarcinoma: [*P < 0.008, Fisher A second study (Snyder
Weanling either conventional or germ- L-485 Conventional AL, 26%; 30% DR, exact test, 2-tailed] et al., 1990), with the same
n = 100 free housing 30% DR for 6, 6%* design, reported similar
up to 41 mo 18, 30 mo or Germ-free AL, 5%; 30% DR, 10% results: 22%, 7% [NS],
Pollard et al. (1989) > 30 mo 7%, 6%
Wistar-Unilever (M) Cyproterone, 4 weekly i.p. AIN-93M for Prostate adenocarcinoma: P = 0.03 Lower bw in DR rats than
6 wk injections; then testosterone, 4 wk; then half Cancer-free survival at 50 wk: in AL rats
n = 194 daily i.p. injection; then of all rats given AL, 35%; 20% DR, 52% Additional groups fed
age 73 wk MNU, 50 mg/kg bw, at age 20% DR lycopene or tomato
Boileau et al. (2003) 9 wk supplements AL or at 20%
DR
Probasin/SV40 T Probasin/SV40 T antigen NIH-07 Adenocarcinoma: NS 100% incidence of PIN in
antigen transgenic (soybean-free) Ratio of epithelial area to whole P < 0.01 for ventral, all groups
on Sprague-Dawley 30% DR prostate area lateral, or dorsal Small n values; bw
background (M) prostate significantly lower in 30%
6 wk DR group vs controls
n = 40
13 wk
Kandori et al. (2005)
Wistar-Unilever (M) Cyproterone, oral gavage Purina 5001 Prostate adenocarcinoma: NS 30% DR had no effect on
7–8 wk for 21 d; then testosterone, laboratory chow Control, 74% bw
n = 159 daily s.c. for 3 d; then MNU, 15% DR 15% DR, 64%
52 wk 30 mg/kg bw, at age ~12 wk 30% DR 30% DR, 72%
McCormick et al.
(2007)
Mouse
TRAMP on C57BL6 TRAMP mice NTP-2000 Lower grade of lesions: *P < 0.05,
background (M) 20% DR for 4 wk 11 wk: ventral***, lateral***, **P < 0.01,
7 wk or 13 wk dorsal***, and anterior** lobes ***P < 0.001,
n = 10 20 wk: ventral, lateral**, dorsal***, Mann–Whitney
13 wk and anterior* lobes U test
Suttie et al. (2003)
Species, strain (sex) Route, dose, and duration of Type of diet, Type of tumours: Statistical Comments
Age at start carcinogen administration dosing regimen, Tumour incidence (number of significance
Number of animals and duration of tumours/effective number of
Duration of study intervention animals, and/or percentage),
specified
TRAMP on C57BL6 TRAMP mice NTP-2000 No effect on grade of lesions NS [Model not adequate for
background (M) 20% DR for 4 wk, Survival at 19 wk: NS studying prostate cancer
20 wk 12 wk, or 19 wk AL, 75%; DR, 90% at later age]
n = 10
19 wk
Suttie et al. (2005)
TRAMP on C57BL6 TRAMP mice Ain-93M fed Adenocarcinoma: P < 0.006, IDR vs Bw of DR mice fairly
background (M) AL, 25% DR, or Latency to detection: AL constant, whereas that of
5 wk 25% IDR (50% AL, 33 wk; 25% DR, 35 wk; IDR, P = 0.39, DR vs AL IDR mice fluctuated
n = 130 DR for 2 wk, 38 wk See also cross-sectional
48–50 wk then AL for 2 wk study by Bonorden et al.
Bonorden et al. for 11 cycles) (2009b)
(2009a)
Hi-Myc FVB/N (M) Transgenic Hi-Myc mice AIN76A, fed AL, Prostate invasive adenocarcinoma: *P = 0.0001, DR vs [Values read from graph]
6–8 wk 30% DR, or AL, 62%; DR, 0%*; HFD, 97% AL or HFD
n = 36 HFD (60% fat) Prostate in situ carcinoma:
age 26 wk AL, 100%; DR, 38%*; HFD, 100%
Blando et al. (2011)
AL, ad libitum; bw, body weight; d, day or days; DR, dietary restriction; F, female; HFD, high-fat diet; IDR, intermittent dietary restriction; i.p., intraperitoneal; M, male; mo, month or
months; MNU, N-methyl-N-nitrosourea; NR, not reported; NS, not significant; PIN, prostatic intraepithelial neoplasia; TRAMP, transgenic adenocarcinoma of the mouse prostate;
vs, versus; wk, week or weeks.

539
540

Table 3.17 Studies on the prevention of lymphomas by dietary/calorie restriction in mice
Strain (sex) Route, dose, Type of diet, dosing regimen, and Type of tumours: Statistical Comments
Age at start and duration duration of intervention Tumour incidence (number of significance
Number of of carcinogen tumours/effective number of
study specified
Reference
B10C3F1 (M) Spontaneous Semi-purified diet, fed AL or 44% Lymphoma (%); mean lifespan with #NS (P < 0.08) Animals were killed
45–50 wk neoplasms DR with supplementation, from lymphoma: *P < 0.01, t test when moribund
67–68/group age 12–13 mo
Lifelong Control (160 kcal/wk) 47%, 31.9 mo
Weindruch & 44% DR (90 kcal/wk) 31%#, 36.2 mo*
Walford (1982)
C57BL/6 (M) Spontaneous AL or 25% DR, from age 12 mo Lymphoma (%): *P < 0.05, χ2 test
45 wk neoplasms Control (AL) 14/72 (19%)
60–72/group 25% DR 3/60 (5%)*
Age 25 mo
Volk et al.
(1994)
C57BL6 (F, M) Spontaneous NIH-31 open formula diet, fed AL Lymphoma: *P < 0.05 vs controls, SS mice were combined
4 wk neoplasms or 40% DR. Mice were SS at 12, 18, Mice SS at 0–24 mo, 30 mo: Fisher exact test because of the low
266/group 24, 30, and 36 mo, or necropsied incidence of lymphoma
SS at 12– when died spontaneously or were
36 mo, or killed when moribund
lifelong Control (AL) F: 11/43, 5/14
Blackwell et al. M: 2/43, 2/14
(1995) 40% DR F: 3/44*, 0/15
M: 2/44, 0/15
Lymphoma (%): *P < 0.01 vs controls,
Mice that died spontaneously or χ2 test
were killed when moribund in the
(0–33 mo):
Control (AL) F: 35/115 (29.6%)
M: 12/156 (7.7%)
40% DR F: 13/100 (13%)*
M: 9/106 (8.5%)
study specified
Reference
Blackwell et al. Histiocytic sarcoma: NS
(1995) SS at 0–24 mo, 30 mo:
(cont.) Control (AL) F: 1/43, 5/14
M: 1/43, 4/14
40% DR F: 3/44, 3/15
M: 1/44, 5/15
Histiocytic sarcoma (%): *P < 0.001 vs
Mice that died spontaneously or controls, Mantel–
were killed when moribund in the Haenszel χ2 test
(0–27 mo, 28–33 mo):
Control (AL) F: 19/75 (25%), 18/40 (45%)
M: 34/83 (41%), 54/73 (74%)
40% DR F: 20/50 (40%), 27/50 (54%)
M: 9/32 (28.1%), 31/74 (41.9%)*
B6C3F1 (F, M) Spontaneous NIH-31 open formula diet, fed AL Lymphoma (%): NS Incidence of malignant
4 wk neoplasms or 40% DR. Mice were SS at 12, 18, SS at 30 mo, 36 mo: lymphoma during the
266/group 24, 30, and 36 mo, or necropsied period 0–24 mo was
SS at 12– when died spontaneously or were < 3% in F and M mice of
36 mo, or killed when moribund AL and DR groups

lifelong Control (AL) F: 4/15 (26.7%), 9/14 (64.3%)
Sheldon et al. M: 3/15 (20.0%), 6/15 (40.0%)
(1995) 40% DR F: 0/15 (0%), 1/15 (6.7%)
M: 0/15 (0%), 0/15 (0%)
Lymphoma (%): NS
Mice that died spontaneously or
(0–27 mo, 28–40 mo):
Control (AL) F: 24/56 (42.9%), 84/161 (52.2%)
M: 8/39 (20.5%), 46/118 (39%)
541
542

study specified
Reference
Sheldon et al. 40% DR F: 2/10 (20%), 28/72 (38.9%)
(1995) M: 0/7 (0%), 13/45 (28.9%)
(cont.) Histiocytic sarcoma (%): NS
Mice that died spontaneously or
(27–33 mo, 28–40 mo):
Control (AL) F: 7/95 (7.4%), 4/36 (11.1%)
M: 9/68 (13.2%), 5/50 (10%)
40% DR F: 3/23 (13%), 3/49 (6.1%)
M: 4/12 (33%), 5/33 (15%)
p53−/−, p53+/+ Spontaneous AIN-76A diet, fed AL or 40% DR Lymphoma; mean time to death by *P < 0.05 vs controls Median time to death:
[94% C57BL/6, neoplasms from 6–9 wk, lifelong lymphoma (n): 16 wk in p53–/– AL, 25 wk
6% 129/Sv] Control (AL) p53–/–: 17/30, 110 ± 50 d (16) in p53–/– DR; 68 wk in
(M) p53+/+: 4/30, 384 ± 187 d (4) p53+/+ AL, 102 wk in
4–7 wk 40% DR p53–/–: 19/28, 162 ± 59 d (16) p53+/+ DR
28–30/group p53+/+: 6/30, 679 ± 198 d* (6)
Lifelong
Hursting et al.
(1997)
p53+/– C57BL6 Spontaneous AIN-76A diet, fed AL or 40% DR, Number of mice that died from *P = 0.001, Mean bw: ~50 g in AL,
(M) neoplasms or 1 d/wk fast (14% DR) lymphoma/effective number of **P = 0.039 vs 27 g in 40% DR, 38 g in
10.5 mo mice; mean lifespan: AL group, Cox 1 d/wk fast
31–32/group Control (AL) 17/32, 313 ± 17 d proportional hazards
12.5 mo 40% DR 15/31, 388 ± 23 d* analysis (1-tailed)
Berrigan et al. 1 d/wk fast (14% DR) 15/31, 357 ± 23 d**
(2002)
AL, ad libitum; bw, body weight; d, day or days; DR, dietary restriction; F, female; M, male; mo, month or months; NS, not significant; SS, scheduled-sacrificed; vs, versus; wk, week or
weeks.
the mean lifespan of mice with lymphoma or 40% DR group than in the AL group (Hursting
reduced the incidence of lymphoma. et al., 1997). In another study, p53+/− mice prone
In another study (Blackwell et al., 1995), the to malignant tumours including lymphoma
incidence of spontaneous malignant lymphoma (mostly histiocytic sarcoma) were subjected to
was significantly lower in the female 40% DR adult-onset 40% DR (Berrigan et al., 2002). The
group than the female AL group at 24 months mean lifespan was longer in the 40% DR group
and 30 months; however, there was no differ- than in the AL group. Even with 1 day of fasting
ence between the male 40% DR and AL groups. per week followed by AL feeding, the regimen
[Lymphoma incidence in male AL mice was reduced the body weight to 76% of that of the
substantially lower than that in female AL mice.] AL group and extended the lifespan, although
In the same study, the incidence of histio- the effect was modest (P = 0.039) compared with
cytic sarcoma [diffuse large B-cell lymphoma] in that observed in the 40% DR group.
scheduled-sacrificed mice did not differ signif-
icantly between the 40% DR and AL groups in (b) Leukaemia
either male or female mice. However, the inci- See Table 3.18.
dence of histiocytic sarcoma in mice that died In long-term studies, F344 rats (particularly
spontaneously or were killed when moribund F344/N) often develop leukaemia, mostly mono-
was significantly lower in the male DR group nuclear (large granular) cell leukaemia.
compared with the male AL group (Blackwell The incidence of leukaemia in rats killed at
et al., 1995). [The Working Group noted that 24 months and 30 months did not differ between
the study provided separate results for sched- AL and 40% DR groups of male and female rats
uled-sacrificed and moribund animals, which (Thurman et al., 1994). The proportion of rats
makes evaluation of the effect difficult.] bearing leukaemia that died spontaneously or
In another study in male and female B6C3F1 were killed when moribund seemed to be greater
mice, the incidence of malignant lymphoma was in the 40% DR groups of male and female rats.
lower than 3% up to 24 months in both AL and However, the mean ages at death in rats found
40% DR groups (Sheldon et al., 1995). At age dead or killed when moribund were higher in the
30 months and 36 months, the incidence was 40% DR groups of male and female rats than in
greater than 20% in both female and male AL the AL groups. [The number of animals may have
mice, whereas it remained at mostly 0% in 40% been too small to allow relevant statistics; also,
DR mice. In mice that died spontaneously or leukaemia incidence may have been increased in
were killed when moribund during the periods the DR rats because they lived longer.]
of 0–27 months and 28–40 months, the inci- Peto et al. (1980) and Gart et al. (1986) have
dence was lower in the DR groups than in the addressed the biases inherent to long-term animal
AL groups of male and female mice, although studies in which lifespan is extended by an inter-
the difference was not statistically significant. No vention such as DR, and have described statistical
difference was observed for histiocytic sarcomas. analyses to circumvent the problem. By following
Tp53-deleted mice display earlier occurrence their statistical procedures, Shimokawa et al.
of spontaneous neoplasms, including malignant (1996) estimated that the onset rate of leukaemia
lymphoma (Hursting et al., 1994). The incidence in F344 rats was reduced by 20% in the 40% DR
of malignant lymphoma in necropsied Tp53−/− rats compared with the AL animals.
mice subjected to 40% DR did not differ from Pathological data of F344/N rats generated
that in their AL counterparts, but the mean in the National Toxicology Program study of
time to death by lymphoma was longer in the butyl benzyl phthalate (NTP, 1997) indicated
543
544

Table 3.18 Studies on the prevention of leukaemia by dietary/calorie restriction in rats
Age at start and duration duration of intervention Tumour incidence (number significance
Number of of carcinogen of tumours/effective
animals administration number of animals, and/or
Duration of study percentage), multiplicity, or
Reference other outcomes as specified
F344 (F, M) Spontaneous NIH-31 open formula diet, fed AL Leukaemia (%) at 24 mo, NS
4 wk neoplasms or 40% DR gradually implemented 30 mo:
54/group over 2 wk from age 14 wk
Lifelong Control (AL) F: 1/12 (8.3%), 7/12 (58.3%)
Thurman et al. M: 6/12 (50%), 5/9 (55.6%)
(1994) 40% DR F: 4/12 (33.3%), 3/12 (25%)
M: 7/12 (58.3%), 6/12 (50%)
F344 (M) Spontaneous Semi-synthetic diet, fed AL or 40% Leukaemia (%); relative onset *P < 0.05, The relative onset rate, defined by
6 wk neoplasms DR rate: Peto test Peto et al. (1980), is a descriptive
153 or 155/group Control (AL) 38/111 (34.2%), 1.00 index useful in determining whether
Shimokawa et al. 40% DR 39/89 (43.8%), 0.80* a dietary modulation influences
(1996) the occurrence of a neoplasm.
The expected number of rats with
leukaemia/lymphoma was calculated
by analysing the death rate and the
prevalence rate separately.
F344/N Spontaneous NIH-07 open formula diet, fed AL; Leukaemia (mostly *P < 0.01 Adjusted rate: Kaplan–Meier-
(F, M) neoplasms M: 20% DR between 14 wk and mononuclear cell leukaemia); estimated neoplasm incidence at the
6 wk 52 wk and 7% DR between 53 wk adjusted rate: end of the study after adjustment for
60/group and 101 wk; intercurrent mortality
104 wk F: 25% DR between 14 wk and 52 wk
NTP (1997) and 30% DR between 53 wk and
104 wk
Control (AL) F: 21/50, 51.7%
M: 31/50, 71.8%
Weight-matched control of a F: 13/50, 28.6%
carcinogen testing protocol (7–30% M: 15/50*, 34.9%
DR) from 14 wk to 104 wk
AL, ad libitum; DR, dietary restriction; F, female; F344, Fischer 344; M, male; mo, month or months; NS, not significant; NTP, National Toxicology Program; wk, week or weeks.
a reduction in the incidence of leukaemia in Birt DF, Pelling JC, White LT, Dimitroff K, Barnett T
F344 rats in which daily feed allocations were (1991). Influence of diet and calorie restriction on the
initiation and promotion of skin carcinogenesis in
restricted to 7–30% less than that of untreated the SENCAR mouse model. Cancer Res, 51(7):1851–4.
AL rats, to weight-match the animals. When the PMID:1900738
rates were adjusted for intercurrent mortality, Birt DF, Pinch HJ, Barnett T, Phan A, Dimitroff K (1993).
Inhibition of skin tumor promotion by restriction of
the Kaplan–Meier-estimated incidences of fat and carbohydrate calories in SENCAR mice. Cancer
leukaemia were approximately 50% less than Res, 53(1):27–31. PMID:8416746
those in AL animals. Birt DF, Pour PM, Nagel DL, Barnett T, Blackwood
D, Duysen E (1997). Dietary energy restriction
does not inhibit pancreatic carcinogenesis by
N-nitrosobis(2-oxopropyl)amine in the Syrian
References hamster. Carcinogenesis, 18(11):2107–11. doi:10.1093/
carcin/18.11.2107 PMID:9395209
Algire C, Amrein L, Zakikhani M, Panasci L, Pollak M Birt DF, Salmasi S, Pour PM (1981). Enhancement of
(2010). Metformin blocks the stimulative effect of a experimental pancreatic cancer in Syrian golden
high-energy diet on colon carcinoma growth in vivo hamsters by dietary fat. J Natl Cancer Inst, 67(6):1327–
and is associated with reduced expression of fatty 32. PMID:6273636
acid synthase. Endocr Relat Cancer, 17(2):351–60. Blackwell BN, Bucci TJ, Hart RW, Turturro A (1995).
doi:10.1677/ERC-09-0252 PMID:20228137 Longevity, body weight, and neoplasia in ad libi-
Anisimov VN, Berstein LM, Egormin PA, Piskunova TS, tum-fed and diet-restricted C57BL6 mice fed NIH-31
Popovich IG, Zabezhinski MA, et al. (2005). Effect of open formula diet. Toxicol Pathol, 23(5):570–82.
metformin on life span and on the development of doi:10.1177/019262339502300503 PMID:8578100
spontaneous mammary tumors in HER-2/neu trans- Blando J, Moore T, Hursting S, Jiang G, Saha A, Beltran L,
genic mice. Exp Gerontol, 40(8–9):685–93. doi:10.1016/j. et al. (2011). Dietary energy balance modulates prostate
exger.2005.07.007 PMID:16125352 cancer progression in Hi-Myc mice. Cancer Prev Res
Anisimov VN, Egormin PA, Piskunova TS, Popovich (Phila), 4(12):2002–14. doi:10.1158/1940-6207.CAPR-
IG, Tyndyk ML, Yurova MN, et al. (2010). Metformin 11-0182 PMID:21952584
extends life span of HER-2/neu transgenic mice and Bobe G, Barrett KG, Mentor-Marcel RA, Saffiotti U,
in combination with melatonin inhibits growth of Young MR, Colburn NH, et al. (2008). Dietary cooked
transplantable tumors in vivo. Cell Cycle, 9(1):188–97. navy beans and their fractions attenuate colon carcino-
doi:10.4161/cc.9.1.10407 PMID:20016287 genesis in azoxymethane-induced ob/ob mice. Nutr
Bergen HT, Mizuno TM, Taylor J, Mobbs CV (1998). Cancer, 60(3):373–81. doi:10.1080/01635580701775142
Hyperphagia and weight gain after gold-thioglucose: PMID:18444172
relation to hypothalamic neuropeptide Y and pro - Boileau TWM, Liao Z, Kim S, Lemeshow S, Erdman
opiomelanocortin. Endocrinology, 139(11):4483–8. JW Jr, Clinton SK (2003). Prostate carcinogenesis in
PMID:9794456 N-methyl-N-nitrosourea (NMU)-testosterone-treated
Berrigan D, Perkins SN, Haines DC, Hursting SD rats fed tomato powder, lycopene, or energy-restricted
(2002). Adult-onset calorie restriction and fasting diets. J Natl Cancer Inst, 95(21):1578–86. doi:10.1093/
delay spontaneous tumorigenesis in p53-deficient jnci/djg081 PMID:14600090
mice. Carcinogenesis, 23(5):817–22. doi:10.1093/ Boissonneault GA, Elson CE, Pariza MW (1986). Net
carcin/23.5.817 PMID:12016155 energy effects of dietary fat on chemically induced
Birt DF, Barnett T, Pour PM, Copenhaver J (1996). mammary carcinogenesis in F344 rats. J Natl Cancer
High-fat diet blocks the inhibition of skin carcinogen- Inst, 76(2):335–8. PMID:3080637
esis and reductions in protein kinase C by moderate Bonorden MJL, Grossmann ME, Ewing SA, Rogozina
energy restriction. Mol Carcinog, 16(2):115–20. OP, Ray A, Nkhata KJ, et al. (2012). Growth and
doi:10.1002/(SICI)1098-2744(199606)16:2<115::AID- progression of TRAMP prostate tumors in relation
MC8>3.0.CO;2-H PMID:8645426 to diet and obesity. Prostate Cancer, 2012:543970.
Birt DF, Pelling JC, Anderson J, Barnett T (1994). doi:10.1155/2012/543970 PMID:23304522
Consumption of reduced-energy/low-fat diet or Bonorden MJL, Rogozina OP, Kluczny CM,
constant-energy/high-fat diet during mezerein Grossmann ME, Grambsch PL, Grande JP, et al.
treatment inhibited mouse skin tumour promo- (2009a). Intermittent calorie restriction delays
tion. Carcinogenesis, 15(10):2341–5. doi:10.1093/ prostate tumor detection and increases survival
carcin/15.10.2341 PMID:7955075 time in TRAMP mice. Nutr Cancer, 61(2):265–75.
doi:10.1080/01635580802419798 PMID:19235043
545
Bonorden MJL, Rogozina OP, Kluczny CM, Grossmann mammary tumors. Breast Cancer Res Treat, 77(3):205–
ME, Grande JP, Lokshin A, et al. (2009b). Cross- 15. doi:10.1023/A:1021891825399 PMID:12602920
sectional analysis of intermittent versus chronic caloric Cowey SL, Quast M, Belalcazar LM, Wei J, Deng X, Given
restriction in the TRAMP mouse. Prostate, 69(3):317– R, et al. (2005). Abdominal obesity, insulin resistance,
26. doi:10.1002/pros.20878 PMID:19016490 and colon carcinogenesis are increased in mutant mice
Boutwell RK, Brush MK, Rusch HP (1949). The stimu- lacking gastrin gene expression. Cancer, 103(12):2643–
lating effect of dietary fat on carcinogenesis. Cancer 53. doi:10.1002/cncr.21094 PMID:15864814
Res, 9(12):741–6. PMID:15395908 Dawson DW, Hertzer K, Moro A, Donald G, Chang
Brandon EL, Gu J-W, Cantwell L, He Z, Wallace G, HH, Go VL, et al. (2013). High-fat, high-calorie diet
Hall JE (2009). Obesity promotes melanoma tumor promotes early pancreatic neoplasia in the condi-
growth: role of leptin. Cancer Biol Ther, 8(19):1871–9. tional KrasG12D mouse model. Cancer Prev Res (Phila),
doi:10.4161/cbt.8.19.9650 PMID:19713740 6(10):1064–73. doi:10.1158/1940-6207.CAPR-13-0065
Buison AM, Pellizzon MA, Brogan KE, Barnes MJ, PMID:23943783
Jen CK-L (2005). Weight cycling did not increase Dinkova-Kostova AT, Fahey JW, Jenkins SN, Wehage SL,
tumor incidence in high fat-fed rats treated with a Talalay P (2008). Rapid body weight gain increases
low-dose 7,12-dimethylbenzyl(1)anthracene. Nutr Res, the risk of UV radiation-induced skin carcinogen-
25(12):1097–108. doi:10.1016/j.nutres.2005.09.003 esis in SKH-1 hairless mice. Nutr Res, 28(8):539–43.
Cleary MP (2013). Impact of obesity on development and doi:10.1016/j.nutres.2008.05.009 PMID:19083457
progression of mammary tumors in preclinical models Dogan S, Hu X, Zhang Y, Maihle NJ, Grande JP, Cleary
of breast cancer. J Mammary Gland Biol Neoplasia, MP (2007). Effects of high-fat diet and/or body weight
18(3–4):333–43. doi:10.1007/s10911-013-9300-x on mammary tumor leptin and apoptosis signaling
PMID:24122258 pathways in MMTV-TGF-α mice. Breast Cancer Res,
Cleary MP, Grande JP, Juneja SC, Maihle NJ (2004b). Diet- 9(6):R91. doi:10.1186/bcr1840 PMID:18162139
induced obesity and mammary tumor development in Duffy PH, Lewis SM, Mayhugh MA, Trotter RW, Hass
MMTV-neu female mice. Nutr Cancer, 50(2):174–80. BS, Latendresse JR, et al. (2008). Neoplastic pathology
doi:10.1207/s15327914nc5002_7 PMID:15623464 in male Sprague-Dawley rats fed AIN-93M diet ad
Cleary MP, Grande JP, Maihle NJ (2004a). Effect of libitum or at restricted intakes. Nutr Res, 28(1):36–42.
high fat diet on body weight and mammary tumor doi:10.1016/j.nutres.2007.09.018 PMID:19083386
latency in MMTV-TGF-α mice. Int J Obes Relat Metab Dunlap SM, Chiao LJ, Nogueira L, Usary J, Perou CM,
Disord, 28(8):956–62. doi:10.1038/sj.ijo.0802664 Varticovski L, et al. (2012). Dietary energy balance
PMID:15254485 modulates epithelial-to-mesenchymal transition and
Cleary MP, Grossmann ME (2011). The manner in which tumor progression in murine claudin-low and basal-
calories are restricted impacts mammary tumor cancer like mammary tumor models. Cancer Prev Res (Phila),
prevention. J Carcinog, 10(1):21. doi:10.4103/1477- 5(7):930–42. doi:10.1158/1940-6207.CAPR-12-0034
3163.85181 PMID:22013391 PMID:22588949
Cleary MP, Hu X, Grossmann ME, Juneja SC, Dogan Ealey KN, Lu S, Archer MC (2008). Development of aber-
S, Grande JP, et al. (2007). Prevention of mammary rant crypt foci in the colons of ob/ob and db/db mice:
tumorigenesis by intermittent caloric restriction: evidence that leptin is not a promoter. Mol Carcinog,
does caloric intake during refeeding modulate the 47(9):667–77. doi:10.1002/mc.20419 PMID:18240295
response? Exp Biol Med (Maywood), 232(1):70–80. Engelman RW, Day NK, Chen RF, Tomita Y, Bauer-
PMID:17202587 Sardiña I, Dao ML, et al. (1990). Calorie consumption
Cleary MP, Jacobson MK, Phillips FC, Getzin SC, Grande level influences development of C3H/Ou breast adeno-
JP, Maihle NJ (2002). Weight-cycling decreases inci- carcinoma with indifference to calorie source. Proc Soc
dence and increases latency of mammary tumor devel- Exp Biol Med, 193(1):23–30. doi:10.3181/00379727-193-
opment to a greater extent than does chronic restriction 42984 PMID:2294519
in mouse mammary tumor virus-transforming growth Engelman RW, Day NK, Good RA (1994). Calorie intake
factor-α female mice. Cancer Epidemiol Biomarkers during mammary development influences cancer risk:
Prev, 11:836–43. PMID:12223427 lasting inhibition of C3H/HeOu mammary tumori-
Cleary MP, Juneja SC, Phillips FC, Hu X, Grande JP, Maihle genesis by peripubertal calorie restriction. Cancer Res,
NJ (2004d). Leptin receptor deficient MMTV-TGF-α/ 54(21):5724–30. PMID:7923222
LeprdbLeprdb female mice do not develop oncogene-in- Ershler WB, Berman E, Moore AL (1986). Slower B16
duced mammary tumors. Exp Biol Med, 229:182–93. melanoma growth but greater pulmonary colonization
doi:10.1177/153537020422900207 PMID:14734797 in calorie-restricted mice. J Natl Cancer Inst, 76(1):81–5.
Cleary MP, Phillips FC, Getzin SC, Jacobson TL, Jacobson PMID:3455745
MK, Christensen TA, et al. (2004c). Genetically obese
MMTV-TGF-α/LepobLepob female mice do not develop
546
Everitt JI, Alder R (2013). Issues in laboratory animal Hayashi K, Suzuki R, Miyamoto S, Shin-Ichiroh Y, Kohno
science: the role of diet in toxicology studies. In: H, Sugie S, et al. (2007). Citrus auraptene suppresses
Hascheck WM, Rousseaux CG, Wallig MA, Bolon azoxymethane-induced colonic preneoplastic lesions
B, Ochoa R, Mahler BW, editors. Haschek and in C57BL/KsJ-db/db mice. Nutr Cancer, 58(1):75–84.
Rousseaux’s handbook of toxicologic pathology. 3rd doi:10.1080/01635580701308216 PMID:17571970
edition. London, UK: Academic Press, Elsevier; pp. Heston WE, Vlahakis G (1961). Influence of the Ay
634–40. gene on mammary-gland tumors, hepatomas, and
Frederich RC, Hamann A, Anderson S, Löllmann B, normal growth in mice. J Natl Cancer Inst, 26:969–83.
Lowell BB, Flier JS (1995). Leptin levels reflect body lipid PMID:13713953
content in mice: evidence for diet-induced resistance Heston WE, Vlahakis G (1962). Genetic obesity
to leptin action. Nat Med, 1(12):1311–4. doi:10.1038/ and neoplasia. J Natl Cancer Inst, 29:197–209.
nm1295-1311 PMID:7489415 PMID:13907025
Gart JJ, Krewski D, Lee PN, Tarone RE, Wahrendorf J Hill-Baskin AE, Markiewski MM, Buchner DA, Shao
(1986). Statistical methods in cancer research. Volume H, DeSantis D, Hsiao G, et al. (2009). Diet-induced
III – The design and analysis of long-term animal hepatocellular carcinoma in genetically predisposed
experiments. IARC Sci Publ, 79:1–219. PMID:3301661 mice. Hum Mol Genet, 18(16):2975–88. doi:10.1093/
Gillette CA, Zhu Z, Westerlind KC, Melby CL, Wolfe hmg/ddp236 PMID:19454484
P, Thompson HJ (1997). Energy availability and Hirose Y, Hata K, Kuno T, Yoshida K, Sakata K, Yamada
mammary carcinogenesis: effects of calorie restriction Y, et al. (2004). Enhancement of development of azoxy
and exercise. Carcinogenesis, 18(6):1183–8. doi:10.1093/ methane-induced colonic premalignant lesions in
carcin/18.6.1183 PMID:9214601 C57BL/KsJ-db/db mice. Carcinogenesis, 25(5):821–5.
Gravaghi C, Bo J, LaPerle KMD, Quimby F, Kucherlapati doi:10.1093/carcin/bgh059 PMID:14729596
R, Edelmann W, et al. (2008). Obesity enhances gastro- Hursting SD, Dunlap SM, Ford NA, Hursting MJ,
intestinal tumorigenesis in Apc-mutant mice Int J Lashinger LM (2013). Calorie restriction and cancer
Obes (London), 32(11):1716–9. doi:10.1038/ijo.2008.149 prevention: a mechanistic perspective. Cancer Metab,
PMID:18725892 1(1):10. doi:10.1186/2049-3002-1-10 PMID:24280167
Hakkak R, Holley AW, Macleod SL, Simpson PM, Hursting SD, Lashinger LM, Brown PH, Perkins SN
Fuchs GJ, Jo CH, et al. (2005). Obesity promotes (2011). The utility of transgenic mouse models for
7,12-dimethylbenz(a)anthracene-induced mammary cancer prevention research. In: Teicher BA, editor.
tumor development in female Zucker rats. Breast Cancer Tumor models in cancer research (cancer drug
Res, 7(5):R627–33. doi:10.1186/bcr1263 PMID:16168107 discovery and development). 2nd edition. New York,
Hakkak R, MacLeod S, Shaaf S, Holley AW, Simpson P, USA: Springer Science+Business Media; pp. 423–43.
Fuchs G, et al. (2007). Obesity increases the incidence of doi:10.1007/978-1-60761-968-0_18
7,12-dimethylbenz(a)anthracene-induced mammary Hursting SD, Perkins SN, Brown CC, Haines DC, Phang
tumors in an ovariectomized Zucker rat model. Int J JM (1997). Calorie restriction induces a p53-independent
Oncol, 30(3):557–63. PMID:17273756 delay of spontaneous carcinogenesis in p53-defi-
Harvell DM, Spady TJ, Strecker TE, Lemus-Wilson AM, cient and wild-type mice. Cancer Res, 57(14):2843–6.
Pennington KL, Shen F, et al. (2001). Dietary energy PMID:9230186
restriction inhibits estrogen induced pituitary tumor- Hursting SD, Perkins SN, Phang JM (1994). Calorie
igenesis in a rat strain specific manner. In: Li JJ, Li restriction delays spontaneous tumorigenesis in
SA, Daling JR, editors. Hormonal carcinogenesis III: p53-knockout transgenic mice. Proc Natl Acad Sci
proceedings of the third international symposium. USA, 91(15):7036–40. doi:10.1073/pnas.91.15.7036
New York, USA: Springer; pp. 496–501. PMID:8041741
Harvell DME, Strecker TE, Xie B, Pennington KL, Hursting SD, Smith SM, Lashinger LM, Harvey AE,
McComb RD, Shull JD (2002). Dietary energy Perkins SN (2010). Calories and carcinogenesis: lessons
restriction inhibits estrogen-induced mammary, learned from 30 years of calorie restriction research.
but not pituitary, tumorigenesis in the ACI rat. Carcinogenesis, 31(1):83–9. doi:10.1093/carcin/bgp280
Carcinogenesis, 23(1):161–9. doi:10.1093/carcin/23.1.161 PMID:19969554
PMID:11756237 IARC (2002). Weight control and physical activity. Lyon,
Harvey AE, Lashinger LM, Otto G, Nunez NP, Hursting France: IARC Press (IARC Handbooks of Cancer
SD (2013). Decreased systemic IGF-1 in response Prevention, Vol. 6). Available from: http://publications.
to calorie restriction modulates murine tumor cell iarc.fr/376.
growth, nuclear factor-κB activation, and inflamma- Jiang W, Zhu Z, Thompson HJ (2008a). Modulation of
tion-related gene expression. Mol Carcinog, 52(12):997– the activities of AMP-activated protein kinase, protein
1006. doi:10.1002/mc.21940 PMID:22778026 kinase B, and mammalian target of rapamycin by
limiting energy availability with 2-deoxyglucose.
547
Mol Carcinog, 47(8):616–28. doi:10.1002/mc.20425 Lagopoulos L, Sunahara GI, Würzner H, Dombrowsky

PMID:18247380 I, Stalder R (1991). The effects of alternating dietary
Jiang W, Zhu Z, Thompson HJ (2008b). Dietary energy restriction and ad libitum feeding of mice on the devel-
restriction modulates the activity of AMP-activated opment of diethylnitrosamine-induced liver tumours
protein kinase, Akt, and mammalian target of rapa- and its correlation to insulinaemia. Carcinogenesis,
mycin in mammary carcinomas, mammary gland, 12(2):311–5. doi:10.1093/carcin/12.2.311 PMID:1847320
and liver. Cancer Res, 68(13):5492–9. doi:10.1158/0008- Lanza-Jacoby S, Yan G, Radice G, LePhong C, Baliff
5472.CAN-07-6721 PMID:18593953 J, Hess R (2013). Calorie restriction delays the
Kanasaki K, Koya D (2011). Biology of obesity: lessons progression of lesions to pancreatic cancer in the
from animal models of obesity. J Biomed Biotechnol, LSL-KrasG12D; Pdx-1/Cre mouse model of pancrea tic
2011:197636. doi:10.1155/2011/197636 PMID:21274264 cancer. Exp Biol Med (Maywood), 238(7):787–97.
Kandori H, Suzuki S, Asamoto M, Murasaki T, Mingxi doi:10.1177/1535370213493727 PMID:23828595
T, Ogawa K, et al. (2005). Influence of atrazine admin- Lashinger LM, Harrison LM, Rasmussen AJ, Logsdon
istration and reduction of calorie intake on prostate CD, Fischer SM, McArthur MJ, et al. (2013). Dietary
carcinogenesis in probasin/SV40 T antigen trans- energy balance modulation of Kras- and Ink4a/Arf+/−-
genic rats. Cancer Sci, 96(4):221–6. doi:10.1111/j.1349- driven pancreatic cancer: the role of insulin-like
7006.2005.00041.x PMID:15819720 growth factor-I. Cancer Prev Res (Phila), 6(10):1046–55.
Keenan KP, Soper KA, Smith PF, Ballam GC, Clark doi:10.1158/1940-6207.CAPR-13-0185 PMID:23980075
RL (1995). Diet, overfeeding, and moderate dietary Lashinger LM, Malone LM, Brown GW, Daniels EA,
restriction in control Sprague-Dawley rats: I. Effects on Goldberg JA, Otto G, et al. (2011). Rapamycin partially
spontaneous neoplasms. Toxicol Pathol, 23(3):269–86. mimics the anticancer effects of calorie restriction in
doi:10.1177/019262339502300305 PMID:7659952 a murine model of pancreatic cancer. Cancer Prev Res
Keenan KP, Wallig MA, Haschek WM (2013). Nature (Phila), 4(7):1041–51. doi:10.1158/1940-6207.CAPR-11-
via nurture: effect of diet on health, obesity, and 0023 PMID:21593197
safety assessment. Toxicol Pathol, 41(2):190–209. Lee WM, Lu S, Medline A, Archer MC (2001). Susceptibility
doi:10.1177/0192623312469857 PMID:23334694 of lean and obese Zucker rats to tumorigenesis induced
Khalid S, Hwang D, Babichev Y, Kolli R, Altamentova S, by N-methyl-N-nitrosourea. Cancer Lett, 162(2):155–60.
Koren S, et al. (2010). Evidence for a tumor promoting doi:10.1016/S0304-3835(00)00635-2 PMID:11146220
effect of high-fat diet independent of insulin resist- Lemke LB, Rogers AB, Nambiar PR, Fox JG (2008). Obesity
ance in HER2/Neu mammary carcinogenesis. Breast and non-insulin-dependent diabetes mellitus in Swiss-
Cancer Res Treat, 122(3):647–59. doi:10.1007/s10549- Webster mice associated with late-onset hepatocellular
009-0586-8 PMID:19851863 carcinoma. J Endocrinol, 199(1):21–32. doi:10.1677/
Klurfeld DM, Lloyd LM, Welch CB, Davis MJ, Tulp OL, JOE-07-0588 PMID:18667586
Kritchevsky D (1991). Reduction of enhanced mammary Lipman NS, Leary SL (2015). Design and management
carcinogenesis in LA/N-cp (corpulent) rats by energy of research facilities. In: Anderson LC, Otto GM,
restriction. Proc Soc Exp Biol Med, 196(4):381–4. Pritchett-Corning KR, Whary MT, Fox JG, editors.
doi:10.3181/00379727-196-43202 PMID:1901169 Laboratory animal medicine. 3rd edition. Elsevier; pp.
Klurfeld DM, Welch CB, Davis MJ, Kritchevsky D (1989a). 1543–97. doi:10.1016/B978-0-12-409527-4.00036-5
Determination of degree of energy restriction necessary Llaverias G, Danilo C, Wang Y, Witkiewicz AK, Daumer
to reduce DMBA-induced mammary tumorigenesis in K, Lisanti MP, et al. (2010). A Western-type diet accel-
rats during the promotion phase. J Nutr, 119(2):286–91. erates tumor progression in an autochthonous mouse
PMID:2493082 model of prostate cancer. Am J Pathol, 177(6):3180–91.
Klurfeld DM, Welch CB, Lloyd LM, Kritchevsky D (1989b). doi:10.2353/ajpath.2010.100568 PMID:21088217
Inhibition of DMBA-induced mammary tumorigen- Maeda H, Gleiser CA, Masoro EJ, Murata I, McMahan
esis by caloric restriction in rats fed high-fat diets. CA, Yu BP (1985). Nutritional influences on aging of
Int J Cancer, 43(5):922–5. doi:10.1002/ijc.2910430532 Fischer 344 rats: II. Pathology. J Gerontol, 40(6):671–
PMID:2497075 88. doi:10.1093/geronj/40.6.671 PMID:4056322
Kritchevsky D (1999). Caloric restriction and experi- Mai V, Colbert LH, Berrigan D, Perkins SN, Pfeiffer R,
mental carcinogenesis. Toxicol Sci, 52(2 Suppl):13–6. Lavigne JA, et al. (2003). Calorie restriction and diet
doi:10.1093/toxsci/52.2.13 PMID:10630585 composition modulate spontaneous intestinal tumori-
Kritchevsky D, Welch CB, Klurfeld DM (1989). Response genesis in ApcMin mice through different mechanisms.
of mammary tumors to caloric restriction for different Cancer Res, 63(8):1752–5. PMID:12702556
time periods during the promotion phase. Nutr Matthews SB, Zhu Z, Jiang W, McGinley JN, Neil ES,
Cancer, 12(3):259–69. doi:10.1080/01635588909514025 Thompson HJ (2014). Excess weight gain accelerates
PMID:2505241 1-methyl-1-nitrosourea-induced mammary carcino-
genesis in a rat model of premenopausal breast cancer.
548
Cancer Prev Res (Phila), 7(3):310–8. doi:10.1158/1940- of ovarian hormones. Nutr Cancer, 60(4):534–41.
6207.CAPR-13-0297 PMID:24441676 doi:10.1080/01635580801966195 PMID:18584488
McCormick DL, Johnson WD, Haryu TM, Bosland Olivo-Marston SE, Hursting SD, Perkins SN, Schetter
MC, Lubet RA, Steele VE (2007). Null effect of A, Khan M, Croce C, et al. (2014). Effects of calorie
dietary restriction on prostate carcinogenesis in the restriction and diet-induced obesity on murine colon
Wistar-Unilever rat. Nutr Cancer, 57(2):194–200. carcinogenesis, growth and inflammatory factors,
doi:10.1080/01635580701277494 PMID:17571953 and microRNA expression. PLoS One, 9(4):e94765.
Minor RK, López M, Younts CM, Jones B, Pearson KJ, doi:10.1371/journal.pone.0094765 PMID:24732966
Anson RM, et al. (2011). The arcuate nucleus and Pandey V, Vijayakumar MV, Ajay AK, Malvi P, Bhat
neuropeptide Y contribute to the antitumorigenic MK (2012). Diet-induced obesity increases mela-
effect of calorie restriction. Aging Cell, 10(3):483–92. noma progression: involvement of Cav-1 and FASN.
doi:10.1111/j.1474-9726.2011.00693.x PMID:21385308 Int J Cancer, 130(3):497–508. doi:10.1002/ijc.26048
Mizuno NK, Rogozina OP, Seppanen CM, Liao DJ, Cleary PMID:21387314
MP, Grossmann ME (2013). Combination of inter- Pape-Ansorge KA, Grande JP, Christensen TA, Maihle NJ,
mittent calorie restriction and eicosapentaenoic acid Cleary MP (2002). Effect of moderate caloric restriction
for inhibition of mammary tumors. Cancer Prev Res and/or weight cycling on mammary tumor incidence
(Phila), 6(6):540–7. doi:10.1158/1940-6207.CAPR-13- and latency in MMTV-Neu female mice. Nutr Cancer,
0033 PMID:23550153 44(2):162–8. doi:10.1207/S15327914NC 4402 _07
Molon-Noblot S, Keenan KP, Coleman JB, Hoe CM, PMID:12734063
Laroque P (2001). The effects of ad libitum over- Parisotto M, Metzger D (2013). Genetically engineered
feeding and moderate and marked dietary restriction mouse models of prostate cancer. Mol Oncol, 7(2):190–
on age-related spontaneous pancreatic islet pathology 205. doi:10.1016/j.molonc.2013.02.005 PMID:23481269
in Sprague-Dawley rats. Toxicol Pathol, 29(3):353–62. Park EJ, Lee JH, Yu GY, He G, Ali SR, Holzer RG, et al.
doi:10.1080/019262301316905318 PMID:11442021 (2010). Dietary and genetic obesity promote liver
Moore T, Beltran L, Carbajal S, Hursting SD, DiGiovanni inflammation and tumorigenesis by enhancing IL-6
J (2012). Energy balance modulates mouse skin and TNF expression. Cell, 140(2):197–208. doi:10.1016/j.
tumour promotion through altered IGF-1R and EGFR cell.2009.12.052 PMID:20141834
crosstalk. Cancer Prev Res (Phila), 5(10):1236–46. Park H, Kim M, Kwon GT, Yu R, Sung MK, Lee KW,
doi:10.1158/1940-6207.CAPR-12-0234 PMID:22896210 et al. (2012). A high-fat diet increases angiogenesis,
Mori A, Sakurai H, Choo MK, Obi R, Koizumi K, Yoshida solid tumor growth, and lung metastasis of CT26
C, et al. (2006). Severe pulmonary metastasis in colon cancer cells in obesity-resistant BALB/c mice.
obese and diabetic mice. Int J Cancer, 119(12):2760–7. Mol Carcinog, 51(11):869–80. doi:10.1002/mc.20856
doi:10.1002/ijc.22248 PMID:16998795 PMID:21919080
Newberne PM, Bueche D, Suphiphat V, Schrager TF, Pashko LL, Schwartz AG (1992). Reversal of food restric-
Sahaphong S (1990). The influence of pre- and post- tion-induced inhibition of mouse skin tumour promo-
natal caloric intake on colon carcinogenesis. Nutr tion by adrenalectomy. Carcinogenesis, 13(10):1925–8.
Cancer, 13(3):165–73. doi:10.1080/01635589009514057 doi:10.1093/carcin/13.10.1925 PMID:1423856
PMID:2308872 Pearson KJ, Lewis KN, Price NL, Chang JW, Perez E,
Nkhata KJ, Ray A, Dogan S, Grande JP, Cleary MP (2009). Cascajo MV, et al. (2008). Nrf2 mediates cancer protec-
Mammary tumor development from T47-D human tion but not prolongevity induced by caloric restriction.
breast cancer cells in obese ovariectomized mice with Proc Natl Acad Sci USA, 105(7):2325–30. doi:10.1073/
and without estradiol supplements. Breast Cancer Res pnas.0712162105 PMID:18287083
Treat, 114(1):71–83. doi:10.1007/s10549-008-9991-7 Peto R, Pike MC, Day NE, Gray RG, Lee PN, Parish S,
PMID:18392696 et al. (1980). Guidelines for simple, sensitive signifi-
Nogueira LM, Dunlap SM, Ford NA, Hursting SD cance tests for carcinogenic effects in long-term animal
(2012). Calorie restriction and rapamycin inhibit experiments. IARC Monogr Eval Carcinog Risk Chem
MMTV-Wnt-1 mammary tumor growth in a mouse Hum Suppl, (Suppl 2):311–426. PMID:6935185
model of postmenopausal obesity. Endocr Relat Cancer, Philip B, Roland CL, Daniluk J, Liu Y, Chatterjee D, Gomez
19(1):57–68. doi:10.1530/ERC-11-0213 PMID:22143497 SB, et al. (2013). A high-fat diet activates oncogenic
NTP (1997). Effect of dietary restriction on toxicology Kras and COX2 to induce development of pancreatic
and carcinogenesis studies in F344/N rats and B6C3F1 ductal adenocarcinoma in mice. Gastroenterology,
mice. Natl Toxicol Program Tech Rep Ser, 460:1–414. 145(6):1449–58. doi:10.1053/j.gast ro.2013.08.018
PMID:12587016 PMID:23958541
Nuñez NP, Perkins SN, Smith NC, Berrigan D, Berendes
DM, Varticovski L, et al. (2008). Obesity acceler-
ates mouse mammary tumor growth in the absence
549
Pollard M, Luckert PH, Pan G-Y (1984). Inhibition Prev Res (Phila), 9(5):339–48. doi:10.1158/1940-6207.
of intestinal tumorigenesis in methylazoxymetha- CAPR-15-0348 PMID:26869351
nol-treated rats by dietary restriction. Cancer Treat Ruggeri BA, Klurfeld DM, Kritchevsky D, Furlanetto RW
Rep, 68(2):405–8. PMID:6697329 (1989a). Caloric restriction and 7,12-dimethylbenz(a)
Pollard M, Luckert PH, Snyder D (1989). Prevention anthracene-induced mammary tumor growth in rats:
of prostate cancer and liver tumors in L-W rats by alterations in circulating insulin, insulin-like growth
moderate dietary restriction. Cancer, 64(3):686–90. factors I and II, and epidermal growth factor. Cancer
doi:10.1002/1097-0142(19890801)64:3<686::AID-CN- Res, 49(15):4130–4. PMID:2501021
CR2820640320>3.0.CO;2-Q PMID:2472862 Ruggeri BA, Klurfeld DM, Kritchevsky D, Furlanetto
Pour PM, Permert J, Mogaki M, Fujii H, Kazakoff K (1993). RW (1989b). Growth factor binding to 7,12-dimethyl
Endocrine aspects of exocrine cancer of the pancreas. benz(a)anthracene-induced mammary tumors from
Their patterns and suggested biologic significance. Am rats subject to chronic caloric restriction. Cancer Res,
J Clin Pathol, 100(3):223–30. doi:10.1093/ajcp/100.3.223 49(15):4135–41. PMID:2501022
PMID:8379530 Sheldon WG, Bucci TJ, Blackwell B, Turturro A (1996).
Raju J, Bird RP (2003). Energy restriction reduces the Effect of ad libitum feeding and 40% feed restriction
number of advanced aberrant crypt foci and attenuates on body weight, longevity, and neoplasms in B6C3F1,
the expression of colonic transforming growth factor C57BL6, and B6D2F1 mice. In: Mohr U, Dungworth
beta and cyclooxygenase isoforms in Zucker obese (fa/ D, Capen C, Carlton W, Sundberg J, Ward J, editors.
fa) rats. Cancer Res, 63(20):6595–601. PMID:14583451 Pathobiology of the aging mouse. Washington (DC),
Ray A, Cleary MP (2013). Animal models to study the USA: ILSI Press; pp. 21–6.
interplay between cancer and obesity. In: Kolomin Sheldon WG, Bucci TJ, Hart RW, Turturro A (1995).
MG, editor. Adipose tissue and cancer. New York, USA: Age-related neoplasia in a lifetime study of ad libi-
Springer; pp. 99–119. doi:10.1007/978-1-4614-7660-3_6 tum-fed and food-restricted B6C3F1 mice. Toxicol
Reddy BS, Wang C-X, Maruyama H (1987). Effect of Pathol, 23(4):458–76. doi:10.1177/019262339502300403
restricted caloric intake on azoxymethane-induced PMID:7501958
colon tumor incidence in male F344 rats. Cancer Res, Shimokawa I, Higami Y, Yu BP, Masoro EJ, Ikeda T (1996).
47(5):1226–8. PMID:3815332 Influence of dietary components on occurrence of and
Ribeiro AM, Andrade S, Pinho F, Monteiro JD, Costa M, mortality due to neoplasms in male F344 rats. Aging
Lopes C, et al. (2010). Prostate cancer cell prolifera- (Milano), 8(4):254–62. PMID:8904955
tion and angiogenesis in different obese mice models. Shull JD, Birt DF, McComb RD, Spady TJ, Pennington
Int J Exp Pathol, 91(4):374–86. doi:10.1111/j.1365- KL, Shaw-Bruha CM (1998). Estrogen induction
2613.2010.00726.x PMID:20666851 of prolactin-producing pituitary tumours in the
Roebuck BD, Baumgartner KJ, MacMillan DL (1993). Fischer 344 rat: modulation by dietary-energy but not
Caloric restriction and intervention in pancreatic protein consumption. Mol Carcinog, 23(2):96–105.
carcinogenesis in the rat. Cancer Res, 53(1):46–52. doi:10.1002/(SICI)1098-2744(199810)23:2<96::AID-
PMID:8416749 MC6>3.0.CO;2-# PMID:9808163
Rogozina OP, Bonorden MJL, Grande JP, Cleary MP Shull JD, Spady TJ, Snyder MC, Johansson SL, Pennington
(2009). Serum insulin-like growth factor-I and KL (1997). Ovary-intact, but not ovariectomized female
mammary tumor development in ad libitum-fed, ACI rats treated with 17β-estradiol rapidly develop
chronic calorie-restricted, and intermittent calorie-re- mammary carcinoma. Carcinogenesis, 18(8):1595–601.
stricted MMTV-TGF-α mice. Cancer Prev Res (Phila), doi:10.1093/carcin/18.8.1595 PMID:9276635
2(8):712–9. doi:10.1158/1940-6207.CAPR-09-0028 Snyder DL, Pollard M, Wostmann BS, Luckert P (1990).
PMID:19654106 Life span, morphology, and pathology of diet-restricted
Rogozina OP, Nkhata KJ, Nagle EJ, Grande JP, Cleary germ-free and conventional Lobund-Wistar rats.
MP (2013). The protective effect of intermittent J Gerontol, 45(2):B52–8. doi:10.1093/geronj/45.2.B52
calorie restriction on mammary tumorigenesis is not PMID:2313040
compromised by consumption of a high fat diet during Soga M, Hashimoto S, Kishimoto Y, Hirasawa T, Makino
refeeding. Breast Cancer Res Treat, 138(2):395–406. S, Inagaki S (2010). Insulin resistance, steatohepatitis,
doi:10.1007/s10549-013-2464-7 PMID:23446811 and hepatocellular carcinoma in a new congenic strain
Ross MH, Bras G (1971). Lasting influence of early caloric of Fatty Liver Shionogi (FLS) mice with the Lepob gene.
restriction on prevalence of neoplasms in the rat. J Natl Exp Anim, 59(4):407–19. doi:10.1538/expanim.59.407
Cancer Inst, 47(5):1095–113. PMID:4330798 PMID:20660987
Rossi EL, de Angel RE, Bowers LW, Khatib SA, Smith LA, Spady TJ, Lemus-Wilson AM, Pennington KL, Blackwood
Van Buren E, et al. (2016). Obesity-associated alter- DJ, Paschall TM, Birt DF, et al. (1998). Dietary energy
ations in inflammation, epigenetics and mammary restriction abolishes development of prolactin-pro-
tumor growth persist in formerly obese mice Cancer ducing pituitary tumours in Fischer 344 rats treated
550
with 17β-estradiol. Mol Carcinog, 23(2):86–95. Thompson HJ, Zhu Z, Jiang W (2003). Dietary
doi:10.1002/(SICI)1098-2744(199810)23:2<86::AID- energy restriction in breast cancer prevention.
MC5>3.0.CO;2-5 PMID:9808162 J Mammary Gland Biol Neoplasia, 8(1):133–42.
Spady TJ, Pennington KL, McComb RD, Birt DF, doi:10.1023/A:1025743607445 PMID:14587868
Shull JD (1999). Estrogen-induced pituitary tumour Thompson HJ, Zhu Z, Jiang W (2004a). Identification of
development in the ACI rat not inhibited by dietary the apoptosis activation cascade induced in mammary
energy restriction. Mol Carcinog, 26(4):239–53. carcinomas by energy restriction. Cancer Res,
doi:10.1002/(SICI)1098-2744(199912)26:4<239::AID- 64(4):1541–5. doi:10.1158/0008-5472.CAN-03-3108
MC3>3.0.CO;2-H PMID:10569801 PMID:14973070
Sundaram S, Le TL, Essaid L, Freemerman AJ, Huang MJ, Thompson JH (2015). Impact of energy balance on chem-
Galanko JA, et al. (2014). Weight loss reversed obesi- ically induced mammary carcinogenesis in a rat. In:
ty-induced HGF/c-Met pathway and basal-like breast Berger AN, editor. Murine models, energy balance, and
cancer progression Front Oncol, 4:175. doi:10.3389/ cancer. Springer International Publishing; pp. 175–96.
fonc.2014.00175 PMID:25072025 doi:10.1007/978-3-319-16733-6_9
Suttie A, Nyska A, Haseman JK, Moser GJ, Hackett TR, Thompson MD, Grubbs CJ, Bode AM, Reid JM,
Goldsworthy TL (2003). A grading scheme for the McGovern R, Bernard PS, et al. (2015). Lack of effect
assessment of proliferative lesions of the mouse pros- of metformin on mammary carcinogenesis in nondia-
tate in the TRAMP model. Toxicol Pathol, 31(1):31–8. betic rat and mouse models. Cancer Prev Res (Phila),
doi:10.1080/01926230390173842 PMID:12597447 8(3):231–9. doi:10.1158/1940-6207.CAPR-14-0181-T
Suttie AW, Dinse GE, Nyska A, Moser GJ, Goldsworthy PMID:25681088
TL, Maronpot RR (2005). An investigation of the Thurman JD, Bucci TJ, Hart RW, Turturro A
effects of late-onset dietary restriction on prostate (1994). Survival, body weight, and spontaneous
cancer development in the TRAMP mouse. Toxicol neoplasms in ad libitum-fed and food-restricted
Pathol, 33(3):386–97. doi:10.1080/01926230590930272 Fischer-344 rats. Toxicol Pathol, 22(1):1–9.
PMID:15805078 doi:10.1177/019262339402200101 PMID:8073218
Sylvester PW, Aylsworth CF, Meites J (1981). Relationship Tucker MJ (1979). The effect of long-term food restric-
of hormones to inhibition of mammary tumor develop- tion on tumours in rodents. Int J Cancer, 23(6):803–7.
ment by underfeeding during the “critical period” after doi:10.1002/ijc.2910230611 PMID:468413
carcinogen administration. Cancer Res, 41(4):1384–8. Volk MJ, Pugh TD, Kim M, Frith CH, Daynes RA, Ershler
PMID:6783295 WB, et al. (1994). Dietary restriction from middle age
Sylvester PW, Aylsworth CF, Van Vugt DA, Meites J (1982). attenuates age-associated lymphoma development and
Influence of underfeeding during the “critical period” interleukin 6 dysregulation in C57BL/6 mice. Cancer
or thereafter on carcinogen-induced mammary tumors Res, 54(11):3054–61. PMID:8187095
in rats. Cancer Res, 42(12):4943–7. PMID:6814746 Waxler SH, Tabar P, Melcher LR (1953). Obesity and
Tannenbaum A (1945a). The dependence of tumor forma- the time of appearance of spontaneous mammary
tion on the composition of the calorie-restricted diet carcinoma in C3H mice. Cancer Res, 13(3):276–8.
as well as on the degree of restriction. Cancer Res, PMID:13042818
5:616–25. Weber RV, Stein DE, Scholes J, Kral JG (2000). Obesity
Tannenbaum A (1945b). The dependence of tumor forma- potentiates AOM-induced colon cancer. Dig Dis
tion on the degree of caloric restriction. Cancer Res, Sci, 45(5):890–5. doi:10.1023/A :10 0556 0 621722
5:609–15. PMID:10795750
Teraoka N, Mutoh M, Takasu S, Ueno T, Nakano K, Weindruch R, Walford RL (1982). Dietary restriction
Takahashi M, et al. (2011). High susceptibility to azoxy in mice beginning at 1 year of age: effect on life-
methane-induced colorectal carcinogenesis in obese span and spontaneous cancer incidence. Science,
KK-Ay mice. Int J Cancer, 129(3):528–35. doi:10.1002/ 215(4538):1415–8. doi:10.1126/science.7063854
ijc.25711 PMID:20886595 PMID:7063854
Thompson HJ, McGinley JN, Spoelstra NS, Jiang W, Wheatley KE, Williams EA, Smith NCP, Dillard A,
Zhu Z, Wolfe P (2004b). Effect of dietary energy Park EY, Nunez NP, et al. (2008). Low-carbohydrate
restriction on vascular density during mammary diet versus caloric restriction: effects on weight loss,
carcinogenesis. Cancer Res, 64(16):5643–50. hormones, and colon tumor growth in obese mice. Nutr
doi:10.1158/0008-5472.CAN-04-0787 PMID:15313902 Cancer, 60(1):61–8. doi:10.1080/01635580701510150
Thompson HJ, Zhu Z, Jiang W (2002). Protection PMID:18444137
against cancer by energy restriction: all experimental White PB, True EM, Ziegler KM, Wang SS, Swartz-Basile
approaches are not equal. J Nutr, 132(5):1047–9. DA, Pitt HA, et al. (2010). Insulin, leptin, and tumoral
PMID:11983835 adipocytes promote murine pancreatic cancer growth.
551
J Gastrointest Surg, 14(12):1888–94. doi:10.1007/s11605- Zhu Z, Jiang W, McGinley J, Wolfe P, Thompson HJ (2005).
010-1349-x PMID:20859700 Effects of dietary energy repletion and IGF-1 infusion
Wolff GL, Kodell RL, Cameron AM, Medina D (1982). on the inhibition of mammary carcinogenesis by
Accelerated appearance of chemically induced dietary energy restriction. Mol Carcinog, 42(3):170–6.
mammary carcinomas in obese yellow (Avy/A) doi:10.1002/mc.20071 PMID:15599926
(BALB/c × VY) F1 hybrid mice. J Toxicol Environ Zhu Z, Jiang W, Thompson HJ (1999a). Effect of energy
Health, 10(1):131–42. doi:10.1080/15287398209530237 restriction on tissue size regulation during chemically
PMID:6290678 induced mammary carcinogenesis. Carcinogenesis,
Wolff GL, Medina D, Umholtz RL (1979). Manifestation 20(9):1721–6.doi:10.1093/carcin/20.9.1721PMID:10469616
of hyperplastic alveolar nodules and mammary tumors Zhu Z, Jiang W, Thompson HJ (1999b). Effect of
in “viable yellow” and non-yellow mice. J Natl Cancer energy restriction on the expression of cyclin D1
Inst, 63(3):781–5. PMID:288935 and p27 during premalignant and malignant stages
Wolff GL, Roberts DW, Mountjoy KG (1999). Physiological of chemically induced mammary carcinogenesis.
consequences of ectopic agouti gene expression: the Mol Carcinog, 24(4):241–5. doi:10.1002/(SICI)1098-
yellow obese mouse syndrome. Physiol Genomics, 274 4(19 9 9 0 4)24:4 <241:: A I D -MC1>3.0.CO; 2-P
1(3):151–63. PMID:11015573 PMID:10326860
Yakar S, Nunez NP, Pennisi P, Brodt P, Sun H, Fallavollita Zhu Z, Jiang W, Thompson HJ (2002). An experimental
L, et al. (2006). Increased tumor growth in mice with paradigm for studying the cellular and molecular
diet-induced obesity: impact of ovarian hormones. mechanisms of cancer inhibition by energy restric-
Endocrinology, 147(12):5826–34. doi:10.1210/en.2006- tion. Mol Carcinog, 35(2):51–6. doi:10.1002/mc.10073
0311 PMID:16959846 PMID:12325034
Yang S, Lin HZ, Hwang J, Chacko VP, Diehl AM (2001). Zhu Z, Jiang W, Thompson MD, Echeverria D, McGinley
Hepatic hyperplasia in noncirrhotic fatty livers: is JN, Thompson HJ (2015). Effects of metformin,
obesity-related hepatic steatosis a premalignant condi- buformin, and phenformin on the post-initiation
tion? Cancer Res, 61(13):5016–23. PMID:11431335 stage of chemically induced mammary carcinogen-
Yu BP, Masoro EJ, McMahan CA (1985). Nutritional influ- esis in the rat. Cancer Prev Res (Phila), 8(6):518–27.
ences on aging of Fischer 344 rats: I. Physical, metabolic, doi:10.1158/1940-6207.CAPR-14-0121 PMID:25804611
and longevity characteristics. J Gerontol, 40(6):657–70. Zucker LM, Zucker TF (1961). Fatty, a new mutation in the
doi:10.1093/geronj/40.6.657 PMID:4056321 rat. J Hered, 52:275–8.
Yu W, Cline M, Maxwell LG, Berrigan D, Rodriguez Zyromski NJ, Mathur A, Pitt HA, Wade TE, Wang S,
G, Warri A, et al. (2010). Dietary vitamin D expo- Nakshatri P, et al. (2009). Obesity potentiates the
sure prevents obesity-induced increase in endome- growth and dissemination of pancreatic cancer.
trial cancer in Pten+/− mice. Cancer Prev Res (Phila), Surgery, 146(2):258–63. doi:10.1016/j.surg.2009.02.024
3(10):1246–58. doi:10.1158/1940-6207.CAPR-10-0088 PMID:19628082
PMID:20858763
Yun JP, Behan JW, Heisterkamp N, Butturini A, Klemm
L, Ji L, et al. (2010). Diet-induced obesity accelerates
acute lymphoblastic leukemia progression in two
murine models. Cancer Prev Res (Phila), 3(10):1259–64.
doi:10.1158/1940-6207.CAPR-10-0087 PMID:20823291
Zhang Y, Proenca R, Maffei M, Barone M, Leopold
L, Friedman JM (1994). Positional cloning of the
mouse obese gene and its human homologue.
Nature, 372(6505):425–32. doi:10.1038/372425a0
PMID:7984236
Zhu P, Frei E, Bunk B, Berger MR, Schmähl D (1991). Effect
of dietary calorie and fat restriction on mammary
tumor growth and hepatic as well as tumor glutathione
in rats. Cancer Lett, 57(2):145–52. doi:10.1016/0304-
3835(91)90208-Y PMID:2025887
Zhu Z, Haegele AD, Thompson HJ (1997). Effect of caloric
restriction on pre-malignant and malignant stages of
mammary carcinogenesis. Carcinogenesis, 18(5):1007–
12. doi:10.1093/carcin/18.5.1007 PMID:9163688
552
4. MECHANISTIC AND
OTHER RELEVANT DATA
4.1 Introduction Fig. 4.1. Briefly, evidence must exist (i) that the

factor plays a significant role in the carcinogenic
The goal of this section is to assess which process (arrow 1 in Fig. 4.1), (ii) that obesity exerts
cellular and molecular processes known to be an effect on that factor (arrow 2), and (iii) that
dysregulated during the carcinogenesis process the factor affects the processes that regulate cell
are causally linked with obesity and, when suffi- proliferation, cell death, and/or angiogenesis with
cient data are available, to identify the organ sites an identifiable molecular basis for the observed
for which cancer risk is increased. This assess- changes in those processes (arrow 3). Although
ment, based primarily on data on obesity, was this approach is based on the traditional concept
extended to consider whether the dysregulation of mediation, it distinguishes itself by extending
observed in obesity was reversed by intentional the assessment to hallmarks of cancer and their
weight loss (IWL). With evidence of resolu- molecular underpinnings.
tion, the argument for a causal association was Factors were grouped as being operative
considered to be strengthened. There are two within the target cell (i.e. intracellular factors)
approaches to IWL with demonstrated efficacy or as external factors to which target cells are
in reducing body mass in obese individuals, i.e. exposed (i.e. host factors). Within the intracel-
dietary/energy restriction and bariatric surgery. lular category, the key characteristics related to
For the purposes of this assessment, the term electrophilic and metabolically activated carcin-
dietary restriction (DR) will be adopted to include ogens of exogenous and endogenous origin, the
both dietary and energy (calorie) restriction. damage they cause, and the mutations induced
The framework for this evaluation resulted (Smith et al., 2016) are considered not individ-
from the integration of the key characteristics ually but rather from the perspective of those
of carcinogenic agents used for identifying and mutations that confer a selective growth advan-
evaluating carcinogenic mechanisms in the IARC tage to a cell (i.e. driver mutations) versus those
Monographs (Smith et al., 2016), the concepts that do not (i.e. passenger mutations) (Vogelstein
arising from genome projects (Vogelstein et al., et al., 2013); also considered within the intra-
2013), and the characteristics of cancer referred cellular category are other key characteristics
to as cancer hallmarks (Hanahan & Weinberg, (Smith et al., 2016) that contribute to the emer-
2000, 2011). gence of driver mutations and their expression,
The approach used to evaluate the evidence including oxidative stress, epigenetic alterations
that a particular factor mediates the effects and various aspects of DNA repair. This assess-
of obesity on cancer development is shown in ment emphasized the cancer hallmarks related
553
Fig. 4.1 Diagram of the paradigm used to establish the mechanisms that mediate the effects of
obesity on cancer risk
Obesity,
Cancer
intentional
risk
weight loss
Epidemiology, 2 Putative 1
Epidemiology,
intervention, mechanism preclinical
preclinical (mediation)
Cellular and molecular

characteristics of
cancer (hallmarks)
Compiled by the Working Group.
to the dysregulation of the balance between cell 4.2 Intracellular factors

proliferation and cell death, clonal expansion,
and angiogenesis. Within the approach outlined in Section 4.1,
The host factors considered in this section the findings are presented by the strength of the
are related to small molecules involved in energy evidence of an effect of obesity or IWL on those
metabolism and macromolecular synthesis, factors.
mediators involved in inflammation, and those
factors that exert their effects via cell surface 4.2.1 Cell proliferation, apoptosis, and
receptors (growth factors, sex hormones, and angiogenesis
cytokines) (receptor-mediated effects). With
Because the time frame for the development
regard to the small molecules, many of which can
of obesity and cancer is long and the imbalance
be considered as energy substrates, the approach
between cell proliferation and cell death is small
was inclusive of the microbiome and of the
at any snapshot in time (Bozic et al., 2010), eval-
intracellular energy sensors that integrate extra-
uation of the impact of obesity on these factors
cellular and intracellular signals by affecting the
is problematic, even though their involvement in
processes that drive clonal expansion and selec-
obesity-induced carcinogenesis is obligatory. The
tion and disease progression. Most of the factors
impact of IWL, particularly via bariatric surgery,
considered are shown in Fig. 4.2.
provides an opportunity to gain insight into how
these processes are regulated in the situation of
negative energy balance.
554
Fig. 4.2 Summary of mechanisms underlying the obesity–cancer link
Obesity Dietary
BMI fat
Waist circumference
Inflammation
A TNF-α
Adipose tissue
d
Adipocytes
IL-6, IL-1β
PGE-2
i MCP-1, CRP FFA Liver
MACp TG ↑gluconeogenesis
Mito
Testosterone
↑PAI-1 ↓VEGF Estradiol Plasma

Glucose
↓Adiponectin Insulin ↑ROS

↑IGF-1 ↓IGFBPs
resistance ↓SHBG
↑Leptin
↑C-peptide ↑Bioactive IGF

Bioactive
↑Insulin
estradiol

DNA damage
Angiogenic Cell migration Mitogenic Anti-apoptotic
mutagenesis
Cancer cell promotion and progression

Factors denoted in bold red text are established features of the obesity–cancer connection. Factors denoted in bold blue text are emerging features.
BMI, body mass index; CRP, C-reactive protein; FFA, free fatty acids; IGF, insulin-like growth factor; IGFBP, IGF binding protein; IL, interleukin; MAC, macrophage; MCP-1, monocyte
chemoattractant protein 1; Mito, mitochondria; PAI-1, plasminogen activator inhibitor 1; PGE2, prostaglandin E2; ROS, reactive oxygen species; SHBG, sex hormone-binding globulin;
TG, triglycerides; TNF-α, tumour necrosis factor alpha; VEGF, vascular endothelial growth factor.
555
(a) Cell proliferation binding of p16 and p19 to CDK4 (Zhu et al.,
The effect of IWL on neoplasia-associated 1999a; Jiang et al., 2003). In addition, DR reduced
cell proliferation has received little attention in epidermal proliferation during tumour promo-
intervention studies, and the evidence is limited tion in mice (Azrad et al., 2011), and endometrial
to effects of bariatric surgery. Whereas IWL via cancer cells grown in sera obtained from women
bariatric surgery, including the Roux-en-Y gastric randomized to calorie restriction were less mito-
bypass, resulted in a reduction in endometrial genic than cells grown in sera obtained from
hyperplasia (Argenta et al., 2013; Modesitt et al., overweight women (Moore et al., 2012).
2015), proliferation was reported to be increased (b) Apoptosis
after Roux-en-Y gastric bypass or jejunoileal
bypass surgery in the rectum (Appleton et al., Intervention studies of IWL that evaluated
1988; Sainsbury et al., 2008; Kant et al., 2011); apoptosis end-points in cancer were not identi-
however, when a sleeve gastrectomy was used, fied. Therefore, studies in rat and mouse models
hyperproliferation was not observed in the rectal with cancer-related end-points were reviewed.
mucosa (Kant et al., 2014). Although the reduc- A dose-dependent relationship between DR and
tion in endometrial hyperplasia is consistent with elevated rates of apoptosis has been reported (Zhu
reduced risk of endometrial cancer, the question et al., 1999b; Thompson et al., 2004a; Tomita,
of how the hyperproliferative state would affect 2012; Olivo-Marston et al., 2014). DR induced a
the risk of cancer of the colon and rectum has pro-apoptotic state via the coordinated regula-
been identified as a concern (Appleton et al., tion of pro- and anti-apoptotic factors involved in
1988; Sainsbury et al., 2008; Kant et al., 2011). the mitochondrial pathway of caspase activation
Another question of interest is how the regu- (Thompson et al., 2004a). Specifically, comple-
lation of cell cycle machinery, which ultimately mentary DNA (cDNA) microarray analysis
accounts for effects on the magnitude of cell identified the Bcl-2, CARD, and IAP functional
proliferation observed in a tissue, is affected by gene groupings as being involved in induction of
IWL; however, that specific question has not apoptosis. Consistent with the microarray data,
been addressed in IWL intervention studies the activities of caspases 9 and 3 were observed
in humans or rodents. What has been done in to be 2-fold higher in carcinomas from DR rats,
experiments in rodents is to investigate the whereas the activity of caspase 8 was similar in
effects of DR, which is protective against cancer carcinomas from DR animals and those fed ad
(see Section 3) and has been reported to decrease libitum. Collectively, this evidence indicated that
cell proliferation in mammary tumours (Zhu et DR-induced apoptosis is mediated by the mito-
al., 1999b; Jiang et al., 2003). Briefly, the studies chondrial pathway.
focused on factors that regulate the G1/S transi- (c) Angiogenesis
tion, which appears to be a target when energy
availability is limited by DR (Jiang et al., 2003). Studies of the effects of IWL on angiogenesis
Observed effects included reductions in levels of in the context of cancer in humans were not iden-
phosphorylated retinoblastoma protein and the tified. However, studies have reported the effects
transcription factor E2F1, decreased activity of of IWL via DR (Rizkalla et al., 2012; Cullberg
cyclin-dependent kinase 2 (CDK2) and CDK4, et al., 2013) or bariatric surgery (Lemoine et
increased concentrations of the CDK inhibitors al., 2012; Moreno-Castellanos et al., 2015) on
Cip1/p21 and Kip1/p27, increased levels of these circulating factors that reflect angiogenic drive
proteins complexed with CDK2, and increased and on gene and protein expression profiles in
adipose tissue sampled before and after weight
556
loss. For either intervention approach, levels of concentrations of growth factors and hormones
circulating factors associated with angiogenesis, that affect cell function as well as the mecha-
for example vascular endothelial growth factor nisms that drive the carcinogenic process. These
A (VEGF-A) and angiopoietin 1 (ANG-1), are IWL-mediated intracellular and systemic effects
reduced by IWL, whereas the level of angio are transduced to signalling pathways that regu-
poietin-like 4 (ANGPTL-4) is increased and the late tissue growth and endothelial homeostasis
pattern of gene or protein expression in adipose via intracellular nutrient and energy sensors.
tissue in response to IWL has been characterized Prominent among these pathways are those
as anti-angiogenic (Cullberg et al., 2013). regulated by adenosine monophosphate (AMP)-
Reduction in tumour vascularization in activated protein kinase (AMPK)–mammalian
response to DR has been reported in rodent models target of rapamycin (mTOR)–protein kinase B
of cancer (Mukherjee et al., 2004; Thompson et (AKT), sirtuins, peroxisome proliferator-acti-
al., 2004a; Higami et al., 2006; Powolny et al., vated receptors (PPARs), and soluble guanylyl
2008; Zhu et al., 2009; De Lorenzo et al., 2011; cyclase (sGC). Most of this discussion focuses on
Kurki et al., 2012), and this has been shown to AMPK–mTOR–AKT (i.e. the mTOR network);
involve many of the same factors identified in the other pathways are briefly discussed, recog-
the clinical studies. These factors play roles at nizing their likely involvement in mediating the
different stages of the angiogenic process, which effects of IWL.
can be divided into endothelial proliferation and
migration, blood coagulation, fibrinolysis, and (a) The mTOR network
the degradation of basement membranes and the IWL can inhibit tumour growth by
extracellular matrix. suppressing the activation of the mTOR signal-
ling network. In this network, mTOR plays a key
(d) Synthesis role in integrating information received from
Alteration of cell proliferation, apoptosis, and the extracellular environment via the binding of
angiogenesis are key characteristics of carcino- growth factors and hormones with their cognate
genesis, and their necessary involvement in the receptor tyrosine kinases (Gwinn et al., 2008).
development of cancer is established. Available Suppression is mediated through the effects of
studies of IWL in humans and rodents, although restricted energy availability on concentrations
limited in number, support the view that obesity of the circulating growth factors and hormones
dysregulates one or more of these processes, and and of the substrates used in intermediary metab-
that IWL can reverse these changes. olism to synthesize high-energy phosphates and
reducing equivalents. As a consequence, the drive
4.2.2 The mTOR network and other energy- for cell proliferation is reduced (Zhu et al., 1999a;
sensor networks Jiang et al., 2003; Moore et al., 2008; Lashinger et
al., 2011; De Angel et al., 2013), a pro-apoptotic
Blood levels of amino acids, carbohydrates, environment is maintained (Zhu et al., 1999b;
and lipids – the primary substrates that are inter- Thompson et al., 2004a), and the stimulus for
converted and metabolized to produce energy – formation of new blood vessels is suppressed
are generally altered in obesity, and are reduced (Thompson et al., 2004b). One or more elements
during IWL, whether it is achieved via bariatric of the mTOR network are dysregulated in the
surgery or DR (Thompson et al., 2012; Fabian et majority of human cancers (Wood et al., 2007).
al., 2013; Modesitt et al., 2015). In addition, IWL AMPK serves as a metabolic checkpoint,
exerts systemic effects by altering circulating downregulating cell growth and cell division
557
in the absence of an adequate supply of biosyn- studies have shown that activation or inhibition
thetic and energy substrates (Gwinn et al., 2008). of histone deacetylases can alter the carcinogenic
AMPK has been shown to be an exquisitely sensi- process (Ahmad et al., 2012; Guo & Zhang, 2012;
tive detector of small changes in the intracellular Jiang et al., 2013; Ravillah et al., 2014; Busch et
ratio of AMP to adenosine triphosphate (ATP), al., 2015).
and some investigators have even proposed that
AMPK plays a central role in homeostatic regula- (c) Peroxisome proliferator-activated receptors
tion of whole-body energy metabolism (Hardie, PPARs are transcription factors that are
2004). activated by long-chain fatty acids and their
IWL by bariatric surgery (Peng et al., 2010) oxidized metabolites, the oxylipins. There are
and DR (Jiang et al., 2008, 2009) results in AMPK three isoforms of PPARs (α, β/δ, and γ), each of
activation. This suggests that either energy which has tissue-specific distribution and activity
availability alters substrate availability (the fuel (Georgiadi & Kersten, 2012; Janani & Ranjitha
mixture presented to tissues throughout the Kumari, 2015). Because the intracellular concen-
body) or activation is being induced via a mech- trations of PPARs are affected by obesity and
anism independent of the AMP-to-ATP ratio. IWL, they are considered to be energy sensors,
In this regard, it is clear that additional factors and their activation or lack thereof regulates not
control the activation of AMPK, including only energy metabolism (lipid metabolism as
various cytokines such as adiponectin (Kahn et well as glucose homeostasis) but also cell growth
al., 2005). and differentiation (Cantó et al., 2015; Cetrullo et
Limiting energy availability, for example by al., 2015; Cao et al., 2016).
DR, has been reported to decrease circulating Studies of the effects of IWL on PPAR expres-
levels of insulin and insulin-like growth factor sion in the context of cancer in humans were not
1 (IGF-1) (Zhu et al., 2005; Jiang et al., 2008; identified. The expression of PPARγ1, which has
Nogueira et al., 2012; Ford et al., 2013; Lashinger been reported to be suppressed in subcutaneous
et al., 2013; Harvey et al., 2014; Olivo-Marston adipose tissue in obesity, is restored by IWL
et al., 2014). Lower levels of these growth factors induced by bariatric surgery (Leyvraz et al.,
downregulate signalling via the pathway of 2012). Many reports in humans and rodents indi-
which IGF-1 receptor (IGF-1R), phosphoinositide cate that suppression of PPAR-related signalling
3-kinase (PI3K), and AKT are components. Of constitutes a link between obesity and cancer
these proteins, activated Akt, a serine/threonine and that pharmacological activation of PPARs is
kinase, is the critical effector molecule (Hursting protective against cancer (Georgiadi & Kersten,
et al., 2003). 2012; Laplante & Sabatini, 2013; Janani &
Ranjitha Kumari, 2015; Kim et al., 2015; Mishra
(b) Sirtuins et al., 2016; Polvani et al., 2016).
Studies of the effects of IWL on histone
deacetylase activity in the context of cancer in (d) Soluble guanylyl cyclase
humans have not been identified. However, it is sGC is the receptor for nitric oxide, which is
widely recognized that the activity of SIRT1 is synthesized and released by various cell types as a
lower in obesity and that sirtuins are activated paracrine–autocrine mechanism that coordinates
by IWL in liver and adipose tissue (Moschen et energy production with consumption, in part by
al., 2013; Xu et al., 2013; Jukarainen et al., 2016; improving the delivery of substrates and oxygen
Rappou et al., 2016). Sirtuins play a significant via the vascular system (Bellamy et al., 2002;
role in altering gene expression, and recent Nossaman et al., 2012). Nitric oxide-mediated
558
signalling has been reported to be suppressed in population, the more than 100 identified loci
obesity and restored by IWL induced by bariatric associated with body mass index (BMI) account
surgery (Felipo et al., 2013; Blum et al., 2015). for only 3% of the inter-individual variation of
Although the activation of sGC by nitric oxide BMI, and genome-wide estimates suggest that
induces tissue-specific responses, its link with common variation accounts for more than 20%
energy metabolism and cancer is attributed to of BMI variation (Speliotes et al., 2010; Locke et
endothelial homeostasis, to induction of angio- al., 2015; Shungin et al., 2015). It is hypothesized
genesis, and to the downstream effects of cyclic that epigenetic mechanisms may be a missing
guanosine monophosphate (cGMP), the product link between the obesity-associated genes and
of sGC; cGMP activates protein kinase GI, which the phenotype, and evidence is beginning to
in turn inhibits RhoA, resulting in the release emerge in this area. Despite the different types
of the RhoA/Rho-associated protein kinase of epigenetic alterations, studies in humans have
(ROCK)-dependent inhibition of the insulin– largely been limited to examining DNA methyl-
insulin receptor substrate 1 (IRS-1)–PI3K–Akt ation. In a few small genome-wide studies, asso-
pathway (Furukawa et al., 2005; Huang et al., ciations between DNA methylation and BMI or
2013). Of additional interest is a recent report other indices of obesity were investigated, but the
that sGC agonists induce brown adipose tissue findings were generally inconclusive (Feinberg et
differentiation and the browning of white al., 2010; Wang et al., 2010; Almén et al., 2012;
adipose tissue in obese mice, effects that result Relton et al., 2012).
in increased energy expenditure and weight loss An epigenome-wide association study
(Hoffmann et al., 2015). Therefore, this little- investigated associations between methylation
studied energy-sensing cascade provides direct patterns in whole blood from 459 European indi-
links between energy metabolism, vascular viduals and BMI. Samples were typed using the
supply, and tumour progression. Infinium HumanMethylation450 array. BMI
was associated with differential methylation at
(e) Synthesis sites cg22891070, cg27146050, and cg16672562
The role of the mTOR network in obesity located in the intron 1 region of HIF3A (Dick et
and cancer is well established and illustrates the al., 2014). A subanalysis of methylation patterns
complex nature of the regulatory cascades that in adipose tissue found a similar association,
underlie this relationship. There are suggestions thus suggesting that this is a BMI-related modifi-
that other energy-sensing networks, such as cation of the epigenome. In a subsequent investi-
sirtuins, PPARs, and sGC, are involved in the gation conducted in 991 individuals in the USA,
association between obesity and cancer; however, with replication sets from other cohorts in the
direct evidence of an effect of IWL is lacking. USA, associations between DNA methylation
and BMI and waist circumference were assessed
4.2.3 Epigenetics, oxidative stress, DNA repair, (Aslibekyan et al., 2015). Differentially methyl-
and telomeres ated loci in CPT1A and PHGDH (genes involved
in energy metabolism) and CD38 were found to
(a) Epigenetics be associated with BMI and waist circumference.
(i) Epigenetics and obesity (ii) Epigenetics and intentional weight loss
Unlike in cancer research, epigenetic inves- Emerging evidence in humans suggests that
tigations are relatively new in the field of obesity IWL is associated with changes in DNA methyl-
research (van Dijk et al., 2015). In the general ation patterns. A small study showed that DNA
559
methylation in adipose tissue after a 6-month DR or the MethyLight assay. Compared with 209
were higher in 7 women who lost 3% or more normal-weight women (BMI, 18.5–24.9 kg/m2),
of their body fat than in 7 women who lost less 167 overweight women (BMI ≥ 25.0 kg/m2) were
than 3% of their body fat (Bouchard et al., 2010). more likely to have methylated promoter regions
Results from other studies using shorter-term for HIN1 (OR, 1.57; 95% CI, 1.03–2.39). No signif-
dietary interventions also suggest that diet- icant associations were detected for the 12 other
induced weight loss causes differential DNA genes that were investigated.
methylation patterns (Milagro et al., 2011; A subsequent study examined methylation at
Mansego et al., 2015). 1505 genes with known relevance to cancer using
breast tumour tissue from women with breast
(iii) Epidemiological evidence of the epigenetic
cancer (Hair et al., 2015). Methylation status of
mediation between obesity and cancer risk
the tissue was assessed using the Cancer Panel
Evidence supporting an epigenetic mediation 1 platform. Although 30 CpG sites were differ-
in the link between obesity and cancer risk is entially methylated among 195 normal-weight
sparse and fragmented, and was identified only women with breast cancer compared with 150
for breast cancer and colorectal cancer (CRC). obese women with breast cancer in unadjusted
In a study of 803 premenopausal and post- analyses, only two sites (on the SH3BP2 and
menopausal women with breast cancer, asso- XIST genes) remained statistically significant in
ciations between BMI and waist-to-hip ratio the final adjusted models (false discovery rate
(WHR) with methylation at the E-cadherin, q < 0.05). In analyses limited to estrogen receptor
p16, and RAR-β2 genes were examined in breast (ER)-positive tumours, differential methylation
tumour tissue (Tao et al., 2011). Promoter at CpG sites was statistically significant on the
methylation was assessed by using real-time SH3BP2, IGFBP6, DNMT3B, and ERCC6 genes.
methylation-specific polymerase chain reaction In a case–case study, associations between
(PCR). Compared with women in the lowest BMI and the CpG island methylator phenotype
quartile of WHR, those in the highest quartile (CIMP) in CRC were investigated using data
were more likely to have methylation at one or from 3119 patients from the Colon Cancer Family
more of the promoter regions that were assessed Registry (Weisenberger et al., 2015). CIMP CRC
(odds ratio [OR], 1.85; 95% confidence interval was more common in women (16.8%) than in
[CI], 1.10–3.11). No significant differences were men (9.3%) (P = 0.0001). However, only among
found in similar case–case comparisons of BMI, women were positive associations between BMI
or weight change (from age 20 years to 1 year and CIMP CRC observed. Compared with
before study enrolment), nor were significant normal-weight women, overweight and obese
trends detected for these indicators of body size women were more likely to have CIMP CRC
and body size history. (OR, 1.42; 95% CI, 1.09–1.86 for overweight
In another study of 532 postmenopausal women and OR, 1.93; 95% CI, 1.09–2.56 for obese
women with breast cancer in the USA, one arm women).
of the investigation examined whether BMI was Evidence on IWL is limited to experimental
associated with promoter methylation status in studies. One study used a diet-induced obesity
13 breast cancer-related genes (APC, BRCA1, (DIO) rodent model followed by DR to inves-
CCND2, CDH1, DAPK1, ESR1, GSTP1, HIN1, tigate the epigenetic effects of DIO and DR on
CDKN2A, PGR, RARβ, RASSF1A, and TWIST1) mammary tissue (Rossi et al., 2016). C57BL/6
(McCullough et al., 2015). Promoter methylation mice were fed a control diet or a DIO regimen, and
status was assessed by methylation-specific PCR mice on the DIO regimen were then randomized
560
to continue the DIO diet or switch to the control (i) Oxidative stress and obesity
diet, resulting in formerly obese mice with In obesity, adipose tissue is characterized
weights comparable to those of the control mice. by chronic, low-grade inflammation, which
Comparisons among control, DIO mice, and promotes oxidative stress. Adipokines can
formerly obese mice both showed that there was also induce the production of reactive oxygen
a persistent effect of obesity on hypermethylation species (ROS), resulting in oxidative stress and,
patterns in mammary tumours, even after DR. in turn, causing production of other adipokines
(iv) Synthesis (Marseglia et al., 2015). Many activated immune
Data on the epigenetics of obesity are cells generate free radicals, and the synthesis of
emerging. Although epigenetic links between ROS further promotes inflammation (Marseglia
obesity and cancer risk are biologically plau- et al., 2015). Obesity-induced oxidative stress may
sible, to date the evidence in support of them is elicit or exacerbate insulin resistance (Marseglia et
sparse and fragmented, and most of the studies al., 2015). In addition, increased ROS production
have investigated only DNA methylation. may promote calcium mishandling by affecting
Epidemiological studies of breast cancer (Tao et the redox state of key proteins implicated in this
al., 2011; Hair et al., 2015; McCullough et al., 2015) process. Levels of ROS are frequently increased
and CRC (Weisenberger et al., 2015) have used in obesity, and obesity induced by a high-fat diet
DNA methylation at known cancer-related genes has been shown to increase oxidative stress in
to investigate associations of BMI with epige- animal models (e.g. Dobrian et al., 2001; Vincent
netic tumour characteristics. Taken together, et al., 2007; Matsuda & Shimomura, 2013; Cerdá
these studies suggest that obesity may contribute et al., 2014).
to carcinogenesis via epigenetic mechanisms, but (ii) Oxidative stress and dietary restriction/
to date few associations have been detected and weight loss
there has been almost no replication of findings One important and consistent effect of DR
among the different investigations. is the ability to reduce oxidative stress and its
resulting damage to macromolecules. Three
(b) Oxidative stress
possible mechanisms have been identified for
Oxidative stress is a well-established mech- the antioxidant effects of DR: DR may (i) reduce
anism of the carcinogenic process and is one the production of ROS, (ii) directly increase the
of the key characteristics as defined by Smith activity of antioxidant enzymes, or (iii) increase
et al. (2016). To date, multiple biomarkers have the turnover of oxidized macromolecules, such
been developed that measure oxidative damage. as oxidized lipids or DNA, which are commonly
A commonly measured marker for whole-body measured as biomarkers. These effects are
oxidative stress is the isoprostane 8-epi-prosta- complicated and are thought to be influenced by
glandin F2α (8-epi-PGF2α), which can be measured several factors, including sex, species, or tissue
in blood and/or urine (Morrow & Roberts, 1997; studied, types of ROS or biomarkers and anti-
Czerska et al., 2015). The activity of antioxidant oxidant enzymes examined, and duration of DR
enzymes and their products (e.g. glutathione (Merry, 2000; Skrha, 2009).
peroxidase, catalase) and 8-hydroxydeoxy- Five recent studies were identified that inves-
guanosine (8-oxo-dG) can also provide some tigated the effect of weight-loss interventions
information about oxidative stress processes in on an individual’s oxidative stress level: three
humans (Roszkowski, 2014). randomized controlled trials (RCTs) (Meydani
et al., 2011; Buchowski et al., 2012; Wegman et
561
al., 2015) and two non-randomized intervention who were randomized to either a high glycaemic
studies (Gutierrez-Lopez et al., 2012; Chae et load or a low glycaemic load regimen with either
al., 2013). All of these studies measured oxida- 10% (n = 12) or 30% (n = 34) reduction in calorie
tive stress by identifying markers (e.g. activity intake for 6 months. Overall, independently of
of enzymes, 8-epi-PGF2α, 8-oxo-dG) in blood the type of calorie-restriction regimen, body
(plasma or serum) or urine samples. weight decreased, plasma glutathione perox-
Buchowski et al. (2012) conducted an RCT idase activity increased (P = 0.04), and plasma
comparing a 25% calorie-restricted diet and a protein carbonyl levels decreased (P = 0.02), with
control (habitual) diet in 40 overweight or obese a concurrent nonsignificant decrease in plasma
women, with direct observation for 28 days and 8-epi-PGF2α levels (P = 0.09) and no changes in
follow-up for the next 90 days. The initial (base- superoxide dismutase and catalase activity.
line) serum F2-isoprostane concentration in the Wegman et al. (2015) recruited a cohort of
calorie-restricted group (median, 57.0 pg/mL; 24 healthy individuals in a double-crossover,
interquartile range, 40.5–79.5 pg/mL) was 1.75 double-blinded RCT of intermittent fasting.
times the average concentration in normal- Study participants underwent two 3-week treat-
weight women (32.5 pg/mL). During calorie ment periods: intermittent fasting and intermit-
restriction (which resulted in a 3.2% reduction tent fasting with antioxidant (vitamins C and
in body weight after 29 days), F2-isoprostane E) supplementation. Despite strict adherence to
levels fell rapidly, resulting in statistically signif- study-provided diets, no change in expression
icant differences from the control group by of oxidative stress markers was observed. Body
day 5 (median, 33.5 pg/mL; interquartile range, weight remained stable over the entire trial
26.0–48.0 pg/mL; P < 0.001). F2-isoprostane period.
levels remained low while the study participants Chae et al. (2013) investigated overweight or
continued on the calorie-restricted diet, but obese participants (BMI, 25–34 kg/m2, n = 122,
returned to the higher baseline concentrations aged 30–59 years) who joined a clinical interven-
in about 80% of the women after 3 months on a tion lasting 3 years and involving daily calorie
habitual diet. deficits of 100 kcal. Body weight changed by 5.4%
In an intervention study of 16 normal-weight (−4.16 ± 0.31 kg) in the group with successful
and 32 obese individuals (BMI, 30–34.9 kg/m2), mild weight loss (n = 50) compared with
Gutierrez-Lopez et al. (2012) studied the effects 0.05 ± 0.14 kg in the unsuccessful group (n = 49).
of a hypocaloric diet and a hypocaloric diet plus Successful mild weight loss was coupled with
regular moderate aerobic exercise on oxidative significantly reduced serum levels of insulin,
stress. Over 90 days, an average weight loss of IL-6 (30% decrease; P = 0.031), IL-1β (45%
7.6% was achieved. Higher levels of oxidative decrease; P < 0.001), and tumour necrosis factor
stress markers and increased molecular damage alpha (TNF-α) (P < 0.001), as well as urinary
and polymerization of insulin were observed in 8-epi-PGF2α (14% decrease; P = 0.036). A posi-
the blood from obese individuals at baseline. tive correlation was reported between IL-1β and
Treatment with a hypocaloric diet significantly urinary 8-epi-PGF2α (r = 0.435, P < 0.001) and
decreased oxidative stress and molecular damage between the corresponding changes in IL-6 and
to values similar to those of normal-weight urinary 8-epi-PGF2α (r = 0.393, P < 0.001).
individuals.
As part of a controlled feeding study, Meydani
et al. (2011) studied 46 moderately overweight
volunteers (BMI, 25–30 kg/m2) aged 20–42 years
562
(iii) Synthesis to a defect in the DNA mismatch repair system),

Oxidative stress is well established as a cellular and in about 30% of endometrial cancer cases,
mechanism that can affect DNA integrity and has sporadic MSI occurs (Mills & Longacre, 2016).
been linked to cancer, metabolic syndrome, and Furthermore, the positive associations observed
obesity. Evidence of the involvement of oxidative between BMI and endometrial cancer are signif-
stress in obesity-induced cancer in humans is icantly stronger among carriers of germline
limited by methodological issues. Results from mutations in the DNA mismatch repair genes
weight-loss intervention trials indicate that MLH1, MSH2, MSH6, or PMS2 than among
oxidative stress can be rapidly reduced and the non-carriers (Win et al., 2011).
lower level sustained through a modest reduc- In a clinical study of 446 women with endo-
tion in calorie intake. metrial cancer (McCourt et al., 2007), women
with MSS tumours were significantly heavier
(c) DNA repair (median BMI, 32.7 kg/m2) than those with MSI
(i) DNA repair mediation in obesity and tumours (median BMI, 30.3 kg/m2) (P = 0.02). A
cancer larger, population-based case–control study of
524 cases and 1032 controls investigated associ-
Elevated BMI is consistently associated with
ations between obesity and endometrial cancer
CRC (see Section 2.2.1). To assess the role of
by microsatellite status (Amankwah et al., 2013).
DNA repair in this association, several studies
Unlike in the previous study, the association
have investigated associations between BMI and
between BMI and MSI tumours was stronger
CRC stratified by tumour microsatellite status
than that between BMI and MSS tumours
(Campbell et al., 2010; Hoffmeister et al., 2013).
(P heterogeneity = 0.05 for overweight and 0.02 for
In a population-based study, CRC cases were
obesity).
divided into those with high-level microsatellite
instability (MSI-high) tumours and microsat- (ii) DNA repair and dietary restriction/weight
ellite-stable (MSS) tumours (Hoffmeister et al., loss
2013). Among the 1215 cases, 67% were over- One recent RCT investigated the effect
weight or obese, and 115 (9.5%) had MSI-high of weight loss on DNA repair capacity. The
tumours. BMI was weakly associated with MSS Nutrition and Exercise for Women RCT
tumours in women (OR, 1.15; 95% CI, 0.97–1.35 recruited 439 women, who were randomized
per 5 kg/m2) and in men (OR, 1.25; 95% CI, to one of four groups: (i) dietary intervention,
1.08–1.45 per 5 kg/m2); in contrast, the asso- (ii) aerobic exercise, (iii) diet plus exercise, or
ciation between BMI and MSI-high CRC was (iv) control (Habermann et al., 2015). The diet
significant only in women (OR, 2.04; 95% CI, intervention was a group-based programme
1.50–2.77 per 5 kg/m2). When the analysis was with a goal of 10% weight loss. The exercise
limited to case–case comparisons, BMI was intervention consisted of moderate to vigorous
more strongly associated with MSI-high than aerobic activity for 45 minutes per day, 5 days
with MSS tumours in women (OR, 1.84; 95% CI, per week. DNA repair capacity was measured
1.34–2.52 per 5 kg/m2), but not in men. in fasting blood samples taken at baseline and
Elevated BMI is consistently associated after 12 months in a subset of 226 women, using
with increased risk of endometrial cancer a modified comet assay conducted in pre- and
(see Section 2.2.9). Endometrial cancer is also post-intervention cryopreserved lymphocytes,
commonly observed in women with Lynch analysed within the same batch. DNA repair
syndrome (hereditary non-polyposis CRC due capacity did not change significantly with any of
563
the diet or exercise interventions compared with of the interventions resulted in any significant
the control group. Similarly, there were no signif- group differences in leukocyte telomere length
icant changes when the analysis was stratified by compared with controls, and there were no
changes in body composition or aerobic fitness differences in telomere length by the degree of
(maximal oxygen consumption, Vo2max). weight loss.
García-Calzón et al. (2014) reported that a
(iii) Synthesis
2-month energy-restricted diet (30% of energy
The role of DNA repair function in cancer from fat, 15% from proteins, and 55% from carbo-
risk is unequivocal and is particularly well hydrates) among overweight or obese adolescents
established for cancers of the colorectum, breast, aged 12–16 years resulted in increased telomere
endometrium, and skin. However, there have length, with a greater effect in those who had the
been few studies investigating functional assays shortest telomeres at baseline (r = 0.96, P < 0.001).
of DNA repair in the context of obesity or weight
reduction. One well-designed RCT showed no
effects, but assay limitations were present. Several 4.3 Receptor-mediated effects
studies point towards a link between BMI and
Adiposity and overweight/obesity are asso-
DNA mismatch repair deficiencies. Overall, a
ciated with significant metabolic and endo-
causal link of obesity and weight control with
crinological changes that are included as key
DNA repair is still lacking.
characteristics of the carcinogenesis process, in
(d) Telomeres particular (i) alterations in sex hormone metab-
olism, (ii) changes in insulin levels and IGF
Multiple studies have reported that obesity signalling, and (iii) chronic inflammation (see
is associated with shorter telomere length in Table 4.1; for a review, see Pischon & Nimptsch,
different cell types. A recent systematic review 2016). A large and growing number of epidemio-
and meta-analysis comprising 119 439 individ logical and experimental studies have measured
uals reported that 39 studies showed weak to biomarkers of these pathways in relation to cancer
moderate correlations between obesity and at different sites. Data in the tables are presented
telomere length (Mundstock et al., 2015). by type of cancer (Tables 4.2–4.11), whereas the
However, there was significant heterogeneity, text summarizes the studies by mechanistic
which suggests that this relationship is still
pathway.
incompletely understood.
In the Nutrition and Exercise for Women
RCT of 439 postmenopausal women random- 4.3.1 Sex hormones
ized to diet, exercise, diet plus exercise, or control Sex hormones are involved in specific cancers,
groups for 1 year (see Section 4.2.3c(ii)), DNA was exemplified by the implications of estrogen in
extracted from isolated leukocytes, and telomere breast and endometrial cancers and of androgen
length was measured by quantitative PCR in prostate cancer. In postmenopausal women,
(Mason et al., 2013a). Baseline telomere length estrogens are synthesized almost exclusively in
was correlated inversely with age (r = −0.12, adipose tissue stromal cells, and consequently
P < 0.01) and positively with Vo2max (r = 0.11, obese postmenopausal women have elevated
P = 0.03), but was not correlated with BMI or levels of estrogens compared with leaner post-
body fat percentage. The change in telomere menopausal women (Key et al., 2003).
length was inversely associated with the telomere
length at baseline (r = −0.47, P < 0.0001). None
564
(a) Cancer of the breast 2002; Zhu & Pollard, 2007). In premenopausal
See Table 4.2. endometrial tissue, the actions of estrogen are
Estrogens stimulate the proliferation of opposed by those of progesterone (Gao & Tseng,
normal breast tissue and neoplastic breast 1997). Consistent with these mechanistic data,
epithelial cells directly and can promote the use of unopposed estrogen postmenopausal
development of ER-positive, estrogen-dependent hormone therapy is associated with a signifi-
breast cancer by both endocrine and paracrine cantly higher risk of endometrial cancer, whereas
mechanisms (Vona-Davis & Rose, 2007; Bulun use of the combined estrogen plus progestogen
et al., 2012). formulation appears to have a protective effect
Elevated levels of circulating estrogens have (Beral et al., 2005).
been linked to breast cancer risk in numerous There are consistent epidemiological data
epidemiological studies (Hankinson et al., 1998a; linking higher circulating estrogen levels with
Kaaks et al., 2005; Tworoger et al., 2011; Zhang et increased risk of endometrial cancer. Five
al., 2013). The Endogenous Hormones and Breast prospective investigations of estradiol concentra-
Cancer Collaborative Group, which pooled data tions and endometrial cancer have all reported
from nine prospective investigations of sex relative risks between 2 and 4 for the compar-
hormone levels and breast cancer comprising ison of high versus low estradiol levels in multi-
individual data from 663 incident breast cancer variate models that controlled for adiposity and
cases and 1765 controls, reported that risk of other established endometrial cancer risk factors
postmenopausal breast cancer is 2-fold higher (Zeleniuch-Jacquotte et al., 2001; Lukanova et al.,
among women in the highest versus the lowest 2004a; Allen et al., 2008; Gunter et al., 2008a;
quintile of estradiol and testosterone levels, Brinton et al., 2016). In addition, higher circu-
as well as for other related sex hormones such lating levels of sex hormone-binding globulin
as estrone, dehydroepiandrosterone (DHEA), (SHBG) were associated with significantly lower
DHEA sulfate, and androstenedione (Key et al., risk of endometrial cancer in three of these
2002). Furthermore, in a subsequent analysis, prospective studies (Zeleniuch-Jacquotte et al.,
the positive association between BMI and risk 2001; Lukanova et al., 2004a; Allen et al., 2008).
of postmenopausal breast cancer was almost (c) Cancer of the colorectum
entirely explained by levels of estradiol (Key et
al., 2003). For premenopausal breast cancer, the See Table 4.4.
risk of breast cancer was 40% higher among The role of sex hormones in CRC develop
women in the highest versus the lowest quintile of ment is unclear and is likely to be complex.
estradiol level. Levels of androstenedione, DHEA The Women’s Health Initiative Clinical Trial
sulfate, and testosterone were also significantly reported a significant reduction in CRC
positively associated with risk of breast cancer incidence among women assigned to the
in multivariate models that included established combined estrogen plus progestogen interven-
breast cancer risk factors (Key et al., 2013). tion arm (Chlebowski et al., 2004); however,
with additional follow-up of the trial partici-
(b) Cancer of the endometrium pants, it was later suggested that this effect may
See Table 4.3. be a consequence of diagnostic delay (Simon
Estrogens play a critical role in the normal et al., 2012). Experimental data also suggest
proliferation of endometrial tissue during the that estrogens may have protective effects
menstrual cycle (Barile et al., 1979; Klotz et al., on CRC development. ERβ has been demon-
strated to play an important role in the anti-
565
proliferative effects of estrogens on colonic tissue Sex steroids, and specifically androgens such
(Hartman et al., 2009). Furthermore, expression as testosterone, play critical roles in the devel-
of ERβ is low in human CRC cells (Waliszewski opment and function of the prostate gland, and
et al., 1997) and is inversely associated with the their involvement in prostate tumorigenesis has
stage of colon cancer (Castiglione et al., 2008), long been hypothesized (Hsing, 2001). However,
suggesting a possible role in disease progression. testosterone levels tend to be lower in obese men
However, investigations of the relationship than in men of normal weight.
between endogenous circulating estrogens and Prospective studies that have investigated
CRC have produced inconsistent results. Of the androgen levels and prostate cancer develop-
five prospective studies published to date, all of ment have reported inconsistent findings. The
which were mainly of postmenopausal women, Endogenous Hormones and Prostate Cancer
three reported null associations between circu- Collaborative Group, which pooled data from
lating estrogens and CRC risk (Clendenen et 18 prospective studies evaluating individual
al., 2009; Lin et al., 2013; Falk et al., 2015) and a data from 3886 incident prostate cancers and
fourth reported a borderline significant positive 6438 men without prostate cancer, reported
association (Gunter et al., 2008b). More recently, null associations between testosterone, DHEA
in a case–control study nested within the non-in- sulfate, androstenedione, or estradiol and inci-
tervention arms of the Women’s Health Initiative dent prostate cancer (Roddam et al., 2008). It has
Clinical Trial that included only postmeno- been hypothesized that a hypoandrogenic envi-
pausal women, higher endogenous levels of free ronment promotes the development of higher-
estradiol were inversely associated with CRC risk grade prostate tumours. At least two prospec-
(OR for highest vs lowest quartile, 0.43; 95% CI, tive studies have reported inverse relationships
0.27–0.69) in a multivariate model that included between serological testosterone levels and high-
established CRC risk factors as well as other obesi- grade prostate cancer (Platz et al., 2005a; Severi et
ty-related hormones such as insulin and IGF-1 al., 2006a). Furthermore, in the Prostate Cancer
(Murphy et al., 2015). Higher levels of SHBG were Prevention Trial, finasteride, which lowers testos-
positively associated with CRC development terone levels, reduced the risk of low-grade pros-
(OR for highest vs lowest quartile, 2.30; 95% CI, tate cancer by 25% but led to a higher incidence
1.51–3.51), and this relationship strengthened of high-grade disease (Thompson et al., 2003).
after statistical adjustment for levels of circulating Interestingly, the association between obesity
estradiol, estrone, insulin, IGF-1, and C-reactive and prostate cancer is stronger for high-grade
protein (CRP) (OR for highest vs lowest quartile, (fatal) tumours (see Section 2.2.14).
2.50; 95% CI, 1.59–3.92). Interestingly, the link Collectively, these data point to a complex
between obesity and CRC is weaker in women relationship between androgen levels and
than in men (see Section 2.2.1). Furthermore, in prostate cancer, with an indication of tumour
the study by Murphy et al. (2015) of postmeno- subtype specificity, but offer limited insight into
pausal women who were non-users of hormone the mechanisms underlying the link between
replacement therapy (HRT), the inclusion of obesity and prostate cancer.
estradiol in the waist circumference–CRC model
strengthened the risk estimate. (e) Cancer at other sites
Sex hormones have been hypothesized to play
(d) Cancer of the prostate a role in ovarian cancer development. Use of oral
See Table 4.5. contraceptives confers a reduced risk of ovarian
cancer, whereas use of postmenopausal HRT
566
is associated with increased risk (Beral et al., exercise (Foster-Schubert et al., 2012). Over a
2008, 2015). However, epidemiological studies 12-month period, women in the diet group and
that have investigated circulating estrogen and the diet plus exercise group lost on average 8.5%
SHBG levels in relation to risk of ovarian cancer and 10.8%, respectively, of their pre-intervention
were generally null (Table 4.6; Helzlsouer et al., weight (Campbell et al., 2012). Compared with
1995; Lukanova et al., 2003a; Rinaldi et al., 2007; the control group, women in these two groups
Trabert et al., 2016). [It is plausible that the asso- had statistically significant reductions in estrone
ciations may be specific to particular ovarian and estradiol levels. SHBG levels increased signif-
cancer subtypes, but to date individual studies icantly in the diet group and the diet plus exer-
have been of insufficient size to address this cise group, and decreased slightly in the control
hypothesis with precision.] group and the exercise group.
For other cancer types, there are intriguing Results from an analysis of overweight post-
data that point to possible sex hormone- menopausal women enrolled in the Diabetes
mediated mechanisms. In a case–control study, Prevention Program who underwent moderate
SHBG levels were strongly associated with weight loss did not reveal significant effects on
risk of hepatocellular carcinoma (HCC) even estradiol or testosterone levels, although DHEA
after adjusting for all established risk factors levels were reduced and there was a statistically
(Lukanova et al., 2014; Table 4.7). Also, there significant increase in SHBG concentrations
are distinct sex differences in the incidence of (Kim et al., 2012).
cancers of the oesophagus, liver, pancreas, and
kidney, all of which occur more frequently in (g) Synthesis
men than in women (see Sections 2.2.2, 2.2.4, Estrogen levels correlate with amount of body
2.2.7, and 2.2.16, respectively). However, to date fat in postmenopausal women. Overall, data
there are no published data on the association from observational and experimental studies
of endogenous sex hormones with these cancers. support clear associations between higher levels
Experimental data, mainly from studies of cell of estrogens and increased risk of breast cancer
lines, indicate possible anti-proliferative and and endometrial cancer. In addition, IWL affects
anti-tumorigenic effects of estrogen in renal cells sex steroid hormones and SHBG levels in post-
(Yu et al., 2013). menopausal women in a direction that would
favour reducing their risk of breast cancer and
(f) Impact of weight loss on sex hormones endometrial cancer. For CRC, estradiol may
Investigations of the effects of IWL on sex be anti-tumorigenic and may in fact lessen the
steroid levels are relatively consistent; however, impact of adiposity on CRC development. For
these studies have largely been restricted to post- cancers of the prostate and ovary, the data are
menopausal women. A comprehensive overview much less consistent and the associations are
of the available literature until 2011 concluded likely to be more complex. For other tumours,
that IWL reduces levels of sex steroid hormones the role of sex hormones in their development is
in postmenopausal women and increases SHBG largely unknown.
levels in premenopausal and postmenopausal
women (Byers & Sedjo, 2011). In the Nutrition 4.3.2 Insulin resistance
and Exercise for Women RCT, 439 overweight or
obese postmenopausal women were randomized Insulin resistance indicates the presence of an
to one of four groups: control, dietary interven- impaired physiological response to insulin, and
tion only, exercise intervention only, or diet plus is manifested by decreased insulin-stimulated
567
glucose transport. Hyperinsulinaemia, which breast cancer risk in this population (Hvidtfeldt
is a consequence of insulin resistance, is more et al., 2012).
common in obese individuals than in those
of normal weight, and metabolic indicators of (b) Cancer of the endometrium
hyperinsulinaemia, such as C-peptide levels, are See Table 4.3.
positively associated with BMI and waist circum- Hyperinsulinaemia, whether assessed by
ference (Bezemer et al., 2005). fasting insulin levels or C-peptide levels, has
Insulin, in addition to its metabolic effects, been associated with increased incidence of
has mitogenic and anti-apoptotic activity and endometrial cancer in several prospective inves-
appears to play a significant role in normal tigations (Lukanova et al., 2004b; Cust et al.,
organogenesis. Insulin has been shown to 2007a; Gunter et al., 2008a). In an analysis in
stimulate cell proliferation in normal tissues the Women’s Health Initiative cohort, baseline
such as breast tissue and in human cancer cell fasting insulin levels among women who were
lines (Ish-Shalom et al., 1997; Chappell et al., non-users of HRT were positively associated
2001), and administration of exogenous insulin with risk of endometrioid adenocarcinoma
promotes tumour growth in animal models after adjusting for estradiol levels and other
(Heuson & Legros, 1972; Shafie & Grantham, factors (HRq4−q1, 2.33; 95% CI, 1.13–4.82), and
1981; Shafie & Hilf, 1981). this association was stronger among women
with BMI ≥ 25 kg/m2 (HRq4−q1, 4.30; 95% CI,
(a) Cancer of the breast 1.62–11.43) (Gunter et al., 2008a). Two additional
See Table 4.2. studies that measured C-peptide concentrations
A number of epidemiological studies have also reported significant positive associations
investigated the association of fasting insulin with endometrial cancer risk (Lukanova et al.,
levels in women with higher BMI with incidence 2004b; Cust et al., 2007a). An analysis from
of breast cancer, with variable results. One study the European Prospective Investigation into
found a positive association between hyperin- Cancer and Nutrition (EPIC) study reported
sulinaemia and postmenopausal breast cancer an increased risk of endometrial cancer among
among women with BMI > 26 kg/m2, but not women with high C-peptide levels compared
among women with BMI ≤ 26 kg/m2 (Muti et al., with those with low levels; this association was
2002). In an analysis conducted in the Women’s independent of obesity, but the risk estimate was
Health Initiative Observational Study, fasting attenuated after adjustment for estradiol (Cust et
insulin levels were positively associated with al., 2007a). Recently, further support for a causal
postmenopausal breast cancer among women role of insulin in endometrial cancer development
who were non-users of HRT, in a multivariate came from a Mendelian randomization analysis
model that controlled for multiple breast cancer conducted in 1287 endometrial cancer cases and
risk factors, including estradiol and BMI (hazard 8273 controls, which identified a robust posi-
ratio for highest vs lowest quartile of insulin level tive association between genetically determined
[HRq4−q1], 2.40; 95% CI, 1.30–4.41; Ptrend < 0.001) insulin levels and endometrial cancer (Nead et
(Gunter et al., 2009, 2015a). In a subsequent al., 2015).
formal mediation analysis, it was demonstrated
(c) Cancer of the colorectum
that insulin, rather than estradiol, explained the
majority of the association between obesity and See Table 4.4.
In laboratory models, high insulin levels
have been shown to promote the development of
568
aberrant crypt foci in the colon (which are posited were at equivalent risk of CRC as overweight
to be CRC precursors), as well as the growth of and obese individuals with higher C-peptide
colon cancer cells (Koohestani et al., 1997; Tran levels. In contrast, overweight or obese partici-
et al., 2006). Furthermore, overexpression of the pants without raised C-peptide levels were not
insulin receptor can induce cell transforma- at increased risk of CRC. These findings support
tion in vitro (Giorgino et al., 1991), and human an association of hyperinsulinaemia with CRC
colorectal adenocarcinomas have been shown to independent of obesity status (Murphy et al.,
express the insulin receptor at high levels, indi- 2016).
cating that these cells may be sensitive to the
growth effects of insulin (Kiunga et al., 2004). (d) Cancer at other sites
Epidemiological data on the association of A number of prospective studies have inves-
hyperinsulinaemia with CRC are somewhat tigated the association of insulin with prostate
inconsistent. Of the five published studies to cancer development (Table 4.5), and the majority
date that directly measured fasting insulin reported null associations (Stattin et al., 2000;
levels (Schoen et al., 1999; Palmqvist et al., 2003; Hubbard et al., 2004; Stocks et al., 2007; Parekh
Saydah et al., 2003; Limburg et al., 2006; Gunter et al., 2013; Lai et al., 2014). A study nested within
et al., 2008b), three reported positive associations the Alpha-Tocopherol, Beta-Carotene Cancer
between hyperinsulinaemia and CRC (Schoen Prevention Study reported a 2-fold higher risk of
et al., 1999; Limburg et al., 2006; Gunter et al., prostate cancer when men in the highest quartile
2008b), but the associations were attenuated after of insulin level were compared with those in the
adjustment for other risk factors. In the largest lowest quartile (Albanes et al., 2009).
of such studies, which was conducted in the A null association was also reported in
Women’s Health Initiative cohort, insulin levels the single study of ovarian cancer (Table 4.6;
were significantly associated with CRC (HRq4−q1, Lukanova et al., 2003b).
1.89; 95% CI, 1.33–2.69; Ptrend = 0.0005); however, Hyperinsulinaemia has also been linked to
adjustment for waist circumference weakened development of liver cancer in a small number of
the association (HRq4−q1, 1.42; 95% CI, 0.91–2.23; prospective studies (Table 4.7). An investigation
Ptrend = 0.11), just as adjustment for insulin also nested within the EPIC cohort demonstrated
attenuated the relationship between obesity and a more than 3-fold greater risk of HCC and an
CRC (Gunter et al., 2008b). The remaining two almost 10-fold greater risk of intrahepatic bile
studies found no association between insulin duct tumour among participants in the highest
and CRC [insulin was measured in non-fasting tertile of C-peptide level compared with the lowest
blood specimens, which complicates the inter- tertile (Aleksandrova et al., 2014). Similarly, in a
pretation] (Palmqvist et al., 2003; Saydah et al., study of men chronically infected with hepatitis
2003). Other prospective studies have assessed B virus, individuals with fasting insulin levels
C-peptide concentrations in relation to CRC higher than 6.1 µU/mL were at more than 2-fold
and have generally reported positive associa- higher risk of HCC compared with those with
tions (Kaaks et al., 2000; Ma et al., 2004; Wei et insulin levels in the range 2.75–4.10 µU/mL
al., 2005a; Jenab et al., 2007; Otani et al., 2007). (Chao et al., 2011).
Most recently, an analysis in the EPIC study For pancreatic cancer (Table 4.8), two studies
demonstrated that individuals with a normal that measured insulin levels in pre-diagnostic
BMI but elevated C-peptide levels were at higher samples both reported statistically significant
risk of CRC compared with those with a normal positive associations between insulin levels and
BMI and without elevated C-peptide levels, and risk of pancreatic cancer (Stolzenberg-Solomon
569
et al., 2005; Wolpin et al., 2013), whereas an cancer (Table 4.2), CRC (Table 4.4), and pros-
investigation nested within the EPIC cohort tate cancer (Table 4.5). However, the evidence
reported no association between C-peptide levels that the IGF axis is dysregulated in obesity and
and pancreatic cancer (Grote et al., 2011). is modified by IWL, although convincing in
A single nested case–control study reported studies in animals, is less convincing in humans,
statistically significant associations between both at least in part due to current challenges in meas-
insulin and C-peptide levels (highest vs lowest uring bioavailable IGF-1 in human biospeci-
tertiles) and risk of stomach cancer (Table 4.9; mens. Although insulin levels generally rise with
Hidaka et al., 2015). increasing BMI and waist circumference, most
large, population-based studies have reported
(e) Synthesis a non-linear relationship between measures of
Hyperinsulinaemia and insulin resistance adiposity and IGF-1 levels. One study found the
are metabolic disturbances commonly observed highest IGF-1 levels among those with a BMI in
in obesity. Insulin, in addition to indirectly the range 26–27.9 kg/m2 (Allen et al., 2003), and
raising free estrogen levels by suppression of other studies suggest decreasing levels of IGF-1
SHBG expression, can directly activate cellular as BMI rises above 25 kg/m2 (Lukanova et al.,
pathways that confer growth and survival advan- 2004c). The non-linearity hypothesis relating
tages to the cell and therefore may promote circulating IGF-1 levels to adiposity is also
cancer development. Experimental data in in supported by findings among women enrolled
vitro and animal models generally support a pro- in the EPIC study, which reported a positive
tumorigenic effect of insulin; studies in humans trend in IGF-1 levels as BMI and waist circum-
generally only support a positive association ference increased, with levels peaking at a BMI
between hyperinsulinaemia and cancers of the of 24.6–26.6 kg/m2, and then declining among
endometrium and colorectum, whereas findings participants with BMI > 26.6 kg/m2 (Gram et
for breast cancer and prostate cancer are more al., 2006). In contrast, linear regression analysis
heterogeneous. There are few data for cancer at of data from the United States National Health
other sites. and Nutrition Examination Survey (NHANES)
suggested an overall inverse relationship between
4.3.3 Insulin-like growth factors circulating total IGF-1 levels and BMI (Faupel-
Badger et al., 2009). This phenomenon may, in
The IGF system comprises two ligands, IGF-1 part, be explained by obesity-induced hyperin-
and IGF-2, as well as at least six binding proteins sulinaemia and growth hormone effects. Insulin
(IGFBPs) that sequester IGF-1 and IGF-2 and inhibits the synthesis of IGFBP-1 and IGFBP-2,
regulate their bioavailability and activity. IGF-1 leading to an increase in unbound IGF-1 (Nam et
and IGF-2 are growth factors that share signifi- al., 1997). Thus, as adiposity increases over time,
cant structural similarities with insulin but have IGF-1 levels rise, but with the development of
much stronger mitogenic and anti-apoptotic obesity, elevated free IGF-1 levels exert a negative
effects. feedback effect on pituitary secretion of growth
A substantial body of epidemiological liter- hormone, with subsequent attenuation of hepatic
ature has now accumulated on the association IGF-1 synthesis (Tannenbaum et al., 1983).
of circulating IGF-1 levels with cancer devel- A comprehensive review of the literature
opment, and several meta-analyses and pooled on IGF-1 and IGFBPs presents an inconsistent
studies have demonstrated robust associations portrait of how IWL affects IGF-1 and IGFBPs
of systemic IGF-1 concentrations with breast (Byers & Sedjo, 2011). In the Nutrition and
570
Exercise for Women RCT, published afterwards discussed here: (i) altered production of inflam-
(Mason et al., 2013b), weight loss was positively matory factors secreted from adipose and other
associated with change in circulating IGF-1 and tissues, (ii) increased adipose tissue inflamma-
in the molar ratio of IGF-1 to IGFBP-3 in the diet tion (as measured by crown-like structures and
group. other measures of infiltration by immune cells),
The IGF axis plays a major role in the regu- and (iii) adipose tissue remodelling. In addi-
lation of cell growth and survival, and increased tion, several emerging contributors to the obesi-
signalling through the IGF system can exert ty-associated pro-inflammatory state, including
a pro-tumorigenic effect. Studies in humans the cyclooxygenase-2 (COX-2)/prostaglandin
support a role for systemic IGF-1 levels in deter- pathway, the nuclear factor kappa-light-chain-
mining risk of breast, prostate, and colorectal enhancer of activated B cells (NF-κB) pathway,
cancer, whereas for other malignancies the rela- and inflammation-related molecules from the
tionship is much less clear. However, the rela- microbiome, are briefly discussed.
tionship between IGF-1/IGFBPs and obesity is
uncertain and is still being investigated. (a) Established markers and mechanisms
(i) Changes in inflammatory markers
4.3.4 Chronic inflammation Leptin, a peptide hormone produced by
Chronic inflammation, a key characteristic adipocytes (and thus referred to as an adipokine),
of carcinogenesis (Hanahan & Weinberg, 2000, is positively correlated with adipose storage and
2011; Smith et al., 2016), has been associated nutritional status, and functions as an energy
with obesity in a large number of epidemiolog- sensor. Leptin interacts directly with periph-
ical and experimental studies. Obesity is consid- eral tissues, interacts indirectly with hypo-
ered a chronic pro-inflammatory state associated thalamic pathways, and modulates immune
with progressive infiltration of adipose tissue by function, cytokine production, angiogenesis,
macrophages and other immune cells that secrete and many other biological processes (Münzberg
pro-inflammatory cytokines (including TNF-α, & Morrison, 2015). At the high levels found with
IL-1β, and IL-6) and other chemical mediators of obesity, leptin also has pro-inflammatory activity
a persistent, subacute (often referred to as smoul- and upregulates the secretion of TNF-α and IL-6
dering) inflammatory response (Renehan et al., (Park & Ahima, 2015).
2008). In addition, several clinical and experi- The leptin receptor is structurally and func-
mental studies suggest that IWL – by behavioural tionally similar to class I cytokine receptors and
interventions, bariatric surgery, or pharmacolog- acts through the signal transducer and activator
ical approaches – can reverse some of the obesi- of transcription (STAT) family of transcrip-
ty-associated changes in certain inflammatory tion factors (Villanueva & Myers, 2008). STATs
factors, particularly CRP. However, it is also clear induce signalling pathways for several cellular
from this literature that the underlying causes, processes, including cell growth, proliferation,
cellular contributors, and molecular and meta- survival, migration, and differentiation. The
bolic factors involved in the obesity–inflamma- activity of STATs is commonly dysregulated in
tion–cancer triad are extremely complex. cancer (Yu et al., 2014).
The increase in white adipose tissue mass Adiponectin is the most abundant hormone
associated with obesity drives chronic inflam- secreted from adipose tissue. In contrast with
mation through at least three established leptin, levels of adiponectin correlate negatively
and interacting mechanisms, which are each with adiposity. Adiponectin can reduce pro-
571
inflammatory cytokine expression and induce Activation of NF-κB by cytokines or Akt can lead
anti-inflammatory cytokine expression via inhi- to the translocation of the active NF-κB subunit,
bition of NF-κB (discussed below) (Fantuzzi, p65, from the cytoplasm to the nucleus (Adli &
2013; Park & Ahima, 2015). Baldwin, 2006), inducing multiple genes associ-
CRP is a non-glycosylated circulating acute- ated with inflammation and cancer, including
phase reactant protein of the pentraxin family IL-6, COX-2, and IL-1β (Karin, 2006).
(Thiele et al., 2015). CRP has long been used as (ii) Increased adipose tissue inflammation
a marker of inflammation in studies in humans
(see Section 4.3.4c), and data are accumulating A new role of adipose tissue inflammation
that CRP (particularly the monomeric form that in obesity and its connection to cancer has been
results from dissociation from the pentameric proposed. Subclinical inflammation in visceral
form on activated macrophages and platelets) and subcutaneous white adipose tissue is char-
may be a mediator of inflammation (Thiele et al., acterized by rings of activated macrophages
2015). surrounding engorged or necrotic adipocytes and
In addition to adiponectin, leptin, and referred to as crown-like structures. Macrophages,
CRP, many other adipokines, cytokines, and T cells, and other immune cells infiltrate adipose
acute-phase reactant proteins can be produced tissue at the onset of weight gain. This adipocyte–
by adipocytes, by other cells in adipose tissue macrophage interaction results in the production
(e.g. macrophages, dendritic cells, fibroblasts, of a pro-inflammatory secretome from both cell
B and T lymphocytes), or by other tissues (e.g. types, which activates the cellular transcription
stomach, skeletal muscle, liver) (Blüher & factor NF-κB, increases levels of cytokines and
Mantzoros, 2015). With increased adiposity, the other inflammatory factors, and triggers inflam-
secretome (the conglomerate of secreted factors) mation. Chronic inflammation eventually leads
can become dysregulated and have significant to systemic insulin resistance and altered levels
biological impacts on insulin sensitivity, inflam- of circulating adipokines, cytokines, and other
matory response, vascular endothelial function, factors that promote the development of obesity,
estrogen metabolism, and cell proliferation. and also plays a role in obesity-associated cancers
At the intracellular level, inflammatory (Wellen & Hotamisligil, 2003; Neels & Olefsky,
signals are transduced through multiple pathways 2006; Subbaramaiah et al., 2011).
to drive cellular responses. For example, NF-κB (iii) Adipose tissue remodelling
is a transcription factor activated in response to Stored triacylglycerides undergo lipol-
various stimuli, including cytokines and other ysis within the cytoplasm of adipocytes and
inflammatory molecules, and is responsible for are released into the bloodstream as free fatty
inducing gene expression associated with cell acids during times of low substrate availability
proliferation, apoptosis, angiogenesis, cytokine or heightened energy requirements (Duncan
secretion, and other responses to inflammatory et al., 2007). Once in the circulation, free fatty
signals (Xia et al., 2014). Activation of NF-κB acids can be used to generate energy. In a state
has been observed in many types of tumour of obesity, white adipose tissue does not respond
cells (Karin, 2006). There is considerable cross- appropriately to changes in energy requirements,
talk between growth factor signalling pathways resulting in elevated production of adipokines
and NF-κB signalling, and obesity and energy and cytokines (Jung & Choi, 2014).
restriction modulate NF-κB activation, possibly When lipid storage capacity in adipose tissue
through alterations in systemic cytokines, growth is exceeded, surplus lipids often accumulate
factors, and Akt signalling (Hursting et al., 2013).
572
within muscle, liver, and pancreatic tissue, (c) Epidemiological evidence for the mediation
leading to impairment of lipid processing and of inflammatory factors between obesity
clearance within these tissues (Henry et al., 2012; and cancer
Suganami et al., 2012). As a result, lipid interme- (i) Cancer of the breast
diates impair the function of cellular organelles
and cause further release of cytokines, which See Table 4.2.
foster inflammation as well as insulin resistance. Epidemiological studies on the association
of adipokines and inflammatory factors with
(b) Emerging markers and mechanisms breast cancer have generally yielded inconsistent
results. Adiponectin levels have been reported
(i) COX-2, prostaglandins, and other lipid
to be inversely associated with breast cancer
mediators
incidence in several prospective investigations
COX-2 can be highly induced in several tissue (Tworoger et al., 2007b; Gross et al., 2013) but not
types as part of the inflammatory response; in others (Cust et al., 2009; Gaudet et al., 2013);
COX-2 levels are increased in many obesity- three recent meta-analyses that included both
associated cancers, including breast, ovarian, prospective cohort and case–control studies
and colorectal tumours, and are associated with reported an overall inverse relationship between
a poor clinical outcome (Eberhart et al., 1994; adiponectin levels and breast cancer risk (Liu et
Howe, 2007; Lee et al., 2013). al., 2013; Macis et al., 2014; Ye et al., 2014). Most
In addition, the increased lipolysis that recently, data from the Women’s Health Initiative
occurs with obesity results in a higher concentra- demonstrated an inverse association between
tion of circulating free fatty acids (Björntorp et adiponectin levels and postmenopausal breast
al., 1969; Jensen et al., 1989; Nicklas et al., 1996), cancer, but this relationship was attenuated after
and saturated fatty acids can stimulate expres- adjustment for insulin (Gunter et al., 2015b).
sion of COX-2 and secretion of prostaglandin E2 Data on the association of other adipokines,
in cultured macrophages via activation of Toll- such as leptin, plasminogen activator inhibitor 1
like receptor 4 and subsequent NF-κB signalling (PAI-1), and resistin, with breast cancer risk are
(Lee et al., 2001; Hellmann et al., 2013). This also mixed (Gaudet et al., 2013; Gross et al., 2013;
may be another contributor to obesity-associ- Gunter et al., 2015b).
ated adipose tissue inflammation. Also, serum CRP, a sensitive but nonspecific marker of the
concentrations of IL-6 and TNF-α are generally inflammatory response, has been investigated in
increased with obesity (Fain, 2006), and these relation to breast cancer risk in a large number of
cytokines have been shown to stimulate COX-2 prospective studies. A recent meta-analysis that
expression and to promote production of prosta- summarized data from 12 prospective studies
glandin E2 (Geng et al., 1995; Maihöfner et al., concluded that moderately elevated CRP levels
2003). were associated with higher risk of breast cancer
(ii) Inflammatory contributions from the such that for every doubling in CRP concen-
microbiome tration, the risk of breast cancer increased by
An emerging field of research is the influ- 7% (Chan et al., 2015). An additional study not
ence of the microbiome – the community of included in the meta-analysis also reported that
commensal, symbiotic, and pathogenic micro higher circulating CRP levels were associated
organisms that inhabit an individual – on obesity, with increased incidence of breast cancer (Gunter
inflammation, and related chronic diseases et al., 2015b). Specifically, the breast cancer inci-
(discussed in Section 4.3.6a). dence in women in the upper two quartiles of
573
CRP levels was twice that of those in the lowest the risk estimates were no longer significant
quartile of CRP levels among women who were after adjusting for BMI (Dossus et al., 2010).
non-users of HRT, even after controlling for However, a subsequent study in the same popul-
estradiol, insulin, BMI, and established breast ation reported significant positive associations
cancer risk factors. Furthermore, in that analysis between circulating TNF-α levels and endome-
CRP appeared to be a significant mediator of the trial cancer, even after controlling for BMI and
relationship between BMI and breast cancer, other endometrial cancer risk factors (Dossus et
along with insulin and estradiol. al., 2011), and a factor analysis of all metabolic
and inflammatory markers revealed a distinct
(ii) Cancer of the endometrium
inflammatory pattern of markers that was
See Table 4.3. predictive of endometrial cancer development
A number of prospective studies have investi- (Dossus et al., 2013).
gated the association of circulating inflammatory
factors and adipokines with endometrial cancer. (iii) Cancer of the colorectum
Within the Prostate, Lung, Colorectal, and Ovar See Table 4.4.
ian Cancer Screening Trial, leptin levels were Several prospective studies have investigated
positively associated with endometrial cancer the association of adipokines and inflammatory
development (ORt3−t1, 2.77; 95% CI, 1.60–4.79), markers with CRC risk. In general, most studies
whereas adiponectin levels were inversely related have reported an inverse association between
to risk (ORt3−t1, 0.48; 95% CI, 0.29–0.80); both adiponectin levels and CRC (Wei et al., 2005b;
associations strengthened when analyses were Aleksandrova et al., 2012a; Song et al., 2013). In
restricted to women who were non-users of HRT a nested case–control study of CRC in Norway,
(Luhn et al., 2013). Leptin levels were also posi- leptin levels were positively associated with colon
tively associated with endometrial cancer risk cancer (ORq4–q1, 2.72; 95% CI, 1.44–5.12) but not
in the B~FIT study, although this relationship with rectal cancer (Stattin et al., 2004b). In the
was no longer significant after adjustment for Women’s Health Initiative Observational Study,
BMI (Dallal et al., 2013). Within the EPIC study, a panel of pro-inflammatory adipokines, namely
adiponectin concentrations were significantly leptin, IL-6, and PAI-1, were associated with
inversely associated with endometrial cancer higher incidence of CRC, whereas adiponectin
risk, even after controlling for BMI (ORq4−q1, 0.56; levels were inversely related to CRC risk (Ho et
95% CI. 0.36–0.86) (Cust et al., 2007a); however, al., 2012). However, the associations were attenu-
other studies did not report significant inverse ated, and only leptin remained significant, after
associations between adiponectin and endome- adjusting for insulin, suggesting that their effects
trial cancer after adjustment for BMI (Soliman on CRC risk may be attributed partly to insulin.
et al., 2011; Dallal et al., 2013). A follow-up study conducted in the same popul-
Data from the Women’s Health Initiative ation reported that higher levels of the soluble
indicated a significant positive association IL-1 receptor were associated with significantly
between CRP levels and endometrial cancer; lower risk of CRC, suggesting that regulators of
however, this relationship was attenuated and cytokine signalling and availability may modify
lost statistical significance after adjustment for CRC development (Ho et al., 2014).
insulin and estradiol levels (Wang et al., 2011). A substantial number of studies have inves-
A case–control study nested within the EPIC tigated the link between circulating CRP levels
cohort found levels of CRP and IL-6 to be posi- and CRC, and the majority have reported posi-
tively associated with endometrial cancer, but tive associations. In a recent meta-analysis that
574
captured data from more than 4500 CRC cases, risk (Table 4.5); one study reported a significant
risk of CRC was increased by 12% for every unit inverse relationship (Stocks et al., 2007), whereas
change in the natural logarithm of CRP concen- other investigations generally reported null asso-
tration (Zhou et al., 2014). ciations for both leptin and adiponectin (Hsing
et al., 2001; Li et al., 2010; Lai et al., 2014).
(iv) Cancer of the ovary
Combined data from five cohorts in the USA
See Table 4.6. yielded a significant inverse association between
A potential role for inflammation in ovarian adiponectin levels and pancreatic cancer
cancer development was hypothesized several (ORq5–q1, 0.63; 95% CI, 0.43–0.92) (Bao et al.,
decades ago, but until recently, data from 2013b). In the EPIC cohort, adiponectin was
prospective cohort studies were sparse. In the inversely related to pancreatic cancer risk, but
EPIC cohort, CRP levels above 10 mg/L were only among never-smokers (Grote et al., 2012b).
indicative of higher risk of epithelial ovarian Interestingly, data from a Japanese cohort
cancer compared with levels of 1 mg/L or below and from the EPIC study reported a consistent
(OR, 1.67; 95% CI, 1.03–2.70), and this relation- positive association between circulating IL-6
ship was more pronounced among overweight levels and HCC development (Aleksandrova et
and obese women (Ose et al., 2015). Similar al., 2014; Ohishi et al., 2014), with relative risks
findings were reported from the Prostate, Lung, between 3 and 5. Levels of other cytokines and
Colorectal, and Ovarian Cancer Screening Trial, CRP were not associated with HCC in these
the Nurses’ Health Study, and a combined studies.
analysis in the New York University Women’s
Health Study, the Northern Sweden Health and (vi) Weight loss
Disease Study, and ORDET Cohort, which all Although the obesity–inflammation–cancer
reported statistically significant positive associ- link is established for some cancers, namely CRC,
ations between CRP levels and ovarian cancer the impact of IWL on inflammation and cancer
risk when comparing levels above 10 mg/L with risk has been much less studied. However, a
levels of 1 mg/L or below (Lundin et al., 2009; consistent picture is emerging that IWL, particu-
Poole et al., 2013; Trabert et al., 2014). larly if more than 10%, can reverse some of the
The Prostate, Lung, Colorectal, and Ovarian pro-inflammatory effects of obesity.
Cancer Screening Trial also identified significant Most of the studies in humans that addressed
associations between specific pro-inflamma- weight loss and inflammation assessed systemic
tory cytokines, namely TNF-α and IL-1α, and markers of inflammation, including CRP (the
ovarian cancer risk (Trabert et al., 2014), and a most consistently changed inflammatory marker
case–control study nested within three prospec- with weight loss), IL-6, and TNF-α. One review
tive cohorts also found a significant link between included about 30 observational cohort studies
specific cytokines related to a pro-inflammatory or RCTs that used various modes of weight loss,
phenotype and ovarian cancer (Clendenen et al., including surgical, dietary, physical activity, or
2011). pharmacological interventions, and encompassed
a large range of weight-loss attainment (Byers &
(v) Cancer at other sites
Sedjo, 2011). Consistent decreases in circulating
For cancer at other sites, studies of circulating levels of CRP, TNF-α, and IL-6 were observed
inflammatory markers and adipokines are more in studies that reported weight loss of more
sparse. Overall, there are inconsistent data for the than 10%; the findings are less clear with more
association of leptin levels with prostate cancer moderate (and more achievable and sustainable)
575
levels of weight loss achieved through diet and/or (vii) Synthesis

exercise interventions. Obesity is associated with a state of chronic,
In the Nutrition and Exercise for Women low-grade inflammation triggered by adipose
RCT, high-sensitivity CRP, serum amyloid A tissue remodelling and reflected in several local
(another acute-phase reactant protein that can and systemic changes in the levels of adipokines
stimulate CRP), IL-6, and neutrophil counts (e.g. increased leptin; decreased adiponectin
decreased statistically significantly in both the released from hypertrophied adipocytes),
diet group and the diet plus exercise group. The cytokines (e.g. increased IL-6, TNF-α, secreted
intervention reduced high-sensitivity CRP by from adipocytes and macrophages), and acute-
36% in the diet group and by 42% in the diet phase reactant proteins (e.g. CRP, secreted
plus exercise group. Women who lost more body primarily from the liver). These secretome
weight (> 5%) experienced the largest reduc- changes are accompanied by changes in tissue
tions (e.g. 50% and 49% for high-sensitivity CRP markers of inflammation (such as crown-like
with diet and diet plus exercise, respectively) structures) and cancer-associated intracellular
(Imayama et al., 2012). In another study (Fabian signals (such as activation of the NF-κB pathway).
et al., 2013), obese women who lost more than Emerging contributors to the obesity–
10% of their body weight had reductions of about inflammation–cancer relationship include
30–50% in the leptin-to-adiponectin ratio, IL-6, (i) the COX-2/prostaglandin pathway, which
and CRP levels in their serum, whereas women can be particularly activated in adipocytes and
who lost less than 10% of their body weight had macrophages in response to an obesity-associ-
little or no change in systemic biomarkers. [There ated abundance of free fatty acids (catalysed to
appears to be a threshold below which inflamma- inflammatory prostaglandins and other lipid
tory markers fail to respond to weight loss; this intermediates); (ii) the microbiome, particularly
may be due partly to sensitivity in the analytical the community of organisms residing in the gut
methods and partly to limitations of statistical that can release short-chain fatty acids and other
power, as well as the heterogeneity in marker metabolites that have pro-inflammatory activity.
levels before weight loss is initiated.] Data from epidemiological studies of inflam-
Few studies investigated weight loss and matory markers and cancer development have
changes in tissue-specific biomarkers of inflam- generally shown consistent positive associa-
mation. Campbell et al. (2013) reported that in tions between circulating levels of CRP, a highly
overweight or obese postmenopausal women, sensitive but nonspecific marker of the inflam-
weight loss through a 6-month dietary inter- matory response, and cancers of the breast and
vention, exercise, or the combination of the colorectum, and suggestive positive associations
two resulted in significant changes in adipose for cancers of the ovary and endometrium.
tissue gene expression; in addition to significant However, the specific inflammatory pathways
reductions in leptin mRNA expression, steroid that mediate this relationship, and the extent
hormone metabolism, inflammatory genes, and to which it might be specific to certain ovarian
IGF signalling also appeared to be altered. In cancer subtypes, remain unknown. In weight-
overweight or obese postmenopausal women who loss intervention trials, CRP concentrations were
underwent a weight-loss intervention (Fabian generally reduced in the individuals who lost
et al., 2013), the adiponectin-to-leptin ratio in more than 5–10% of their initial body weight.
fine-needle aspirate of breast tissue increased in Data on the associations of specific circulating
response to weight loss. cytokines and adipokines with obesity-related
576
cancers are generally more limited, possibly activity, which tend to be poorly measured)
because of technical challenges in measuring cannot be excluded.] Further background infor-
these proteins, which typically circulate at very mation about vitamin D and cancer can be found
low concentrations. A number of studies have in the IARC Working Group Report on vitamin
investigated these markers in relation to endo- D and cancer (IARC, 2008).
metrial cancer, with inconsistent results.
(b) Vitamin D and obesity
4.3.5 Vitamin D Increasing BMI has been consistently associ-
ated with lower serum 25(OH)D concentrations
(a) Vitamin D and cancer and higher parathyroid hormone concentra-
Vitamin D can induce cell differentiation tions (Vanlint, 2013; Pereira-Santos et al., 2015).
and apoptosis and can also inhibit proliferation, Moreover, body fat content is inversely associ-
inflammation, and angiogenesis, as well as inva- ated with serum 25(OH)D concentrations, and
sion and metastasis (Fleet et al., 2012; Feldman this association may be stronger than that with
et al., 2014; Castronovo et al., 2015; Davis-Yadley BMI and body weight alone (Arunabh et al.,
& Malafa, 2015; Christakos et al., 2016; Meeker 2003; Vanlint, 2013).
et al., 2016) and thus may have cancer-preventive In a recent meta-analysis, the prevalence of
effects. Despite these strong experimental data, vitamin D deficiency was 35% higher in obese
the epidemiological data on Vitamin D levels and individuals compared with a normal-weight
cancer risk have been limited and heterogeneous. group, and 24% higher compared with over-
Cohort studies that measure the biomarker weight individuals. There were no significant
25-hydroxyvitamin D (25(OH)D) pre-diagnosti- differences in this proportion between children
cally have shown a consistent reduction of CRC and adults; however, there was a significant
risk in the range of 30–40% among individuals degree of heterogeneity between studies overall
with high versus low levels (Feldman et al., 2014). (Pereira-Santos et al., 2015). [A challenge of this
However, a large RCT of vitamin D supplemen- meta-analysis was the change in the definitions
tation showed no effects in preventing the recur- of vitamin D deficiency over time, although the
rence of colorectal adenoma (Baron et al., 2015). results appeared to be consistent independent of
Similarly, a meta-analysis of prospective studies the cut-off points used.]
of prostate cancer showed no inverse association There are multiple potential reasons for
(Gilbert et al., 2011), and one study even suggested the inverse associations between obesity and
an increased risk with higher vitamin D levels vitamin D (Soares et al., 2012; Pereira-Santos et
(Brändstedt et al., 2012). For breast cancer, only al., 2015). One theory is that because of issues of
some studies observed inverse associations, and low social acceptance, obese individuals reduce
these were not linear and were limited to post- their exposure to sunlight, cover up their bodies
menopausal women (Chlebowski et al., 2008; more, and are less active outdoors. Nevertheless,
Bauer et al., 2013). [This pattern of cancer risk – in the Framingham Heart Study cohort, adjust-
preventive for colon cancer and, to some extent, ment for outdoor physical activity did not entirely
postmenopausal breast cancer – mimics the attenuate this association (Cheng et al., 2010).
associations of physical activity with cancer risk It has also been proposed that vitamin D
(IARC, 2002), and 25(OH)D is strongly associ- metabolites are retained by excess body fat, and
ated with physical activity. Therefore, a direct that cholecalciferol that is synthesized in the skin
interrelation or residual confounding by physical or taken up through the diet is in part seques-
activity (particularly levels of outdoor physical tered by the body fat before transport to the
577
liver for hydroxylation (Wortsman et al., 2000). women. At baseline, 25(OH)D levels were lower
Adipocytes of obese individuals show significant in the obese women than in 25 normal-weight
levels of 1-α-hydroxylase, which activates vitamin controls (P < 0.001). The 20-week low-calorie
D; this could explain the greater local use of diet (26 completers) resulted in reductions of
vitamin D. This hypothesis is consistent with the body weight and BMI by 10% and an increase in
observation that after exposure to sunlight, obese 25(OH)D (15.4 ± 6.0 ng/mL vs 18.3 ± 5.1 ng/mL,
individuals have shown a 53% lower increase in P < 0.05), compared with baseline. This increase
25(OH)D compared with non-obese individuals, was also associated with improvement in insulin
independent of the amount of the cutaneous resistance and the homeostasis model assess-
precursor of vitamin D (Pereira-Santos et al., ment index.
2015). Rock et al. (2012) prospectively examined the
Alternatively, some experimental data suggest effects of weight loss on serum 25(OH)D concen-
that vitamin D deficiency may facilitate adiposity trations in 383 overweight or obese women who
by causing higher parathyroid hormone levels participated in a 2-year clinical trial of a weight-
and greater influx of calcium into adipocytes, loss programme and recommendation to increase
thereby increasing lipogenesis (Pereira-Santos physical activity. More than half of the women
et al., 2015). There are several additional mech- lost at least 5% of baseline weight by 24 months,
anisms, investigated mainly in experiments in and serum 25(OH)D levels increased at the end
animals, that link vitamin D, through vitamin of the intervention period, with a linear trend
D receptor-mediated activity, directly to energy towards greater increases in women who lost
regulation and effects in adipocytes (Martini & more weight; 25(OH)D increased by 5.0 ng/mL
Wood, 2006). for those who lost more than 10% of baseline
Finally, there is increasing experimental weight (P = 0.014). [Although the programme
evidence that vitamin D may also have anti- included some increase in physical activity,
inflammatory properties, presumably via effects this was not a major component of the inter-
on the state of low-grade chronic inflamma- vention, and the resulting greater sun exposure
tion in adipose tissues (Fleet et al., 2012; Song during outdoor activity is unlikely to explain the
& Sergeev, 2012; Feldman et al., 2014). In an observed effect.]
RCT of 218 postmenopausal women with BMI In 192 obese patients with knee osteoarthritis,
≥ 25 kg/m2 who underwent 12 months of weight- Christensen et al. (2012) tested an 8-week formula
loss intervention plus either 2000 IU/day of weight-loss diet of 415–810 kcal/day, followed by
oral vitamin D3 or daily placebo, significantly 8 weeks on a hypo-energetic 1200 kcal/day diet
decreased circulating levels of IL-6 were reported combining normal food and formula products.
with vitamin D in an analysis stratified by weight They reported that this intensive programme
loss (P = 0.004) (Duggan et al., 2015). increased bone mineral density and improved
25(OH)D concentrations. [It is not clear whether
(c) Vitamin D and weight loss this increase in 25(OH)D was attributable to the
Several studies have demonstrated effects of effects of the calorie restriction or the supplemen-
weight loss on improving vitamin D biomarker tation with vitamin D as part of the formula.]
status in obese individuals. Several studies have also investigated the
Tzotzas et al. (2010) investigated changes of effects of bariatric surgery on vitamin D status,
25(OH)D at 4 weeks and 20 weeks after intro- and suggest decreases in vitamin D status with
duction of a weight-loss programme (low- surgery (Karefylakis et al., 2014; Costa et al.,
calorie diet of ~1000 kcal/day) among 44 obese 2015; Luger et al., 2015). [This type of weight-loss
578
intervention can alter the resorption of dietary lipopolysaccharide, a component of the outer
vitamin D, and therefore is not considered membrane of Gram-negative bacteria (Cani et
informative.] al., 2008); lipopolysaccharide induces meta-
bolic endotoxaemia, characterized by elevated
(d) Synthesis infiltration of adipose tissue by macrophages
Vitamin D status can directly affect many and elevated expression of pro-inflammatory
cellular processes relevant to cancer preven- cytokines (Cani et al., 2007), thus inducing
tion. Prospective studies of the blood biomarker chronic systemic and adipose tissue inflamma-
25(OH)D have found consistent inverse asso- tion. These effects were completely prevented
ciations with CRC, and to a lesser extent with by treatment with a broad-spectrum antibiotic
postmenopausal breast cancer. There is a clear (Cani et al., 2008). Given the known role that this
inverse relationship between obesity and vitamin type of inflammation plays in the progression of
D status, but the causes for this association are many cancer types (see Section 4.3.4), it is plau-
not well defined and may range from societal sible that obesity-induced perturbations of the
factors and links via physical activity to physio- gut microbiota are a contributing factor in the
logical changes in the adipose tissue that result in obesity–cancer link.
a sequestering of vitamin D metabolites; weight
loss appears to improve 25(OH)D status. The (b) Gut hormones
experimental data are limited and do not further The role of gut hormones and appetite regula-
inform the role of vitamin D as a mediator in the tory factors in cancer development is an emerging
effect of obesity on cancer risk. area of research, and may be a mechanism
linking obesity with cancer. Ghrelin, a hormone
4.3.6 Other factors produced in the gastric fundic glands, is known
to mediate appetite and fatty acid metabolism
This section summarizes factors that may and to promote fat storage (Higgins et al., 2007).
play a role in mediating the obesity–cancer Ghrelin can also inhibit the expression and/or
connection but for which there are limited data. production of pro-inflammatory cytokines, and
(a) The gut microbiome ghrelin treatment increases anti-inflammatory
cytokines (Gonzalez-Rey et al., 2006; Baatar et
Obesity is associated with an overall reduc- al., 2011) (see Section 1.3.1).
tion in bacterial diversity in the gut microbiota In the three small prospective studies of the
(Turnbaugh et al., 2009) (see Section 1.3.8), and association of ghrelin with gastrointestinal cancer
decreased bacterial richness has been linked to development, individuals in the lowest quartile
elevated systemic inflammation, measured by of serum ghrelin at baseline, compared with
plasma CRP levels and white blood cell counts those in the highest quartile, had an increased
(Le Chatelier et al., 2013). Furthermore, weight risk of oesophageal adenocarcinoma (31 cases;
loss does not significantly improve CRP levels in OR, 5.55; 95% CI, 1.28–25.0) (de Martel et al.,
obese individuals with low microbiome richness 2007), oesophageal squamous cell carcinoma (82
(Cotillard et al., 2013), suggesting that resistance cases; OR, 6.83; 95% CI, 1.46–31.84) (Murphy
to the inflammation-reducing effects of weight et al., 2012), gastric cardia cancer (98 cases; OR,
loss may be mediated by differences in micro- 4.90; 95% CI, 2.11–11.35), and gastric non-cardia
biome richness. Feeding mice a high-fat diet cancer (261 cases; OR, 5.63; 95% CI, 3.16–10.03)
is accompanied by impairments in gut barrier (Murphy et al., 2011). There is considerable
function, including increased plasma levels of cross-talk between ghrelin and other hormones
579
involved in energy metabolism, such as leptin, (d) Immune function

adiponectin, and insulin, and as more data The major mechanisms relating immunity to
become available on the association of ghrelin obesity focus on the inflammatory response that
with cancer development, the gut hormones may originates in the adipose tissue (see Section 4.3.4).
emerge as an important pathway linking obesity For a description of the innate immune response
with cancer development. to obesity, see Lumeng (2013).
(c) Non-alcoholic fatty liver and pancreatic Studies investigating immune competence
diseases in relation to calorie restriction or IWL are few.
One cross-sectional study in 114 overweight or
Obesity is the most common cause of non- obese postmenopausal women reported that
alcoholic fatty liver disease (NAFLD), a spec- natural killer cell cytotoxicity (assessed by flow
trum of diseases including variable degrees of cytometry at four effector-to-target ratios) was
simple steatosis, non-alcoholic steatohepatitis inversely associated with increasing frequency
(NASH), and cirrhosis (Papandreou & Andreou, of prior IWL (Ptrend = 0.003) (Shade et al., 2004).
2015). Simple steatosis is benign, whereas NASH Conversely, longer duration of recent weight
is characterized by hepatocyte injury, inflamma- stability was associated with significantly greater
tion, and/or fibrosis, which can lead to cirrhosis, natural killer cell cytotoxicity (21.6% ± 11.9%,
liver failure, and HCC (Hui et al., 2003). About 24.4% ± 11.0%, and 31.9% ± 14.4% for ≤ 2, > 2 to
80% of cases of cryptogenic cirrhosis [end-stage ≤ 5, and > 5 years of weight stability, respectively;
liver disease of unknown etiology] present with Ptrend = 0.0002).
NASH, and 0.5% of these cases will progress to In one RCT, 91 obese women were random
HCC, a percentage that increases significantly ized to control (n = 22), exercise (n = 21), diet
with hepatitis C-associated cirrhosis (White et (n = 26), or exercise plus diet (n = 22) groups.
al., 2012). The prevalence of NAFLD has increased After 12 weeks of calorie restriction (1200–
concomitantly with the increase in childhood 1300 kcal/day) with weight loss of about 9%,
obesity over the past 30 years (Berardis & Sokal, mitogen-stimulated lymphocyte proliferation
2014). was significantly reduced, whereas no changes
Adipocyte infiltration and fat accumulation were observed in natural killer cell cytotoxicity,
in the pancreas appear to be early events in obesi- monocyte and granulocyte phagocytosis and
ty-associated pancreatic endocrine dysfunction, oxidative burst activity, or the number of days
and can trigger pancreatic steatosis, non-alcoholic with upper respiratory tract infections (Nieman
fatty pancreatic disease (NAFPD), and subclin- et al., 1998).
ical pancreatitis. In addition, “fatty pancreas”
has been positively associated with visceral (e) Cancer stem cells
white adipose tissue mass and systemic insulin A link between obesity and cancer stem cells
resistance (van Geenen et al., 2010; Smits & van has been identified by Zheng et al. (2011), who
Geenen, 2011). Together, pancreatic steatosis and showed that spontaneous tumours derived from
NAFPD contribute to the already complex meta- mouse mammary tumour virus (MMTV)-Wnt-1
bolic and inflammatory perturbations associated transgenic mice, when transplanted, were highly
with obesity and metabolic syndrome. dependent on leptin for growth. Thus, when
these tumours were transplanted into obese,
leptin receptor-deficient (Leprdb/Leprdb) mice
with high leptin concentrations, they grew to 8
580
times the volume of those tumours transplanted into mice rapidly formed claudin-low tumours
into wild-type mice, whereas in leptin-deficient that were highly responsive to the tumour-
(Lepob/Lepob) mice, tumour growth was impaired. enhancing effects of obesity, as well as the anti-
The residual tumours in Lepob/Lepob mice were cancer effects of DR.
found to have fewer “cancer stem cells”, and
these cells were characterized by flow cytom- (f) Synthesis
etry to express leptin receptor. When isolated by Emerging factors that are likely to contribute
leptin receptor expression, these cells exhibited to the obesity–cancer link, but for which there is
stem cell properties based on the ability to form currently insufficient data, include the gut micro-
tumourspheres in vitro, and their survival was biome, gut hormones (such as ghrelin produced
regulated by leptin. Dunlap et al. (2012) used by the stomach), NAFLD (which drives secretion
two types of cells – mesenchymal (M-Wnt) or of CRP and other inflammation-related factors),
epithelial (E-Wnt) – derived from spontaneous the immune function, and cancer stem cells.
mammary tumours in MMTV-Wnt-1 mice,
transplanted into ovariectomized C57BL/6
mice to emulate human claudin-low and basal-
like breast tumours, respectively. They reported
that M-Wnt, but not E-Wnt, mammary tumour
cells were stably enriched in breast cancer cell
markers, and exhibited stem cell properties. In
addition, M-Wnt cells orthotopically injected
Table 4.1 Effect of obesity and weight reduction on selected serological factors involved in the
carcinogenesis process
Serological factor Obesity Weight reduction

Sex hormones
Estradiol Increase Decrease
Sex hormone-binding globulin Decrease Increase
Testosterone Decrease (men) Increase (men)
Increase (women) Decrease (women)
Insulin and IGF-1
Insulin Increase Decrease
IGF-1 Increase (overweight) Decrease
Decrease (obese)
IGFBP-1 Decrease Increase
IGFBP-3 — —
Inflammation
Adiponectin Decrease Increase
Leptin Increase Decrease
C-reactive protein Increase Decrease
IGF-1, insulin-like growth factor 1; IGFBP, IGF binding protein.
581
Table 4.2 Molecular epidemiological studies of obesity–cancer related mechanisms for cancer of
the breast
Reference; country; Design; number Assays Biomarker, categories; RR, highest vs

study of cases, controls; lowest (95% CI), Ptrend
menopausal status
Sex hormones
Hankinson et al. (1998b); Nested case–control; Steroids: extraction, Estradiol, quartiles
USA; Nurses’ Health 156, 312 separation by column RR = 1.91 (1.06–3.46), Ptrend = 0.03
Study chromatography, Estrone, quartiles
radioimmunoassay RR = 1.96 (1.05–3.65), Ptrend = 0.01
Estrone sulfate: enzyme Estrone sulfate, quartiles
hydrolysis, organic RR = 2.25 (1.23–4.12), Ptrend = 0.01
extraction, separation by DHEAS, quartiles
column chromatography, RR = 2.15 (1.11–4.17), Ptrend = 0.01
radioimmunoassay Percentage free or percentage bioavailable
DHEAS: radioimmunoassay estradiol, androstenedione, testosterone,
DHEAS: NS
Key et al. (2002); USA, Pooled analysis; 663, NR Estradiol, quintiles
Japan, Italy; 1765; postmenopausal RR = 2.00 (1.47–2.71), Ptrend < 0.001
9 prospective studies Free estradiol, quintiles
RR = 2.58 (1.76–3.78), Ptrend < 0.001
Estrone, quintiles
RR = 2.19 (1.48–3.22), Ptrend < 0.001
Estrone sulfate
RR = 2.00 (1.26–3.16), Ptrend < 0.001
DHEA, quintiles
RR = 2.04 (1.21–3.45), Ptrend = 0.18
DHEAS, quintiles
RR = 1.75 (1.26–2.43), Ptrend = 0.002
Testosterone, quintiles
RR = 2.22 (1.59–3.10), Ptrend < 0.001
SHBG, quintiles
RR = 0.66 (0.43–1.00), Ptrend = 0.041
Kaaks et al. (2005); Nested case– DHEAS and testosterone: Estradiol, quintiles
several European control; 677, 1309; radioimmunoassay RR = 2.28 (1.61–3.23), Ptrend < 0.0001
countries; EPIC postmenopausal Androstenedione, estrone, Free estradiol, quintiles
estradiol, and SHBG: double- RR = 2.13 (1.52–2.98), Ptrend < 0.0001
antibody radioimmunoassay Estrone, quintiles
RR = 2.07 (1.42–3.02), Ptrend = 0.0001
SHBG, quintiles
RR = 0.61 (0.44–0.84), Ptrend = 0.004
RR = 1.85 (1.33–2.57), Ptrend < 0.0001
Free testosterone, quintiles
RR = 2.50 (1.76–3.55), Ptrend < 0.0001
Androstenedione, quintiles
RR = 1.94 (1.40–2.69), Ptrend < 0.0001
DHEAS, quintiles
RR = 1.69 (1.23–2.33), Ptrend = 0.0002
Gunter et al. (2009); Case–cohort; 835, 816 Vitros ECi Estradiol, tertiles
USA; Women’s Health immunodiagnostic assay All, HR = 1.59 (1.00–2.55), Ptrend = 0.04
Initiative Observational Non-HRT users, HR = 1.59 (1.00–2.55),
Study Ptrend = 0.04
Further adjustments, HR = 1.87 (1.11–3.15),
Ptrend = 0.03
582

menopausal status
Baglietto et al. (2010); Case–cohort; 197, 857; Estradiol and testosterone: Total estradiol, quartiles
Australia; Melbourne postmenopausal electrochemiluminescence HR = 1.44 (0.89–2.35), Ptrend = 0.12
Collaborative Cohort immunoassay Free estradiol, quartiles
Study Estrone sulfate: HR = 1.75 (1.06–2.89), Ptrend = 0.01
radioimmunoassay Estrone sulfate, quartiles
DHEAS: competitive HR = 2.05 (1.24–3.37), Ptrend < 0.01
immunoassay Testosterone, quartiles
Androstenedione: HR = 1.25 (0.78–2.01), Ptrend = 0.37
radioimmunoassay DHEAS, quartiles
SHBG: immunometric assay HR = 1.41 (0.88–2.27), Ptrend = 0.17
Androstenedione, quartiles
HR = 1.49 (0.91–2.44), Ptrend = 0.08
SHBG, quartiles
HR = 0.33 (0.19–0.55), Ptrend < 0.01
Farhat et al. (2011); Case–cohort; 317, 594; Radioimmunoassay Estradiol, quartiles
USA; Women’s Health postmenopausal ER+, HR = 1.86 (1.00–3.45), Ptrend = 0.08
Initiative Observational ER−, HR = 0.83 (0.43–1.61), Ptrend = 0.60
Study Testosterone, quartiles
ER+, HR = 1.55 (0.92–1.61), Ptrend = 0.04
ER−, HR = 0.51 (0.28–0.94), Ptrend = 0.03
James et al. (2011); Nested case– Radioimmunoassay Estradiol, ER+
several European control; 554, 821; Tertiles, OR = 2.58 (1.69–3.92),
countries; EPIC postmenopausal Ptrend < 0.0001
Continuous, OR = 1.63 (1.29–2.107)
Estradiol, ER+/PR+
Tertiles, OR = 2.91 (1.62–5.23),
Ptrend = 0.002
Continuous, OR = 1.58 (1.17–2.12)
Free estradiol, ER+
Tertiles, OR = 2.05 (1.39–3.02),
Ptrend = 0.003
Continuous, OR = 1.63 (1.31–2.02)
Free estradiol, ER+/PR+
Tertiles, OR = 2.09 (1.23–3.54),
Ptrend = 0.01
Continuous, OR = 1.61 (1.21–2.13)
Testosterone, ER+
Tertiles, OR = 1.68 (1.16–2.44),
Ptrend = 0.006
Continuous, OR = 1.54 (1.27–1.87)
Testosterone, ER+/PR+
Tertiles, OR = 2.27 (1.35–3.81),
Ptrend = 0.002
Continuous, OR = 1.79 (1.36–2.36)
583

menopausal status
Tworoger et al. (2011); Nested case– Radioimmunoassay Estrone, quintiles
USA; Nurses’ Health control; 265, 541; All, RR = 2.1 (1.3–3.4)
Study postmenopausal ER+, RR = 2.8 (1.5–5.3)
Estradiol, quintiles
All, RR = 2.4 (1.4–4.1)
ER+, RR = 2.9 (1.4–5.9)
Estrone sulfate, quintiles
All, RR = 2.4 (1.5–3.9)
ER+, RR = 2.2 (1.2–4.0)
All, RR = 1.8 (1.1–2.9)
ER+, RR = 2.0 (1.0–3.7)
All, RR = 2.1 (1.3–3.6)
ER+, RR = 2.6 (1.3–5.0)
DHEA, quintiles
All, RR = 1.5 (0.9–2.4)
ER+, RR = 1.6 (0.9–2.9)
DHEAS, quintiles
All, RR = 2.5 (1.4–4.2)
ER+, RR = 2.0 (1.0–3.8)
Fuhrman et al. (2012); Nested case– LC-MS Unconjugated estradiol, continuous
USA; Prostate, Lung, control; 227, 423; HR = 2.07 (1.19–3.62), Ptrend = 0.01
Colorectal, and Ovarian postmenopausal 2-Hydroxylation pathway, continuous
Cancer Screening Trial HR = 0.66 (0.51–0.87), Ptrend = 0.003
2/16-Hydroxylation pathway, continuous
HR = 0.62 (0.45–0.86), Ptrend = 0.05
Sieri et al. (2012); Italy; Nested case–control; Chemiluminescence SHBG, quartiles, diagnosis > 55 yr
ORDET Cohort 356; 1537 immunoassay RR = 0.60 (0.36–0.99), Ptrend = 0.059
Würtz et al. (2012); Nested case– Radioimmunoassay Non-HRT users
Denmark; Diet, Cancer control; 348, 348; Estradiol, tertiles
and Health Cohort postmenopausal RR = 1.56 (0.70–3.51), Ptrend = 0.55
Estrone, tertiles
RR = 2.02 (0.83–4.89), Ptrend = 0.06
Estrone sulfate, tertiles
RR = 4.21 (1.81–9.81), Ptrend = 0.01
584

menopausal status
Fortner et al. (2013); Nested case–control; Estrogens and testosterone: Follicular phase
USA; Nurses’ Health 634, 1264 (514, 1030 organic extraction, Estradiol, all
Study II with timed samples) celite chromatography, Quintiles, OR = 1.0 (0.7–1.5), Ptrend = 0.76
radioimmunoassay Doubling, OR = 0.6 (0.4–0.9), Ptrend < 0.01
SHBG and progesterone: Estradiol, invasive
automated immunoassay Doubling, OR = 0.6 (0.4–0.9), Ptrend < 0.01
or chemiluminescence Estradiol, ER+/PR+
immunometric assay Doubling, OR = 0.6 (0.4–0.9), Ptrend = 0.01
Free estradiol, all
Quintiles, OR = 0.8 (0.5–1.3), Ptrend = 0.48
Doubling, OR = 0.5 (0.4–0.8), Ptrend < 0.01
Free estradiol, invasive
Free estradiol, ER+/PR+
Testosterone, all
SHBG, all
Key et al. (2013); USA, Pooled analysis; 767, Radioimmunoassay, Estradiol, quintiles
United Kingdom, several 1699 competitive immunoassay, OR = 1.41 (1.02–1.95), Ptrend = 0.0042
European countries; LC-MS Calculated free estradiol, quintiles
7 prospective studies OR = 1.19 (0.86–1.64), Ptrend = 0.014
Estrone, quintiles
OR = 1.50 (1.02–2.19), Ptrend = 0.014
OR = 1.68 (1.18–2.39), Ptrend = 0.0026
DHEAS, quintiles
OR = 1.45 (1.07–1.95), Ptrend = 0.010
OR = 1.32 (0.98–1.76), Ptrend = 0.018
Calculated free testosterone, quintiles
OR = 1.25 (0.94–1.66), Ptrend = 0.15
Schernhammer et al. Nested case– Estradiol: radioimmunoassay Free testosterone, tertiles
(2013); Italy; ORDET control; 104, 225; Testosterone and OR = 2.43 (1.15–5.10), Ptrend = 0.03
Cohort premenopausal free testosterone: Total testosterone, tertiles
radioimmunoassay OR = 1.27 (0.62–2.61), Ptrend = 0.51
SHBG: chemiluminescence Progesterone, tertiles
immunometric assay OR = 1.16 (0.60–2.27), Ptrend = 0.75
Estradiol, tertiles
OR = 0.69 (0.35–1.35), Ptrend = 025
SHBG, tertiles
OR = 0.93 (0.50–1.72), Ptrend = 0.78
585

menopausal status
Zhang et al. (2013); USA; Nested case–control; Radioimmunoassay Estradiol, quartiles
Nurses’ Health Study 707; 1414 or LC-MS or solid- RR = 2.1 (1.6–2.7), Ptrend < 0.001
phase competitive Free estradiol, quartiles
chemiluminescence enzyme RR = 1.9 (1.4–2.4), Ptrend < 0.001
immunoassay Testosterone, quartiles
RR = 1.5 (1.2–1.9), Ptrend < 0.001
Free testosterone, quartiles
RR = 1.9 (1.5–2.5), Ptrend < 0.001
DHEAS, quartiles
RR = 1.7 (1.3–2.3), Ptrend = 0.001
SHBG, quartiles
RR = 0.68 (0.52–0.88), Ptrend = 0.004
Dallal et al. (2014); USA; Case–cohort; 407, LC-MS Estradiol, quintiles
Breast and Bone Follow- 496; postmenopausal HR = 1.86 (1.19–2.90), Ptrend = 0.04
up to the Fracture 2-Hydroxylation pathway, quintiles
Intervention Trial HR = 0.69 (0.46–1.05), Ptrend = 0.01
4-Hydroxylation pathway, quintiles
HR = 0.61 (0.40–0.93), Ptrend = 0.004
2/16-Hydroxylation pathway, quintiles
HR = 0.60 (0.40–0.90), Ptrend = 0.002
Kaaks et al. (2014b); Nested case–control; Estradiol: immunoassay Estradiol, quartiles
several European 801, 1132 Progesterone OR = 1.04 (0.93–1.15), Ptrend = 0.52
countries; EPIC and testosterone: Progesterone, quartiles
SHBG: sandwich SHBG, quartiles
immunoradiometric assay OR = 0.98 (0.88–1.08), Ptrend = 0.64
Testosterone, quartiles
OR = 1.56 (1.15–2.13), Ptrend = 0.02
Free testosterone, quartiles
OR = 1.33 (0.99–1.79), Ptrend = 0.04
Insulin
Toniolo et al. (2000); Nested case–control; Radioimmunoassay C-peptide, quartiles
USA; New York premenopausal: 172, Premenopausal, RR = 0.76 (0.44–1.31),
University Women’s 486; postmenopausal: Ptrend = 0.90
Health Study 115, 220 Postmenopausal, RR = 1.24 (0.66–2.34),
Ptrend = 0.58
Kaaks et al. (2002); Nested case–control; Double-antibody Insulin, quartiles
Sweden; Umeå Cohort 246, 454 immunoradiometric assay OR = 0.59 (0.30–1.18), Ptrend = 0.88
Mink et al. (2002); USA; Cohort; 189, 7705 Radioimmunoassay Insulin, quartiles
Atherosclerosis Risk RR = 1.01 (0.55–1.86), Ptrend = 0.87
in Communities Study
Cohort
Muti et al. (2002); Italy; Nested case–control; Double-antibody Insulin, quartiles
ORDET Cohort premenopausal: 69, radioimmunoassay Premenopausal, RR = 1.72 (0.71–4.15),
265; postmenopausal: Ptrend = 0.14
64, 238 Postmenopausal, RR = 0.85 (0.36–2.00),
Ptrend = 0.76
586

menopausal status
Keinan-Boker et al. Nested case–control; Competitive C-peptide, quartiles
(2003); The Netherlands; EPIC: 71, 163; radioimmunoassay OR = 1.3 (0.7–2.7)
EPIC, PPHV PPHV: 78, 170;
postmenopausal
Verheus et al. (2006); Nested case–control; Radioimmunoassay C-peptide, quintiles
several European 1141, 2204 Non-fasting
countries; EPIC ≤ 50 yr, OR = 0.74 (0.30–1.82), Ptrend = 0.35
50–60 yr, OR = 1.08 (0.56–2.08),
Ptrend = 0.51
> 60 yr, OR = 1.69 (0.97–2.95), Ptrend = 0.22
Fasting, all ORs: NS
Cust et al. (2009); Nested case–control; Radioimmunoassay C-peptide, tertiles
Sweden; Northern 561, 561 < 55 yr, OR = 0.75 (0.44–1.29), Ptrend = 0.34
Sweden Health and ≥ 55 yr, OR = 1.32 (0.84–2.05), Ptrend = 0.20
Disease Study Cohort
Gunter et al. (2009); Case–cohort; 835, 816; ELISA Insulin, quartiles
USA; Women’s Health postmenopausal HR = 1.46 (1.00–2.13), Ptrend = 0.2
Study Ptrend < 0.001
Adjusted also for estradiol, HR = 2.40
(1.30–4.41), Ptrend < 0.001
Kabat et al. (2009); Longitudinal ELISA Insulin, tertiles
USA; Women’s Health study; 190, 5450; HR = 2.22 (1.39–3.53), Ptrend = 0.0008
Initiative postmenopausal
Tworoger et al. (2011); Nested case–control; ELISA C-peptide, quintiles
USA; Nurses’ Health 265, 541 All, RR = 1.4 (0.8–2.4)
Study
Sieri et al. (2012); Italy; Nested case–control; Chemiluminescence Insulin, quartiles
ORDET Cohort 356, 1537 immunoassay Premenopausal, RR = 1.52 (0.92–2.51),
Ptrend = 0.08
Postmenopausal, RR = 1.31 (0.81–2.12),
Ptrend = 0.25
Parekh et al. (2013); Cohort; 217, 2152 NR Insulin, tertiles
USA; Framingham Heart HR = 1.41 (0.88–2.24), Ptrend = 0.33
Study-Offspring Cohort
IGFs
Hankinson et al. (1998b); Nested case–control; ELISA IGF-1
USA; Nurses’ Health 397, 620 Premenopausal, tertiles
Study RR = 2.33 (1.06–5.16), Ptrend = 0.08
Adjusted for IGFBP-3, RR = 2.88
(1.21–6.85), Ptrend = 0.02
Premenopausal, < 50 yr at blood collection,
tertiles
RR = 4.58 (1.75–12.0), Ptrend = 0.02
(2.40–22.0), Ptrend = 0.01
Postmenopausal, quintiles
RR = 0.85 (0.53–1.39), Ptrend = 0.63
587

menopausal status
Toniolo et al. (2000); Nested case–control; Radioimmunoassay IGF-1
USA; New York premenopausal: 172, Premenopausal, quartiles
University Women’s 486; postmenopausal: OR = 1.60 (0.91–2.81), Ptrend = 0.09
Health Study 115, 220 Adjusted for IGFBP-3, OR = 1.49
(0.80–2.79)
Premenopausal, < 50 yr at blood collection,
quartiles
OR = 2.30 (1.07–4.94), Ptrend = 0.03
Adjusted for IGFBP-3, OR = 1.90
(0.82–4.42)
Postmenopausal, quintiles
OR = 0.95 (0.49–1.86), Ptrend = 0.87
Kaaks et al. (2002); Nested case–control; Double-antibody IGF-1, quartiles
Sweden; Umeå and 513, 987 immunoradiometric assay OR = 1.17 (0.84–1.63), Ptrend = 0.55
Malmö Cohorts < 50 yr at recruitment, OR = 0.63
(0.29–2.39)
≥ 50 yr at recruitment, OR = 1.29
(0.80–2.07)
Muti et al. (2002); Italy; Nested case–control; Double-antibody IGF-1, quartiles
ORDET Cohort premenopausal: 69, immunoradiometric assay Premenopausal, RR = 3.12 (1.13–8.60),
265; postmenopausal: Ptrend = 0.01
64, 238 Postmenopausal, RR = 0.58 (0.24–1.36),
Ptrend = 0.25
All other analytes: NS
Keinan-Boker et al. Nested case– Immunoradiometric assay IGF-1, quartiles
(2003); The Netherlands; control; 149, 333; OR = 1.1 (0.6–2.1)
EPIC, PPHV postmenopausal
Allen et al. (2005); Nested case–control; Double-antibody ELISA IGF-1, tertiles
United Kingdom; premenopausal: 70, Premenopausal, OR = 1.71 (0.74–3.95),
Guernsey Cohort 209; postmenopausal: Ptrend = 0.21
47, 141 Postmenopausal, OR = 0.73 (0.29–1.84),
Ptrend = 0.52
Rinaldi et al. (2005); Nested case– ELISA IGF-1, quintiles
USA, Sweden, Italy; control; 220, 434; OR = 1.41 (0.75–2.63), Ptrend = 0.15
New York University premenopausal
Women’s Health Study,
Northern Sweden Health
and Disease Study
Cohort, ORDET Cohort
Lukanova et al. (2006); Nested case–control; Immunoradiometric assay IGF-1, tertiles
Sweden; Northern 212, 369 OR = 1.7 (1.1–2.7), Ptrend = 0.02
Sweden Maternity
Cohort
588

menopausal status
Baglietto et al. (2007); Case–cohort; 423, ELISA IGF-1, quartiles
Australia; Melbourne 1901 All, HR = 1.20 (0.87–1.65), Ptrend = 0.38
Collaborative Cohort Premenopausal, HR = 0.83 (0.49–1.38),
Study Ptrend = 0.29
Postmenopausal, HR = 1.59 (1.03–2.44),
Ptrend = 0.05
≥ 60 yr at follow-up, HR = 1.61 (1.04–2.51),
Ptrend = 0.06
Vatten et al. (2008); Nested case–control; Double-antibody IGF-1, quintiles, adjusted for IGFBP-3
Norway; Janus Biobank 325, 647 radioimmunoassay OR = 1.46 (0.93–2.32), Ptrend = 0.15
IGFBP-3, quintiles, adjusted for IGF-1
OR = 0.78 (0.49–1.23), Ptrend = 0.12
T3 for IGF-1 and T1 for IGFBP-3 (tertiles)
OR = 2.00 (1.01–3.96)
Gunter et al. (2009); Case–cohort; 835, 816; ELISA Total IGF-1, quartiles
USA; Women’s Health postmenopausal All, HR = 1.21 (0.85–1.72), Ptrend = 0.92
Study Ptrend = 0.72
Sakauchi et al. Nested case–control; Immunoradiometric assay IGF-1, tertiles
(2009); Japan; Japan 63, 187 Premenopausal, OR = 1.2 (0.32–4.09),
Collaborative Cohort Ptrend = 0.81
Study Postmenopausal, OR = 2.8 (0.73–10.6),
Ptrend = 0.17
Key et al. (2010); 12 Pooled analysis; 4790, NR IGF-1, quintiles
countries; 17 prospective 9428 All, OR = 1.28 (1.14–1.44),
studies Ptrend < 0.0001
Premenopausal, OR = 1.21 (1.00–1.45),
Ptrend = 0.50
Postmenopausal, OR = 1.33 (1.14–1.55),
Ptrend = 0.0002
IGFBP-3, quintiles
All, OR = 1.3 (0.99–1.28),
Ptrend < 0.062
IGFBP-3, quintiles
Premenopausal, OR = 1.00 (0.82–1.22),
Ptrend = 0.921
Postmenopausal, OR = 1.23 (1.04–1.45),
Ptrend = 0.012
Tworoger et al. (2011); Nested case– ELISA IGF-1, quintiles
USA; Nurses’ Health control; 265, 541; RR = 1.1 (0.6–2.0)
Study postmenopausal
589

menopausal status
Kaaks et al. (2014a); Nested case–control; ELISA IGF-1, all
several European 938, 1394 Quartiles, OR = 1.34 (1.0–1.78)
countries; EPIC IGF-1, ER+
Quartiles, OR = 1.41 (1.01–1.98)
Continuous, OR = 1.17 (1.04–1.33),
Ptrend = 0.01
IGF-1, ER+, diagnosis ≥ 50 yr
Tertiles, OR = 1.38 (1.01–1.89)
Continuous, OR = 1.19 (1.04–1.36),
Ptrend = 0.01
Inflammatory factors
Krajcik et al. (2003); Nested case–control; ELISA TNF-α, quartiles
USA; Kaiser Permanente 81, 81; premenopausal OR = 0.60 (0.15–2.31), Ptrend = 0.45
Medical Care Program sTNFR1, quartiles
OR = 0.67 (0.20–2.25), Ptrend = 0.78
sTNFR2, quartiles
OR = 0.46 (0.06–3.50), Ptrend = 0.63
Stattin et al. (2004a); Nested case– Double-antibody Leptin, quartiles
Sweden; Northern control; 149, 258; radioimmunoassay OR = 0.94 (0.53–1.67), Ptrend = 0.54
Sweden Health and postmenopausal
Siemes et al. (2006); The Cohort; 184, 3790 Particle immunoassay CRP
Netherlands; Rotterdam Tertiles, HR = 1.59 (1.05–2.41)
Study > 3 vs < 1 mg/L, HR = 1.68 (1.14–2.47)
Continuous, HR = 1.28 (1.07–1.54)
Tworoger et al. (2007a); Nested case–control; Radioimmunoassay Adiponectin, quartiles
USA; Nurses’ Health 1477, 296 RR = 0.89 (0.71–1.11), Ptrend = 0.54
Study and Nurses’ Postmenopausal, RR = 0.73 (0.55–0.98),
Health Study II Cohorts Ptrend = 0.08
Allin et al. (2009); Cohort; 202, 1624 Turbidimetry/nephelometry CRP
Denmark; Copenhagen Quintiles, OR = 0.9 (0.5–1.7)
City Health Study > 3 vs < 1 mg/L, HR = 0.7 (0.4–1.4)
Cust et al. (2009); Nested case–control; Double-antibody Leptin, tertiles
Sweden; Northern 561, 561 radioimmunoassay < 55 yr, OR = 0.80 (0.52–1.22), Ptrend = 0.29
Sweden Health and ≥ 55 yr, OR = 1.15 (0.76–1.74), Ptrend = 0.53
Disease Study Cohort Stage I, OR = 0.64 (0.41–1.00), Ptrend = 0.06
Stage II–IV, OR = 1.37 (0.91–1.06),
Ptrend = 0.14
Adiponectin, tertiles
< 55 yr, OR = 0.56 (0.28–1.11), Ptrend = 0.08
≥ 55 yr, OR = 0.96 (0.55–1.65), Ptrend = 0.95
Stage I, OR = 0.74 (0.40–1.38), Ptrend = 0.42
Stage II–IV, OR = 0.83 (0.46–1.51),
Ptrend = 0.53
Harris et al. (2011); USA; Nested case– Enzyme immunoassay Leptin, quartiles
Nurses’ Health Study II control; 330, 636; OR = 0.55 (0.31–0.99), Ptrend = 0.04
premenopausal
590

menopausal status
Gaudet et al. (2013); Nested case– ELISA Total adiponectin, tertiles
Cancer Prevention control; 302, 302; OR = 0.84 (0.54–1.30), Ptrend = 0.38
Study-II Nutrition postmenopausal CRP, tertiles
Cohort OR = 1.09 (0.70–1.70), Ptrend = 0.16
Gross et al. (2013); USA; Nested case– ELISA Leptin, tertiles
CLUE II Cohort control; 272, 272; OR = 1.98 (1.20–3.29), Ptrend = 0.05
postmenopausal Adiponectin, tertiles
OR = 1.63 (1.02–2.60), Ptrend = 0.08
sTNFR2, tertiles
OR = 2.44 (1.30–4.58), Ptrend = 0.008
Liu et al. (2013); Meta-analysis; 3578, NR Adiponectin
13 studies 4363 OR = 0.838 (0.744–0.943)
Prizment et al. (2013); Cohort; 176, 7603 Immunoturbidimetric assay CRP, continuous
USA; Atherosclerosis HR = 1.27 (1.07–1.51)
Risk in Communities
Study Cohort
Touvier et al. (2013); Nested case–control; ELISA hsCRP, quartiles
France; Supplémentation 218, 436 OR = 1.25 (0.73–2.14), Ptrend = 0.7
en Vitamines et Leptin, quartiles
Minéraux Antioxydants OR = 0.64 (0.34–1.20), Ptrend = 0.1
Trial Adiponectin, quartiles
OR = 1.13 (0.68–1.87), Ptrend = 0.4
Dossus et al. (2014); Nested case– Particle-enhanced CRP, all
France; Etude control; 549, 1040; immunoturbidimetric assay < 1.5 vs 2.5–10 mg/L, OR = 1.24
Epidémiologique postmenopausal (0.92–1.66)
auprès de femmes de la Continuous, OR = 1.13 (0.98–1.29),
Mutuelle Générale de Ptrend = 0.09
l’Éducation Nationale CRP, BMI ≥ 25 kg/m2
< 1.5 vs 2.5–10 mg/L, OR = 1.92 (1.20–3.08)
Continuous, OR = 1.52 (1.16–2.00),
Ptrend = 0.003
Macis et al. (2014); Meta-analysis; 4249 NR Adiponectin
15 studies SRR = 0.66 (0.50–87)
Chan et al. (2015); Meta-analysis; 3522, NR CRP, doubling concentration
12 studies 69 610 RR = 1.07 (10.2–1.12)
591

menopausal status
Gunter et al. (2015b); Case–cohort; 875, CRP: latex-enhanced CRP, quartiles
USA; Women’s Health 839; postmenopausal immunonephelometry All, HR = 1.24 (0.86–1.80), Ptrend = 0.12
Initiative Leptin and TNF-α: Milliplex Non-HRT users, HR = 1.67 (1.04–2.68),
Human Adipokine Panel B Ptrend = 0.029
Adiponectin, PAI-1, and Leptin, quartiles
resistin: Milliplex Human All, HR = 1.39 (0.93–2.09), Ptrend = 0.279
Adipokine Panel A Adiponectin, quartiles
IL-6: ELISA All, HR = 0.76 (0.55–1.06), Ptrend = 0.78
Resistin, quartiles
All, HR = 0.93 (0.68–1.27), Ptrend = 0.664
PAI-1, quartiles
All, HR = 1.33 (0.96–1.86), Ptrend = 0.145
Non-HRT users, HR = 1.71 (1.02–2.89),
Ptrend = 0.077
IL-6, quartiles
All, HR = 1.20 (0.85–1.69), Ptrend = 0.528
TNF-α, quartiles
All, HR = 0.82 (0.59–1.14), Ptrend = 0.292
Wang et al. (2015); Cohort; 87, 19 437 Nephelometric assay hsCRP, < 1 vs > 3 mg/L
China; Kailuan Female All, HR = 1.74 (1.01–2.97), Ptrend = 0.047
Cohort Women < 50 yr, HR = 2.76 (1.18–6.48)
Excluding CRP > 10 mg/L, HR = 1.89
(1.08–3.32), Ptrend = 0.029
BMI, body mass index; CI, confidence interval; CRP, C-reactive protein; DHEA, dehydroepiandrosterone; DHEAS, DHEA sulfate; ELISA,
enzyme-linked immunosorbent assay; EPIC, European Prospective Investigation into Cancer and Nutrition; ER+, estrogen receptor-positive;
HR, hazard ratio; HRT, hormone replacement therapy; hsCRP, high-sensitivity C-reactive protein; IGF, insulin growth factor; IGFBP, IGF
binding protein; IL, interleukin; LC-MS, liquid chromatography–mass spectrometry; NR, not reported; NS, no significant association; OR, odds
ratio; PAI-1, plasminogen activator inhibitor 1; PPHV, Monitoring Project on Cardiovascular Disease Risk Factors; PR+, progesterone receptor-
positive; RR, relative risk; SHBG, sex hormone-binding globulin; SRR, summary relative risk; sTNFR, soluble tumour necrosis factor receptor;
TNF, tumour necrosis factor; yr, year or years.
592
the endometrium

menopausal status
Sex hormones
Zeleniuch-Jacquotte Nested case– Estradiol and estrone: Estradiol, tertiles
et al. (2001); USA; control; 57, 222; organic extraction, OR = 1.8 (0.75–4.2), Ptrend = 0.19
New York University postmenopausal celite chromatography, Free estradiol, tertiles
Women’s Health Study radioimmunoassay OR = 2.8 (1.3–6.4), Ptrend = 0.004
Cohort SHBG: chemiluminescence SHBG-bound estradiol, tertiles
immunometric assay OR = 0.60 (0.26–1.4), Ptrend = 0.22
Free estradiol: ultrafiltration Estrone, tertiles
method OR = 3.2 (1.3–7.8), Ptrend = 0.008
Estradiol bound to SHBG: SHBG, tertiles
concanavalin A–agarose OR = 0.49 (0.22–1.1), Ptrend = 0.08
binding assay
Lukanova et al. (2004a); Nested case– Estrone: radioimmunoassay Estrone, quartiles
USA, Sweden, Italy; control; 124, 236; or double-antibody OR = 3.67 (1.71–7.88), Ptrend = 0.0007
New York University postmenopausal, radioimmunoassay Estradiol, quartiles
Women’s Health Study, non-HRT users Estradiol: radioimmuno OR = 4.13 (1.76–9.72), Ptrend = 0.0008
Northern Sweden assay or ultrasensitive SHBG, quartiles
Health and Disease double-antibody radio OR = 0.46 (0.20–1.05), Ptrend = 0.01
Study, ORDET Cohort immunoassay
SHBG: immunometric
chemiluminescence assay or
immunoradiometric assay
Allen et al. (2008); Nested case– Estrone and estradiol: Postmenopausal women:
several European control; 247, 481; radioimmunoassay with Estrone, tertiles
countries; EPIC premenopausal double-antibody system OR = 2.66 (1.50–4.72),
(55, 107) and SHBG: solid-phase sandwich Ptrend (continuous) = 0.002
postmenopausal immunoradiometric assay Estradiol, tertiles
(192, 374) OR = 2.07 (1.20–3.60),
Ptrend (continuous) = 0.001
Free estradiol, tertiles
OR = 1.66 (0.98–2.82),
SHBG, tertiles
OR = 0.57 (0.34–0.95),
Gunter et al. (2008a); Case–cohort; 250, Vitros ECi immunodiagnos Estradiol, tertiles
USA; Women’s Health 465; postmenopausal tic assay HR = 3.16 (1.71–5.81), Ptrend < 0.001
Initiative
Brinton et al. (2016); Nested case–control; Stable-isotope dilution liquid Estrone, quintiles
USA; Women’s Health 313 (271 type I, chromatography-tandem OR = 3.19 (1.69–6.04), Ptrend = 0.0001
Initiative Observational 42 type II), 354; mass spectrometry Estradiol, quintiles
Study postmenopausal OR = 1.41 (0.75–2.67), Ptrend = 0.4531
Unconjugated estradiol, quintiles
OR = 6.19 (2.95–13.03), Ptrend = 0.0001
Conjugated estradiol, quintiles
0.95 (0.51–1.77), Ptrend = 0.6747
593

menopausal status
Insulin
Lukanova et al. (2004b); Nested case– Radioimmunoassay C-peptide, quintiles
USA, Sweden, Italy; control; 166, 315; OR = 4.76 (1.91–11.8), Ptrend = 0.0002
New York University premenopausal and OR = 4.40 (1.65–11.7), Ptrend = 0.003, adjusted
Women’s Health Study, postmenopausal for BMI, other confounders
Northern Sweden
Health and Disease
Study, ORDET Cohort
Cust et al. (2007a); Nested case– Immunoradiometric assay C-peptide, quartiles
several European control; 286, 555; All, RR = 2.13 (1.33–3.41), Ptrend = 0.001
countries; EPIC premenopausal and Postmenopausal, RR = 1.28 (0.67–2.45),
postmenopausal Ptrend = 0.42, adjusted for free estradiol
Gunter et al. (2008a); Case–cohort; 250 ELISA Insulin, quartiles
USA; Women’s Health (205 endometrioid Endometrioid adenocarcinoma, non-HRT
Initiative adenocarcinoma), users, HR = 2.33 (1.13–4.82), Ptrend = 0.02,
465; postmenopausal adjusted for age, estradiol
BMI ≥ 25 kg/m2, HR = 4.30 (1.62–11.43),
Ptrend = 0.001
IGFs
Lukanova et al. (2004b); Nested case– Immunoradiometric assay IGFBP-1, quintiles
USA, Sweden, Italy; control; 166, 315; (IGF-1 after acid–ethanol OR = 0.30 (0.15–0.62), Ptrend = 0.002
New York University premenopausal and precipitation of IGFBPs) OR = 0.49 (0.22–1.07), Ptrend = 0.06, adjusted
Women’s Health Study, postmenopausal for BMI, other confounders
Northern Sweden IGF-1: NS
Health and Disease
Study, ORDET Cohort
Cust et al. (2007a); Nested case– IGFBP-1: immunoradio IGFBP-1, quartiles
several European control; 286, 555; metric assay RR = 0.76 (0.47–1.21), Ptrend = 0.25
countries; EPIC premenopausal and IGFBP-2: radioimmunoassay IGFBP-2, quartiles
postmenopausal RR = 0.56 (0.35–0.90), Ptrend = 0.03
Gunter et al. (2008a); Case–cohort; 250 ELISA Free IGF, quartiles
USA; Women’s Health (205 endometrioid Endometrioid adenocarcinoma,
Initiative adenocarcinoma), HR = 0.53 (0.31–0.90), Ptrend = 0.05, adjusted
465; postmenopausal for age, HRT, estradiol
Overweight or obese, HR = 0.43 (0.20–0.97),
adjusted for age, HRT, estradiol
Total IGF-1: NS
Cust et al. (2007b); Nested case–control; ELISA Adiponectin, quartiles
several European 284, 548 RR = 0.56 (0.36–0.86), Ptrend = 0.006, adjusted
countries; EPIC for BMI
Dossus et al. (2010); Nested case–control; CRP and IL-6: ELISA CRP, quartiles
several European 305, 574 IL-1Ra: bead-based OR = 1.58 (1.03–2.41), Ptrend = 0.02
countries; EPIC immunoassay IL-6, quartiles
OR = 1.66 (1.08–2.54), Ptrend = 0.008
IL-1Ra, quartiles
OR = 1.82 (1.22–2.73), Ptrend = 0.004
All ORs adjusted for BMI: NS
594

menopausal status
Dossus et al. (2011); Nested case–control: ELISA TNF-α, quartiles
several European 270, 518 OR = 1.73 (1.09–2.73), Ptrend = 0.01
countries; EPIC
Soliman et al. (2011); Nested case–control; ELISA Adiponectin, > 15 µg/mL
USA; Nurses’ Health 146; 377 RR = 0.86 (0.53–1.39), Ptrend = 0.48, adjusted
Study for BMI
Wang et al. (2011); Case–cohort; CRP: high-sensitivity CRP, quartiles
USA; Women’s Health 151, 301; latex-enhanced HR = 2.29 (1.13–4.65), Ptrend = 0.012, adjusted
Initiative postmenopausal, immunonephelometry for age, BMI
non-HRT users IL-6: ultrasensitive solid- HR = 1.70 (0.78–3.68), Ptrend = 0.127, adjusted
phase ELISA also for estradiol, insulin
TNF-α: Milliplex Human IL-6, TNF-α: NS
Adipokine Panel B
Dallal et al. (2013); USA; Nested case–control; ELISA Leptin, tertiles
Breast and Bone Follow- 62, 124 OR = 2.96 (1.21–7.25), Ptrend < 0.01, adjusted
up to the Fracture for estradiol, C-peptide
Intervention Trial OR = 2.11 (0.69–6.44), Ptrend = 0.18, adjusted
also for BMI
Adiponectin: NS
Luhn et al. (2013); Nested case–control; Radioimmunoassay Leptin, tertiles
USA; Prostate, Lung, 167, 327 All, OR = 2.77 (1.60–4.79), Ptrend < 0.01
Colorectal, and Ovarian Non-HRT users, OR = 4.72 (1.15–19.38),
Cancer Screening Trial Ptrend = 0.02
All, OR = 0.48 (0.29–0.80), Ptrend < 0.01
Non-HRT users, OR = 0.25 (0.08–0.75),
Ptrend = 0.01
BMI, body mass index; CI, confidence interval; CRP, C-reactive protein; ELISA, enzyme-linked immunosorbent assay; EPIC, European
Prospective Investigation into Cancer and Nutrition; HR, hazard ratio; HRT, hormone replacement therapy; IGF, insulin growth factor; IGFBP,
IGF binding protein; IL, interleukin; IL-1Ra, IL-1 receptor antagonist; NR, not reported; NS, no significant association; OR, odds ratio; RR,
relative risk; SHBG, sex hormone-binding globulin; TNF-α, tumour necrosis factor alpha.
595
the colorectum
Reference; country; study Design; number of Assays Biomarker, categories; RR, highest
cases, controls; sex vs lowest (95% CI), Ptrend
Sex hormones
Gunter et al. (2008b); USA; Case–cohort; 273, Vitros ECi immunodiagnostic Estradiol, tertiles
Women’s Health Initiative 442; F assay HR = 1.43 (0.95–2.16), Ptrend = 0.09
Clendenen et al. (2009); USA; Nested case–control; Estrone and estradiol: Estrone, quartiles
New York University Women’s 148, 293; F radioimmunoassay OR = 1.6 (0.8–3.0), Ptrend = 0.09
Health Study SHBG: immunometric Estradiol, tertiles
chemiluminescence assay OR = 0.8 (0.4–1.7), Ptrend = 0.43
SHBG, quartiles
OR = 0.8 (0.4–1.4), Ptrend = 0.48
Hyde et al. (2012); Australia; Cohort; 104, 3416; M Chemiluminescence SHBG, 60 vs 40 nmol/L
Health in Men Study immunoassay sub-HR = 0.98 (0.62–1.56),
Ptrend = 0.84
Lin et al. (2013); USA; Nurses’ Nested case–control; Estrone and estradiol: liquid Estradiol, quartiles
Health Study, Women’s Health M: 439, 719; F: 293, chromatography-tandem M: RR = 1.15 (0.73–1.81), Ptrend = 0.67
Study, Health Professionals 437 mass spectrometry F: RR = 1.12 (0.62–2.03), Ptrend = 0.93
Follow-up Study, Physicians’ SHBG: electrochemilumi Estrone, quartiles
Health Study II nescence immunoassay M: RR = 1.04 (0.68–1.62), Ptrend = 0.96
F: RR = 1.30 (0.74–2.26), Ptrend = 0.55
SHBG, quartiles
M: RR = 0.65 (0.42–0.99), Ptrend = 0.02
F: RR = 1.17 (0.63–2.20), Ptrend = 0.68
Falk et al. (2015); USA; Breast Case–cohort; 187, NR Estradiol, quartiles
and Bone Follow-up to the 501; F OR = 0.98 (0.58–1.64), Ptrend = 1.00
Fracture Intervention Trial Estrone, quartiles
OR = 1.15 (0.69–1.93), Ptrend = 0.54
Murphy et al. (2015); USA; Nested case–control; Estrone and estradiol: Estradiol quartiles
Women’s Health Initiative 401, 802; F radioimmunoassay OR = 0.64 (0.43–0.97), Ptrend = 0.12
Clinical Trial SHBG: immunometric Estrone, quartiles
chemiluminescence assay OR = 0.50 (0.33–0.75), Ptrend = 0.002
SHBG, quartiles
OR = 2.30 (1.51–3.51), Ptrend < 0.0001
Insulin
Schoen et al. (1999); USA; Cohort; 102, 5747; Solid-phase Insulin, quartiles
Cardiovascular Health Study M&F radioimmunoassay RR = 1.2 (0.7–2.1)
Kaaks et al. (2000); USA; New Nested case–control; Radioimmunoassay C-peptide, quintiles
York University Women’s 102, 200; F OR = 2.92 (1.26–6.75), Ptrend = 0.001
Health Study
Palmqvist et al. (2003); Nested case–control; Double-antibody Insulin, quartiles
Sweden; Northern Sweden 168, 376; M&F immunoradiometric assay OR = 1.22 (0.64–2.31), Ptrend = 0.41
Health and Disease Study
Cohort
Saydah et al. (2003); USA; Nested case–control; Ultrasensitive ELISA Insulin, quartiles
CLUE II Cohort colon: 132, rectum: OR = 0.78 (0.45–1.35), Ptrend = 0.24
41, 346; M&F
Ma et al. (2004); USA; Nested case–control; ELISA C-peptide, quintiles
Physicians’ Health Study 176, 294; M RR = 2.7 (1.2–6.2), Ptrend = 0.047
596
Stattin et al. (2004b); Norway; Nested case–control; Radioimmunoassay C-peptide, quartiles
Janus Biobank colon: 235, 235; Colon: OR = 1.82 (0.67–4.86),
rectum: 143, 143; M Ptrend = 0.19
Rectum: OR = 0.44 (0.10–1.99),
Ptrend = 0.21
Wei et al. (2005a); USA; Nested case–control; ELISA C-peptide, quartiles
Nurses’ Health Study 182, 350; F RR = 1.17 (0.63–2.20), Ptrend = 0.94
Limburg et al. (2006); Finland; Case–cohort; 134, Two-site immunoenzymatic Insulin, quartiles, age-adjusted
ATBC 399; M assay HR = 1.84 (1.03–3.30), Ptrend = 0.12
Insulin, quartiles, multivariate
HR = 1.74 (0.74–4.07), Ptrend = 0.40
Jenab et al. (2007); several Nested case–control; Radioimmunoassay C-peptide, quintiles
European countries; EPIC 1078, 1078; M&F OR = 1.37 (1.00–1.88), Ptrend = 0.03
Otani et al. (2007); Japan; Nested case–control; Radioimmunoassay C-peptide, quartiles
Japan Public Health Center- M: 196, 392, F: 179, M: OR = 3.2 (1.4–7.6), Ptrend = 0.0072
based Prospective Study 35 F: OR = 0.78 (0.38–1.6), Ptrend = 0.49
Gunter et al. (2008b); USA; Case–cohort; 429, ELISA Insulin, quartiles
Women’s Health Initiative 800; F HR = 1.89 (1.33–2.69), Ptrend = 0.0005
Adjusted also for waist
circumference, HR = 1.42 (0.91–2.23),
Ptrend = 0.11
Stocks et al. (2008); Sweden; Nested case–control; Immunoradiometric assay C-peptide, quartiles
Northern Sweden Health and 306, 595; M&F OR = 0.94 (0.62–1.41), Ptrend = 0.82
Kabat et al. (2012); USA; Case–cohort; 80, ELISA Insulin, ≥ 11.85 vs < 7.75 μU/mL
Ollberding et al. (2012); USA; Nested case–control; ELISA Insulin, tertiles
Multiethnic Cohort Study 249, 1571; M&F OR = 1.21 (0.84–1.75), Ptrend = 0.29
Lin et al. (2013); USA; Nurses’ Nested case–control; ELISA or C-peptide, quartiles
Health Study, Women’s Health M: 439, 719; F: 293, electrochemiluminescence M: RR = 1.29 (0.80–2.08), Ptrend = 0.27
Study, Health Professionals 437 immunoassay F: RR = 1.73 (0.94–3.18), Ptrend = 0.09
Follow-up Study, Physicians’
Health Study II
Parekh et al. (2013); USA; Cohort; 71, 3433; NR Insulin, ≥ 10.09 vs < 4.94 pmol/L
Framingham Heart Study- M&F HR = 2.10 (1.12–3.93), P = 0.0354
Offspring Cohort
Murphy et al. (2015); USA; Nested case–control; ELISA Insulin, quartiles
Women’s Health Initiative 401, 802; F OR = 0.76 (0.50–1.14), Ptrend = 0.21
Clinical Trial
IGFs
Ma et al. (1999); USA; Nested case–control; ELISA IGF-1, quintiles
Physicians’ Health Study 193, 318; M RR = 1.36 (0.72–2.55), Ptrend = 0.51
(1.15–5.46), Ptrend = 0.02
IGFBP-3, quintiles
RR = 0.47 (0.23–0.95), Ptrend = 0.07
Giovannucci et al. (2000); Nested case–control; ELISA IGF-1, tertiles
USA; Nurses’ Health Study 79, 158; F RR = 2.18 (0.94–5.08), Ptrend = 0.10
Cohort IGFBP-3, tertiles
RR = 0.28 (0.10–0.83), Ptrend = 0.05
597
Kaaks et al. (2000); USA; New Nested case–control; Double-antibody IGF-1, quintiles
York University Women’s 102, 200; F immunoradiometric assay OR = 1.88 (0.72–4.91), Ptrend = 0.25
Health Study IGFBP-3, quintiles
OR = 2.46 (1.09–5.57), Ptrend = 0.19
Probst-Hensch et al. (2001); Nested case–control; IGF-1: radioimmunoassay IGF-1, quintiles
China; Shanghai Cohort Study 135, 661; M IGFBP-3: immunoradiometric OR = 1.52 (0.82–2.85), Ptrend = 0.34
assay IGFBP-3, quintiles
OR = 1.72 (0.91–3.25), Ptrend = 0.07
Palmqvist et al. (2002); Nested case–control; Double-antibody IGF-1, quartiles
Sweden; Northern Sweden 168, 336; M&F immunoradiometric assay Colorectum: OR = 1.27 (0.65–2.47),
Health and Disease Study Ptrend = 0.51
Cohort Colon: OR = 2.66 (1.09–6.50),
Ptrend = 0.03
IGFBP-3, quartiles
Colorectum: OR = 1.23 (0.68–2.22),
Ptrend = 0.24
Colon: OR = 1.93 (0.92–4.06),
Ptrend = 0.02
Nomura et al. (2003); USA; Nested case–control; ELISA IGF-1, quartiles
Honolulu Heart Program 282, 282; M OR = 1.5 (0.8–2.8), Ptrend = 0.13
IGFBP-3, quartiles
OR = 0.8 (0.4–1.6), Ptrend = 0.45
Wei et al. (2005a); USA; Nested case–control; ELISA IGF-1, quartiles, colon
Nurses’ Health Study Cohort 137, 262; F RR = 1.95 (0.97–3.91), Ptrend = 0.09
IGFBP-3, quartiles, colon
RR = 1.20 (0.62–2.30), Ptrend = 0.62
Morris et al. (2006); United Nested case–control; ELISA IGF-1, quartiles
Kingdom; British United 147, 440; M OR = 1.10 (0.56–2.18), Ptrend = 0.65
Provident Association Study
Otani et al. (2007); Japan; Nested case–control; Immunoradiometric assay IGF-1, quartiles
Japan Public Health Center- M: 196, 392; F: 179, M: OR = 0.83 (0.40–1.7), Ptrend = 0.91
based Prospective Study 358 F: OR = 0.83 (0.38–1.8), Ptrend = 0.60
IGFBP-3, quartiles
M: OR = 1.4 (0.65–2.8), Ptrend = 0.60
F: OR = 1.1 (0.53–2.3), Ptrend = 0.73
Gunter et al. (2008b); USA; Case–cohort; 438, ELISA Total IGF-1, quartiles
Free IGF-1, quartiles
HR = 1.21 (0.86–1.72), Ptrend = 0.16
Max et al. (2008); Finland; Case–cohort; 134, ELISA IGF-1, quartiles
ATBC 399; M RR = 0.92 (0.49–1.70), Ptrend = 0.90
IGFBP-3, quartiles
RR = 0.98 (0.51–1.88), Ptrend = 0.85
Suzuki et al. (2009); Japan; Nested case–control; Immunoradiometric assay IGF-1, tertiles
Japan Collaborative Cohort 101, 302; M&F OR = 1.01 (0.49–2.10), Ptrend = 0.35
Study
Rinaldi et al. (2010); several Nested case–control; ELISA IGF-1, quintiles
European countries; EPIC 1121, 1121; M&F OR = 1.07 (0.81–1.40)
IGFBP-3, quintiles
OR = 1.17 (0.87–1.56)
598
Ollberding et al. (2012); USA; Nested case–control; ELISA IGF-1, tertiles
Multiethnic Cohort Study IGF-1: 258, 1701; OR = 0.84 (0.60–1.17), Ptrend = 0.30
IGF-2: 255, 1571; IGFBP-3, tertiles
M&F OR = 0.63 (0.45–0.88), Ptrend = 0.48
Murphy et al. (2015); USA; Nested case–control; ELISA IGF-1, quartiles
Women’s Health Initiative 401, 802; F OR = 0.70 (0.48–1.03), Ptrend = 0.15
Clinical Trial
Stattin et al. (2003); Sweden; Nested case–control; Double-antibody Leptin, quartiles
Northern Sweden Health and 168, 327; M&F immunoradiometric assay OR = 2.28 (1.09–4.76)
Stattin et al. (2004b); Norway; Nested case–control; Radioimmunoassay Leptin, quartiles
Janus Biobank colon: 235, 235; Colon: OR = 2.72 (1.44–5.12),
rectum: 143, 143; M Ptrend = 0.008
Rectum: OR = 0.91 (0.49–1.70),
Ptrend = 0.68
Tamakoshi et al. (2005); Nested case–control; Immunometric sandwich Leptin, quintiles
Japan; Japan Collaborative 58, 145; F enzyme immunoassay OR = 3.94 (1.04–14.9), Ptrend = 0.02
Cohort Study
Wei et al. (2005b); USA; Nested case–control; Radioimmunoassay Adiponectin, quintiles
Health Professionals Follow- 179, 356; M RR = 0.42 (0.23–0.78), Ptrend = 0.01
up Study
Stocks et al. (2008); Sweden; Nested case–control; Leptin: radioimmunoassay Leptin, quartiles
Northern Sweden Health and 306, 595; M&F Adiponectin: ELISA OR = 1.09 (0.74–1.61), Ptrend = 0.29
Disease Study Cohort Adiponectin, quartiles
OR = 0.95 (0.63–1.44), Ptrend = 0.61
Heikkilä et al. (2009); United Cohort; M: CRP: 41, CRP: nephelometric assay CRP, continuous
Kingdom; British Women’s 897; IL-6: 30, 845; F: IL-6: ELISA M: HR = 0.89 (0.66–1.22), P = 0.5
Heart and Health Study, 32, 3074 F: HR = 0.97 (0.70–1.34), P = 0.8
Caerphilly Cohort IL-6, continuous
M: HR = 0.71 (0.41–1.23), P = 0.2
F: HR = 0.92 (0.53–1.60), P = 0.8
Chan et al. (2011); USA; Nested case–control; CRP: immunoturbidimetric CRP, quartiles
Nurses’ Health Study 280, 560; F assay RR = 0.65 (0.40–1.05), Ptrend = 0.17
IL-6 and sTNFR2: ELISA IL-6, quartiles
RR = 1.18 (0.75–1.85), Ptrend = 0.55
sTNFR2, quartiles
RR = 1.67 (1.05–2.68), Ptrend = 0.03
Aleksandrova et al. (2012a); Nested case–control; Multimeric ELISA Adiponectin, quintiles
several European countries; 1206, 1206; M&F OR = 0.71 (0.53–0.95), Ptrend = 0.03
EPIC
Aleksandrova et al. (2012b); Nested case–control; ELISA Leptin, quintiles
several European countries; 1129, 1129; M&F OR = 1.14 (0.81–1.61), Ptrend = 0.85
EPIC
599
Ho et al. (2012); USA; Nested case–cohort; Leptin, adiponectin, PAI-1, Leptin, quartiles
Women’s Health Initiative 457, 841; F resistin, HGF, and TNF-α: HR = 2.50 (1.70–3.67), Ptrend < 0.001
Observational Study multiplex assay Adiponectin, quartiles
IL-6: ultrasensitive solid- HR = 0.65 (0.45–0.94), Ptrend = 0.015
phase ELISA PAI-1, quartiles
HR = 1.87 (1.27–2.76), Ptrend = 0.006
Resistin, quartiles
HR = 1.16 (0.81–1.65), Ptrend = 0.329
HGF, quartiles
HR = 1.26 (0.87–1.82), Ptrend = 0.232
TNF-α, quartiles
HR = 0.97 (0.66–1.42), Ptrend = 0.969
IL-6, quartiles
HR = 1.41 (0.97–2.06), Ptrend = 0.043
Adjusted for insulin, HR = 1.04
(0.68–1.58), Ptrend = 0.662
Song et al. (2013); USA; Nested case–control; ELISA Adiponectin, quartiles
Nurses’ Health Study, Health 616, 1205; M&F M: RR = 0.55 (0.35–0.86), Ptrend = 0.02
Professionals Follow-up Study F: RR = 0.96 (0.67–1.39), Ptrend = 0.74
Ho et al. (2014); USA; Nested case–cohort; Milliplex Human Cytokine/ sIL-6R, quartiles
Women’s Health Initiative 433, 821; F Chemokine Panel RR = 0.56 (0.38–0.83), Ptrend = 0.007
Observational Study sIL-1R2, quartiles
RR = 0.44 (0.29–0.67); Ptrend < 0.001
IL-1Ra, sgp130, sTNFR1, sTNFR2: NS
Zhou et al. (2014); CRP: 18 Meta-analysis; CRP: NR CRP, 1 unit change in natural
studies; IL-6: 6 studies 4706 cases, IL-6: logarithm
1068 cases; M&F RR = 1.12 (1.05–1.21)
IL-6, 1 unit change in natural
logarithm
RR = 1.10 (0.88–1.36)
Murphy et al. (2015); USA; Nested case– Chemiluminescence CRP, quartiles
Women’s Health Initiative control; 401, 802; F immunometric assay OR = 0.89 (0.60–1.34), Ptrend = 0.47
Clinical Trial (postmenopausal)
ATBC, Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study; CI, confidence interval; CRP, C-reactive protein; ELISA, enzyme-linked
immunosorbent assay; EPIC, European Prospective Investigation into Cancer and Nutrition; F, female; HGF, hepatocyte growth factor; HR,
hazard ratio; IGF, insulin growth factor; IGFBP, IGF binding protein; IL, interleukin; IL-1Ra, IL-1 receptor antagonist; M, male; NR, not
reported; NS, no significant association; OR, odds ratio; PAI-1, plasminogen activator inhibitor 1; RR, relative risk; SHBG, sex hormone-binding
globulin; sTNFR, soluble tumour necrosis factor receptor; TNF-α, tumour necrosis factor alpha.
600
the prostate
Reference; country; Design; number of Assays Biomarker, categories; RR, highest vs

study cases, controls lowest (95% CI), Ptrend
Sex hormones
Gann et al. (1996); USA; Nested case– Testosterone, 3α-diolG, Estradiol, quartiles
Physicians’ Health Study control; 222, 390 DHT, and estradiol: OR = 0.56 (0.32–0.98), Ptrend = 0.03
radioimmunoassay Testosterone, quartiles
SHBG: radioimmunometric OR = 2.60 (1.34–5.02), Ptrend = 0.004
assay DHT, quartiles
OR = 0.71 (0.34–1.48), Ptrend = 0.30
3α-diolG, quartiles
OR = 1.60 (0.93–2.76), Ptrend = 0.09
SHBG, quartiles
OR = 0.46 (0.24–0.89), Ptrend = 0.01
Platz et al. (2005b); USA; Nested case– Testosterone: Total prostate cancer
Health Professionals control; 460, 460 chemiluminescence Testosterone, quartiles
Follow-up Study immunoassay OR = 0.79 (0.48–1.31), Ptrend = 0.79
SHBG: coated-tube SHBG, quartiles
non-competitive OR = 1.09 (0.66–1.82), Ptrend = 0.97
immunoradiometric assay Gleason score ≥ 7 (n = 148)
Testosterone, quartiles
OR = 0.26 (1.0–0.66), Ptrend = 0.01
SHBG, quartiles
OR = 2.72 (1.02–7.24), Ptrend = 0.05
Severi et al. (2006a); Case–cohort; 524, Testosterone: electrochemi Aggressive prostate cancer
Australia; Melbourne 1859 luminescence immunoassay Total testosterone, doubling of
Collaborative Cohort SHBG: immunometric assay concentration
Study DHEAS: competitive HR = 0.55 (0.32–0.95)
immunoassay Total testosterone, quartiles
Androstenedione: HR = 0.53 (0.28–1.03), Ptrend = 0.03
radioimmunoassay SHBG, quartiles
HR = 0.54 (0.28–1.04), Ptrend = 0.1
DHEAS, quartiles
HR = 0.38 (0.15–0.95), Ptrend = 0.005
Androstenedione, quartiles
HR = 0.46 (0.24–0.88), Ptrend = 0.007
Wirén et al. (2007); Nested case– Testosterone: coated-tube Total testosterone, quartiles
Sweden; Västerbotten control; 392, 392 radioimmunoassay OR = 1.02 (0.62–1.68), Ptrend = 0.83
Intervention Project SHBG: time-resolved Free testosterone, quartiles
immunofluorometric assay OR = 1.09 (0.67–1.78), Ptrend = 0.92
3α-diolG: direct SHBG, quartiles
OR = 0.92 (0.60–1.41), Ptrend = 1.00
Roddam et al. (2008); Pooled analysis; NR SHBG, quintiles
18 prospective studies 3886, 6438 RR = 0.86 (0.75–0.98), Ptrend = 0.01
Testosterone, calculated free testosterone,
DHT, DHEAS, androstenedione,
androstanediol glucuronide, estradiol,
calculated free estradiol: NS
601

Weiss et al. (2008); Nested case– Androstenedione and Androstenedione, quartiles
USA; Prostate, Lung, control; 727, 889 3α-diolG: direct double- OR = 0.96 (0.70–1.32), Ptrend = 0.76
Colorectal, and Ovarian antibody radioimmunoassay Testosterone, quartiles
Cancer Screening Trial Testosterone: direct OR = 1.39 (0.92–2.08), Ptrend = 0.22
radioimmunoassay Free testosterone, quartiles
SHBG: sandwich OR = 1.20 (0.87–1.65), Ptrend = 0.36
immunoradiometric assay SHBG, quartiles
OR = 0.76 (0.52–1.10), Ptrend = 0.22
OR = 0.87 (0.60–1.18), Ptrend = 0.31
Sawada et al. (2010); Nested case– Testosterone: electrochemi Total testosterone, quartiles
Japan; Japan Public control; 201, 402 luminescence immunoassay OR = 0.71 (0.36–1.41), Ptrend = 0.43
Health Center-based SHBG: immunoradiometric Free testosterone, quartiles
Prospective Study assay OR = 0.70 (0.39–1.27), Ptrend = 0.08
SHBG, quartiles
OR = 1.38 (0.69–2.77), Ptrend = 0.23
Hyde et al. (2012); Cohort; 297, 3338 Chemiluminescence Total testosterone, continuous
Australia; Health in Men immunoassay HR = 1.10 (0.97–1.25), P = 0.140
Study Free testosterone, continuous
HR = 1.13 (1.03–1.24), P = 0.013
SHBG, continuous
HR = 0.97 (0.84–1.11), P = 0.615
Insulin
Stattin et al. (2000); Nested case– Double-antibody Insulin, quartiles
Sweden; Northern control; 149, 298 radioimmunoassay OR = 0.98 (0.53–1.81), Ptrend = 0.23
Sweden Health and
Hubbard et al. (2004); Longitudinal study; Radioimmunoassay Fasting insulin, quartiles
USA; Baltimore 87, 823 RR = 0.72 (0.34–1.54), Ptrend = 0.56
Longitudinal Study of 2-Hour insulin, quartiles
Aging RR = 0.64 (0.32–1.31), Ptrend = 0.04
Stocks et al. (2007); Nested case– Immunoradiometric assay C-peptide, continuous
Sweden; Västerbotten control; 392, 392 OR = 0.96 (0.79–1.16), P = 0.65
Intervention Project
Albanes et al. (2009); Case–cohort; 100, Double-antibody immuno Insulin, quartiles
Finland; ATBC 400 chemiluminometric assay OR = 2.55 (1.18–5.51), Ptrend = 0.2
Schenk et al. (2009); Nested case– Multiplex sandwich ELISA C-peptide, quartiles
USA; Prostate Cancer control; 698, 709 OR = 0.80 (0.59–1.08), Ptrend = 0.31
Prevention Trial
Parekh et al. (2013); Cohort; 152, 1493 Radioimmunoassay Insulin, tertiles
USA; Framingham HR = 1.21 (0.78–1.88), Ptrend = 0.32
Heart Study-Offspring
Cohort
Lai et al. (2014); USA; Nested case– ELISA C-peptide, continuous
Health Professionals control; 1314, 1314 OR = 1.00 (0.93–1.08), Ptrend = 0.99
Follow-up Study C-peptide, quartiles
OR = 1.05 (0.83–1.33)
602

IGFs
Chan et al. (1998); USA; Nested case– ELISA IGF-1, quartiles
Physicians’ Health Study control; 152, 152 RR = 2.41 (1.23–4.74), Ptrend = 0.006
Stattin et al. (2000); Nested case– Double-antibody IGF-1, quartiles
Sweden; Northern control; 149, 298 immunoradiometric assay OR = 1.72 (0.93–3.19), Ptrend = 0.006
Sweden Health and < 59 yr, IGF-1, tertiles
Disease Study Cohort OR = 4.30 (1.19–15.50), Ptrend = 0.01
Woodson et al. (2003); Case–cohort; 100, ELISA IGF-1, quartiles
Finland; ATBC 400 OR = 1.00 (0.54–1.87), Ptrend = 0.76
Stattin et al. (2004c); Nested case– Immunoradiometric assay IGF-1, quartiles
Sweden; Northern control; 281, 560 OR = 1.67 (1.02–2.71), Ptrend = 0.05
Sweden Health and
Meyer et al. (2005); Nested case– Chemiluminescence IGF-1, quartiles
France; Supplémentation control; 100, 400 immunoassay on an Immulite OR = 1.80 (0.76–4.27), Ptrend = 0.13
en Vitamines et analyser
Minéraux Antioxydants
Trial
Platz et al. (2005b); USA; Nested case– ELISA IGF-1, quartiles
Follow-up Study
Severi et al. (2006b); Case–cohort; 524, ELISA IGF-1, quartiles
Australia; Melbourne 1826 HR = 1.07 (0.79–1.46), Ptrend = 0.5
Collaborative Cohort
Study
Allen et al. (2007); Nested case– ELISA plus acid–ethanol IGF-1, tertiles
several European control; 630, 630 precipitation OR = 1.35 (0.99–1.82), Ptrend = 0.08
countries; EPIC Adjusted for IGFBP-3, OR = 1.39 (1.02–1.89),
Ptrend = 0.12
Weiss et al. (2007); Nested case– ELISA IGF-1, quartiles
USA; Prostate, Lung, control; 727, 887 OR = 1.14 (0.86–1.51), Ptrend = 0.18
Colorectal, and Ovarian
Cancer Screening Trial
Mucci et al. (2010); USA; Nested case– ELISA Free IGF-1, quartiles
Physicians’ Health Study control; 545, 545 RR = 0.9 (0.6–1.3), Ptrend = 0.78
Price et al. (2012); several Nested case– DSL-10-5600 ACTIVE ELISA IGF-1, quartiles
European countries; control; 1542, 1542 or IDS-iSYS immunoassay OR = 1.69 (1.35–2.13), Ptrend = 0.0002
EPIC system IGF-1, doubling
OR = 1.38 (1.17–1.64), Ptrend = 0.0002
603

Travis et al. (2016); 17 Pooled analysis; NR Prospective studies
prospective and 2 cross- 10 554, 13 618 IGF-1, quintiles
sectional studies OR = 1.29 (1.16–1.43), Ptrend < 0.001
IGF-2, quintiles
OR = 1.20 (1.00–1.43), Ptrend = 0.038
IGFBP-1, quintiles
OR = 0.81 (0.68–0.96), Ptrend = 0.053
IGFBP-2, quintiles
OR = 1.26 (1.03–1.54), Ptrend < 0.001
IGFBP-3, quintiles
OR = 1.25 (1.12–1.40), Ptrend < 0.001
Stocks et al. (2007); Nested case– Double-antibody Leptin, continuous
Sweden; Västerbotten control; 392, 392 radioimmunoassay OR = 0.93 (0.89–0.97), P = 0.002
Intervention Project
Heikkilä et al. (2009); Cohort; CRP: 36, CRP: nephelometric assay CRP, continuous
United Kingdom; 897; IL-6: 40, 845 IL-6: ELISA HR = 1.12 (0.81–1.56), P = 0.5
Caerphilly Cohort IL-6, continuous
HR = 0.61 (0.40–0.96), P = 0.031
Schenk et al. (2009); Nested case– Multiplex sandwich ELISA Leptin, quartiles
USA; Prostate Cancer control; 698, 709 OR = 1.05 (0.73–1.50), Ptrend = 0.48
Prevention Trial Adiponectin, quartiles
OR = 0.65 (0.47–0.87), Ptrend = 0.004
Li et al. (2010); USA; Nested case– Competitive Leptin, quartiles
Physicians’ Health Study control; 654, 644 radioimmunoassay RR = 1.06 (0.65–1.72), Ptrend = 0.8
Adiponectin, quartiles
RR = 0.73 (0.46–1.14), Ptrend = 0.38
Touvier et al. (2013); Nested case– ELISA Leptin, quartiles
France; Supplémentation control; 156, 312 OR = 0.69 (0.27–1.75), Ptrend = 0.9
en Vitamines et Adiponectin, quartiles
Minéraux Antioxydants OR = 1.34 (0.69–2.61), Ptrend = 0.3
Trial hsCRP, quartiles
OR = 2.52 (1.18–5.39), Ptrend = 0.03
Lai et al. (2014); USA; Nested case– ELISA Leptin, continuous
Follow-up Study
3α-diolG, 5α-androstane-3α,17β-diol glucuronide; ATBC, Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study; CI, confidence interval;
CRP, C-reactive protein; DHEAS, dehydroepiandrosterone sulfate; DHT, dihydrotestosterone; ELISA, enzyme-linked immunosorbent assay;
EPIC, European Prospective Investigation into Cancer and Nutrition; HR, hazard ratio; hsCRP, high-sensitivity C-reactive protein; IGF, insulin
growth factor; IGFBP, IGF binding protein; IL, interleukin; NR, not reported; NS, no significant association; OR, odds ratio; RR, relative risk;
SHBG, sex hormone-binding globulin; yr, year or years.
604
the ovary
Reference; country; study Design; number of Assays Biomarker, categories; RR, highest vs
cases, controls lowest (95% CI), Ptrend
Sex hormones
Helzlsouer et al. (1995); USA; Nested case–control; Estrone and estradiol: Estrone, estradiol, progesterone: NS
population-based serum 31, 62 solvent extraction,
bank celite chromatography,
radioimmunoassay
Progesterone:
radioimmunoassay
Lukanova et al. (2003a); Nested case–control; Estrone: double-antibody Estrone, quartiles
USA, Sweden, Italy; New 132, 258; radioimmunoassay OR = 1.15 (0.47–2.82), Ptrend = 0.47
York University Women’s postmenopausal SHBG: immunoradiometric SHBG, quartiles
Health Study, Northern women assay OR = 1.66 (0.67–4.09), Ptrend < 0.19
Sweden Health and Disease
Study, ORDET Cohort
Rinaldi et al. (2007); several Nested case–control; Sandwich immunoradio SHBG, continuous log2 scale
European countries; EPIC 192, 346 metric assay All cases: NS
BMI < 26.8, OR = 0.31 (0.14–0.68)
BMI ≥ 26.8, OR = 2.48 (1.31–4.71)
P heterogeneity = 0.0001
Trabert et al. (2016); USA; Nested case–control; Stable-isotope dilution liquid Estrone, quintiles
Women’s Health Initiative 169, 412 chromatography-tandem OR = 1.54 (0.82–2.90), Ptrend = 0.05
mass spectrometry 2-Methoxyestrone metabolites,
quintiles
OR = 2.03 (1.06–3.88), Ptrend = 0.02
4-Methoxyestrone metabolites,
quintiles
OR = 1.86 (0.98–3.56), Ptrend = 0.01
Insulin
Lukanova et al. (2003b); Nested case–control; Radioimmunoassay C-peptide, quartiles
USA, Sweden, Italy; New 132, 263 OR = 0.89 (0.44–1.81), Ptrend = 0.92
York University Women’s
Health Study, Northern
Study, ORDET Cohort
IGFs
Lukanova et al. (2002); USA, Nested case–control; Peptides: double-antibody IGF-1, tertiles
Sweden, Italy; New York 132, 263 immunoradiometric assay All cases: NS
University Women’s Health IGF-1: acid–ethanol < 55 yr, OR = 4.97 (1.22–20.2)
Study, Northern Sweden precipitation of IGFBPs IGFBP-3, tertiles
Health and Disease Study, All cases: NS
ORDET Cohort
Lukanova et al. (2003b); Nested case–control; IGFBP-1: immunoradio IGFBP-1, quartiles
USA, Sweden, Italy; New 132, 263 metric assay OR = 0.79 (0.38–1.62)
York University Women’s IGFBP-2: radioimmunoassay IGFBP-2, quartiles
Health Study, Northern OR = 0.87 (0.45–1.68)
Study, ORDET Cohort
605
Peeters et al. (2007); several Nested case–control; Peptides: ELISA IGF-1, tertiles
European countries; EPIC 214, 388 IGF-1: acid–ethanol All, OR = 1.1 (0.7–1.7), Ptrend = 0.94
precipitation of IGFBPs Diagnosis ≤ 55 yr, OR = 2.4 (0.9–6.4),
Ptrend = 0.08
Diagnosis > 55 yr, OR = 0.9 (0.5–1.6),
Ptrend = 0.74
IGFBP-3, tertiles
All, OR = 1.1 (0.7–1.8), Ptrend = 0.65
Diagnosis ≤ 55 yr, OR = 2.1 (0.8–5.4),
Ptrend = 0.12
Diagnosis > 55 yr, OR = 1.0 (0.6–1.7),
Ptrend = 0.91
Tworoger et al. (2007b); Nested case–control; ELISA after acid extraction IGF-1, quartiles
USA; Nurses’ Health Study, 222, 599 RR = 0.56 (0.32–0.97), Ptrend = 0.14
Nurses’ Health Study II, IGFBP-2, IGFBP-3, IGF-1 ratio to
Women’s Health Study IGFBPs: NS
Lundin et al. (2009); USA, Nested case–control; High-sensitivity CRP, > 10 vs ≤ 1 mg/L
Sweden, Italy; New York 237, 427 immunoturbidimetric assay All, OR = 4.4 (1.8–10.9)
University Women’s Health Diagnosis > 2 yr after blood donation,
Study, Northern Sweden OR = 3.0 (1.2–8.0)
Health and Disease Study, Diagnosis > 5 yr after blood donation,
ORDET Cohort OR = 3.6 (1.0–13.2)
Clendenen et al. (2011); USA, Nested case–control; Luminex xMAP technology IL-2, quartiles
Sweden, Italy; New York 230, 432 OR = 1.57 (0.98–2.52), Ptrend = 0.07
University Women’s Health IL-4, quartiles
Study, Northern Sweden OR = 1.50 (0.95–2.38), Ptrend = 0.06
Health and Disease Study, IL-6, quartiles
ORDET Cohort OR = 1.63 (1.03–2.58), Ptrend = 0.03
IL-12p40, quartiles
OR = 1.60 (1.02–2.51), Ptrend = 0.06
IL-13, quartiles
OR = 1.42 (0.90–2.26), Ptrend = 0.11
Poole et al. (2013); USA; Nested case–control; CRP: validated CRP
Nurses’ Health Study, Nurses’ Health immunoturbidimetric Quartiles, RR = 0.53 (1.05–2.23),
Nurses’ Health Study II, Studies: 217, 434; method Ptrend = 0.01
Women’s Health Study Women’s Health IL-6: quantitative sandwich > 10 vs ≤ 1 mg/L, RR = 2.16 (1.23–3.78)
Study: 159, NR enzyme immunoassay IL-6, TNF-α-R2, Nurses’ Health
TNF-α-R2: ELISA Studies: NS
606
Trabert et al. (2014); USA; Nested case–control; Luminex bead-based assay CRP, tertiles
Prostate, Lung, Colorectal, 149, 149 OR = 2.04 (1.06–3.93), Ptrend = 0.03
and Ovarian Cancer IL-1α, detectable vs undetectable
Screening Trial OR = 2.23 (1.14–4.34)
TNF-α, tertiles
OR = 2.21 (1.06–4.63), Ptrend = 0.04
IL-8, tertiles
OR = 1.86 (0.96–3.61), Ptrend = 0.05
Serous ovarian cancer (n = 83)
CRP, tertiles
OR = 3.96 (1.14–11.14), Ptrend = 0.008
IL-8, tertiles
OR = 3.05 (1.09–8.51), Ptrend = 0.03
Ose et al. (2015); several Nested case–control; CRP: high-sensitivity CRP
European countries; EPIC 754, 1497 immunoassay All cases: NS
IL-6: high-sensitivity > 10 vs ≤ 1 mg/L, OR = 1.67 (1.03–2.70)
quantitative sandwich IL-6
enzyme immunoassay All cases: NS
Waist circumference ≤ 80 cm,
OR log2 = 0.97 (0.81–1.16)
Waist circumference 80–88 cm,
OR log2 = 0.85 (0.66–1.11)
Waist circumference > 88 cm,
OR log2 = 1.78 (1.28–2.48)
P heterogeneity ≤ 0.01
BMI, body mass index; CI, confidence interval; CRP, C-reactive protein; ELISA, enzyme-linked immunosorbent assay; EPIC, European
Prospective Investigation into Cancer and Nutrition; IGF, insulin growth factor; IGFBP, IGF binding protein; IL, interleukin; NR, not reported;
NS, no significant association; OR, odds ratio; RR, relative risk; SHBG, sex hormone-binding globulin; TNF-α, tumour necrosis factor alpha;
TNF-α-R, tumour necrosis factor alpha receptor; yr, year or years.
607
the liver (including the biliary tract)

study cases, controls; sex lowest (95% CI), Ptrend
Sex hormones
Lukanova et al. (2014); Nested case–control; 125, Radioimmunoassay SHBG, tertiles
several European 247; M&F HCC: OR = 6.64 (2.58–17.1), Ptrend < 0.001
countries; EPIC
Insulin
Chao et al. (2011); Case–cohort; 124, 1084; Radioimmunoassay Insulin, > 6.10 vs 2.75–4.10 μU/mL
Taiwan, China; M HCC: HR = 2.36 (1.43–3.90), Ptrend = 0.014
Hepatitis B virus-
positive Cohort
Aleksandrova et Nested case–control; Immulite 2000 C-peptide, tertiles
al. (2014); several HCC: 125, 250; IBDC: 34, HCC: RR = 3.13 (1.20–8.12), Ptrend = 0.009
European countries; 68; GBTC: 137, 274; M&F IBDC: RR = 9.89 (1.21–80.45), Ptrend = 0.03
EPIC GBTC: RR = 0.77 (0.39–1.52), Ptrend = 0.58
IGFs
Mazziotti et al. (2002); Cohort; 20, 84; M&F Immunoradiometric assay IGF-1
Italy; Hepatitis C virus- HCC: significantly lower levels, P < 0.001
related cirrhosis Cohort
Lukanova et al. (2014); Nested case–control; 125, ELISA IGF-1, tertiles
several European 247; M&F HCC: OR = 0.21 (0.09–0.50), Ptrend < 0.001
countries; EPIC
Aleksandrova et Nested case–control; CRP: Turbidimetric CRP, tertiles
al. (2014); several HCC: 125, 250; IBDC: 34, Modular system HCC: RR = 1.41 (0.67–2.96), Ptrend = 0.05
European countries; 68; GBTC: 137, 274; M&F Leptin and adiponectin: IBDC: RR = 3.92 (0.78–19.68), Ptrend = 0.05
EPIC ELISA GBTC: RR = 2.26 (1.26–4.07), Ptrend = 0.009
IL-6: ECLIA Modular Leptin, tertiles
system HCC: RR = 1.18 (0.43–3.26), Ptrend = 0.94
IBDC: RR = 3.73 (0.36–38.47), Ptrend = 0.14
GBTC: RR = 0.52 (0.24–1.13), Ptrend = 0.05
HCC: RR = 1.50 (0.69–3.28), Ptrend = 0.29
IBDC: RR = 0.42 (0.11–1.29), Ptrend = 0.23
GBTC: RR = 1.82 (0.93–3.53), Ptrend = 0.04
IL-6, tertiles
HCC: RR = 3.85 (1.31–11.38), Ptrend = 0.004
IBDC: RR = 1.87 (0.43–8.12), Ptrend = 0.22
GBTC: RR = 1.19 (0.54–2.62), Ptrend = 0.68
Ohishi et al. (2014); Nested case–control; 188, CRP: autoanalyser and CRP, tertiles
Japan; Adult Health 605; M&F high-sensitivity assay kit HCC: RR = 1.94 (0.72–5.51)
Study Cohort IL-6: multiplex bead array IL-6, tertiles
assay HCC: RR = 5.12 (1.54–20.1)
CI, confidence interval; CRP, C-reactive protein; ELISA, enzyme-linked immunosorbent assay; EPIC, European Prospective Investigation into
Cancer and Nutrition; F, female; GBTC, gall bladder and biliary tract cancers outside of the liver; HCC, hepatocellular carcinoma; HR, hazard
ratio; IBDC, intrahepatic bile duct cancer; IGF, insulin growth factor; IL, interleukin; M, male; OR, odds ratio; RR, relative risk; SHBG, sex
hormone-binding globulin.
608
the pancreas
Insulin
Stolzenberg-Solomon et al. (2005); Case–cohort; 169, 2-site Insulin, quartiles
Finland; ATBC 400; M immunoenzymatic HR = 2.01 (1.03–3.93), Ptrend = 0.03
assay
Grote et al. (2011); several European Nested case–control; Double-antibody C-peptide, quartiles
countries; EPIC 466, 466; M&F radioimmunoassay OR = 1.15 (0.70–1.91), Ptrend = 0.886
Wolpin et al. (2013); USA; Nested case–control; NR Insulin, quintiles
5 prospective studies 449, 982; M&F OR = 1.57 (1.08–2.30), Ptrend = 0.002
Proinsulin, quintiles
OR = 2.22 (1.50–3.29), Ptrend < 0.001
IGFs
Stolzenberg-Solomon et al. (2004); Case–cohort; 93, 400; ELISA IGF-1, tertiles
Finland; ATBC M OR = 0.67 (0.37–1.21), Ptrend = 0.17
Wolpin et al. (2007); USA; Nested case–control; ELISA IGF-1, quartiles
4 prospective studies 212, 635; M&F OR = 0.94 (0.60–1.48), Ptrend = 0.97
IGF-2, quartiles
OR = 0.96 (0.61–1.52), Ptrend = 0.93
Douglas et al. (2010); USA; Prostate, Nested case–control; ELISA IGF-1, quartiles
Lung, Colorectal, and Ovarian Cancer 187, 374; M&F OR = 1.58 (0.91–2.76), Ptrend = 0.25
Screening Trial IGF-2, quartiles
OR = 0.86 (0.49–1.50), Ptrend = 0.31
Rohrmann et al. (2012); several Nested case–control; ELISA IGF-1, quartiles
European countries; EPIC 422, 422; M&F OR = 1.21 (0.75–1.93), Ptrend = 0.30
Stolzenberg-Solomon et al. (2008); Case–cohort; 311, ELISA Adiponectin, quintiles
Finland; ATBC 510; M OR = 0.65 (0.39–1.07), Ptrend = 0.04
Grote et al. (2012a); several European Nested case–control; CRP: multiplex CRP, quartiles
countries; EPIC 455, 455; M&F immunoassay OR = 1.02 (0.66–1.57), Ptrend = 0.6
IL-6: ELISA IL-6, quartiles
OR = 1.01 (0.64–1.61), Ptrend = 0.7
Grote et al. (2012b); several European Nested case–control; Multiplex Adiponectin, quartiles
countries; EPIC 452, 452; M&F immunoassay OR = 1.10 (0.69–1.75), Ptrend = 0.71
Bao et al. (2013a); USA; 5 prospective Nested case–control; NR CRP, quintiles
studies 470, 1094; M&F OR = 1.10 (0.74–1.63), Ptrend = 0.81
IL-6, quintiles
OR = 1.19 (0.81–1.76), Ptrend = 0.08
Bao et al. (2013b); USA; 5 prospective Nested case–control; ELISA Adiponectin, quintiles
studies 468, 1080; M&F OR = 0.63 (0.43–0.92), Ptrend = 0.01
Stolzenberg-Solomon et al. (2015); Nested case–control; ELISA Leptin, quintiles
USA, Finland; 3 prospective studies 731, 909; M&F OR = 1.13 (0.75–1.71), Ptrend = 0.38
ATBC, Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study; CI, confidence interval; CRP, C-reactive protein; ELISA, enzyme-linked
immunosorbent assay; EPIC, European Prospective Investigation into Cancer and Nutrition; F, female; HR, hazard ratio; IGF, insulin growth
factor; IL, interleukin; M, male; NR, not reported; OR, odds ratio; RR, relative risk.
609
the stomach
cases, controls; sex lowest (95% CI), Ptrend
Insulin
Hidaka et al. (2015); Japan; Nested case– Human Endocrine Insulin, tertiles
Japan Public Health Center- control; 77, 477; Milliplex kit OR = 1.91 (1.15–3.18), Ptrend = 0.01
based Prospective Study M&F C-peptide, tertiles
OR = 1.31 (0.82–2.11), Ptrend = 0.26
IGFs
Yatsuya et al. (2005); Japan; Nested case– Immunoradiometric assay IGF-1, mean cases/controls ± SD
Japan Collaborative Cohort control; 210, 410; M: 127 ± 52 vs 131 ± 54 ng/mL,
Study M&F P = 0.70
F: 121 ± 53 vs 117 ± 53 ng/mL, P = 0.41
IGF-2, mean cases/controls ± SD
M: 548 ± 127.4 vs 571 ± 139.2 ng/mL,
P = 0.13
F: 618 ± 122 vs 607 ± 118 ng/mL,
P = 0.40
Wong et al. (2011); China; Nested case– LINCOplex kit IL-6, > 4.06 vs < 1.76 pg/mL
Shanghai Women’s Health Study control; 141, 282; F OR = 1.73 (1.00–3.00), Ptrend = 0.04
TNF-α, > 7.17 vs < 4.86 pg/mL
OR = 0.74 (0.42–1.30), Ptrend = 0.27
Epplein et al. (2013); China; Nested case– Milliplex MAP high- IL-8, quartiles
Shanghai Men’s Health Study control; 180, 358; M sensitivity Human OR = 2.30 (1.26–4.19), Ptrend = 0.008
Cytokine Magnetic Bead TNF-α, quartiles
Panel assay kit OR = 1.37 (0.77–2.44), Ptrend = 0.22
CI, confidence interval; ELISA, enzyme-linked immunosorbent assay; F, female; IGF, insulin growth factor; IL, interleukin; M, male; OR, odds
ratio; RR, relative risk; SD. standard deviation; TNF-α, tumour necrosis factor alpha.
Table 4.10 Molecular epidemiological studies of obesity–cancer related mechanisms for cancer
of the kidney
Reference; country; study Design; number of cases, Assays Biomarker, categories; RR, highest vs
controls; sex lowest (95% CI), Ptrend
IGFs
Major et al. (2010); Finland; ATBC Nested case–control; 100, 400; M ELISA IGF-1, quartiles
OR = 0.40 (0.18–0.90), Ptrend = 0.03
Liao et al. (2013); Finland; ATBC Nested case–control; 273, 273; M ELISA Leptin, continuous
OR = 0.93 (0.84–1.03)
Adiponectin, continuous
OR = 0.87 (0.78–0.97)
Resistin, continuous
OR = 1.04 (0.94–1.16)
ATBC, Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study; CI, confidence interval; ELISA, enzyme-linked immunosorbent assay; IGF,
insulin growth factor; M, male; OR, odds ratio; RR, relative risk.
610
Table 4.11 Molecular epidemiological studies of obesity–cancer related mechanisms for cancer
of the oesophagus
Reference; country; study Design; number of cases, Assays Biomarker, categories; RR,
controls; sex highest vs lowest (95% CI), Ptrend
Hardikar et al. (2014); USA; Case–cohort; CRP: 43, 386; CRP: immunonephelometric CRP, quartiles
Seattle Barrett’s Esophagus IL-6: 45, 394; M&F assay HR = 1.55 (0.56–4.24), Ptrend = 0.04
Study IL-6: ELISA IL-6, quartiles
HR = 1.17 (0.42–3.26), Ptrend = 0.87
Keeley et al. (2014); Islamic Nested case–control; 36, Luminex xMAP multiplex Interferon-γ, quartiles
Republic of Iran; Golestan 81; M&F assay OR = 5 (1.87–13.36)
Cohort Study TNF-α, quartiles
OR = 8.2 (2.66–25.31)
CI, confidence interval; CRP, C-reactive protein; ELISA, enzyme-linked immunosorbent assay; F, female; HR, hazard ratio; IL, interleukin; M,
male; OR, odds ratio; RR, relative risk; TNF-α, tumour necrosis factor alpha.
References Hepatology,
PMID:24443059
60(3):858–71. doi:10.1002/hep.27016
Allen NE, Appleby PN, Kaaks R, Rinaldi S, Davey GK, Key

Adli M, Baldwin AS (2006). IKK-i/IKKε controls consti- TJ (2003). Lifestyle determinants of serum insulin-like
tutive, cancer cell-associated NF-κB activity via regu- growth-factor-I (IGF-I), C-peptide and hormone
lation of Ser-536 p65/RelA phosphorylation. J Biol binding protein levels in British women. Cancer Causes
Chem, 281(37):26976–84. doi:10.1074/jbc.M603133200 Control, 14(1):65–74. doi:10.1023/A:1022518321634
PMID:16840782 PMID:12708727
Ahmad M, Hamid A, Hussain A, Majeed R, Qurishi Y, Bhat Allen NE, Key TJ, Appleby PN, Travis RC, Roddam AW,
JA, et al. (2012). Understanding histone deacetylases Rinaldi S, et al. (2007). Serum insulin-like growth factor
in the cancer development and treatment: an epige- (IGF)-I and IGF-binding protein-3 concentrations
netic perspective of cancer chemotherapy. DNA Cell and prostate cancer risk: results from the European
Biol, 31(Suppl 1):S62–71. doi:10.1089/dna.2011.1575 Prospective Investigation into Cancer and Nutrition.
PMID:22462686 Cancer Epidemiol Biomarkers Prev, 16(6):1121–7.
Albanes D, Weinstein SJ, Wright ME, Männistö S, doi:10.1158/1055-9965.EPI-06-1062 PMID:17548673
Limburg PJ, Snyder K, et al. (2009). Serum insulin, Allen NE, Key TJ, Dossus L, Rinaldi S, Cust A, Lukanova
glucose, indices of insulin resistance, and risk of A, et al. (2008). Endogenous sex hormones and
prostate cancer. J Natl Cancer Inst, 101(18):1272–9. endometrial cancer risk in women in the European
doi:10.1093/jnci/djp260 PMID:19700655 Prospective Investigation into Cancer and Nutrition
Aleksandrova K, Boeing H, Jenab M, Bueno-de-Mesquita (EPIC). Endocr Relat Cancer, 15(2):485–97.
HB, Jansen E, van Duijnhoven FJ, et al. (2012a). doi:10.1677/ERC-07-0064 PMID:18509001
Total and high-molecular weight adiponectin and Allen NE, Roddam AW, Allen DS, Fentiman IS, Dos
risk of colorectal cancer: the European Prospective Santos Silva I, Peto J, et al. (2005). A prospective
Investigation into Cancer and Nutrition study. study of serum insulin-like growth factor-I (IGF-I),
Carcinogenesis, 33(6):1211–8. doi:10.1093/carcin/ IGF-II, IGF-binding protein-3 and breast cancer risk.
bgs133 PMID:22431719 Br J Cancer, 92(7):1283–7. doi:10.1038/sj.bjc.6602471
Aleksandrova K, Boeing H, Jenab M, Bueno-de-Mesquita PMID:15756268
HB, Jansen E, van Duijnhoven FJ, et al. (2012b). Leptin Allin KH, Bojesen SE, Nordestgaard BG (2009). Baseline
and soluble leptin receptor in risk of colorectal cancer C-reactive protein is associated with incident cancer
in the European Prospective Investigation into Cancer and survival in patients with cancer. J Clin Oncol,
and Nutrition cohort. Cancer Res, 72(20):5328–37. 27(13):2217–24. doi :10.12 0 0/ JCO. 2 0 0 8 .19. 8 4 4 0
doi:10.1158/0008-5472.CAN-12-0465 PMID:22926557 PMID:19289618
Aleksandrova K, Boeing H, Nöthlings U, Jenab M, Fedirko Almén MS, Jacobsson JA, Moschonis G, Benedict C,
V, Kaaks R, et al. (2014). Inflammatory and metabolic Chrousos GP, Fredriksson R, et al. (2012). Genome
biomarkers and risk of liver and biliary tract cancer. wide analysis reveals association of a FTO gene variant
611
with epigenetic changes. Genomics, 99(3):132–7. Barile G, Sica G, Montemurro A, Iacobelli S, Corradini
doi:10.1016/j.ygeno.2011.12.007 PMID:22234326 M (1979). Levels of estrogen and progesterone receptor
Amankwah EK, Friedenreich CM, Magliocco AM, in human endometrium during the menstrual
Brant R, Courneya KS, Speidel T, et al. (2013). cycle. Eur J Obstet Gynecol Reprod Biol, 9(4):243–6.
Anthropometric measures and the risk of endometrial doi:10.1016/0028-2243(79)90062-5 PMID:264091
cancer, overall and by tumor microsatellite status and Baron JA, Barry EL, Mott LA, Rees JR, Sandler RS, Snover
histological subtype. Am J Epidemiol, 177(12):1378–87. DC, et al. (2015). A trial of calcium and vitamin D
doi:10.1093/aje/kws434 PMID:23673247 for the prevention of colorectal adenomas. N Engl J
Appleton GV, Wheeler EE, Al-Mufti R, Challacombe DN, Med, 373(16):1519–30. doi:10.1056/NEJMoa1500409
Williamson RC (1988). Rectal hyperplasia after jejun- PMID:26465985
oileal bypass for morbid obesity. Gut, 29(11):1544–8. Bauer SR, Hankinson SE, Bertone-Johnson ER, Ding EL
doi:10.1136/gut.29.11.1544 PMID:3209111 (2013). Plasma vitamin D levels, menopause, and risk of
Argenta PA, Kassing M, Truskinovsky AM, Svendsen CA breast cancer: dose-response meta-analysis of prospec-
(2013). Bariatric surgery and endometrial pathology in tive studies. Medicine (Baltimore), 92(3):123–31.
asymptomatic morbidly obese women: a prospective, doi:10.1097/MD.0b013e3182943bc2 PMID:23625163
pilot study. BJOG, 120(7):795–800. doi:10.1111/1471- Bellamy TC, Wood J, Garthwaite J (2002). On the acti-
0528.12100 PMID:23231632 vation of soluble guanylyl cyclase by nitric oxide.
Arunabh S, Pollack S, Yeh J, Aloia JF (2003). Body fat Proc Natl Acad Sci USA, 99(1):507–10. doi:10.1073/
content and 25-hydroxyvitamin D levels in healthy pnas.012368499 PMID:11752394
women. J Clin Endocrinol Metab, 88(1):157–61. Beral V, Bull D, Reeves G; Million Women Study
doi:10.1210/jc.2002-020978 PMID:12519845 Collaborators (2005). Endometrial cancer and
Aslibekyan S, Demerath EW, Mendelson M, Zhi D, Guan hormone-replacement therapy in the Million Women
W, Liang L, et al. (2015). Epigenome-wide study iden- Study. Lancet, 365(9470):1543–51. doi:10.1016/S0140-
tifies novel methylation loci associated with body mass 6736(05)66455-0 PMID:15866308
index and waist circumference. Obesity (Silver Spring), Beral V, Doll R, Hermon C, Peto R, Reeves G; Collaborative
23(7):1493–501. doi:10.1002/oby.21111 PMID:26110892 Group on Epidemiological Studies of Ovarian Cancer
Azrad M, Chang PL, Gower BA, Hunter GR, Nagy TR (2008). Ovarian cancer and oral contraceptives:
(2011). Reduced mitogenicity of sera following weight collaborative reanalysis of data from 45 epidemio-
loss in premenopausal women. Nutr Cancer, 63(6):916– logical studies including 23,257 women with ovarian
23. doi:10.1080/01635581.2011.594209 PMID:21774593 cancer and 87,303 controls. Lancet, 371(9609):303–14.
Baatar D, Patel K, Taub DD (2011). The effects of doi:10.1016/S0140-6736(08)60167-1 PMID:18294997
ghrelin on inflammation and the immune system. Beral V, Gaitskell K, Hermon C, Moser K, Reeves G, Peto
Mol Cell Endocrinol, 340(1):44–58. doi:10.1016/j. R; Collaborative Group on Epidemiological Studies
mce.2011.04.019 PMID:21565248 of Ovarian Cancer (2015). Menopausal hormone
Baglietto L, English DR, Hopper JL, Morris HA, Tilley use and ovarian cancer risk: individual participant
WD, Giles GG (2007). Circulating insulin-like growth meta-analysis of 52 epidemiological studies. Lancet,
factor-I and binding protein-3 and the risk of breast 385(9980):1835–42. doi:10.1016/S0140-6736(14)61687-1
cancer. Cancer Epidemiol Biomarkers Prev, 16(4):763–8. PMID:25684585
doi:10.1158/1055-9965.EPI-06-0960 PMID:17416768 Berardis S, Sokal E (2014). Pediatric non-alcoholic fatty
Baglietto L, Severi G, English DR, Krishnan K, Hopper liver disease: an increasing public health issue. Eur J
JL, McLean C, et al. (2010). Circulating steroid Pediatr, 173(2):131–9. doi:10.1007/s00431-013-2157-6
hormone levels and risk of breast cancer for postmeno- PMID:24068459
pausal women. Cancer Epidemiol Biomarkers Prev, Bezemer ID, Rinaldi S, Dossus L, Gils CH, Peeters PH,
19(2):492–502. doi:10.1158/1055-9965.EPI-09-0532 Noord PA, et al. (2005). C-peptide, IGF-I, sex-steroid
PMID:20086116 hormones and adiposity: a cross-sectional study in
Bao Y, Giovannucci EL, Kraft P, Qian ZR, Wu C, Ogino healthy women within the European Prospective
S, et al. (2013a). Inflammatory plasma markers and Investigation into Cancer and Nutrition (EPIC). Cancer
pancreatic cancer risk: a prospective study of five Causes Control, 16(5):561–72. doi:10.1007/s10552-004-
U.S. cohorts. Cancer Epidemiol Biomarkers Prev, 7472-9 PMID:15986111
22(5):855–61. doi:10.1158/1055-9965.EPI-12-1458 Björntorp P, Bergman H, Varnauskas E (1969). Plasma
PMID:23462920 free fatty acid turnover rate in obesity. Acta Med Scand,
Bao Y, Giovannucci EL, Kraft P, Stampfer MJ, Ogino 185(4):351–6. PMID:5806343
S, Ma J, et al. (2013b). A prospective study of plasma Blüher M, Mantzoros CS (2015). From leptin to other
adiponectin and pancreatic cancer risk in five US adipokines in health and disease: facts and expecta-
cohorts. J Natl Cancer Inst, 105(2):95–103. doi:10.1093/ tions at the beginning of the 21st century. Metabolism,
jnci/djs474 PMID:23243202
612
64(1):131–45. doi:10.1016/j.metabol.2014.10.016 exercise-induced weight loss. Cancer Prev Res (Phila),

PMID:25497344 6(3):217–31. doi:10.1158/1940-6207.CAPR-12-0212
Blum A, Ginat-Maimon L, Yehuda H, Geron N, Ben Ami PMID:23341572
M, Tamir S (2015). Inhibition of inflammation may Campbell PT, Jacobs ET, Ulrich CM, Figueiredo JC,
enhance nitric oxide availability in patients under- Poynter JN, McLaughlin JR, et al.; Colon Cancer
going bariatric surgery for weight loss. J Intern Med, Family Registry (2010). Case-control study of over-
278(4):401–9. doi:10.1111/joim.12379 PMID:26123268 weight, obesity, and colorectal cancer risk, overall
Bouchard L, Rabasa-Lhoret R, Faraj M, Lavoie ME, Mill and by tumor microsatellite instability status. J Natl
J, Pérusse L, et al. (2010). Differential epigenomic and Cancer Inst, 102(6):391–400. doi:10.1093/jnci/djq011
transcriptomic responses in subcutaneous adipose PMID:20208017
tissue between low and high responders to caloric Cani PD, Amar J, Iglesias MA, Poggi M, Knauf C, Bastelica
restriction. Am J Clin Nutr, 91(2):309–20. doi:10.3945/ D, et al. (2007). Metabolic endotoxemia initiates
ajcn.2009.28085 PMID:19939982 obesity and insulin resistance. Diabetes, 56(7):1761–72.
Bozic I, Antal T, Ohtsuki H, Carter H, Kim D, Chen S et al. doi:10.2337/db06-1491 PMID:17456850
(2010). Accumulation of driver and passenger muta- Cani PD, Bibiloni R, Knauf C, Waget A, Neyrinck AM,
tions during tumor progression. Proc Natl Acad Sci Delzenne NM, et al. (2008). Changes in gut microbiota
USA, 107(43):18545–50. doi:10.1073/pnas.1010978107 control metabolic endotoxemia-induced inflamma-
PMID:20876136 tion in high-fat diet-induced obesity and diabetes in
Brändstedt J, Almquist M, Manjer J, Malm J (2012). mice. Diabetes, 57(6):1470–81. doi:10.2337/db07-1403
Vitamin D, PTH, and calcium and the risk of pros- PMID:18305141
tate cancer: a prospective nested case-control study. Cantó C, Menzies KJ, Auwerx J (2015). NAD+ metabolism
Cancer Causes Control, 23(8):1377–85. doi:10.1007/ and the control of energy homeostasis: a balancing
s10552-012-9948-3 PMID:22706676 act between mitochondria and the nucleus. Cell
Brinton LA, Trabert B, Anderson GL, Falk RT, Felix AS, Metab, 22(1):31–53. doi:10.1016/j.cmet.2015.05.023
Fuhrman BJ, et al. (2016). Serum estrogens and estrogen PMID:26118927
metabolites and endometrial cancer risk among post- Cao Y, Jiang X, Ma H, Wang Y, Xue P, Liu Y (2016). SIRT1
menopausal women. Cancer Epidemiol Biomarkers and insulin resistance. J Diabetes Complications,
Prev, 25(7):1081–9. doi:10.1158/1055-9965.EPI-16-0225 30(1):178–83. doi:10.1016/j.jdiacomp.2015.08.022
PMID:27197275 PMID:26422395
Buchowski MS, Hongu N, Acra S, Wang L, Warolin J, Castiglione F, Taddei A, Rossi Degl’Innocenti D,
Roberts LJ 2nd (2012). Effect of modest caloric restric- Buccoliero AM, Bechi P, Garbini F, et al. (2008).
tion on oxidative stress in women, a randomized Expression of estrogen receptor beta in colon cancer
trial. PLoS One, 7(10):e47079. doi:10.1371/journal. progression. Diagn Mol Pathol, 17(4):231–6. doi:10.1097/
pone.0047079 PMID:23071718 PDM.0b013e3181656d67 PMID:19034156
Bulun SE, Chen D, Moy I, Brooks DC, Zhao H (2012). Castronovo C, Castronovo V, Nikkels A, Peulen O (2015).
Aromatase, breast cancer and obesity: a complex [Vitamin D anti-cancer activities: observations, doubts
interaction. Trends Endocrinol Metab, 23(2):83–9. and certainties.] Rev Med Liege, 70(10):495–500. [in
doi:10.1016/j.tem.2011.10.003 PMID:22169755 French] PMID:26727838
Busch C, Burkard M, Leischner C, Lauer UM, Frank Cerdá C, Sánchez C, Climent B, Vázquez A, Iradi A, El
J, Venturelli S (2015). Epigenetic activities of flavo- Amrani F, et al. (2014). Oxidative stress and DNA
noids in the prevention and treatment of cancer. Clin damage in obesity-related tumorigenesis. Adv Exp
Epigenetics, 7(1):64. doi:10.1186/s13148-015-0095-z Med Biol, 824:5–17. doi:10.1007/978-3-319-07320-0_2
PMID:26161152 PMID:25038989
Byers T, Sedjo RL (2011). Does intentional weight Cetrullo S, D’Adamo S, Tantini B, Borzi RM, Flamigni
loss reduce cancer risk? Diabetes Obes Metab, F (2015). mTOR, AMPK, and Sirt1: key players in
13(12):1063–72. doi:10.1111/j.1463-1326.2011.01464.x metabolic stress management. Crit Rev Eukaryot Gene
PMID:21733057 Expr, 25(1):59–75. doi:10.1615/CritRevEukaryotGene
Campbell KL, Foster-Schubert KE, Alfano CM, Wang CC, Expr.2015012975 PMID:25955819
Wang CY, Duggan CR, et al. (2012). Reduced-calorie Chae JS, Paik JK, Kang R, Kim M, Choi Y, Lee SH,
dietary weight loss, exercise, and sex hormones in post- et al. (2013). Mild weight loss reduces inflammatory
menopausal women: randomized controlled trial. J Clin cytokines, leukocyte count, and oxidative stress in over-
Oncol, 30(19):2314–26. doi:10.1200/JCO.2011.37.9792 weight and moderately obese participants treated for 3
PMID:22614972 years with dietary modification. Nutr Res, 33(3):195–
Campbell KL, Foster-Schubert KE, Makar KW, Kratz 203. doi:10.1016/j.nutres.2013.01.005 PMID:23507225
M, Hagman D, Schur EA, et al. (2013). Gene expres-
sion changes in adipose tissue with diet- and/or
613
Chan AT, Ogino S, Giovannucci EL, Fuchs CS (2011). Clendenen TV, Koenig KL, Shore RE, Levitz M, Arslan
Inflammatory markers are associated with risk of AA, Zeleniuch-Jacquotte A (2009). Postmenopausal
colorectal cancer and chemopreventive response levels of endogenous sex hormones and risk of
to anti-inflammatory drugs. Gastroenterology, colorectal cancer. Cancer Epidemiol Biomarkers Prev,
140(3):799–808, quiz e11. doi:10.1053/j.gastro. 18(1):275–81. doi:10.1158/1055-9965.EPI-08-0777
2010.11.041 PMID:21115010 PMID:19124509
Chan DS, Bandera EV, Greenwood DC, Norat T (2015). Clendenen TV, Lundin E, Zeleniuch-Jacquotte A,
Circulating C-reactive protein and breast cancer Koenig KL, Berrino F, Lukanova A, et al. (2011).
risk – systematic literature review and meta-ana- Circulating inflammation markers and risk of epithe-
lysis of prospective cohort studies. Cancer Epidemiol lial ovarian cancer. Cancer Epidemiol Biomarkers Prev,
Biomarkers Prev, 24(10):1439–49. doi:10.1158/1055- 20(5):799–810. doi:10.1158/1055-9965.EPI-10-1180
9965.EPI-15-0324 PMID:26224798 PMID:21467242
Chan JM, Stampfer MJ, Giovannucci E, Gann PH, Ma Costa TL, Paganotto M, Radominski RB, Kulak CM,
J, Wilkinson P, et al. (1998). Plasma insulin-like Borba VC (2015). Calcium metabolism, vitamin D
growth factor-I and prostate cancer risk: a prospec- and bone mineral density after bariatric surgery.
tive study. Science, 279(5350):563–6. doi:10.1126/ Osteoporos Int, 26(2):757–64. doi:10.1007/s00198-014-
science.279.5350.563 PMID:9438850 2962-4 PMID:25388022
Chao LT, Wu CF, Sung FY, Lin CL, Liu CJ, Huang CJ, Cotillard A, Kennedy SP, Kong LC, Prifti E, Pons N,
et al. (2011). Insulin, glucose and hepatocellular Le Chatelier E, et al.; ANR MicroObes consortium
carcinoma risk in male hepatitis B carriers: results (2013). Dietary intervention impact on gut microbial
from 17-year follow-up of a population-based cohort. gene richness. Nature, 500(7464):585–8. doi:10.1038/
Carcinogenesis, 32(6):876–81. doi:10.1093/carcin/ nature12480 PMID:23985875
bgr058 PMID:21464041 Cullberg KB, Christiansen T, Paulsen SK, Bruun JM,
Chappell J, Leitner JW, Solomon S, Golovchenko I, Pedersen SB, Richelsen B (2013). Effect of weight loss
Goalstone ML, Draznin B (2001). Effect of insulin and exercise on angiogenic factors in the circula-
on cell cycle progression in MCF-7 breast cancer tion and in adipose tissue in obese subjects. Obesity
cells. Direct and potentiating influence. J Biol Chem, (Silver Spring), 21(3):454–60. doi:10.1002/oby.20060
276(41):38023–8. PMID:11500498 PMID:23401397
Cheng S, Massaro JM, Fox CS, Larson MG, Keyes MJ, Cust AE, Allen NE, Rinaldi S, Dossus L, Friedenreich
McCabe EL, et al. (2010). Adiposity, cardiometabolic C, Olsen A, et al. (2007a). Serum levels of C-peptide,
risk, and vitamin D status: the Framingham Heart IGFBP-1 and IGFBP-2 and endometrial cancer risk;
Study. Diabetes, 59(1):242–8. doi:10.2337/db09-1011 results from the European Prospective Investigation
PMID:19833894 into Cancer and Nutrition. Int J Cancer, 120(12):2656–
Chlebowski RT, Johnson KC, Kooperberg C, Pettinger M, 64. doi:10.1002/ijc.22578 PMID:17285578
Wactawski-Wende J, Rohan T, et al.; Women’s Health Cust AE, Kaaks R, Friedenreich C, Bonnet F, Laville M,
Initiative Investigators (2008). Calcium plus vitamin D Lukanova A, et al. (2007b). Plasma adiponectin levels
supplementation and the risk of breast cancer. J Natl and endometrial cancer risk in pre- and postmeno-
Cancer Inst, 100(22):1581–91. doi:10.1093/jnci/djn360 pausal women. J Clin Endocrinol Metab, 92(1):255–63.
PMID:19001601 doi:10.1210/jc.2006-1371 PMID:17062769
Chlebowski RT, Wactawski-Wende J, Ritenbaugh C, Cust AE, Stocks T, Lukanova A, Lundin E, Hallmans G,
Hubbell FA, Ascensao J, Rodabough RJ, et al.; Women’s Kaaks R, et al. (2009). The influence of overweight and
Health Initiative Investigators (2004). Estrogen plus insulin resistance on breast cancer risk and tumour
progestin and colorectal cancer in postmenopausal stage at diagnosis: a prospective study. Breast Cancer
women. N Engl J Med, 350(10):991–1004. doi:10.1056/ Res Treat, 113(3):567–76. doi:10.1007/s10549-008-
NEJMoa032071 PMID:14999111 9958-8 PMID:18330696
Christakos S, Dhawan P, Verstuyf A, Verlinden L, Czerska M, Mikołajewska K, Zieliński M, Gromadzińska
Carmeliet G (2016). Vitamin D: metabolism, molecular J, Wąsowicz W (2015). Today’s oxidative stress markers.
mechanism of action, and pleiotropic effects. Physiol Med Pr, 66(3):393–405. doi:10.13075/mp.5893.00137
Rev, 96(1):365–408. doi:10.1152/physrev.00014.2015 PMID:26325052
PMID:26681795 Dallal CM, Brinton LA, Bauer DC, Buist DS, Cauley JA,
Christensen P, Bartels EM, Riecke BF, Bliddal H, Leeds Hue TF, et al.; B~FIT Research Group (2013). Obesity-
AR, Astrup A, et al. (2012). Improved nutritional status related hormones and endometrial cancer among post-
and bone health after diet-induced weight loss in seden- menopausal women: a nested case-control study within
tary osteoarthritis patients: a prospective cohort study. the B~FIT cohort. Endocr Relat Cancer, 20(1):151–60.
Eur J Clin Nutr, 66(4):504–9. doi:10.1038/ejcn.2011.201 doi:10.1530/ERC-12-0229 PMID:23222000
PMID:22190136
614
Dallal CM, Tice JA, Buist DS, Bauer DC, Lacey JV Jr, Douglas JB, Silverman DT, Pollak MN, Tao Y, Soliman AS,
Cauley JA, et al.; B~FIT Research Group (2014). Stolzenberg-Solomon RZ (2010). Serum IGF-I, IGF-II,
Estrogen metabolism and breast cancer risk among IGFBP-3, and IGF-I/IGFBP-3 molar ratio and risk of
postmenopausal women: a case-cohort study within pancreatic cancer in the Prostate, Lung, Colorectal,
B~FIT. Carcinogenesis, 35(2):346–55. doi:10.1093/ and Ovarian Cancer Screening Trial. Cancer Epidemiol
carcin/bgt367 PMID:24213602 Biomarkers Prev, 19(9):2298–306. doi:10.1158/1055-
Davis-Yadley AH, Malafa MP (2015). Vitamins in 9965.EPI-10-0400 PMID:20699371
pancreatic cancer: a review of underlying mecha- Duggan C, de Dieu Tapsoba J, Mason C, Imayama I,
nisms and future applications. Adv Nutr, 6(6):774–802. Korde L, Wang CY, et al. (2015). Effect of vitamin D3
doi:10.3945/an.115.009456 PMID:26567201 supplementation in combination with weight loss on
De Angel RE, Conti CJ, Wheatley KE, Brenner AJ, Otto inflammatory biomarkers in postmenopausal women:
G, Degraffenried LA, et al. (2013). The enhancing a randomized controlled trial. Cancer Prev Res (Phila),
effects of obesity on mammary tumor growth and Akt/ 8(7):628–35. doi:10.1158/1940-6207.CAPR-14-0449
mTOR pathway activation persist after weight loss and PMID:25908506
are reversed by RAD001. Mol Carcinog, 52(6):446–58. Duncan RE, Ahmadian M, Jaworski K, Sarkadi-Nagy E,
doi:10.1002/mc.21878 PMID:22290600 Sul HS (2007). Regulation of lipolysis in adipocytes.
De Lorenzo MS, Baljinnyam E, Vatner DE, Abarzúa Annu Rev Nutr, 27(1):79–101. doi:10.1146/annurev.
P, Vatner SF, Rabson AB (2011). Caloric restriction nutr.27.061406.093734 PMID:17313320
reduces growth of mammary tumors and metastases. Dunlap SM, Chiao LJ, Nogueira L, Usary J, Perou CM,
Carcinogenesis, 32(9):1381–7. doi:10.1093/carcin/ Varticovski L, et al. (2012). Dietary energy balance
bgr107 PMID:21665891 modulates epithelial-to-mesenchymal transition and
de Martel C, Haggerty TD, Corley DA, Vogelman JH, tumor progression in murine claudin-low and basal-
Orentreich N, Parsonnet J (2007). Serum ghrelin levels like mammary tumor models. Cancer Prev Res (Phila),
and risk of subsequent adenocarcinoma of the esoph- 5(7):930–42. doi:10.1158/1940-6207.CAPR-12-0034
agus. Am J Gastroenterol, 102(6):1166–72. doi:10.1111/ PMID:22588949
j.1572-0241.2007.01116.x PMID:17378911 Eberhart CE, Coffey RJ, Radhika A, Giardiello FM,
Dick KJ, Nelson CP, Tsaprouni L, Sandling JK, Aïssi Ferrenbach S, DuBois RN (1994). Up-regulation of
D, Wahl S, et al. (2014). DNA methylation and cyclooxygenase 2 gene expression in human colorectal
body-mass index: a genome-wide analysis. Lancet, adenomas and adenocarcinomas. Gastroenterology,
383(9933):1990–8. doi:10.1016/S0140-6736(13)62674-4 107(4):1183–8. doi:10.1016/0016-5085(94)90246-1
PMID:24630777 PMID:7926468
Dobrian AD, Davies MJ, Schriver SD, Lauterio TJ, Epplein M, Xiang YB, Cai Q, Peek RM Jr, Li H, Correa
Prewitt RL (2001). Oxidative stress in a rat model of P, et al. (2013). Circulating cytokines and gastric
obesity-induced hypertension. Hypertension, 37(2 Pt 2): cancer risk. Cancer Causes Control, 24(12):2245–50.
554–60. doi:10.1161/01.HYP.37.2.554 PMID:11230334 doi:10.1007/s10552-013-0284-z PMID:24052422
Dossus L, Becker S, Rinaldi S, Lukanova A, Tjønneland A, Fabian CJ, Kimler BF, Donnelly JE, Sullivan DK, Klemp
Olsen A, et al. (2011). Tumor necrosis factor (TNF)-α, JR, Petroff BK, et al. (2013). Favorable modulation
soluble TNF receptors and endometrial cancer risk: the of benign breast tissue and serum risk biomarkers
EPIC study. Int J Cancer, 129(8):2032–7. doi:10.1002/ is associated with > 10 % weight loss in postmeno-
ijc.25840 PMID:21154749 pausal women. Breast Cancer Res Treat, 142(1):119–32.
Dossus L, Jimenez-Corona A, Romieu I, Boutron-Ruault doi:10.1007/s10549-013-2730-8 PMID:24141897
MC, Boutten A, Dupré T, et al. (2014). C-reactive protein Fain JN (2006). Release of interleukins and other inflam-
and postmenopausal breast cancer risk: results from the matory cytokines by human adipose tissue is enhanced
E3N cohort study. Cancer Causes Control, 25(4):533–9. in obesity and primarily due to the nonfat cells. Vitam
doi:10.1007/s10552-014-0355-9 PMID:24504436 Horm, 74:443–77. doi:10.1016/S0083-6729(06)74018-3
Dossus L, Lukanova A, Rinaldi S, Allen N, Cust AE, Becker PMID:17027526
S, et al. (2013). Hormonal, metabolic, and inflam- Falk RT, Dallal CM, Lacey JV Jr, Bauer DC, Buist DS,
matory profiles and endometrial cancer risk within Cauley JA, et al.; B~FIT Research Group (2015).
the EPIC cohort – a factor analysis. Am J Epidemiol, Estrogen metabolites are not associated with
177(8):787–99. doi:10.1093/aje/kws309 PMID:23492765 colorectal cancer risk in postmenopausal women.
Dossus L, Rinaldi S, Becker S, Lukanova A, Tjonneland A, Cancer Epidemiol Biomarkers Prev, 24(9):1419–22.
Olsen A, et al. (2010). Obesity, inflammatory markers, doi:10.1158/1055-9965.EPI-15-0541 PMID:26104910
and endometrial cancer risk: a prospective case-control Fantuzzi G (2013). Adiponectin in inflammatory and
study. Endocr Relat Cancer, 17(4):1007–19. doi:10.1677/ immune-mediated diseases. Cytokine, 64(1):1–10.
ERC-10-0053 PMID:20843938 doi:10.1016/j.cyto.2013.06.317 PMID:23850004
615
Farhat GN, Cummings SR, Chlebowski RT, Parimi N, Gann PH, Hennekens CH, Ma J, Longcope C, Stampfer
Cauley JA, Rohan TE, et al. (2011). Sex hormone levels MJ (1996). Prospective study of sex hormone levels and
and risks of estrogen receptor-negative and estrogen risk of prostate cancer. J Natl Cancer Inst, 88(16):1118–
receptor-positive breast cancers. J Natl Cancer Inst, 26. doi:10.1093/jnci/88.16.1118 PMID:8757191
103(7):562–70. doi:10.1093/jnci/djr031 PMID:21330633 Gao J, Tseng L (1997). Progesterone receptor (PR) inhibits
Faupel-Badger JM, Berrigan D, Ballard-Barbash R, expression of insulin-like growth factor-binding
Potischman N (2009). Anthropometric correlates of protein-1 (IGFBP-1) in human endometrial cell line
insulin-like growth factor 1 (IGF-1) and IGF binding HEC-1B: characterization of the inhibitory effect of PR
protein-3 (IGFBP-3) levels by race/ethnicity and on the distal promoter region of the IGFBP-1 gene. Mol
gender. Ann Epidemiol, 19(12):841–9. doi:10.1016/j. Endocrinol, 11(7):973–9. doi:10.1210/mend.11.7.9932
annepidem.2009.08.005 PMID:19944347 PMID:9178756
Feinberg AP, Irizarry RA, Fradin D, Aryee MJ, Murakami García-Calzón S, Moleres A, Marcos A, Campoy C,
P, Aspelund T, et al. (2010). Personalized epigenomic Moreno LA, Azcona-Sanjulián MC, et al.; EVASYON
signatures that are stable over time and covary with Study Group (2014). Telomere length as a biomarker for
body mass index. Sci Transl Med, 2(49):49ra67. adiposity changes after a multidisciplinary interven-
doi:10.1126/scitranslmed.3001262 PMID:20844285 tion in overweight/obese adolescents: the EVASYON
Feldman D, Krishnan AV, Swami S, Giovannucci E, study. PLoS One, 9(2):e89828. doi:10.1371/journal.
Feldman BJ (2014). The role of vitamin D in reducing pone.0089828 PMID:24587065
cancer risk and progression. Nat Rev Cancer, 14(5):342– Gaudet MM, Patel AV, Teras LR, Sun J, Campbell PT,
57. doi:10.1038/nrc3691 PMID:24705652 Stevens VL, et al. (2013). Obesity-related markers and
Felipo V, Urios A, García-Torres ML, El Mlili N, del Olmo breast cancer in CPS-II Nutrition Cohort. Int J Mol
JA, Civera M, et al. (2013). Alterations in adipocytokines Epidemiol Genet, 4(3):156–66. PMID:24046808
and cGMP homeostasis in morbid obesity patients Geng Y, Blanco FJ, Cornelisson M, Lotz M (1995).
reverse after bariatric surgery. Obesity (Silver Spring), Regulation of cyclooxygenase-2 expression in normal
21(2):229–37. doi:10.1002/oby.20008 PMID:23404955 human articular chondrocytes. J Immunol, 155(2):796–
Fleet JC, DeSmet M, Johnson R, Li Y (2012). Vitamin D and 801. PMID:7608556
cancer: a review of molecular mechanisms. Biochem J, Georgiadi A, Kersten S (2012). Mechanisms of gene regu-
441(1):61–76. doi:10.1042/BJ20110744 PMID:22168439 lation by fatty acids. Adv Nutr, 3(2):127–34. doi:10.3945/
Ford NA, Nunez NP, Holcomb VB, Hursting SD (2013). an.111.001602 PMID:22516720
IGF1 dependence of dietary energy balance effects on Gilbert R, Martin RM, Beynon R, Harris R, Savovic J,
murine Met1 mammary tumor progression, epitheli- Zuccolo L, et al. (2011). Associations of circulating and
al-to-mesenchymal transition, and chemokine expres- dietary vitamin D with prostate cancer risk: a system-
sion. Endocr Relat Cancer, 20(1):39–51. doi:10.1530/ atic review and dose-response meta-analysis. Cancer
ERC-12-0329 PMID:23152442 Causes Control, 22(3):319–40. doi:10.1007/s10552-010-
Fortner RT, Eliassen AH, Spiegelman D, Willett WC, 9706-3 PMID:21203822
Barbieri RL, Hankinson SE (2013). Premenopausal Giorgino F, Belfiore A, Milazzo G, Costantino A,
endogenous steroid hormones and breast cancer risk: Maddux B, Whittaker J, et al. (1991). Overexpression
results from the Nurses’ Health Study II. Breast Cancer of insulin receptors in fibroblast and ovary cells
Res, 15(2):R19. doi:10.1186/bcr3394 PMID:23497468 induces a ligand-mediated transformed phenotype.
Foster-Schubert KE, Alfano CM, Duggan CR, Xiao Mol Endocrinol, 5(3):452–9. doi:10.1210/mend-5-3-452
L, Campbell KL, Kong A, et al. (2012). Effect of diet PMID:1653897
and exercise, alone or combined, on weight and Giovannucci E, Pollak MN, Platz EA, Willett WC,
body composition in overweight-to-obese postmeno- Stampfer MJ, Majeed N, et al. (2000). A prospective
pausal women. Obesity (Silver Spring), 20(8):1628–38. study of plasma insulin-like growth factor-1 and
doi:10.1038/oby.2011.76 PMID:21494229 binding protein-3 and risk of colorectal neoplasia in
Fuhrman BJ, Schairer C, Gail MH, Boyd-Morin J, Xu X, women. Cancer Epidemiol Biomarkers Prev, 9(4):345–9.
Sue LY, et al. (2012). Estrogen metabolism and risk PMID:10794477
of breast cancer in postmenopausal women. J Natl Gonzalez-Rey E, Chorny A, Delgado M (2006).
Cancer Inst, 104(4):326–39. doi:10.1093/jnci/djr531 Therapeutic action of ghrelin in a mouse model of
PMID:22232133 colitis. Gastroenterology, 130(6):1707–20. doi:10.1053/j.
Furukawa N, Ongusaha P, Jahng WJ, Araki K, Choi CS, gastro.2006.01.041 PMID:16697735
Kim HJ, et al. (2005). Role of Rho-kinase in regula- Gram IT, Norat T, Rinaldi S, Dossus L, Lukanova A, Téhard
tion of insulin action and glucose homeostasis. Cell B, et al. (2006). Body mass index, waist circumference
Metab, 2(2):119–29. doi:10.1016/j.cmet.2005.06.011 and waist-hip ratio and serum levels of IGF-I and
PMID:16098829 IGFBP-3 in European women. Int J Obes, 30(11):1623–
31. doi:10.1038/sj.ijo.0803324 PMID:16552400
616
Gross AL, Newschaffer CJ, Hoffman-Bolton J, Rifai N, exercise is an effective strategy to reduce anthro-
Visvanathan K (2013). Adipocytokines, inflammation, pometric parameters and oxidative stress in obese
and breast cancer risk in postmenopausal women: a patients. Obes Facts, 5(1):12–22. doi:10.1159/000336526
prospective study. Cancer Epidemiol Biomarkers Prev, PMID:22433613
22(7):1319–24. doi:10.1158/1055-9965.EPI-12-1444 Gwinn DM, Shackelford DB, Egan DF, Mihaylova MM,
PMID:23651666 Mery A, Vasquez DS, et al. (2008). AMPK phosphoryl-
Grote VA, Kaaks R, Nieters A, Tjønneland A, Halkjær J, ation of raptor mediates a metabolic checkpoint. Mol
Overvad K, et al. (2012a). Inflammation marker and Cell, 30(2):214–26. doi:10.1016/j.molcel.2008.03.003
risk of pancreatic cancer: a nested case-control study PMID:18439900
within the EPIC cohort. Br J Cancer, 106(11):1866–74. Habermann N, Makar KW, Abbenhardt C, Xiao L,
PMID:22617158 Wang CY, Utsugi HK, et al. (2015). No effect of caloric
Grote VA, Rohrmann S, Dossus L, Nieters A, Halkjaer J, restriction or exercise on radiation repair capacity.
Tjønneland A, et al. (2012b). The association of circu- Med Sci Sports Exerc, 47(5):896–904. doi:10.1249/
lating adiponectin levels with pancreatic cancer risk: a MSS.0000000000000480 PMID:25160845
study within the prospective EPIC cohort. Int J Cancer, Hair BY, Troester MA, Edmiston SN, Parrish EA, Robinson
130(10):2428–37. doi:10.1002/ijc.26244 PMID:21681743 WR, Wu MC, et al. (2015). Body mass index is associ-
Grote VA, Rohrmann S, Nieters A, Dossus L, Tjønneland ated with gene methylation in estrogen receptor-pos-
A, Halkjær J, et al. (2011). Diabetes mellitus, glycated itive breast tumors. Cancer Epidemiol Biomarkers
haemoglobin and C-peptide levels in relation to Prev, 24(3):580–6. doi:10.1158/1055-9965.EPI-14-1017
pancreatic cancer risk: a study within the European PMID:25583948
Prospective Investigation into Cancer and Nutrition Hanahan D, Weinberg RA (2000). The hallmarks of cancer.
(EPIC) cohort. Diabetologia, 54(12):3037–46. Cell, 100(1):57–70. doi:10.1016/S0092-8674(00)81683-9
doi:10.1007/s00125-011-2316-0 PMID:21953276 PMID:10647931
Gunter MJ, Hoover DR, Yu H, Wassertheil-Smoller S, Hanahan D, Weinberg RA (2011). Hallmarks of cancer:
Manson JE, Li J, et al. (2008a). A prospective evalu- the next generation. Cell, 144(5):646–74. doi:10.1016/j.
ation of insulin and insulin-like growth factor-I as cell.2011.02.013 PMID:21376230
risk factors for endometrial cancer. Cancer Epidemiol Hankinson SE, Willett WC, Colditz GA, Hunter DJ,
Biomarkers Prev, 17(4):921–9. doi:10.1158/1055-9965. Michaud DS, Deroo B, et al. (1998b). Circulating
EPI-07-2686 PMID:18398032 concentrations of insulin-like growth factor-I and risk
Gunter MJ, Hoover DR, Yu H, Wassertheil-Smoller S, of breast cancer. Lancet, 351(9113):1393–6. doi:10.1016/
Rohan TE, Manson JE, et al. (2008b). Insulin, insu- S0140-6736(97)10384-1 PMID:9593409
lin-like growth factor-I, endogenous estradiol, and risk Hankinson SE, Willett WC, Manson JE, Colditz GA,
of colorectal cancer in postmenopausal women. Cancer Hunter DJ, Spiegelman D, et al. (1998a). Plasma sex
Res, 68(1):329–37. doi:10.1158/0008-5472.CAN-07- steroid hormone levels and risk of breast cancer in post-
2946 PMID:18172327 menopausal women. J Natl Cancer Inst, 90(17):1292–9.
Gunter MJ, Hoover DR, Yu H, Wassertheil-Smoller S, doi:10.1093/jnci/90.17.1292 PMID:9731736
Rohan TE, Manson JE, et al. (2009). Insulin, insu- Hardie DG (2004). The AMP-activated protein kinase
lin-like growth factor-I, and risk of breast cancer in pathway – new players upstream and downstream.
postmenopausal women. J Natl Cancer Inst, 101(1):48– J Cell Sci, 117(Pt 23):5479–87. doi:10.1242/jcs.01540
60. doi:10.1093/jnci/djn415 PMID:19116382 PMID:15509864
Gunter MJ, Wang T, Cushman M, Xue X, Wassertheil- Hardikar S, Onstad L, Song X, Wilson AM, Montine TJ,
Smoller S, Strickler HD, et al. (2015b). Circulating Kratz M, et al. (2014). Inflammation and oxidative stress
adipokines and inflammatory markers and post- markers and esophageal adenocarcinoma incidence
menopausal breast cancer risk. J Natl Cancer Inst, in a Barrett’s esophagus cohort. Cancer Epidemiol
107(9):djv169. doi:10.1093/jnci/djv169 PMID:26185195 Biomarkers Prev, 23(11):2393–403. doi:10.1158/1055-
Gunter MJ, Xie X, Xue X, Kabat GC, Rohan TE, 9965.EPI-14-0384 PMID:25106775
Wassertheil-Smoller S, et al. (2015a). Breast cancer Harris HR, Tworoger SS, Hankinson SE, Rosner BA,
risk in metabolically healthy but overweight post- Michels KB (2011). Plasma leptin levels and risk of
menopausal women. Cancer Res, 75(2):270–4. breast cancer in premenopausal women. Cancer Prev
doi:10.1158/0008-5472.CAN-14-2317 PMID:25593034 Res (Phila), 4(9):1449–56. doi:10.1158/1940-6207.
Guo SQ, Zhang YZ (2012). Histone deacetylase inhibi- CAPR-11-0125 PMID:21680707
tion: an important mechanism in the treatment of Hartman J, Edvardsson K, Lindberg K, Zhao C, Williams
lymphoma. Cancer Biol Med, 9(2):85–9. PMID:23691460 C, Ström A, et al. (2009). Tumor repressive functions
Gutierrez-Lopez L, Garcia-Sanchez JR, Rincon-Viquez MJ, of estrogen receptor beta in SW480 colon cancer cells.
Lara-Padilla E, Sierra-Vargas MP, Olivares-Corichi IM Cancer Res, 69(15):6100–6. doi:10.1158/0008-5472.
(2012). Hypocaloric diet and regular moderate aerobic CAN-09-0506 PMID:19602591
617
Harvey AE, Lashinger LM, Hays D, Harrison LM, Lewis Prev, 23(1):179–88. doi:10.1158/1055-9965.EPI-13-0545
K, Fischer SM, et al. (2014). Calorie restriction decreases PMID:24192010
murine and human pancreatic tumor cell growth, Hoffmann LS, Etzrodt J, Willkomm L, Sanyal A,
nuclear factor-κB activation, and inflammation-related Scheja L, Fischer AWC, et al. (2015). Stimulation of
gene expression in an insulin-like growth factor-1- soluble guanylyl cyclase protects against obesity by
dependent manner. PLoS One, 9(5):e94151. doi:10.1371/ recruiting brown adipose tissue. Nat Commun, 6:7235.
journal.pone.0094151 PMID:24804677 doi:10.1038/ncomms8235 PMID:26011238
Heikkilä K, Harris R, Lowe G, Rumley A, Yarnell J, Hoffmeister M, Bläker H, Kloor M, Roth W, Toth C,
Gallacher J, et al. (2009). Associations of circulating Herpel E, et al. (2013). Body mass index and micro-
C-reactive protein and interleukin-6 with cancer risk: satellite instability in colorectal cancer: a popula-
findings from two prospective cohorts and a meta-ana- tion-based study. Cancer Epidemiol Biomarkers Prev,
lysis. Cancer Causes Control, 20(1):15–26. doi:10.1007/ 22(12):2303–11. doi:10.1158/1055-9965.EPI-13-0239
s10552-008-9212-z PMID:18704713 PMID:24127414
Hellmann J, Zhang MJ, Tang Y, Rane M, Bhatnagar A, Howe LR (2007). Inflammation and breast cancer.
Spite M (2013). Increased saturated fatty acids in obesity Cyclooxygenase/prostaglandin signaling and breast
alter resolution of inflammation in part by stimulating cancer. Breast Cancer Res, 9(4):210. doi:10.1186/
prostaglandin production. J Immunol, 191(3):1383–92. bcr1678 PMID:17640394
doi:10.4049/jimmunol.1203369 PMID:23785121 Hsing AW (2001). Hormones and prostate cancer:
Helzlsouer KJ, Alberg AJ, Gordon GB, Longcope C, what’s next? Epidemiol Rev, 23(1):42–58. doi:10.1093/
Bush TL, Hoffman SC, et al. (1995). Serum gonado- oxfordjournals.epirev.a000795 PMID:11588854
tropins and steroid hormones and the development of Hsing AW, Chua S Jr, Gao YT, Gentzschein E, Chang L,
ovarian cancer. JAMA, 274(24):1926–30. doi:10.1001/ Deng J, et al. (2001). Prostate cancer risk and serum
jama.1995.03530240036037 PMID:8568986 levels of insulin and leptin: a population-based
Henry SL, Bensley JG, Wood-Bradley RJ, Cullen-McEwen study. J Natl Cancer Inst, 93(10):783–9. doi:10.1093/
LA, Bertram JF, Armitage JA (2012). White adipo- jnci/93.10.783 PMID:11353789
cytes: more than just fat depots. Int J Biochem Cell Huang H, Lee DH, Zabolotny JM, Kim YB (2013).
Biol, 44(3):435–40. doi:10.1016/j.biocel.2011.12.011 Metabolic actions of Rho-kinase in periphery and
PMID:22222895 brain. Trends Endocrinol Metab, 24(10):506–14.
Heuson JC, Legros N (1972). Influence of insulin depri- doi:10.1016/j.tem.2013.06.003 PMID:23938132
vation on growth of the 7,12-dimethylbenz(a)anthra- Hubbard JS, Rohrmann S, Landis PK, Metter EJ, Muller
cene-induced mammary carcinoma in rats subjected DC, Andres R, et al. (2004). Association of prostate
to alloxan diabetes and food restriction. Cancer Res, cancer risk with insulin, glucose, and anthropometry
32(2):226–32. PMID:5058183 in the Baltimore longitudinal study of aging. Urology,
Hidaka A, Sasazuki S, Goto A, Sawada N, Shimazu T, 63(2):253–8. doi:10.1016/j.urology.2003.09.060
Yamaji T, et al.; JPHC Study Group (2015). Plasma PMID:14972466
insulin, C-peptide and blood glucose and the risk Hui JM, Kench JG, Chitturi S, Sud A, Farrell GC, Byth
of gastric cancer: the Japan Public Health Center- K, et al. (2003). Long-term outcomes of cirrhosis
based prospective study. Int J Cancer, 136(6):1402–10. in nonalcoholic steatohepatitis compared with
doi:10.1002/ijc.29098 PMID:25066446 hepatitis C. Hepatology, 38(2):420–7. doi:10.1053/
Higami Y, Barger JL, Page GP, Allison DB, Smith SR, Prolla jhep.2003.50320 PMID:12883486
TA, et al. (2006). Energy restriction lowers the expres- Hursting SD, Dunlap SM, Ford NA, Hursting MJ,
sion of genes linked to inflammation, the cytoskeleton, Lashinger LM (2013). Calorie restriction and cancer
the extracellular matrix, and angiogenesis in mouse prevention: a mechanistic perspective. Cancer Metab,
adipose tissue. J Nutr, 136(2):343–52. PMID:16424110 1(1):10. doi:10.1186/2049-3002-1-10 PMID:24280167
Higgins SC, Gueorguiev M, Korbonits M (2007). Ghrelin, Hursting SD, Lavigne JA, Berrigan D, Perkins SN, Barrett
the peripheral hunger hormone. Ann Med, 39(2):116– JC (2003). Calorie restriction, aging, and cancer
36. doi:10.1080/07853890601149179 PMID:17453675 prevention: mechanisms of action and applicability
Ho GY, Wang T, Gunter MJ, Strickler HD, Cushman M, to humans. Annu Rev Med, 54(1):131–52. doi:10.1146/
Kaplan RC, et al. (2012). Adipokines linking obesity annurev.med.54.101601.152156 PMID:12525670
with colorectal cancer risk in postmenopausal women. Hvidtfeldt UA, Gunter MJ, Lange T, Chlebowski RT,
Cancer Res, 72(12):3029–37. doi:10.1158/0008-5472. Lane D, Farhat GN, et al. (2012). Quantifying medi-
CAN-11-2771 PMID:22511581 ating effects of endogenous estrogen and insulin in
Ho GY, Wang T, Zheng SL, Tinker L, Xu J, Rohan TE, the relation between obesity, alcohol consumption,
et al. (2014). Circulating soluble cytokine receptors and and breast cancer. Cancer Epidemiol Biomarkers Prev,
colorectal cancer risk. Cancer Epidemiol Biomarkers 21(7):1203–12. doi:10.1158/1055-9965.EPI-12-0310
PMID:22564867
618
Hyde Z, Flicker L, McCaul KA, Almeida OP, Hankey GJ, Jiang W, Zhu Z, Thompson HJ (2009). Effects of physical
Chubb SA, et al. (2012). Associations between testos- activity and restricted energy intake on chemically
terone levels and incident prostate, lung, and colorectal induced mammary carcinogenesis. Cancer Prev Res
cancer. A population-based study. Cancer Epidemiol (Phila), 2(4):338–44. doi:10.1158/1940-6207.CAPR-08-
Biomarkers Prev, 21(8):1319–29. doi:10.1158/1055-9965. 0169 PMID:19336733
EPI-12-0129 PMID:22828207 Jiang W, Zhu Z, Thompson HJ (2013). Effects of limiting
IARC (2002). Weight control and physical activity. Lyon, energy availability via diet and physical activity on
France: IARC Press (IARC Handbooks of Cancer mammalian target of rapamycin-related signalling in
Prevention, Vol. 6). Available from: http://publications. rat mammary carcinomas. Carcinogenesis, 34(2):378–
iarc.fr/376. 87. doi:10.1093/carcin/bgs350 PMID:23125225
IARC (2008). Vitamin D and cancer. Lyon, France: IARC Jukarainen S, Heinonen S, Rämö JT, Rinnankoski-Tuikka
Press (IARC Working Group Reports, No. 5). Available R, Rappou E, Tummers M, et al. (2016). Obesity is asso-
from: http://publications.iarc.fr/388. ciated with low NAD+/SIRT pathway expression in
Imayama I, Ulrich CM, Alfano CM, Wang C, Xiao L, adipose tissue of BMI-discordant monozygotic twins.
Wener MH, et al. (2012). Effects of a caloric restric- J Clin Endocrinol Metab, 101(1):275–83. doi:10.1210/
tion weight loss diet and exercise on inflammatory jc.2015-3095 PMID:26574954
biomarkers in overweight/obese postmenopausal Jung UJ, Choi MS (2014). Obesity and its metabolic
women: a randomized controlled trial. Cancer Res, complications: the role of adipokines and the relation-
72(9):2314–26. doi:10.1158/0008-5472.CAN-11-3092 ship between obesity, inflammation, insulin resistance,
PMID:22549948 dyslipidemia and nonalcoholic fatty liver disease. Int
Ish-Shalom D, Christoffersen CT, Vorwerk P, Sacerdoti- J Mol Sci, 15(4):6184–223. doi:10.3390/ijms15046184
Sierra N, Shymko RM, Naor D, et al. (1997). Mitogenic PMID:24733068
properties of insulin and insulin analogues mediated Kaaks R, Johnson T, Tikk K, Sookthai D, Tjønneland A,
by the insulin receptor. Diabetologia, 40(Suppl 2):S25– Roswall N, et al. (2014a). Insulin-like growth factor I
31. doi:10.1007/s001250051393 PMID:9248698 and risk of breast cancer by age and hormone receptor
James RE, Lukanova A, Dossus L, Becker S, Rinaldi S, status – a prospective study within the EPIC cohort.
Tjønneland A, et al. (2011). Postmenopausal serum Int J Cancer, 134(11):2683–90. doi:10.1002/ijc.28589
sex steroids and risk of hormone receptor-positive PMID:24248481
and -negative breast cancer: a nested case-con- Kaaks R, Lundin E, Rinaldi S, Manjer J, Biessy C, Söderberg
trol study. Cancer Prev Res (Phila), 4(10):1626–35. S, et al. (2002). Prospective study of IGF-I, IGF-binding
doi:10.1158/1940-6207.CAPR-11-0090 PMID:21813404 proteins, and breast cancer risk, in northern and
Janani C, Ranjitha Kumari BD (2015). PPAR gamma southern Sweden. Cancer Causes Control, 13(4):307–16.
gene – a review. Diabetes Metab Syndr, 9(1):46–50. doi:10.1023/A:1015270324325 PMID:12074500
doi:10.1016/j.dsx.2014.09.015 PMID:25450819 Kaaks R, Rinaldi S, Key TJ, Berrino F, Peeters PH, Biessy
Jenab M, Riboli E, Cleveland RJ, Norat T, Rinaldi S, C, et al. (2005). Postmenopausal serum androgens,
Nieters A, et al. (2007). Serum C-peptide, IGFBP-1 and oestrogens and breast cancer risk: the European
IGFBP-2 and risk of colon and rectal cancers in the Prospective Investigation into Cancer and Nutrition.
European Prospective Investigation into Cancer and Endocr Relat Cancer, 12(4):1071–82. doi:10.1677/
Nutrition. Int J Cancer, 121(2):368–76. doi:10.1002/ erc.1.01038 PMID:16322344
ijc.22697 PMID:17372899 Kaaks R, Tikk K, Sookthai D, Schock H, Johnson T,
Jensen MD, Haymond MW, Rizza RA, Cryer PE, Miles JM Tjønneland A, et al. (2014b). Premenopausal serum sex
(1989). Influence of body fat distribution on free fatty hormone levels in relation to breast cancer risk, overall
acid metabolism in obesity. J Clin Invest, 83(4):1168–73. and by hormone receptor status – results from the
doi:10.1172/JCI113997 PMID:2649512 EPIC cohort. Int J Cancer, 134(8):1947–57. doi:10.1002/
Jiang W, Zhu Z, Thompson HJ (2003). Effect of energy ijc.28528 PMID:24155248
restriction on cell cycle machinery in 1-methyl-1-nitro- Kaaks R, Toniolo P, Akhmedkhanov A, Lukanova A, Biessy
sourea-induced mammary carcinomas in rats. Cancer C, Dechaud H, et al. (2000). Serum C-peptide, insu-
Res, 63(6):1228–34. PMID:12649181 lin-like growth factor (IGF)-I, IGF-binding proteins,
Jiang W, Zhu Z, Thompson HJ (2008). Dietary energy and colorectal cancer risk in women. J Natl Cancer
restriction modulates the activity of AMP-activated Inst, 92(19):1592–600. doi:10.1093/jnci/92.19.1592
protein kinase, Akt, and mammalian target of rapa- PMID:11018095
mycin in mammary carcinomas, mammary gland, Kabat GC, Kim M, Caan BJ, Chlebowski RT, Gunter
and liver. Cancer Res, 68(13):5492–9. doi:10.1158/0008- MJ, Ho GY, et al. (2009). Repeated measures of
5472.CAN-07-6721 PMID:18593953 serum glucose and insulin in relation to postmeno-
pausal breast cancer. Int J Cancer, 125(11):2704–10.
doi:10.1002/ijc.24609 PMID:19588485
619
Kabat GC, Kim MY, Strickler HD, Shikany JM, Lane Key TJ, Appleby PN, Reeves GK, Roddam AW; Endog
D, Luo J, et al. (2012). A longitudinal study of serum enous Hormones and Breast Cancer Collaborative
insulin and glucose levels in relation to colorectal Group (2010). Insulin-like growth factor 1 (IGF1),
cancer risk among postmenopausal women. Br J IGF binding protein 3 (IGFBP3), and breast cancer
Cancer, 106(1):227–32. doi:10.1038/bjc.2011.512 risk: pooled individual data analysis of 17 prospective
PMID:22127286 studies. Lancet Oncol, 11(6):530–42. doi:10.1016/S1470-
Kahn BB, Alquier T, Carling D, Hardie DG (2005). 2045(10)70095-4 PMID:20472501
AMP-activated protein kinase: ancient energy gauge Key TJ, Appleby PN, Reeves GK, Travis RC, Alberg AJ,
provides clues to modern understanding of metabolism. Barricarte A, et al.; Endogenous Hormones and Breast
Cell Metab, 1(1):15–25. doi:10.1016/j.cmet.2004.12.003 Cancer Collaborative Group (2013). Sex hormones
PMID:16054041 and risk of breast cancer in premenopausal women:
Kant P, Perry SL, Dexter SP, Race AD, Loadman PM, a collaborative reanalysis of individual participant
Hull MA (2014). Mucosal biomarkers of colorectal data from seven prospective studies. Lancet Oncol,
cancer risk do not increase at 6 months following 14(10):1009–19. doi:10.1016/S1470-2045(13)70301-2
sleeve gastrectomy, unlike gastric bypass. Obesity PMID:23890780
(Silver Spring), 22(1):202–10. doi:10.1002/oby.20493 Kim C, Kong S, Laughlin GA, Golden SH, Mather KJ, Nan
PMID:23625552 B, et al. (2012). Endogenous sex hormone changes in
Kant P, Sainsbury A, Reed KR, Pollard SG, Scott N, Clarke postmenopausal women in the Diabetes Prevention
AR, et al. (2011). Rectal epithelial cell mitosis and Program. J Clin Endocrinol Metab, 97(8):2853–61.
expression of macrophage migration inhibitory factor doi:10.1210/jc.2012-1233 PMID:22689695
are increased 3 years after Roux-en-Y gastric bypass Kim JH, Song J, Park KW (2015). The multifaceted factor
(RYGB) for morbid obesity: implications for long-term peroxisome proliferator-activated receptor γ (PPARγ)
neoplastic risk following RYGB. Gut, 60(7):893–901. in metabolism, immunity, and cancer. Arch Pharm
doi:10.1136/gut.2010.230755 PMID:21303912 Res, 38(3):302–12. doi:10.1007/s12272-015-0559-x
Karefylakis C, Näslund I, Edholm D, Sundbom M, PMID:25579849
Karlsson FA, Rask E (2014). Vitamin D status 10 years Kiunga GA, Raju J, Sabljic N, Bajaj G, Good CK, Bird RP
after primary gastric bypass: gravely high prevalence (2004). Elevated insulin receptor protein expression
of hypovitaminosis D and raised PTH levels. Obes in experimentally induced colonic tumors. Cancer
Surg, 24(3):343–8. doi:10.1007/s11695-013-1104-y Lett, 211(2):145–53. doi:10.1016/j.canlet.2004.02.015
PMID:24163201 PMID:15219938
Karin M (2006). Nuclear factor-κB in cancer develop- Klotz DM, Hewitt SC, Ciana P, Raviscioni M, Lindzey
ment and progression. Nature, 441(7092):431–6. JK, Foley J, et al. (2002). Requirement of estrogen
doi:10.1038/nature04870 PMID:16724054 receptor-alpha in insulin-like growth factor-1
Keeley BR, Islami F, Pourshams A, Poustchi H, Pak JS, (IGF-1)-induced uterine responses and in vivo
Brennan P, et al. (2014). Prediagnostic serum levels of evidence for IGF-1/estrogen receptor cross-talk. J Biol
inflammatory biomarkers are correlated with future Chem, 277(10):8531–7. doi:10.1074/jbc.M109592200
development of lung and esophageal cancer. Cancer Sci, PMID:11751931
105(9):1205–11. doi:10.1111/cas.12485 PMID:25040886 Koohestani N, Tran TT, Lee W, Wolever TM, Bruce WR
Keinan-Boker L, Bueno De Mesquita HB, Kaaks R, (1997). Insulin resistance and promotion of aberrant
Van Gils CH, Van Noord PA, Rinaldi S, et al. (2003). crypt foci in the colons of rats on a high-fat diet. Nutr
Circulating levels of insulin-like growth factor I, its Cancer, 29(1):69–76. doi:10.1080/01635589709514604
binding proteins -1,-2, -3, C-peptide and risk of post- PMID:9383787
menopausal breast cancer. Int J Cancer, 106(1):90–5. Krajcik RA, Massardo S, Orentreich N (2003). No associ-
doi:10.1002/ijc.11193 PMID:12794762 ation between serum levels of tumor necrosis factor-α
Key T, Appleby P, Barnes I, Reeves G; Endogenous (TNF-α) or the soluble receptors sTNFR1 and sTNFR2
Hormones and Breast Cancer Collaborative Group and breast cancer risk. Cancer Epidemiol Biomarkers
(2002). Endogenous sex hormones and breast cancer in Prev, 12(9):945–6. PMID:14504210
postmenopausal women: reanalysis of nine prospective Kurki E, Shi J, Martonen E, Finckenberg P, Mervaala E
studies. J Natl Cancer Inst, 94(8):606–16. doi:10.1093/ (2012). Distinct effects of calorie restriction on adipose
jnci/94.8.606 PMID:11959894 tissue cytokine and angiogenesis profiles in obese and
Key TJ, Appleby PN, Reeves GK, Roddam A, Dorgan JF, lean mice. Nutr Metab (Lond), 9(1):64. doi:10.1186/1743-
Longcope C, et al.; Endogenous Hormones and Breast 7075-9-64 PMID:22748184
Cancer Collaborative Group (2003). Body mass index, Lai GY, Giovannucci EL, Pollak MN, Peskoe SB, Stampfer
serum sex hormones, and breast cancer risk in post- MJ, Willett WC, et al. (2014). Association of C-peptide
menopausal women. J Natl Cancer Inst, 95(16):1218–26. and leptin with prostate cancer incidence in the
doi:10.1093/jnci/djg022 PMID:12928347 Health Professionals Follow-up Study. Cancer Causes
620
Control, 25(5):625–32. doi:10.1007/s10552-014-0369-3 colorectal cancer in male smokers. Clin Gastroenterol

PMID:24664287 Hepatol, 4(12):1514–21. doi:10.1016/j.cgh.2006.09.014
Laplante M, Sabatini DM (2013). Regulation of mTORC1 PMID:17162243
and its impact on gene expression at a glance. Lin JH, Zhang SM, Rexrode KM, Manson JE, Chan AT,
J Cell Sci, 126(Pt 8):1713–9. doi:10.1242/jcs.125773 Wu K, et al. (2013). Association between sex hormones
PMID:23641065 and colorectal cancer risk in men and women. Clin
Lashinger LM, Harrison LM, Rasmussen AJ, Logsdon Gastroenterol Hepatol, 11(4):419–424.e1. doi:10.1016/j.
CD, Fischer SM, McArthur MJ, et al. (2013). Dietary cgh.2012.11.012 PMID:23200979
energy balance modulation of Kras- and Ink4a/Arf+/−- Liu LY, Wang M, Ma ZB, Yu LX, Zhang Q, Gao DZ,
driven pancreatic cancer: the role of insulin-like et al. (2013). The role of adiponectin in breast cancer:
growth factor-I. Cancer Prev Res (Phila), 6(10):1046–55. a meta-analysis. PLoS One, 8(8):e73183. doi:10.1371/
doi:10.1158/1940-6207.CAPR-13-0185 PMID:23980075 journal.pone.0073183 PMID:23991180
Lashinger LM, Malone LM, Brown GW, Daniels EA, Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH,
Goldberg JA, Otto G, et al. (2011). Rapamycin partially Day FR, et al.; LifeLines Cohort Study; ADIPOGen
mimics the anticancer effects of calorie restriction in Consortium; AGEN-BMI Working Group;
a murine model of pancreatic cancer. Cancer Prev Res CARDIOGRAMplusC4D Consortium; CKDGen
(Phila), 4(7):1041–51. doi:10.1158/1940-6207.CAPR-11- Consortium; GLGC; ICBP; MAGIC Investigators;
0023 PMID:21593197 MuTHER Consortium; MIGen Consortium; PAGE
Le Chatelier E, Nielsen T, Qin J, Prifti E, Hildebrand F, Consortium; ReproGen Consortium; GENIE
Falony G, et al.; MetaHIT consortium (2013). Richness Consortium; International Endogene Consortium
of human gut microbiome correlates with meta- (2015). Genetic studies of body mass index yield new
bolic markers. Nature, 500(7464):541–6. doi:10.1038/ insights for obesity biology. Nature, 518(7538):197–206.
nature12506 PMID:23985870 doi:10.1038/nature14177 PMID:25673413
Lee JY, Myung SK, Song YS (2013). Prognostic role Luger M, Kruschitz R, Langer F, Prager G, Walker M,
of cyclooxygenase-2 in epithelial ovarian cancer: Marculescu R, et al. (2015). Effects of omega-loop gastric
a meta-analysis of observational studies. Gynecol bypass on vitamin D and bone metabolism in morbidly
Oncol, 129(3):613–9. doi:10.1016/j.ygyno.2013.02.011 obese bariatric patients. Obes Surg, 25(6):1056–62.
PMID:23422504 doi:10.1007/s11695-014-1492-7 PMID:25381120
Lee JY, Sohn KH, Rhee SH, Hwang D (2001). Saturated Luhn P, Dallal CM, Weiss JM, Black A, Huang WY,
fatty acids, but not unsaturated fatty acids, induce Lacey JV Jr, et al. (2013). Circulating adipokine levels
the expression of cyclooxygenase-2 mediated through and endometrial cancer risk in the Prostate, Lung,
Toll-like receptor 4. J Biol Chem, 276(20):16683–9. Colorectal, and Ovarian Cancer Screening Trial.
doi:10.1074/jbc.M011695200 PMID:11278967 Cancer Epidemiol Biomarkers Prev, 22(7):1304–12.
Lemoine AY, Ledoux S, Quéguiner I, Caldérari S, Mechler doi:10.1158/1055-9965.EPI-13-0258 PMID:23696194
C, Msika S, et al. (2012). Link between adipose tissue Lukanova A, Becker S, Hüsing A, Schock H, Fedirko V,
angiogenesis and fat accumulation in severely obese Trepo E, et al. (2014). Prediagnostic plasma testos-
subjects. J Clin Endocrinol Metab, 97(5):E775–80. terone, sex hormone-binding globulin, IGF-I and
doi:10.1210/jc.2011-2649 PMID:22419723 hepatocellular carcinoma: etiological factors or risk
Leyvraz C, Verdumo C, Suter M, Paroz A, Calmes JM, markers? Int J Cancer, 134(1):164–73. doi:10.1002/
Marques-Vidal PM, et al. (2012). Changes in gene ijc.28342 PMID:23801371
expression profile in human subcutaneous adipose Lukanova A, Lundin E, Akhmedkhanov A, Micheli
tissue during significant weight loss. Obes Facts, A, Rinaldi S, Zeleniuch-Jacquotte A, et al. (2003a).
5(3):440–51. doi:10.1159/000341137 PMID:22797372 Circulating levels of sex steroid hormones and risk of
Li H, Stampfer MJ, Mucci L, Rifai N, Qiu W, Kurth T, ovarian cancer. Int J Cancer, 104(5):636–42. doi:10.1002/
et al. (2010). A 25-year prospective study of plasma ijc.10990 PMID:12594820
adiponectin and leptin concentrations and prostate Lukanova A, Lundin E, Micheli A, Akhmedkhanov A,
cancer risk and survival. Clin Chem, 56(1):34–43. Rinaldi S, Muti P, et al. (2003b). Risk of ovarian cancer
doi:10.1373/clinchem.2009.133272 PMID:19910504 in relation to prediagnostic levels of C-peptide, insu-
Liao LM, Weinstein SJ, Pollak M, Li Z, Virtamo J, Albanes lin-like growth factor binding proteins-1 and -2 (USA,
D, et al. (2013). Prediagnostic circulating adipokine Sweden, Italy). Cancer Causes Control, 14(3):285–92.
concentrations and risk of renal cell carcinoma in male doi:10.1023/A:1023688603547 PMID:12814208
smokers. Carcinogenesis, 34(1):109–12. doi:10.1093/ Lukanova A, Lundin E, Micheli A, Arslan A, Ferrari P,
carcin/bgs322 PMID:23042303 Rinaldi S, et al. (2004a). Circulating levels of sex steroid
Limburg PJ, Stolzenberg-Solomon RZ, Vierkant RA, hormones and risk of endometrial cancer in postmeno-
Roberts K, Sellers TA, Taylor PR, et al. (2006). pausal women. Int J Cancer, 108(3):425–32. doi:10.1002/
Insulin, glucose, insulin resistance, and incident ijc.11529 PMID:14648710
621
Lukanova A, Lundin E, Toniolo P, Micheli A, Major JM, Pollak MN, Snyder K, Virtamo J, Albanes D
Akhmedkhanov A, Rinaldi S, et al. (2002). Circulating (2010). Insulin-like growth factors and risk of kidney
levels of insulin-like growth factor-I and risk of ovarian cancer in men. Br J Cancer, 103(1):132–5. doi:10.1038/
cancer. Int J Cancer, 101(6):549–54. doi:10.1002/ sj.bjc.6605722 PMID:20517306
ijc.10613 PMID:12237896 Mansego ML, Milagro FI, Zulet MA, Martinez JA
Lukanova A, Lundin E, Zeleniuch-Jacquotte A, Muti P, (2015). SH2B1 CpG-SNP is associated with body
Mure A, Rinaldi S, et al. (2004c). Body mass index, weight reduction in obese subjects following a dietary
circulating levels of sex-steroid hormones, IGF-I restriction program. Ann Nutr Metab, 66(1):1–9.
and IGF-binding protein-3: a cross-sectional study doi:10.1159/000368425 PMID:25471250
in healthy women. Eur J Endocrinol, 150(2):161–71. Marseglia L, Manti S, D’Angelo G, Nicotera A, Parisi E, Di
doi:10.1530/eje.0.1500161 PMID:14763914 Rosa G, et al. (2015). Oxidative stress in obesity: a critical
Lukanova A, Toniolo P, Zeleniuch-Jacquotte A, Grankvist component in human diseases. Int J Mol Sci, 16(1):378–
K, Wulff M, Arslan AA, et al. (2006). Insulin-like 400. doi:10.3390/ijms16010378 PMID:25548896
growth factor I in pregnancy and maternal risk of Martini LA, Wood RJ (2006). Vitamin D status and
breast cancer. Cancer Epidemiol Biomarkers Prev, the metabolic syndrome. Nutr Rev, 64(11):479–
15(12):2489–93. doi:10.1158/1055-9965.EPI-06-0625 86. doi:10.1111/j.1753-4887.2006.tb00180.x
PMID:17132766 PMID:17131943
Lukanova A, Zeleniuch-Jacquotte A, Lundin E, Micheli Mason C, Risques RA, Xiao L, Duggan CR, Imayama I,
A, Arslan AA, Rinaldi S, et al. (2004b). Prediagnostic Campbell KL, et al. (2013a). Independent and combined
levels of C-peptide, IGF-I, IGFBP -1, -2 and -3 and effects of dietary weight loss and exercise on leukocyte
risk of endometrial cancer. Int J Cancer, 108(2):262–8. telomere length in postmenopausal women. Obesity
doi:10.1002/ijc.11544 PMID:14639613 (Silver Spring), 21(12):E549–54. doi:10.1002/oby.20509
Lumeng CN (2013). Innate immune activation in PMID:23640743
obesity. Mol Aspects Med, 34(1):12–29. doi:10.1016/j. Mason C, Xiao L, Duggan C, Imayama I, Foster-Schubert
mam.2012.10.002 PMID:23068074 KE, Kong A, et al. (2013b). Effects of dietary weight
Lundin E, Dossus L, Clendenen T, Krogh V, Grankvist loss and exercise on insulin-like growth factor-I and
K, Wulff M, et al. (2009). C-reactive protein and insulin-like growth factor-binding protein-3 in post-
ovarian cancer: a prospective study nested in menopausal women: a randomized controlled trial.
three cohorts (Sweden, USA, Italy). Cancer Causes Cancer Epidemiol Biomarkers Prev, 22(8):1457–63.
Control, 20(7):1151–9. doi:10.1007/s10552-009-9330-2 doi:10.1158/1055-9965.EPI-13-0337 PMID:23756654
PMID:19301134 Matsuda M, Shimomura I (2013). Increased oxidative
Ma J, Giovannucci E, Pollak M, Leavitt A, Tao Y, Gaziano stress in obesity: implications for metabolic syndrome,
JM, et al. (2004). A prospective study of plasma diabetes, hypertension, dyslipidemia, atheroscle-
C-peptide and colorectal cancer risk in men. J Natl rosis, and cancer. Obes Res Clin Pract, 7(5):e330–41.
Cancer Inst, 96(7):546–53. doi:10.1093/jnci/djh082 doi:10.1016/j.orcp.2013.05.004 PMID:24455761
PMID:15069117 Max JB, Limburg PJ, Ogunseitan A, Stolzenberg-
Ma J, Pollak MN, Giovannucci E, Chan JM, Tao Y, Solomon RZ, Vierkant RA, Pollak MJ, et al. (2008).
Hennekens CH, et al. (1999). Prospective study of IGF-I, IGFBP-3, and IGF-I/IGFBP-3 ratio: no asso-
colorectal cancer risk in men and plasma levels of ciation with incident colorectal cancer in the Alpha-
insulin-like growth factor (IGF)-I and IGF-binding Tocopherol, Beta-Carotene Cancer Prevention Study.
protein-3. J Natl Cancer Inst, 91(7):620–5. doi:10.1093/ Cancer Epidemiol Biomarkers Prev, 17(7):1832–4.
jnci/91.7.620 PMID:10203281 doi:10.1158/1055-9965.EPI-08-0345 PMID:18628438
Macis D, Guerrieri-Gonzaga A, Gandini S (2014). Mazziotti G, Sorvillo F, Morisco F, Carbone A, Rotondi M,
Circulating adiponectin and breast cancer risk: a Stornaiuolo G, et al. (2002). Serum insulin-like growth
systematic review and meta-analysis. Int J Epidemiol, factor I evaluation as a useful tool for predicting the
43(4):1226–36. doi:10.1093/ije/dyu088 PMID:24737805 risk of developing hepatocellular carcinoma in patients
Maihöfner C, Charalambous MP, Bhambra U, Lightfoot with hepatitis C virus-related cirrhosis: a prospective
T, Geisslinger G, Gooderham NJ; Colorectal Cancer study. Cancer, 95(12):2539–45. doi:10.1002/cncr.11002
Group (2003). Expression of cyclooxygenase-2 PMID:12467068
parallels expression of interleukin-1beta, inter- McCourt CK, Mutch DG, Gibb RK, Rader JS, Goodfellow
leukin-6 and NF-kappaB in human colorectal cancer. PJ, Trinkaus K, et al. (2007). Body mass index: rela-
Carcinogenesis, 24(4):665–71. doi:10.1093/carcin/ tionship to clinical, pathologic and features of micro-
bgg006 PMID:12727794 satellite instability in endometrial cancer. Gynecol
Oncol, 104(3):535–9. doi:10.1016/j.ygyno.2006.09.019
PMID:17109938
622
McCullough LE, Chen J, White AJ, Xu X, Cho YH, crosstalk. Cancer Prev Res (Phila), 5(10):1236–46.
Bradshaw PT, et al. (2015). Gene-specific promoter doi:10.1158/1940-6207.CAPR-12-0234 PMID:22896210
methylation status in hormone-receptor-positive Moore T, Beltran L, Carbajal S, Strom S, Traag J, Hursting
breast cancer associates with postmenopausal body SD, et al. (2008). Dietary energy balance modulates
size and recreational physical activity. Int J Cancer Clin signalling through the Akt/mammalian target of rapa-
Res, 2(1):013. PMID:26005715 mycin pathways in multiple epithelial tissues. Cancer
Meeker S, Seamons A, Maggio-Price L, Paik J (2016). Prev Res (Phila), 1(1):65–76. doi:10.1158/1940-6207.
Protective links between vitamin D, inflamma- CAPR-08-0022 PMID:19138937
tory bowel disease and colon cancer. World J Moreno-Castellanos N, Guzmán-Ruiz R, Cano DA,
Gastroenterol, 22(3):933–48. doi:10.3748/wjg.v22. Madrazo-Atutxa A, Peinado JR, Pereira-Cunill JL,
i3.933 PMID:26811638 et al. (2015). The effects of bariatric surgery-induced
Merry BJ (2000). Calorie restriction and age-related weight loss on adipose tissue in morbidly obese
oxidative stress. Ann N Y Acad Sci, 908(1):180–98. women depends on the initial metabolic status. Obes
doi:10.1111/j.1749-6632.2000.tb06646.x PMID:10911958 Surg, 26(8):1757–67. doi:10.1007/s11695-015-1995-x
Meydani M, Das S, Band M, Epstein S, Roberts S (2011). PMID:26678755
The effect of caloric restriction and glycemic load Morris JK, George LM, Wu T, Wald NJ (2006). Insulin-
on measures of oxidative stress and antioxidants in like growth factors and cancer: no role in screening.
humans: results from the CALERIE Trial of Human Evidence from the BUPA study and meta-analysis
Caloric Restriction. J Nutr Health Aging, 15(6):456–60. of prospective epidemiological studies. Br J Cancer,
doi:10.1007/s12603-011-0002-z PMID:21623467 95(1):112–7. doi:10.1038/sj.bjc.6603200 PMID:16804529
Meyer F, Galan P, Douville P, Bairati I, Kegle P, Bertrais Morrow JD, Roberts LJ (1997). The isoprostanes: unique
S, et al. (2005). A prospective study of the insulin-like bioactive products of lipid peroxidation. Prog Lipid
growth factor axis in relation with prostate cancer in Res, 36(1):1–21. doi:10.1016/S0163-7827(97)00001-5
the SU.VI.MAX trial. Cancer Epidemiol Biomarkers PMID:9373618
Prev, 14(9):2269–72. doi:10.1158/1055-9965.EPI-05- Moschen AR, Wieser V, Gerner RR, Bichler A, Enrich
0303 PMID:16172243 B, Moser P, et al. (2013). Adipose tissue and liver
Milagro FI, Campión J, Cordero P, Goyenechea E, expression of SIRT1, 3, and 6 increase after extensive
Gómez-Uriz AM, Abete I, et al. (2011). A dual epig- weight loss in morbid obesity. J Hepatol, 59(6):1315–22.
enomic approach for the search of obesity biomarkers: doi:10.1016/j.jhep.2013.07.027 PMID:23928404
DNA methylation in relation to diet-induced weight Mucci LA, Stark JR, Pollak MN, Li H, Kurth T, Stampfer
loss. FASEB J, 25(4):1378–89. doi:10.1096/fj.10-170365 MJ, et al. (2010). Plasma levels of acid-labile subunit,
PMID:21209057 free insulin-like growth factor-I, and prostate cancer
Mills AM, Longacre TA (2016). Lynch syndrome risk: a prospective study. Cancer Epidemiol Biomarkers
screening in the gynecologic tract: current state of Prev, 19(2):484–91. doi:10.1158/1055-9965.EPI-09-0836
the art. Am J Surg Pathol, 40(4):e35–44. doi:10.1097/ PMID:20142246
PAS.0000000000000608 PMID:26872009 Mukherjee P, Abate LE, Seyfried TN (2004).
Mink PJ, Shahar E, Rosamond WD, Alberg AJ, Folsom Antiangiogenic and proapoptotic effects of dietary
AR (2002). Serum insulin and glucose levels and restriction on experimental mouse and human
breast cancer incidence: the Atherosclerosis Risk in brain tumors. Clin Cancer Res, 10(16):5622–9.
Communities study. Am J Epidemiol, 156(4):349–52. doi:10.1158/1078-0432.CCR-04-0308 PMID:15328205
doi:10.1093/aje/kwf050 PMID:12181105 Mundstock E, Sarria EE, Zatti H, Mattos Louzada F,
Mishra AK, Dubey V, Ghosh AR (2016). Obesity: an Kich Grun L, Herbert Jones M, et al. (2015). Effect
overview of possible role(s) of gut hormones, lipid of obesity on telomere length: systematic review and
sensing and gut microbiota. Metabolism, 65(1):48–65. meta-analysis. Obesity (Silver Spring), 23(11):2165–74.
doi:10.1016/j.metabol.2015.10.008 PMID:26683796 doi:10.1002/oby.21183 PMID:26407932
Modesitt SC, Hallowell PT, Slack-Davis JK, Michalek RD, Münzberg H, Morrison CD (2015). Structure, production
Atkins KA, Kelley SL, et al. (2015). Women at extreme and signaling of leptin. Metabolism, 64(1):13–23.
risk for obesity-related carcinogenesis: baseline endo- doi:10.1016/j.metabol.2014.09.010 PMID:25305050
metrial pathology and impact of bariatric surgery on Murphy G, Kamangar F, Albanes D, Stanczyk FZ,
weight, metabolic profiles and quality of life. Gynecol Weinstein SJ, Taylor PR, et al. (2012). Serum ghrelin
Oncol, 138(2):238–45. doi:10.1016/j.ygyno.2015.05.015 is inversely associated with risk of subsequent oesoph-
PMID:26013696 ageal squamous cell carcinoma. Gut, 61(11):1533–7.
Moore T, Beltran L, Carbajal S, Hursting SD, DiGiovanni doi:10.1136/gutjnl-2011-300653 PMID:22180062
J (2012). Energy balance modulates mouse skin Murphy G, Kamangar F, Dawsey SM, Stanczyk FZ,
tumor promotion through altered IGF-1R and EGFR Weinstein SJ, Taylor PR, et al. (2011). The relationship
between serum ghrelin and the risk of gastric and
623
esophagogastric junctional adenocarcinomas. J Natl Nomura AM, Stemmermann GN, Lee J, Pollak MN (2003).
Cancer Inst, 103(14):1123–9. doi:10.1093/jnci/djr194 Serum insulin-like growth factor I and subsequent risk
PMID:21693726 of colorectal cancer among Japanese-American men.
Murphy N, Cross AJ, Abubakar M, Jenab M, Aleksandrova Am J Epidemiol, 158(5):424–31. doi:10.1093/aje/kwg176
K, Boutron-Ruault MC, et al. (2016). A nested case-con- PMID:12936897
trol study of metabolically defined body size pheno- Nossaman B, Pankey E, Kadowitz P (2012). Stimulators
types and risk of colorectal cancer in the European and activators of soluble guanylate cyclase: review and
Prospective Investigation into Cancer and Nutrition potential therapeutic indications. Crit Care Res Pract,
(EPIC). PLoS Med, 13(4):e1001988. doi:10.1371/journal. 2012:290805. doi:10.1155/2012/290805 PMID:22482042
pmed.1001988 PMID:27046222 Ohishi W, Cologne JB, Fujiwara S, Suzuki G, Hayashi T,
Murphy N, Strickler HD, Stanczyk FZ, Xue X, Wassertheil- Niwa Y, et al. (2014). Serum interleukin-6 associated
Smoller S, Rohan TE, et al. (2015). A prospective with hepatocellular carcinoma risk: a nested case-con-
evaluation of endogenous sex hormone levels and trol study. Int J Cancer, 134(1):154–63. doi:10.1002/
colorectal cancer risk in postmenopausal women. ijc.28337 PMID:23784949
J Natl Cancer Inst, 107(10):djv210. doi:10.1093/jnci/ Olivo-Marston SE, Hursting SD, Perkins SN, Schetter
djv210 PMID:26232761 A, Khan M, Croce C, et al. (2014). Effects of calorie
Muti P, Quattrin T, Grant BJ, Krogh V, Micheli A, restriction and diet-induced obesity on murine colon
Schünemann HJ, et al. (2002). Fasting glucose is a carcinogenesis, growth and inflammatory factors,
risk factor for breast cancer: a prospective study. and microRNA expression. PLoS One, 9(4):e94765.
Cancer Epidemiol Biomarkers Prev, 11(11):1361–8. doi:10.1371/journal.pone.0094765 PMID:24732966
PMID:12433712 Ollberding NJ, Cheng I, Wilkens LR, Henderson BE,
Nam SY, Lee EJ, Kim KR, Cha BS, Song YD, Lim SK, Pollak MN, Kolonel LN, et al. (2012). Genetic variants,
et al. (1997). Effect of obesity on total and free insu- prediagnostic circulating levels of insulin-like growth
lin-like growth factor (IGF)-1, and their relationship factors, insulin, and glucose and the risk of colorectal
to IGF-binding protein (BP)-1, IGFBP-2, IGFBP-3, cancer: the Multiethnic Cohort study. Cancer Epidemiol
insulin, and growth hormone. Int J Obes Relat Biomarkers Prev, 21(5):810–20. doi:10.1158/1055-9965.
Metab Disord, 21(5):355–9. doi:10.1038/sj.ijo.0800412 EPI-11-1105 PMID:22354904
PMID:9152736 Ose J, Schock H, Tjønneland A, Hansen L, Overvad K,
Nead KT, Sharp SJ, Thompson DJ, Painter JN, Savage DB, Dossus L, et al. (2015). Inflammatory markers and
Semple RK, et al.; Australian National Endometrial risk of epithelial ovarian cancer by tumor subtypes:
Cancer Study Group (ANECS) (2015). Evidence of a the EPIC cohort. Cancer Epidemiol Biomarkers Prev,
causal association between insulinemia and endo- 24(6):951–61. doi:10.1158/1055-9965.EPI-14-1279-T
metrial cancer: a Mendelian randomization analysis. PMID:25855626
J Natl Cancer Inst, 107(9):djv178. doi:10.1093/jnci/ Otani T, Iwasaki M, Sasazuki S, Inoue M, Tsugane S; Japan
djv178 PMID:26134033 Public Health Center-based Prospective Study Group
Neels JG, Olefsky JM (2006). Inflamed fat: what starts the (2007). Plasma C-peptide, insulin-like growth factor-I,
fire? J Clin Invest, 116(1):33–5. doi:10.1172/JCI27280 insulin-like growth factor binding proteins and risk of
PMID:16395402 colorectal cancer in a nested case-control study: the
Nicklas BJ, Rogus EM, Colman EG, Goldberg AP (1996). Japan Public Health Center-based Prospective Study.
Visceral adiposity, increased adipocyte lipolysis, and Int J Cancer, 120(9):2007–12. doi:10.1002/ijc.22556
metabolic dysfunction in obese postmenopausal PMID:17266031
women. Am J Physiol, 270(1 Pt 1):E72–8. PMID:8772476 Palmqvist R, Hallmans G, Rinaldi S, Biessy C, Stenling
Nieman DC, Nehlsen-Cannarella SL, Henson DA, R, Riboli E, et al. (2002). Plasma insulin-like growth
Koch AJ, Butterworth DE, Fagoaga OR, et al. (1998). factor 1, insulin-like growth factor binding protein
Immune response to exercise training and/or energy 3, and risk of colorectal cancer: a prospective study
restriction in obese women. Med Sci Sports Exerc, in northern Sweden. Gut, 50(5):642–6. doi:10.1136/
30(5):679–86. doi:10.1097/00005768-199805000-00006 gut.50.5.642 PMID:11950809
PMID:9588608 Palmqvist R, Stattin P, Rinaldi S, Biessy C, Stenling R,
Nogueira LM, Lavigne JA, Chandramouli GV, Lui H, Riboli E, et al. (2003). Plasma insulin, IGF-binding
Barrett JC, Hursting SD (2012). Dose-dependent effects proteins-1 and -2 and risk of colorectal cancer: a
of calorie restriction on gene expression, metabolism, prospective study in northern Sweden. Int J Cancer,
and tumor progression are partially mediated by insu- 107(1):89–93. doi:10.1002/ijc.11362 PMID:12925961
lin-like growth factor-1. Cancer Med, 1(2):275–88. Papandreou D, Andreou E (2015). Role of diet on non-al-
doi:10.1002/cam4.23 PMID:23342276 coholic fatty liver disease: an updated narrative review.
World J Hepatol, 7(3):575–82. doi:10.4254/wjh.v7.i3.575
PMID:25848481
624
Parekh N, Lin Y, Vadiveloo M, Hayes RB, Lu-Yao GL Powolny AA, Wang S, Carlton PS, Hoot DR, Clinton SK
(2013). Metabolic dysregulation of the insulin-glu- (2008). Interrelationships between dietary restriction,
cose axis and risk of obesity-related cancers in the IGF-I axis, and expression of vascular endothelial
the Framingham Heart Study-Offspring Cohort growth factor by prostate adenocarcinoma in rats.
(1971–2008). Cancer Epidemiol Biomarkers Prev, Mol Carcinog, 47(6):458–65. doi:10.1002/mc.20403
22(10):1825–36. doi:10.1158/1055-9965.EPI-13-0330 PMID:18058807
PMID:24064521 Price AJ, Allen NE, Appleby PN, Crowe FL, Travis RC,
Park HK, Ahima RS (2015). Physiology of leptin: Tipper SJ, et al. (2012). Insulin-like growth factor-I
energy homeostasis, neuroendocrine function and concentration and risk of prostate cancer: results from
metabolism. Metabolism, 64(1):24–34. doi:10.1016/j. the European Prospective Investigation into Cancer
metabol.2014.08.004 PMID:25199978 and Nutrition. Cancer Epidemiol Biomarkers Prev,
Peeters PH, Lukanova A, Allen N, Berrino F, Key T, 21(9):1531–41. doi:10.1158/1055-9965.EPI-12-0481-T
Dossus L, et al. (2007). Serum IGF-I, its major binding PMID:22761305
protein (IGFBP-3) and epithelial ovarian cancer risk: Prizment AE, Folsom AR, Dreyfus J, Anderson KE,
the European Prospective Investigation into Cancer Visvanathan K, Joshu CE, et al. (2013). Plasma
and Nutrition (EPIC). Endocr Relat Cancer, 14(1):81– C-reactive protein, genetic risk score, and risk
90. doi:10.1677/erc.1.01264 PMID:17395977 of common cancers in the Atherosclerosis Risk
Peng Y, Rideout DA, Rakita SS, Gower WR Jr, You M, in Communities study. Cancer Causes Control,
Murr MM (2010). Does LKB1 mediate activation of 24(12):2077–87. doi:10.1007/s10552-013-0285-y
hepatic AMP-protein kinase (AMPK) and sirtuin1 PMID:24036889
(SIRT1) after Roux-en-Y gastric bypass in obese rats? Probst-Hensch NM, Yuan JM, Stanczyk FZ, Gao YT,
J Gastrointest Surg, 14(2):221–8. doi:10.1007/s11605- Ross RK, Yu MC (2001). IGF-1, IGF-2 and IGFBP-3
009-1102-5 PMID:19937189 in prediagnostic serum: association with colorectal
Pereira-Santos M, Costa PR, Assis AM, Santos CA, Santos cancer in a cohort of Chinese men in Shanghai. Br J
DB (2015). Obesity and vitamin D deficiency: a system- Cancer, 85(11):1695–9. doi:10.1054/bjoc.2001.2172
atic review and meta-analysis. Obes Rev, 16(4):341–9. PMID:11742490
doi:10.1111/obr.12239 PMID:25688659 Rappou E, Jukarainen S, Rinnankoski-Tuikka R, Kaye S,
Pischon T, Nimptsch K (2016). Obesity and risk of cancer: Heinonen S, Hakkarainen A, et al. (2016). Weight loss
an introductory overview. Recent Results Cancer is associated with increased NAD+/SIRT1 expression
Res, 208:1–15. doi:10.1007/978-3-319-42542-9_1 but reduced PARP activity in white adipose tissue.
PMID:27909899 J Clin Endocrinol Metab, 101(3):1263–73. doi:10.1210/
Platz EA, Leitzmann MF, Rifai N, Kantoff PW, Chen YC, jc.2015-3054 PMID:26760174
Stampfer MJ, et al. (2005a). Sex steroid hormones and Ravillah D, Mohammed A, Qian L, Brewer M, Zhang
the androgen receptor gene CAG repeat and subsequent Y, Biddick L, et al. (2014). Chemopreventive effects of
risk of prostate cancer in the prostate-specific antigen an HDAC2-selective inhibitor on rat colon carcino-
era. Cancer Epidemiol Biomarkers Prev, 14(5):1262–9. genesis and APCmin/+ mouse intestinal tumorigenesis.
doi:10.1158/1055-9965.EPI-04-0371 PMID:15894683 J Pharmacol Exp Ther, 348(1):59–68. doi:10.1124/
Platz EA, Pollak MN, Leitzmann MF, Stampfer MJ, Willett jpet.113.208645 PMID:24218540
WC, Giovannucci E (2005b). Plasma insulin-like Relton CL, Groom A, St Pourcain B, Sayers AE, Swan DC,
growth factor-1 and binding protein-3 and subsequent Embleton ND, et al. (2012). DNA methylation patterns
risk of prostate cancer in the PSA era. Cancer Causes in cord blood DNA and body size in childhood. PLoS
Control, 16(3):255–62. doi:10.1007/s10552-004-3484-8 One, 7(3):e31821. doi:10.1371/journal.pone.0031821
PMID:15947877 PMID:22431966
Polvani S, Tarocchi M, Tempesti S, Bencini L, Galli A Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen
(2016). Peroxisome proliferator activated receptors M (2008). Body-mass index and incidence of cancer:
at the crossroad of obesity, diabetes, and pancre- a systematic review and meta-analysis of prospec-
atic cancer. World J Gastroenterol, 22(8):2441–59. tive observational studies. Lancet, 371(9612):569–78.
doi:10.3748/wjg.v22.i8.2441 PMID:26937133 doi:10.1016/S0140-6736(08)60269-X PMID:18280327
Poole EM, Lee IM, Ridker PM, Buring JE, Hankinson SE, Rinaldi S, Cleveland R, Norat T, Biessy C, Rohrmann S,
Tworoger SS (2013). A prospective study of circulating Linseisen J, et al. (2010). Serum levels of IGF-I, IGFBP-3
C-reactive protein, interleukin-6, and tumor necrosis and colorectal cancer risk: results from the EPIC
factor α receptor 2 levels and risk of ovarian cancer. Am cohort, plus a meta-analysis of prospective studies. Int
J Epidemiol, 178(8):1256–64. doi:10.1093/aje/kwt098 J Cancer, 126(7):1702–15. PMID:19810099
PMID:23966559 Rinaldi S, Dossus L, Lukanova A, Peeters PH, Allen NE,
Key T, et al. (2007). Endogenous androgens and risk
of epithelial ovarian cancer: results from the European
625
Prospective Investigation into Cancer and Nutrition Sawada N, Iwasaki M, Inoue M, Sasazuki S, Yamaji T,
(EPIC). Cancer Epidemiol Biomarkers Prev, 16(1):23–9. Shimazu T, et al.; Japan Public Health Center-based
doi:10.1158/1055-9965.EPI-06-0755 PMID:17220328 Prospective Study Group (2010). Plasma testosterone
Rinaldi S, Toniolo P, Muti P, Lundin E, Zeleniuch- and sex hormone-binding globulin concentrations
Jacquotte A, Arslan A, et al. (2005). IGF-I, IGFBP-3 and the risk of prostate cancer among Japanese men:
and breast cancer in young women: a pooled re-ana- a nested case-control study. Cancer Sci, 101(12):2652–7.
lysis of three prospective studies. Eur J Cancer Prev, doi:10.1111/j.1349-7006.2010.01721.x PMID:20942896
14(6):493–6. doi:10.1097/00008469-200512000-00001 Saydah SH, Platz EA, Rifai N, Pollak MN, Brancati FL,
PMID:16284492 Helzlsouer KJ (2003). Association of markers of insulin
Rizkalla SW, Prifti E, Cotillard A, Pelloux V, Rouault C, and glucose control with subsequent colorectal cancer
Allouche R, et al. (2012). Differential effects of macro- risk. Cancer Epidemiol Biomarkers Prev, 12(5):412–8.
nutrient content in 2 energy-restricted diets on cardi- PMID:12750235
ovascular risk factors and adipose tissue cell size in Schenk JM, Kristal AR, Neuhouser ML, Tangen CM,
moderately obese individuals: a randomized controlled White E, Lin DW, et al. (2009). Serum adiponectin,
trial. Am J Clin Nutr, 95(1):49–63. doi:10.3945/ C-peptide and leptin and risk of symptomatic benign
ajcn.111.017277 PMID:22170375 prostatic hyperplasia: results from the Prostate Cancer
Rock CL, Emond JA, Flatt SW, Heath DD, Karanja N, Prevention Trial. Prostate, 69(12):1303–11. doi:10.1002/
Pakiz B, et al. (2012). Weight loss is associated with pros.20974 PMID:19475640
increased serum 25-hydroxyvitamin D in overweight Schernhammer ES, Sperati F, Razavi P, Agnoli C, Sieri
or obese women. Obesity (Silver Spring), 20(11):2296– S, Berrino F, et al. (2013). Endogenous sex steroids
301. doi:10.1038/oby.2012.57 PMID:22402737 in premenopausal women and risk of breast cancer:
Roddam AW, Allen NE, Appleby P, Key TJ; Endogenous the ORDET cohort. Breast Cancer Res, 15(3):R46.
Hormones and Prostate Cancer Collaborative Group doi:10.1186/bcr3438 PMID:23777922
(2008). Endogenous sex hormones and prostate cancer: Schoen RE, Tangen CM, Kuller LH, Burke GL, Cushman
a collaborative analysis of 18 prospective studies. J Natl M, Tracy RP, et al. (1999). Increased blood glucose and
Cancer Inst, 100(3):170–83. doi:10.1093/jnci/djm323 insulin, body size, and incident colorectal cancer. J Natl
PMID:18230794 Cancer Inst, 91(13):1147–54. doi:10.1093/jnci/91.13.1147
Rohrmann S, Grote VA, Becker S, Rinaldi S, Tjønneland PMID:10393723
A, Roswall N, et al. (2012). Concentrations of IGF-I and Severi G, Morris HA, MacInnis RJ, English DR, Tilley W,
IGFBP-3 and pancreatic cancer risk in the European Hopper JL, et al. (2006a). Circulating steroid hormones
Prospective Investigation into Cancer and Nutrition. and the risk of prostate cancer. Cancer Epidemiol
Br J Cancer, 106(5):1004–10. doi:10.1038/bjc.2012.19 Biomarkers Prev, 15(1):86–91. doi:10.1158/1055-9965.
PMID:22315049 EPI-05-0633 PMID:16434592
Rossi EL, de Angel RE, Bowers LW, Khatib SA, Smith LA, Severi G, Morris HA, MacInnis RJ, English DR, Tilley
Van Buren E, et al. (2016). Obesity-associated alter- WD, Hopper JL, et al. (2006b). Circulating insulin-like
ations in inflammation, epigenetics and mammary growth factor-I and binding protein-3 and risk of
tumor growth persist in formerly obese mice. Cancer prostate cancer. Cancer Epidemiol Biomarkers Prev,
Prev Res (Phila), doi:10.1158/1940-6207.CAPR-15-0348 15(6):1137–41. doi:10.1158/1055-9965.EPI-05-0823
PMID:26869351 PMID:16775172
Roszkowski K (2014). Oxidative DNA damage – the Shade ED, Ulrich CM, Wener MH, Wood B, Yasui Y,
possible use of biomarkers as additional prognostic Lacroix K, et al. (2004). Frequent intentional weight
factors in oncology. Front Biosci (Landmark Ed), loss is associated with lower natural killer cell cyto-
19(5):808–17. doi:10.2741/4248 PMID:24389225 toxicity in postmenopausal women: possible long-
Sainsbury A, Goodlad RA, Perry SL, Pollard SG, Robins term immune effects. J Am Diet Assoc, 104(6):903–12.
GG, Hull MA (2008). Increased colorectal epithelial cell doi:10.1016/j.jada.2004.03.018 PMID:15175588
proliferation and crypt fission associated with obesity Shafie SM, Grantham FH (1981). Role of hormones in the
and Roux-en-Y gastric bypass. Cancer Epidemiol growth and regression of human breast cancer cells
Biomarkers Prev, 17(6):1401–10. doi:10.1158/1055-9965. (MCF-7) transplanted into athymic nude mice. J Natl
EPI-07-2874 PMID:18559555 Cancer Inst, 67(1):51–6. PMID:6265682
Sakauchi F, Nojima M, Mori M, Wakai K, Suzuki S, Shafie SM, Hilf R (1981). Insulin receptor levels and magni-
Tamakoshi A, et al. (2009). Serum insulin-like growth tude of insulin-induced responses in 7,12-dimethyl
factors I and II, insulin-like growth factor binding benz(a)anthracene-induced mammary tumors in rats.
protein-3 and risk of breast cancer in the Japan Cancer Res, 41(3):826–9. PMID:6780189
Collaborative Cohort study. Asian Pac J Cancer Prev, Shungin D, Winkler TW, Croteau-Chonka DC,
10(Suppl):51–5. PMID:20553082 Ferreira T, Locke AE, Mägi R, et al.; ADIPOGen
Consortium; CARDIOGRAMplusC4D Consortium;
626
CKDGen Consortium; GEFOS Consortium; individuals reveal 18 new loci associated with body
GENIE Consortium; GLGC; ICBP; International mass index. Nat Genet, 42(11):937–48. doi:10.1038/
Endogene Consortium; LifeLines Cohort Study; ng.686 PMID:20935630
MAGIC Investigators; MuTHER Consortium; PAGE Stattin P, Bylund A, Rinaldi S, Biessy C, Déchaud H,
Consortium; ReproGen Consortium (2015). New Stenman UH, et al. (2000). Plasma insulin-like growth
genetic loci link adipose and insulin biology to body factor-I, insulin-like growth factor-binding proteins,
fat distribution. Nature, 518(7538):187–96. doi:10.1038/ and prostate cancer risk: a prospective study. J Natl
nature14132 PMID:25673412 Cancer Inst, 92(23):1910–7. doi:10.1093/jnci/92.23.1910
Siemes C, Visser LE, Coebergh JW, Splinter TA, PMID:11106682
Witteman JC, Uitterlinden AG, et al. (2006). C-reactive Stattin P, Lukanova A, Biessy C, Söderberg S, Palmqvist
protein levels, variation in the C-reactive protein R, Kaaks R, et al. (2004b). Obesity and colon cancer:
gene, and cancer risk: the Rotterdam Study. J Clin does leptin provide a link? Int J Cancer, 109(1):149–52.
Oncol, 24(33):5216–22. doi:10.1200/JCO.2006.07.1381 doi:10.1002/ijc.11668 PMID:14735482
PMID:17114654 Stattin P, Palmqvist R, Söderberg S, Biessy C, Ardnor B,
Sieri S, Muti P, Claudia A, Berrino F, Pala V, Grioni S, Hallmans G, et al. (2003). Plasma leptin and colorectal
et al. (2012). Prospective study on the role of glucose cancer risk: a prospective study in Northern Sweden.
metabolism in breast cancer occurrence. Int J Cancer, Oncol Rep, 10(6):2015–21. PMID:14534736
130(4):921–9. doi:10.1002/ijc.26071 PMID:21413010 Stattin P, Rinaldi S, Biessy C, Stenman UH, Hallmans
Simon MS, Chlebowski RT, Wactawski-Wende J, Johnson G, Kaaks R (2004c). High levels of circulating insu-
KC, Muskovitz A, Kato I, et al. (2012). Estrogen lin-like growth factor-I increase prostate cancer risk:
plus progestin and colorectal cancer incidence and a prospective study in a population-based nonscreened
mortality. J Clin Oncol, 30(32):3983–90. doi:10.1200/ cohort. J Clin Oncol, 22(15):3104–12. doi:10.1200/
JCO.2012.42.7732 PMID:23008295 JCO.2004.10.105 PMID:15284261
Skrha J (2009). Effect of caloric restriction on oxidative Stattin P, Söderberg S, Biessy C, Lenner P, Hallmans
markers. Adv Clin Chem, 47:223–47. doi:10.1016/ G, Kaaks R, et al. (2004a). Plasma leptin and breast
S0065-2423(09)47008-2 PMID:19634782 cancer risk: a prospective study in northern Sweden.
Smith MT, Guyton KZ, Gibbons CF, Fritz JM, Portier CJ, Breast Cancer Res Treat, 86(3):191–6. doi:10.1023/
Rusyn I, et al. (2016). Key characteristics of carcino- B:BREA.0000036782.11945.d7 PMID:15567935
gens as a basis for organizing data on mechanisms of Stocks T, Lukanova A, Johansson M, Rinaldi S,
carcinogenesis. Environ Health Perspect, 124(6):713–21. Palmqvist R, Hallmans G, et al. (2008). Components
doi:10.1289/ehp.1509912 PMID:26600562 of the metabolic syndrome and colorectal cancer risk;
Smits MM, van Geenen EJ (2011). The clinical signifi- a prospective study. Int J Obes, 32(2):304–14.
cance of pancreatic steatosis. Nat Rev Gastroenterol doi:10.1038/sj.ijo.0803713 PMID:17878894
Hepatol, 8(3):169–77. doi:10.1038/nrgastro.2011.4 Stocks T, Lukanova A, Rinaldi S, Biessy C, Dossus L,
PMID:21304475 Lindahl B, et al. (2007). Insulin resistance is inversely
Soares MJ, Murhadi LL, Kurpad AV, Chan She Ping- related to prostate cancer: a prospective study in
Delfos WL, Piers LS (2012). Mechanistic roles for Northern Sweden. Int J Cancer, 120(12):2678–86.
calcium and vitamin D in the regulation of body doi:10.1002/ijc.22587 PMID:17278097
weight. Obes Rev, 13(7):592–605. doi:10.1111/j.1467- Stolzenberg-Solomon RZ, Graubard BI, Chari S,
789X.2012.00986.x PMID:22385576 Limburg P, Taylor PR, Virtamo J, et al. (2005). Insulin,
Soliman PT, Cui X, Zhang Q, Hankinson SE, Lu KH glucose, insulin resistance, and pancreatic cancer in
(2011). Circulating adiponectin levels and risk of endo- male smokers. JAMA, 294(22):2872–8. doi:10.1001/
metrial cancer: the prospective Nurses’ Health Study. jama.294.22.2872 PMID:16352795
Am J Obstet Gynecol, 204(2):167.e1–5. doi:10.1016/j. Stolzenberg-Solomon RZ, Limburg P, Pollak M, Taylor
ajog.2010.08.045 PMID:21047616 PR, Virtamo J, Albanes D (2004). Insulin-like growth
Song M, Zhang X, Wu K, Ogino S, Fuchs CS, Giovannucci factor (IGF)-1, IGF-binding protein-3, and pancreatic
EL, et al. (2013). Plasma adiponectin and soluble leptin cancer in male smokers. Cancer Epidemiol Biomarkers
receptor and risk of colorectal cancer: a prospec- Prev, 13(3):438–44. PMID:15006921
tive study. Cancer Prev Res (Phila), 6(9):875–85. Stolzenberg-Solomon RZ, Newton CC, Silverman DT,
doi:10.1158/1940-6207.CAPR-13-0169 PMID:23872505 Pollak M, Nogueira LM, Weinstein SJ, et al. (2015).
Song Q, Sergeev IN (2012). Calcium and vitamin D Circulating leptin and risk of pancreatic cancer:
in obesity. Nutr Res Rev, 25(1):130–41. doi:10.1017/ a pooled analysis from 3 cohorts. Am J Epidemiol,
S0954422412000029 PMID:22588363 182(3):187–97. doi:10.1093/aje/kwv041 PMID:26085045
Speliotes EK, Willer CJ, Berndt SI, Monda KL, Stolzenberg-Solomon RZ, Weinstein S, Pollak M, Tao
Thorleifsson G, Jackson AU, et al.; MAGIC; Procardis Y, Taylor PR, Virtamo J, et al. (2008). Prediagnostic
Consortium (2010). Association analyses of 249,796 adiponectin concentrations and pancreatic cancer
627
risk in male smokers. Am J Epidemiol, 168(9):1047–55. Thompson IM, Goodman PJ, Tangen CM, Lucia MS,
doi:10.1093/aje/kwn221 PMID:18801887 Miller GJ, Ford LG, et al. (2003). The influence of
Subbaramaiah K, Howe LR, Bhardwaj P, Du B, Gravaghi finasteride on the development of prostate cancer.
C, Yantiss RK, et al. (2011). Obesity is associated with N Engl J Med, 349(3):215–24. doi:10.1056/
inflammation and elevated aromatase expression in NEJMoa030660 PMID:12824459
the mouse mammary gland. Cancer Prev Res (Phila), Tomita M (2012). Caloric restriction reduced 1, 2-
4(3):329–46. doi:10.1158/1940-6207.CAPR-10-0381 dimethylhydrazine-induced aberrant crypt foci and
PMID:21372033 induces the expression of Sirtuins in colonic mucosa
Suganami T, Tanaka M, Ogawa Y (2012). Adipose of F344 rats. J Carcinog, 11(1):10. doi:10.4103/1477-
tissue inflammation and ectopic lipid accumulation. 3163.99176 PMID:22919283
Endocr J, 59(10):849–57. doi:10.1507/endocrj.EJ12-0271 Toniolo P, Bruning PF, Akhmedkhanov A, Bonfrer
PMID:22878669 JM, Koenig KL, Lukanova A, et al. (2000). Serum
Suzuki S, Kojima M, Tokudome S, Suzuki K, Ozasa K, Ito insulin-like growth factor-I and breast cancer.
Y, et al.; JACC Study Group (2009). Insulin-like growth Int J Cancer, 88(5):828–32. doi:10.1002/1097-
factor (IGF)-I, IGF-II, IGF binding protein-3, and risk 0215(20001201)88:5<828::AID-IJC22>3.0.CO;2-8
of colorectal cancer: a nested case-control study in the PMID:11072255
Japan Collaborative Cohort study. Asian Pac J Cancer Touvier M, Fezeu L, Ahluwalia N, Julia C, Charnaux N,
Prev, 10(Suppl):45–9. PMID:20553081 Sutton A, et al. (2013). Association between prediag-
Tamakoshi K, Toyoshima H, Wakai K, Kojima M, Suzuki nostic biomarkers of inflammation and endothelial
K, Watanabe Y, et al. (2005). Leptin is associated with function and cancer risk: a nested case-control study.
an increased female colorectal cancer risk: a nested Am J Epidemiol, 177(1):3–13. doi:10.1093/aje/kws359
case-control study in Japan. Oncology, 68(4–6):454–61. PMID:23171880
doi:10.1159/000086988 PMID:16020976 Trabert B, Brinton LA, Anderson GL, Pfeiffer RM, Falk
Tannenbaum GS, Guyda HJ, Posner BI (1983). Insulin- RT, Strickler HD, et al. (2016). Circulating estro-
like growth factors: a role in growth hormone nega- gens and postmenopausal ovarian cancer risk in
tive feedback and body weight regulation via brain. the Women’s Health Initiative Observational Study.
Science, 220(4592):77–9. doi:10.1126/science.6338593 Cancer Epidemiol Biomarkers Prev, 25(4):648–56.
PMID:6338593 doi:10.1158/1055-9965.EPI-15-1272-T PMID:26908437
Tao MH, Marian C, Nie J, Ambrosone C, Krishnan SS, Trabert B, Pinto L, Hartge P, Kemp T, Black A, Sherman
Edge SB, et al. (2011). Body mass and DNA promoter ME, et al. (2014). Pre-diagnostic serum levels of
methylation in breast tumors in the Western New inflammation markers and risk of ovarian cancer in
York Exposures and Breast Cancer Study. Am J the Prostate, Lung, Colorectal and Ovarian Cancer
Clin Nutr, 94(3):831–8. doi:10.3945/ajcn.110.009365 (PLCO) Screening Trial. Gynecol Oncol, 135(2):297–
PMID:21775555 304. doi:10.1016/j.ygyno.2014.08.025 PMID:25158036
Thiele JR, Zeller J, Bannasch H, Stark GB, Peter K, Tran TT, Naigamwalla D, Oprescu AI, Lam L, McKeown-
Eisenhardt SU (2015). Targeting C-reactive protein Eyssen G, Bruce WR, et al. (2006). Hyperinsulinemia,
in inflammatory disease by preventing conforma- but not other factors associated with insulin resist-
tion changes. Mediators Inflamm, 2015:372432. ance, acutely enhances colorectal epithelial prolifera-
doi:10.1155/2015/372432 PMID:26089599 tion in vivo. Endocrinology, 147(4):1830–7. doi:10.1210/
Thompson HJ, McGinley JN, Spoelstra NS, Jiang W, Zhu en.2005-1012 PMID:16410309
Z, Wolfe P (2004b). Effect of dietary energy restriction Travis RC, Appleby PN, Martin RM, Holly JM, Albanes
on vascular density during mammary carcinogenesis. D, Black A, et al.; Endogenous Hormones, Nutritional
Cancer Res, 64(16):5643–50. doi:10.1158/0008-5472. Biomarkers and Prostate Cancer Collaborative
CAN-04-0787 PMID:15313902 Group (2016). A meta-analysis of individual partici-
Thompson HJ, Sedlacek SM, Paul D, Wolfe P, McGinley pant data reveals an association between circulating
JN, Playdon MC, et al. (2012). Effect of dietary patterns levels of IGF-I and prostate cancer risk. Cancer Res,
differing in carbohydrate and fat content on blood 76(8):2288–300. doi:10.1158/0008-5472.CAN-15-1551
lipid and glucose profiles based on weight-loss success PMID:26921328
of breast-cancer survivors. Breast Cancer Res, 14(1):R1. Turnbaugh PJ, Hamady M, Yatsunenko T, Cantarel BL,
doi:10.1186/bcr3082 PMID:22225711 Duncan A, Ley RE, et al. (2009). A core gut microbiome
Thompson HJ, Zhu Z, Jiang W (2004a). Identification of in obese and lean twins. Nature, 457(7228):480–4.
the apoptosis activation cascade induced in mammary doi:10.1038/nature07540 PMID:19043404
carcinomas by energy restriction. Cancer Res, Tworoger SS, Eliassen AH, Kelesidis T, Colditz GA, Willett
64(4):1541–5. doi:10.1158/0008-5472.CAN-03-3108 WC, Mantzoros CS, et al. (2007b). Plasma adiponectin
PMID:14973070 concentrations and risk of incident breast cancer. J Clin
628
Endocrinol Metab, 92(4):1510–6. doi:10.1210/jc.2006- Vona-Davis L, Rose DP (2007). Adipokines as endocrine,

1975 PMID:17213279 paracrine, and autocrine factors in breast cancer risk
Tworoger SS, Lee IM, Buring JE, Pollak MN, Hankinson and progression. Endocr Relat Cancer, 14(2):189–206.
SE (2007a). Insulin-like growth factors and ovarian doi:10.1677/ERC-06-0068 PMID:17639037
cancer risk: a nested case-control study in three cohorts. Waliszewski P, Blaszczyk M, Wolinska-Witort E, Drews
Cancer Epidemiol Biomarkers Prev, 16(8):1691–5. M, Snochowski M, Hurst RE (1997). Molecular study
doi:10.1158/1055-9965.EPI-07-0319 PMID:17684148 of sex steroid receptor gene expression in human colon
Tworoger SS, Rosner BA, Willett WC, Hankinson SE and in colorectal carcinomas. J Surg Oncol, 64(1):3–11.
(2011). The combined influence of multiple sex and doi:10.1002/(SICI)1096-9098(199701)64:1<3::AID-
growth hormones on risk of postmenopausal breast JSO2>3.0.CO;2-G PMID:9040793
cancer: a nested case-control study. Breast Cancer Res, Wang G, Li N, Chang S, Bassig BA, Guo L, Ren J, et al.
13(5):R99. doi:10.1186/bcr3040 PMID:22017816 (2015). A prospective follow-up study of the relationship
Tzotzas T, Papadopoulou FG, Tziomalos K, Karras between C-reactive protein and human cancer risk in
S, Gastaris K, Perros P, et al. (2010). Rising serum the Chinese Kailuan Female Cohort. Cancer Epidemiol
25-hydroxy-vitamin D levels after weight loss in Biomarkers Prev, 24(2):459–65. doi:10.1158/1055-9965.
obese women correlate with improvement in insulin EPI-14-1112 PMID:25490990
resistance. J Clin Endocrinol Metab, 95(9):4251–7. Wang T, Rohan TE, Gunter MJ, Xue X, Wactawski-Wende
doi:10.1210/jc.2010-0757 PMID:20534751 J, Rajpathak SN, et al. (2011). A prospective study of
van Dijk SJ, Molloy PL, Varinli H, Morrison JL, inflammation markers and endometrial cancer risk in
Muhlhausler BS, Buckley M, et al.; Members of postmenopausal hormone nonusers. Cancer Epidemiol
EpiSCOPE (2015). Epigenetics and human obesity. Biomarkers Prev, 20(5):971–7. doi:10.1158/1055-9965.
Int J Obes (Lond), 39(1):85–97. doi:10.1038/ijo.2014.34 EPI-10-1222 PMID:21415362
PMID:24566855 Wang X, Zhu H, Snieder H, Su S, Munn D, Harshfield
van Geenen EJ, Smits MM, Schreuder TC, van der Peet G, et al. (2010). Obesity related methylation changes
DL, Bloemena E, Mulder CJ (2010). Nonalcoholic in DNA of peripheral blood leukocytes. BMC Med,
fatty liver disease is related to nonalcoholic fatty 8(1):87. doi:10.1186/1741-7015-8-87 PMID:21176133
pancreas disease. Pancreas, 39(8):1185–90. doi:10.1097/ Wegman MP, Guo MH, Bennion DM, Shankar MN,
MPA.0b013e3181f6fce2 PMID:20871475 Chrzanowski SM, Goldberg LA, et al. (2015).
Vanlint S (2013). Vitamin D and obesity. Nutrients, Practicality of intermittent fasting in humans and its
5(3):949–56. doi:10.3390/nu5030949 PMID:23519290 effect on oxidative stress and genes related to aging
Vatten LJ, Holly JM, Gunnell D, Tretli S (2008). Nested and metabolism. Rejuvenation Res, 18(2):162–72.
case-control study of the association of circulating doi:10.1089/rej.2014.1624 PMID:25546413
levels of serum insulin-like growth factor I and insu- Wei EK, Giovannucci E, Fuchs CS, Willett WC, Mantzoros
lin-like growth factor binding protein 3 with breast CS (2005b). Low plasma adiponectin levels and risk of
cancer in young women in Norway. Cancer Epidemiol colorectal cancer in men: a prospective study. J Natl
Biomarkers Prev, 17(8):2097–100. doi:10.1158/1055- Cancer Inst, 97(22):1688–94. doi:10.1093/jnci/dji376
9965.EPI-08-0212 PMID:18708402 PMID:16288122
Verheus M, Peeters PH, Rinaldi S, Dossus L, Biessy Wei EK, Ma J, Pollak MN, Rifai N, Fuchs CS, Hankinson
C, Olsen A, et al. (2006). Serum C-peptide levels SE, et al. (2005a). A prospective study of C-peptide,
and breast cancer risk: results from the European insulin-like growth factor-I, insulin-like growth factor
Prospective Investigation into Cancer and Nutrition binding protein-1, and the risk of colorectal cancer in
(EPIC). Int J Cancer, 119(3):659–67. doi:10.1002/ women. Cancer Epidemiol Biomarkers Prev, 14(4):850–5.
ijc.21861 PMID:16572422 doi:10.1158/1055-9965.EPI-04-0661 PMID:15824155
Villanueva EC, Myers MG Jr (2008). Leptin receptor sign- Weisenberger DJ, Levine AJ, Long TI, Buchanan DD,
aling and the regulation of mammalian physiology. Walters R, Clendenning M, et al.; for the Colon
Int J Obes, 32(Suppl 7):S8–12. doi:10.1038/ijo.2008.232 Cancer Family Registry (2015). Association of the
PMID:19136996 colorectal CpG island methylator phenotype with
Vincent HK, Innes KE, Vincent KR (2007). Oxidative molecular features, risk factors, and family history.
stress and potential interventions to reduce oxidative Cancer Epidemiol Biomarkers Prev, 24(3):512–9.
stress in overweight and obesity. Diabetes Obes Metab, doi:10.1158/1055-9965.EPI-14-1161 PMID:25587051
9(6):813–39. doi:10.1111/j.1463-1326.2007.00692.x Weiss JM, Huang WY, Rinaldi S, Fears TR, Chatterjee N,
PMID:17924865 Chia D, et al. (2007). IGF-1 and IGFBP-3: risk of prostate
Vogelstein B, Papadopoulos N, Velculescu VE, Zhou cancer among men in the Prostate, Lung, Colorectal
S, Diaz LA Jr, Kinzler KW (2013). Cancer genome and Ovarian Cancer Screening Trial. Int J Cancer,
landscapes. Science, 339(6127):1546–58. doi:10.1126/ 121(10):2267–73. doi:10.1002/ijc.22921 PMID:17597108
science.1235122 PMID:23539594
629
Weiss JM, Huang WY, Rinaldi S, Fears TR, Chatterjee Xia Y, Shen S, Verma IM (2014). NF-κB, an active player
N, Hsing AW, et al. (2008). Endogenous sex hormones in human cancers. Cancer Immunol Res, 2(9):823–30.
and the risk of prostate cancer: a prospective study. doi:10.1158/2326-6066.CIR-14-0112 PMID:25187272
Int J Cancer, 122(10):2345–50. doi:10.1002/ijc.23326 Xu XJ, Pories WJ, Dohm LG, Ruderman NB (2013).
PMID:18172860 What distinguishes adipose tissue of severely obese
Wellen KE, Hotamisligil GS (2003). Obesity-induced humans who are insulin sensitive and resistant?
inflammatory changes in adipose tissue. J Clin Invest, Curr Opin Lipidol, 24(1):49–56. doi:10.1097/
112(12):1785–8. doi:10.1172/JCI20514 PMID:14679172 MOL.0b013e32835b465b PMID:23298959
White DL, Kanwal F, El-Serag HB (2012). Association Yatsuya H, Toyoshima H, Tamakoshi K, Tamakoshi A,
between nonalcoholic fatty liver disease and risk for Kondo T, Hayakawa N, et al.; JACC Study Group (2005).
hepatocellular cancer, based on systematic review. Serum levels of insulin-like growth factor I, II, and
Clin Gastroenterol Hepatol, 10(12):1342–1359.e2. binding protein 3, transforming growth factor beta-1,
doi:10.1016/j.cgh.2012.10.001 PMID:23041539 soluble fas ligand and superoxide dismutase activity in
Win AK, Dowty JG, Antill YC, English DR, Baron JA, stomach cancer cases and their controls in the JACC
Young JP, et al. (2011). Body mass index in early adult- Study. J Epidemiol, 15(Suppl 2):S120–5. doi:10.2188/
hood and endometrial cancer risk for mismatch repair jea.15.S120 PMID:16127223
gene mutation carriers. Obstet Gynecol, 117(4):899–905. Ye J, Jia J, Dong S, Zhang C, Yu S, Li L, et al. (2014).
doi:10.1097/AOG.0b013e3182110ea3 PMID:21422863 Circulating adiponectin levels and the risk of breast
Wirén S, Stocks T, Rinaldi S, Hallmans G, Bergh A, cancer: a meta-analysis. Eur J Cancer Prev, 23(3):158–65.
Stenman UH, et al. (2007). Androgens and prostate doi:10.1097/CEJ.0b013e328364f293 PMID:23929213
cancer risk: a prospective study. Prostate, 67(11):1230–7. Yu CP, Ho JY, Huang YT, Cha TL, Sun GH, Yu DS,
doi:10.1002/pros.20588 PMID:17562541 et al. (2013). Estrogen inhibits renal cell carcinoma
Wolpin BM, Bao Y, Qian ZR, Wu C, Kraft P, Ogino S, et al. cell progression through estrogen receptor-β acti-
(2013). Hyperglycemia, insulin resistance, impaired vation. PLoS One, 8(2):e56667. doi:10.1371/journal.
pancreatic β-cell function, and risk of pancreatic pone.0056667 PMID:23460808
cancer. J Natl Cancer Inst, 105(14):1027–35. doi:10.1093/ Yu H, Lee H, Herrmann A, Buettner R, Jove R (2014).
jnci/djt123 PMID:23847240 Revisiting STAT3 signalling in cancer: new and
Wolpin BM, Michaud DS, Giovannucci EL, Schernhammer unexpected biological functions. Nat Rev Cancer,
ES, Stampfer MJ, Manson JE, et al. (2007). Circulating 14(11):736–46. doi:10.1038/nrc3818 PMID:25342631
insulin-like growth factor axis and the risk of pancreatic Zeleniuch-Jacquotte A, Akhmedkhanov A, Kato I, Koenig
cancer in four prospective cohorts. Br J Cancer, 97(1):98– KL, Shore RE, Kim MY, et al. (2001). Postmenopausal
104. doi:10.1038/sj.bjc.6603826 PMID:17533398 endogenous oestrogens and risk of endometrial cancer:
Wong HL, Rabkin CS, Shu XO, Pfeiffer RM, Cai Q, Ji BT, results of a prospective study. Br J Cancer, 84(7):975–81.
et al. (2011). Systemic cytokine levels and subsequent doi:10.1054/bjoc.2001.1704 PMID:11286480
risk of gastric cancer in Chinese women. Cancer Sci, Zhang X, Tworoger SS, Eliassen AH, Hankinson SE (2013).
102(10):1911–5. doi:10.1111/j.1349-7006.2011.02033.x Postmenopausal plasma sex hormone levels and breast
PMID:21740481 cancer risk over 20 years of follow-up. Breast Cancer
Wood LD, Parsons DW, Jones S, Lin J, Sjöblom T, Leary Res Treat, 137(3):883–92. doi:10.1007/s10549-012-
RJ, et al. (2007). The genomic landscapes of human 2391-z PMID:23283524
breast and colorectal cancers. Science, 318(5853):1108– Zheng Q, Dunlap SM, Zhu J, Downs-Kelly E, Rich J,
13. doi:10.1126/science.1145720 PMID:17932254 Hursting SD, et al. (2011). Leptin deficiency suppresses
Woodson K, Tangrea JA, Pollak M, Copeland TD, Taylor MMTV-Wnt-1 mammary tumor growth in obese mice
PR, Virtamo J, et al. (2003). Serum insulin-like growth and abrogates tumor initiating cell survival. Endocr
factor I: tumor marker or etiologic factor? A prospec- Relat Cancer, 18(4):491–503. doi:10.1530/ERC-11-0102
tive study of prostate cancer among Finnish men. PMID:21636700
Cancer Res, 63(14):3991–4. PMID:12873996 Zhou B, Shu B, Yang J, Liu J, Xi T, Xing Y (2014). C-reactive
Wortsman J, Matsuoka LY, Chen TC, Lu Z, Holick MF protein, interleukin-6 and the risk of colorectal cancer:
(2000). Decreased bioavailability of vitamin D in a meta-analysis. Cancer Causes Control, 25(10):1397–
obesity. Am J Clin Nutr, 72(3):690–3. PMID:10966885 405. doi:10.1007/s10552-014-0445-8 PMID:25053407
Würtz AM, Tjønneland A, Christensen J, Dragsted LO, Zhu L, Pollard JW (2007). Estradiol-17β regulates mouse
Aarestrup J, Kyrø C, et al. (2012). Serum estrogen and uterine epithelial cell proliferation through insu-
SHBG levels and breast cancer incidence among users lin-like growth factor 1 signaling. Proc Natl Acad Sci
and never users of hormone replacement therapy. USA, 104(40):15847–51. doi:10.1073/pnas.0705749104
Cancer Causes Control, 23(10):1711–20. doi:10.1007/
s10552-012-0050-7 PMID:22926300
630
Zhu Z, Jiang W, McGinley J, Wolfe P, Thompson HJ (2005).

Effects of dietary energy repletion and IGF-1 infusion
on the inhibition of mammary carcinogenesis by
dietary energy restriction. Mol Carcinog, 42(3):170–6.
doi:10.1002/mc.20071 PMID:15599926
Zhu Z, Jiang W, McGinley JN, Thompson HJ (2009).
Energetics and mammary carcinogenesis: effects of
moderate-intensity running and energy intake on
cellular processes and molecular mechanisms in
rats. J Appl Physiol (1985), 106(3):911–8. doi:10.1152/
japplphysiol.91201.2008 PMID:19095749
Zhu Z, Jiang W, Thompson HJ (1999a). Effect of
energy restriction on the expression of cyclin D1
and p27 during premalignant and malignant stages
of chemically induced mammary carcinogenesis.
Mol Carcinog, 24(4):241–5. doi:10.1002/(SICI)1098-
274 4(19 9 9 0 4)24:4 <241:: A I D -MC1>3.0.CO; 2-P
PMID:10326860
Zhu Z, Jiang W, Thompson HJ (1999b). Effect of
energy restriction on tissue size regulation
during chemically induced mammary carcino-
genesis. Carcinogenesis, 20(9):1721–6. doi:10.1093/
carcin/20.9.1721 PMID:10469616
631
5. SUMMARY OF DATA
5.1 Exposure data estimates have highlighted the double burden of

malnutrition (the coexistence of undernutrition
Obesity is the accumulation of excess body and overnutrition) in some low-income coun-
fat. Body mass index (BMI) is commonly used tries. Perceptions about the magnitude of body
as a proxy measure of body fatness, because it weight and about body weight change may vary
correlates strongly with both absolute body fat according to personal beliefs, contextual factors,
and body fat percentage. BMI is calculated by and cultural norms.
dividing the body weight (in kilograms) by the Body weight appears to have a very modest
square of the height (in metres). In adults, over- genetic component; it is estimated that about
weight is defined as BMI ≥ 25 kg ⁄m2 and obesity 3% of the population variation in BMI can be
as BMI ≥ 30 kg ⁄m2. Obesity can be further classi- explained by known genetic variants, high-
fied by severity into class I (30–34.9 kg ⁄m2), class lighting the important role of modifiable risk
II (35–39.9 kg ⁄m2), and class III (≥ 40 kg ⁄m2). In factors in the development of overweight and
children younger than 5 years, overweight and obesity. Current evidence indicates that excess
obesity are defined as a weight-for-height more energy intake (food and drinks) and, to a lesser
than 2 standard deviations (SD) and more than extent, physical inactivity are the major risk
3 SD, respectively, above the WHO Child Growth factors for excess body fatness and body weight
Standards median. In children and adolescents gain throughout life. Global economic develop-
from age 5 years to younger than 19 years, over- ment has an impact on the built environment
weight and obesity are defined as a BMI-for-age and on sociocultural factors, with major effects
more than 1 SD and more than 2 SD, respectively, on food availability, patterns of food intake,
above the WHO Growth Reference median. and levels of physical activity. The most notable
In 2014, an estimated 640 million adults stimuli for excess energy intake are the availa-
were obese, 6 times the number in 1975. In bility, the frequency of consumption, and the
2013, 110 million children and adolescents aged portion sizes of energy-dense foods and drinks.
2–19 years were obese, twice the number in 1980. Enhanced urbanization, more labour-saving
The estimated prevalence of obesity in 2014 devices, and increased concerns about outdoor
was 10.8% in men, 14.9% in women, and 5.0% safety have led to decreases in physical activity
in children. Obesity in adults and children is and increases in time spent sedentary, in the
prevalent in countries with high, middle, or low household, in transportation, and in the employ-
income, and globally more people are overweight ment (or unemployment) and leisure domains,
or obese than are underweight. Inequalities thereby decreasing overall energy expenditure.
between and within countries play a major role
in the burden of obesity. In addition, recent
633
Studies investigating dietary patterns in of which vary between individuals and with age,
relation to body weight control and obesity have sex, and race/ethnicity. Waist circumference is
found that diets characterized by increased intake widely used as an indirect measure of abdom-
of energy-dense and highly processed foods that inal obesity, because it is strongly correlated with
are high in added sugars, fat, and salt and low in total abdominal fat mass and with abdominal
fibre are positively related to weight gain, whereas visceral fat, which is highly metabolically active
diets that consist largely of nutrient-dense foods, and is more difficult to assess.
such as the traditional Mediterranean diet, are More sophisticated assessment methods
inversely related to weight gain and obesity. Meal (e.g. bioelectrical impedance, dual-energy
patterns and sleep duration may also affect the X-ray absorptiometry, and magnetic resonance
risk of excess body fatness. Other factors that imaging) can provide more accurate estimates
have the potential to influence energy balance, of body composition and body fat distribution,
apparently to a lesser extent, are internal regu- but their use in epidemiological studies is limited
latory control of hunger, satiety, and metabolic because of associated costs and concerns about
homeostasis, the fermentation activity of the radiation exposure, and they are therefore not
microbiome and its impact on the metabolism, typically used in population-based assessments.
production, and storage of fatty acids, and endo- In children and adolescents, BMI (referenced to
crine disruptors. appropriate growth standards and recommended
Throughout the life-course, there are a range cut-offs) is the preferred measure.
of critical time points and transition events that
affect body weight and body weight change.
Perinatal factors, including maternal body weight
5.2 Cancer-preventive effects
and weight gain during pregnancy, birth weight, in humans
infant feeding practices, and early growth trajec- The evidence from studies addressing body
tories, have been consistently shown to affect fatness and cancer risk has rapidly expanded.
body fatness in infancy, childhood, and later life. With continued follow-up of cohorts around the
Body weight changes have also been observed world, there are now data for many cancer sites
during the transition from school to higher from hundreds of prospective studies and case–
education or to employment, the transition from control studies. Pooled analyses and meta-ana-
single status to marriage or cohabitation, the lyses have also been carried out, facilitating the
postpartum period, and changes in employment evaluation of associations with less common
or unemployment status. Body weight gain is cancers. Most studies have measured body
also associated with smoking cessation, a range fatness using BMI. A smaller number have used
of comorbidities, and use of certain medications. other measures, most importantly waist circum-
There are a large number of anthropometric ference, and fewer still have assessed changes in
measurement techniques for body fatness. BMI weight over time.
is the most widely used measure to assess overall
body fatness, because weight and height are easy
5.2.1 Cancer of the colorectum
and inexpensive to measure and can be assessed
accurately (even by self-report), and because For cancers of the colon and rectum, evidence
BMI is strongly correlated with overall body from more than 30 prospective studies and
fatness and enables comparisons across studies. about 10 case–control studies published after
However, BMI does not differentiate between 2000 confirmed a positive dose–response rela-
lean mass and fat mass, the relative proportions tionship between BMI and risk. This association
634
was observed consistently across studies and 5.2.3 Cancer of the stomach
geographical regions. The association was
weaker in women than in men, and was weaker (a) Cancer of the gastric cardia
for cancer of the rectum than for cancer of the For cancer of the gastric cardia, evidence from
colon. For cancer of the colon, there was a statis- 10 prospective studies and several case–control
tically significant increase in risk of about 10% studies indicated a statistically significant posi-
per 5 kg/m2 increase in BMI in women and of tive dose–response relationship between BMI
25% per 5 kg/m2 increase in men. Waist circum- and risk. This association was observed in men
ference was also positively associated with risk and women and across geographical regions.
of cancer of the colon (and was less consistently Compared with normal body weight, the rela-
associated with cancer of the rectum). Results tive risk was about 1.2 for overweight and about
from two studies using Mendelian randomiza- 1.8 for obesity, estimated from a meta-analysis of
tion were consistent with these findings. seven prospective studies.
(b) Non-cardia gastric cancer

5.2.2 Cancer of the oesophagus
Findings from more than 10 prospective
(a) Adenocarcinoma of the oesophagus studies and several case–control studies showed
For adenocarcinoma of the oesophagus, evi- a weak relationship, or no relationship, between
dence from 10 prospective studies and 10 case– BMI and risk of non-cardia gastric cancer.
control studies published after 2000 confirmed
a statistically significant positive dose–response 5.2.4 Cancer of the liver (hepatocellular
relationship between BMI and risk. This associ- carcinoma)
ation was observed in almost all studies, in men
and women, and across geographical regions. There was evidence from more than 20
Compared with BMI < 25 kg/m2, the relative risk prospective studies and several case–control
was about 1.5 for overweight, 2.4 for obesity class studies that BMI is positively associated with risk
I, 2.8 for obesity class II, and 4.8 for obesity class of either hepatocellular carcinoma or cancer of
III, estimated from a pooled analysis of 10 case– the liver overall. This association was reported
control studies and 2 cohort studies. Results in studies from Asia, Europe, and the USA.
from a study using Mendelian randomization Compared with normal body weight, the rela-
were consistent with these findings. tive risk was about 1.5 for overweight and about
1.8 for obesity, estimated from a meta-analysis of
(b) Squamous cell carcinoma of the 26 prospective studies of cohorts of the general
oesophagus population.
Squamous cell carcinoma of the oesophagus
was examined in nine individual prospective 5.2.5 Cancer of the gall bladder
studies, several case–control studies, and one For cancer of the gall bladder, evidence from
meta-analysis published after 2000. In all studies, more than 10 individual prospective studies and
BMI was inversely associated with risk of cancer. a comprehensive meta-analysis of 12 prospective
Residual confounding by tobacco smoking is and 8 case–control studies indicated a statistically
likely to account for the inverse associations. significant positive dose–response relationship
between BMI and risk. Compared with normal
body weight, the relative risk was about 1.2 for
635
overweight and about 1.6 for obesity, estimated 5.2.9 Cancer of the breast in women
from the meta-analysis.
More than 30 prospective studies and about
400 case–control studies published after 2000
5.2.6 Cancers of the biliary tract provided data on the association between BMI
For cancers of the biliary tract, the evidence and risk of cancer of the breast in women. In
was inconsistent. postmenopausal women, very consistent posi-
tive associations were observed with BMI meas-
5.2.7 Cancer of the pancreas ured in adulthood. This association was most
pronounced in women not using hormone
For cancer of the pancreas, evidence from replacement therapy (HRT) and for estrogen
more than 20 prospective studies, more than 10 receptor-positive tumours. This association was
case–control studies, and several large pooled not consistently observed in Hispanic women. A
analyses of cohorts indicated a statistically large meta-analysis in postmenopausal women
significant positive dose–response relationship estimated a statistically significant relative risk
between BMI and risk. This association was of about 1.12 per 5 kg/m2 in women not using
observed in the large majority of studies and was HRT, but no association was found in women
found in both men and women. Compared with using HRT. Waist circumference and adult body
normal body weight, the relative risk was about weight gain, both from age 18 years and from
1.2 for overweight and about 1.5 for obesity, age 50 years, were also positively associated with
estimated from a pooled analysis of 14 cohorts. risk of cancer of the breast in postmenopausal
women.
5.2.8 Cancer of the lung In premenopausal women, consistent inverse
For cancer of the lung, the results of about associations were observed between BMI and
20 prospective studies and about 10 case–control risk; however, positive associations between waist
studies consistently suggested an inverse asso- circumference and body weight gain and risk
ciation between BMI and risk, but studies in have been reported. Results from a study using
non-smokers generally showed no association. Mendelian randomization were not consistent
Because tobacco smoking is strongly related to with a positive association between adult BMI
both cancer of the lung and reduced body weight, and risk of cancer of the breast in postmeno-
residual confounding by tobacco smoking is pausal women.
likely to account for the inverse associations.
Results from two studies using Mendelian rand- 5.2.10 Cancer of the breast in men
omization were inconsistent with these find- For cancer of the breast in men, results from
ings in that they showed a positive association a pooled analysis of 11 case–control studies
between BMI and risk of cancer of the lung; indicated an association between BMI and risk,
however, these results are difficult to interpret whereas pooled risk estimates based on 10 cohort
because of concerns about failure to account for studies did not.
smoking status.
5.2.11 Cancer of the endometrium
For cancer of the endometrium, evidence
from more than 20 prospective studies and
30 case–control studies published after 2000
636
confirmed a statistically significant posi- association between BMI and risk of fatal cancer
tive, exponential dose–response relationship of the prostate. There was no consistent asso-
between BMI and risk. This association was ciation between BMI and incidence of total,
observed in all cohort and case–control studies non-aggressive (non-advanced), or aggressive
and was consistent across geographical regions. (advanced) cancer of the prostate. Results from
The association was particularly pronounced for three studies using Mendelian randomization
type 1 cancer of the endometrium: compared were also inconsistent.
with normal body weight, the relative risk for
type 1 endometrial cancer was about 1.5 for over- 5.2.15 Cancer of the testis
weight, about 2.5 for obesity class I, about 4.5 for
obesity class II, and about 7.1 for obesity class III, One cohort study and more than 10 case–
estimated from the most recent pooled analysis control studies have addressed the relationship
of 10 cohorts and 14 case–control studies. Meta- between BMI and risk of cancer of the testis. The
analyses showed a stronger association between association between BMI and risk of cancer of
BMI and risk of cancer of the endometrium in the testis was inconsistent, and a meta-analysis
never-users of HRT than in ever-users (rela- did not identify sources of heterogeneity.
tive risk per 5 kg/m2, 1.18 in ever-users vs 1.90
in never-users). Results from a study using 5.2.16 Cancer of the kidney (renal cell
Mendelian randomization were consistent with carcinoma)
these findings. For cancer of the kidney (renal cell carci-
noma), evidence from about 20 prospective
5.2.12 Cancer of the cervix studies and 10 case–control studies published
For cancer of the cervix, the evidence was after 2000 confirmed a positive dose–response
inconsistent. relationship between BMI and risk. This asso-
ciation was observed in almost all studies and
was consistent in men and women and across
5.2.13 Cancer of the ovary
geographical regions. Compared with normal
Evidence from more than 15 prospective body weight, there was a statistically significant
studies and more than 30 case–control studies relative risk of about 1.3 for overweight and about
indicated a positive dose–response relationship 1.8 for obesity, estimated from the most recent
between BMI and risk of epithelial cancer of the meta-analysis of 21 cohort studies. Results from
ovary. Based on a pooled analysis of 47 studies, a study using Mendelian randomization were
the relative risk in never-users of HRT was about consistent with an association between BMI and
1.1 for overweight and about 1.2 for obesity, risk of cancer of the kidney.
compared with normal body weight. There was
no association in users of HRT. Results from 5.2.17 Cancer of the urinary bladder
a study using Mendelian randomization were
consistent with these findings. Findings from more than 20 prospective
cohorts and 4 case–control studies indicated
inconsistent relationships between BMI and
5.2.14 Cancer of the prostate
risk of cancer of the urinary bladder. Residual
For cancer of the prostate, evidence from confounding by tobacco smoking could not be
about 50 prospective studies and more than excluded.
40 case–control studies suggested a positive
637
5.2.18 Primary tumours of the brain and overweight and about 1.3 for obesity, estimated
central nervous system from a meta-analysis of 10 cohort studies.
For Hodgkin lymphoma, cohort studies
For meningioma, five prospective studies and generally found non-significant positive asso-
two case–control studies showed a consistent ciations with obesity compared with normal
positive association between BMI and risk. BMI; the relative risk was about 1.4, estimated
For glioma, five cohort studies and two case– from a meta-analysis of five prospective studies.
control studies, with only moderate sample sizes, Findings from case–control studies were largely
reported inconsistent associations between BMI null.
and risk. For non-Hodgkin lymphoma and B-cell
lymphoma as a group, findings for an association
5.2.19 Cancer of the thyroid between BMI and risk from individual studies
For cancer of the thyroid, evidence from more and meta-analyses were inconsistent. The incon-
than 10 prospective studies and 10 case–control sistency within the broader category of B-cell
studies indicated a positive dose–response rela- lymphoma may be due to heterogeneity among
tionship between BMI and risk. The relative risk subtypes. There were too few studies on T-cell
per 5 kg/m2 was 1.17 in men and 1.04 in women, lymphoma to enable conclusions to be drawn.
both statistically significant, estimated from a (b) Other haematopoietic malignancies
pooled analysis of 22 prospective studies.
For total leukaemia and myeloid leukaemia,
5.2.20 Tumours of the haematopoietic findings for an association between BMI and risk
from individual studies were inconsistent.
system
(a) Lymphoid tumours 5.2.21 Cancers of the head and neck
For multiple myeloma, there was substantial Epidemiological studies on this heteroge-
evidence from at least 20 prospective studies and neous group of cancers have examined associa-
several case–control studies and meta-analyses tions between BMI and risk of cancers of the oral
or pooled analyses showing positive associations cavity, pharynx (i.e. nasopharynx, oropharynx,
between BMI at baseline and risk. The association and hypopharynx), larynx, and salivary glands.
appeared to be dose-related and was observed for Evidence from five prospective studies, two
overweight and obesity. From a pooled analysis of case–control studies, and four meta-analyses or
20 cohorts, the relative risk of multiple myeloma pooled analyses that examined BMI in relation to
mortality was 1.15–1.24 for overweight, about cancers of the head and neck overall was incon-
1.23 for obesity class I, and about 1.52 for obesity sistent. Several studies examined associations
class II or higher, compared with normal body between BMI and risk of cancer of the oral cavity,
weight. pharynx, or larynx specifically, and the findings
For diffuse large B-cell lymphoma, findings from these studies were also inconsistent. Some
from nine individual prospective studies and of the inconsistencies for these cancers might be
two case–control studies, as well as meta-ana- explained by residual confounding by tobacco
lyses or pooled analyses, suggested a positive use and/or alcohol consumption.
association between BMI and risk, but the results
were not fully consistent. Compared with normal
body weight, the relative risk was about 1.1 for
638
5.2.22 Malignant melanoma 5.2.25 Sustained weight loss and cancer risk
For cutaneous malignant melanoma, eight The few observational studies that have
prospective studies showed no clear relationship evaluated body weight loss, and in particular
between BMI and risk. A weak positive relation- sustained body weight loss, in relation to subse-
ship was suggested by the results of nine case– quent cancer risk are limited to observational
control studies and one pooled analysis of eight studies on weight loss in relation to incidence
case–control studies. of cancer of the breast and on the impact of
intentional weight loss after bariatric surgery on
5.2.23 Excess body fatness in early life and cancer risk in morbidly obese patients. Findings
subsequent cancer risk from cohort studies of weight loss and cancer of
the breast were inconsistent, in part reflecting
Studies that have evaluated relationships the problem of distinguishing between inten-
between excess body fatness in childhood, tional and unintentional weight loss.
adolescence, and early adulthood (age ≤ 25 years) In studies of large series of morbidly obese
and subsequent cancer risk include studies that patients who underwent bariatric surgery and
directly measured weight and height in child- with sufficient follow-up, sustained substantial
hood, studies that determined body shape in early body weight loss is associated with reduced risk
adulthood by recall, and studies that determined of subsequent cancer, especially for cancer of the
trajectories of body shape from childhood to late endometrium.
adulthood. Collectively, these studies indicated
positive associations with several cancer types
known to be associated with excess body fatness 5.3 Cancer-preventive effects in
in middle and later adulthood, except for cancer experimental animals
of the breast in postmenopausal women (see
Section 5.2.9); there was some evidence for an 5.3.1 Excess body weight
inverse association between excess body fatness Numerous models in experimental animals
in early life and subsequent risk of cancer of the have been developed to study the association
breast in postmenopausal women. between obesity and cancer of the mammary
gland, colon, liver, prostate, skin, pancreas,
5.2.24 Excess body fatness in cancer survivors endometrium of the uterus, and haematopoietic
system. Most such animal models are genetically
A large number of studies have evaluated the
manipulated (transgenic) animals: animals are
relationship between BMI at the time of diag-
either genetically modified to induce carcino-
nosis of cancer and cancer-related mortality. The
genicity and fed a modified diet to induce obesity,
data were most consistent for cancer of the breast,
or genetically modified to induce obesity and
for which high BMI has been associated with
administered chemicals to induce cancer.
an increased risk of cancer-related mortality in
For cancer of the mammary gland, the asso-
individual reports and meta-analyses. Data were
ciation between obesity and cancer was tested in
fewer and/or less consistent for other malignan-
five studies in genetically obese mice, five studies
cies. The effect of intentional body weight loss
of diet-induced obesity in mice, two studies of
after cancer diagnosis on cancer mortality has
chemically induced obesity in mice, four studies
been tested in one intervention trial.
in genetically obese rats, and one study in obesi-
ty-prone rats. In all studies except one, obesity
639
increased the incidence of hyperplastic alveolar studies of diet-induced obesity in mice and one
nodules and/or of tumours of the mammary study in two models of genetically obese mice. In
gland, shortened tumour latency, and/or the three studies of genetically induced tumours
increased tumour volume and growth rate. of the pancreas in the diet-induced obesity model,
For cancer of the colon, the association obesity increased the incidence of pancreatic
between obesity and cancer was tested in six intraepithelial neoplasia and of pancreatic ductal
studies in genetically obese mice, including one adenocarcinoma. In the other two studies (one in
study using a transgenic model of carcinogen- transgenic mice and one of diet-induced obesity),
icity, one study of diet-induced obesity in mice, subcutaneous injection of syngeneic pancreatic
and three studies in transgenic obese rats. In all tumour cells led to the development of signifi-
studies, obesity increased the incidence of pre-ne- cantly larger tumours and higher metastatic
oplastic aberrant crypt foci and/or of tumours of rates in obese mice than in lean mice.
the colon (primarily adenocarcinoma), and/or For cancer of the endometrium of the uterus,
increased tumour size and multiplicity. the association between obesity and cancer was
For cancer of the liver, the association tested in one study of diet-induced obesity in
between obesity and cancer was tested in five mice. In that study, obesity increased the inci-
studies in genetically obese mice, four studies dence of pre-neoplastic glandular epithelial
of diet-induced obesity in mice, and one study hyperplasia and adenocarcinoma.
in diabetic obese rats. In all studies except one, For cancers of the haematopoietic system,
obesity increased the incidence of hepatocellular the association between obesity and cancer was
tumours (adenoma and carcinoma), shortened tested in two studies of diet-induced obesity in
tumour latency, and/or increased tumour volume mice. In both studies, obesity shortened latency
and growth rate. for the development of acute lymphoblastic
For cancer of the prostate, the association leukaemia.
between obesity and cancer was tested in five Overall, the data showed that obesity in
studies of diet-induced obesity in mice, including rodents promotes tumorigenesis and increases
three studies using a transgenic model of carcin- the age-specific incidence of cancers of the
ogenicity, two studies in genetically obese mice, mammary gland, colon, liver, pancreas, prostate
and one study of chemically induced obesity (advanced stage cancer), and skin.
in mice. In most studies, obesity enhanced the
development of pre-neoplastic prostatic intraepi- 5.3.2 Dietary/calorie restriction
thelial neoplasia and of adenocarcinoma, leading
to more advanced disease, and/or increased (a) Cancer of the mammary gland
tumour volume. More than 40 studies in several different
For cancer of the skin, the association mouse and rat models have evaluated the effect
between obesity and cancer was tested in four of dietary restriction on the development or
studies in genetically obese mice and two studies progression of tumours of the mammary gland.
of diet-induced obesity in mice. In all studies, Overall, most studies showed that dietary
obesity shortened latency, increased multiplicity, restriction decreased the incidence of mammary
and/or accelerated the progression of subcutane- tumours, extended latency, and/or decreased
ously injected melanoma cells or of tumours of tumour burden.
the skin induced by ultraviolet light. Six recent studies in various transgenic
For cancer of the pancreas, the association mouse models indicated that the pattern of
between obesity and cancer was tested in four restriction is also important in the protective
640
effect of dietary restriction. Periods of inter- tumours. In all three studies using allografts,
mittent restriction had a stronger effect in the dietary restriction significantly reduced tumour
prevention of mammary tumours than did the growth. In three of four studies using a model of
same overall degree of restriction implemented chemically induced tumours, dietary restriction
in a prolonged fashion. One study using a model reduced the incidence of adenoma and carci-
of chemically induced mammary tumours in noma of the colon. In one study in the genetically
rats showed similar results. obese Zucker rat, dietary restriction did not have
In general, dietary restriction interventions an impact on body weight, and had no protective
were implemented in young animals (shortly effect on the development of chemically induced
after weaning, or up to age 9–10 weeks), and aberrant crypt foci. Similarly, no effect was
then maintained throughout the course of the observed in one study using a transgenic mouse
study. This approach usually led to a lower rate of model.
body weight gain than in animals fed ad libitum.
Only two studies have addressed the issue of (c) Cancer of the liver
body weight loss induced by dietary restriction In three lifespan studies in different strains
in obese animals and its impact on the develop- of male and female mice, 40% dietary restric-
ment of tumours of the mammary gland. In both tion reduced the incidence of spontaneous liver
studies, body weight loss reduced the develop- tumours, mostly hepatocellular adenoma or
ment or progression of tumours. carcinoma. The reduction did not always reach
Recently, several studies in mice and rats statistical significance in all analyses (adenomas,
have used chemical mimetics of calorie restric- carcinomas, or adenomas and carcinomas
tion (metformin, buformin, phenformin, and combined), because of the small numbers of
2-deoxyglucose) to assess prevention of tumours animals and the low incidence of tumours in
of the mammary gland. Protective effects were animals fed ad libitum. In two studies of chem-
observed in two of three studies in the HER2/neu ically induced liver tumours in mice, 30% or
mouse model with metformin, as well as in three 40% dietary restriction significantly reduced
of five studies in the rapidly emerging tumour the incidence of hepatocellular tumours, mostly
model in rats (one study each using 2-deoxyglu- carcinomas.
cose, buformin, or phenformin); metformin had
no effect in the remaining two studies. (d) Cancer of the pancreas
In addition, several studies were conducted in In all three studies using transgenic mouse
strains that have different responses to high-fat models to induce tumours of the pancreatic
diets with regard to the rate of body weight gain, duct, 25–30% dietary restriction decreased the
thus providing the opportunity to evaluate the incidence and severity of pre-neoplastic pancre-
effect of body weight independent of diet. In atic lesions or carcinoma and/or increased
these studies, lower body weight was accompa- survival. In one study using a model of chem-
nied by longer tumour latency. ically induced carcinogenesis in rats, in which
animals were “meal-fed” (i.e. fed ad libitum for
(b) Cancer of the colon 5–6 hours per day, resulting in 10–15% dietary
Several models in rats and mice using either restriction), similar results were observed. In one
chemical carcinogens or allografts to induce of three lifespan study in rats, dietary restric-
tumours of the colon have been developed. Nine tion reduced the incidence of spontaneously
studies have assessed the effect of dietary restric- occurring islet cell tumours. In one study using
tion on the development or progression of such a model of chemically induced carcinogenesis
641
in Syrian golden hamsters, 20% or 40% dietary studies examined the impact of dietary restric-
restriction had no effect. In one study in mice tion on development of cancer of the prostate: five
injected with pancreatic tumour cells, dietary in transgenic animals, two in models of hormo-
restriction inhibited tumour growth. nally induced tumours, and one of spontaneous
tumours. In three studies in transgenic animals,
(e) Cancer of the skin dietary restriction reduced the incidence of
Lifespan studies in mice and rats that have adenocarcinoma or high-grade lesions. In one
assessed the effect of dietary restriction on study using a model of hormonally induced
tumours of the skin gave inconclusive results cancer, dietary restriction reduced the incidence
because of the low incidence of spontaneously of adenocarcinoma. The one study of sponta-
occurring tumours. Nine studies using carcino- neous tumours showed a reduction in incidence
gen-induced models have assessed the effect of a of adenocarcinoma with dietary restriction. All
range of levels (15–50%) of dietary restriction at studies initiated dietary restriction in young
the initiation, promotion, or progression phase. animals (aged 3–9 weeks) and reported attenu-
In all studies, all levels of dietary restriction ated weight gain compared with control animals.
inhibited the development of skin papilloma,
the progression of papilloma to carcinoma, or (h) Cancers of the haematopoietic system
the multiplicity of these tumours when dietary Malignant lymphoma and histiocytic sar-
restriction was imposed at the promotion phase coma commonly occur in old mice. In three of
and/or thereafter. five lifespan studies in male or female mice,
In one study using a B16 melanoma cell dietary restriction reduced the incidence of
line injected subcutaneously into mice, dietary lymphoma and/or histiocytic sarcoma. In
restriction inhibited tumour growth. one study using knockout p53−/− mice (prone
to cancer in many organs), dietary restriction
(f) Cancer of the pituitary gland resulted in a moderate reduction in the incidence
Tumours of the anterior pituitary gland are of lymphoma, and a significant delay of death due
prevalent in old female mice and in old male and to lymphoma. In one lifespan study in B10C3F1
female rats. In all five lifespan studies in mice or mice, dietary restriction also increased the mean
rats, 35% and 40% dietary restriction reduced the lifespan of mice with lymphoma.
incidence of spontaneous tumours of the pitui- Mononuclear (large granular) cell leukaemia
tary gland. is prevalent in old F344/N rats, and the incidence
In two studies using the estrogen-induced and severity of disease increase with increased
prolactinoma models in female and male F344 longevity. One 2-year study in F344/N rats
rats, 40% dietary restriction inhibited the increase showed a significant reduction in the incidence
in the weight of the pituitary gland, which is used of large granular cell leukaemia with 7–20%
as an index of tumour growth in this model. dietary restriction. In another lifespan study,
Dietary restriction had no effect in three studies 40% dietary restriction had no significant effect.
using this model in either male Holtzman rats or To address the issue of increased lifetime inci-
female ACI ovariectomized rats. dence of leukaemia, one study assessed the onset
rate of leukaemia, and reported a significant 20%
(g) Cancer of the prostate reduction with 40% dietary restriction, although
The transgenic animal models used to study the lifetime incidence did not differ from that in
cancer of the prostate are characterized by the the group fed ad libitum.
development of highly aggressive disease. Eight
642
5.4 Mechanistic and other relevant other obesity-related cancers, such as those of the
data kidney, pancreas, oesophagus, or liver.
Overall, there is strong evidence that the sex
A short summary of the data is presented at hormone-mediated pathway is a major mecha-
the end of each chapter of Section 4. nism underlying the link between obesity and
The Working Group assessed which cellular certain cancers.
and molecular mechanisms known to be dysreg-
ulated during the carcinogenesis process are 5.4.2 Inflammation
causally linked with obesity, and assessed the
relevance of each mechanism for cancer overall, Obesity leads to subclinical inflammation.
as well as – when sufficient data were available Several clinical and experimental studies indi-
– for individual organ sites. The findings and cate that intentional weight loss by behavioural
levels of evidence are summarized below, by the interventions, bariatric surgery, or pharmaco-
strength of the evidence of the mechanism. logical approaches can reverse obesity-associated
The currently available data in humans inflammatory changes. The most established
and experimental models are consistent with marker of inflammation in these studies, and
the effects of intentional weight loss on cancer the most consistently responsive to intentional
risk being mediated, at least in part, by regu- weight loss, is C-reactive protein, but it is unclear
lation of the balance between cell proliferation whether C-reactive protein is a true biolog-
and apoptosis in carcinogenic progression. The ical mediator of inflammation and cancer or a
cellular machinery that accounts for such regu- marker of other aspects of inflammation. Other
lation includes proteins involved in the G1/S markers related to inflammation – including
cell cycle transition and apoptotic induction, interleukin-6, tumour necrosis factor alpha
whether via the intrinsic (mitochondrial) or (TNF-α), prostaglandins, cyclooxygenase-2
extrinsic pathways. (COX-2), leptin, and adiponectin – either have
inconsistent associations or have not yet been
adequately studied. The obesity-associated
5.4.1 Sex hormone metabolism pro-inflammatory state appears to be triggered
Estrogen levels correlate with amount of by adipose tissue dysregulation resulting from
body fat in postmenopausal women. Estrogens excess triglyceride accumulation in adipocytes,
play a significant role in cancers of the breast leading to the recruitment and reprogramming
and endometrium, and there are consistent data of macrophages and other immune cells that
in humans to demonstrate that women with interact with the lipid-engorged adipocytes to
higher levels of estrogen have an increased risk of increase secretion of multiple cytokines and
these malignancies. For other tumours, the role other inflammatory mediators. The chronic
of sex hormones is less clear. For cancer of the reinforcement of this pro-inflammatory state
colorectum, estrogen may be anti-tumorigenic leads to remodelling of adipose tissue, including
and therefore would not represent a mechanism infiltration of lipids into the liver, pancreas, and
linking adiposity with this cancer. Data linking other tissues to create a pro-tumorigenic envi-
sex hormones with cancers of the prostate and ronment. In addition, several emerging contrib-
ovary are inconsistent and may be dependent on utors to the obesity-associated pro-inflammatory
tumour subtype. There was little evidence that state, including activation of the COX-2/prosta-
sex hormones play a role in the development of glandin pathway as a result of increased cytokine
levels, and the obesity-induced increase in
643
inflammation-related molecules from the micro- effects of obesity on cancer. Currently, there is
biome, also probably play an important role. weak evidence.
The findings support a role for the inflam-
matory process in the development of cancers of 5.4.5 Oxidative stress
the breast and colorectum, and to a lesser extent
of cancer of the ovary. Data for other sites are Oxidative stress can affect DNA integrity and
sparse. Overall, there is strong evidence that has been linked to obesity, metabolic syndrome,
inflammation is a major mechanism underlying and cancer. However, evidence of the involvement
the link between obesity and certain cancers. of oxidative stress in obesity-induced cancer is
limited by methodological issues. Currently,
5.4.3 Insulin and insulin-like growth factor there is weak evidence.
Insulin and insulin-like growth factor 1 5.4.6 DNA repair

(IGF-1) are growth factors that activate the
mammalian target of rapamycin (mTOR)/ The role of DNA repair function in cancer risk
phosphoinositide 3-kinase (PI3K) and mito- is well established for cancers of the colorectum,
gen-activated protein kinase (MAPK) pathways, breast, endometrium, and skin. Several studies
which have mitogenic and anti-apoptotic effects. point towards a link between increased BMI and
Hyperinsulinaemia and insulin resistance are DNA mismatch repair deficiencies. In spite of
common in obese individuals and can raise this, a causal link with obesity and weight control
levels of bioavailable IGF-1 through suppression is lacking, because of methodological challenges.
of IGF-binding proteins. Currently, there is weak evidence.
From epidemiological studies, there is strong
evidence for a role for insulin in the development 5.4.7 Telomeres
of cancer of the endometrium. There is moderate Telomere maintenance is directly linked to
evidence linking insulin to cancer of the breast, immortalization. Likewise, inherited disrup-
whereas recent pooled analyses and meta-an- tions in telomere maintenance have emerged as
alytical approaches indicate a role for IGF-1 in predictors of cancer predisposition at numerous
the development of cancer of the breast. There cancer sites. Evidence from several studies indi-
is moderate to strong evidence that both insulin cates that obesity is inversely associated with
and IGF-1 play a role in cancer of the colorectum. telomere length. Overall, telomere shortening
For prostate cancer, there is moderate evidence may be a relevant emerging mechanism linking
that higher levels of IGF-1 increase the risk of this obesity to risk of cancer. Currently, there is weak
malignancy, whereas the evidence for insulin is evidence.
heterogeneous. For other tumour sites, the data
are much more limited and inconsistent. Overall,
there is moderate evidence that insulin and IGF-1
5.4.8 Other mechanisms
play a role in obesity-induced cancer. For several mechanisms or mechanistically
linked conditions that are potentially related
5.4.4 Epigenetic alterations to obesity and cancer, i.e. vitamin D status, the
gut microbiome, gut hormones, non-alcoholic
Structural modifications of DNA, including fatty liver disease, immune function, and cancer
epigenetic alterations, play an important role in stem cell enrichment, currently, there is weak
tumorigenesis. However, few studies have inves- evidence.
tigated the role of epigenetics in mediating the
644
6. EVALUATION
6.1 Cancer-preventive effects in 6.3 Mechanistic and other relevant

humans data
There is sufficient evidence in humans for There is strong evidence that sex hormone
a cancer-preventive effect of absence of excess metabolism and inflammation are major mech-
body fatness. Absence of excess body fatness anisms underlying the link between excess body
prevents cancers of the colon and rectum, fatness and certain cancers, whereas there is
oesophagus (adenocarcinoma), stomach (gastric moderate evidence for the role of insulin and
cardia), liver (hepatocellular carcinoma), gall insulin-like growth factor. The effects were not
bladder, pancreas, breast in postmenopausal uniform across the organ sites considered.
women, endometrium, ovary, kidney (renal cell There was generally convincing evidence
carcinoma), and thyroid, as well as meningioma that a reduction in excess body fatness through
and multiple myeloma. In addition, inverse asso- intentional weight loss positively affects these
ciations have been observed between absence of biomarkers and mechanisms.
excess body fatness and fatal prostate cancer,
diffuse large B-cell lymphoma, and cancer of the
breast in men.
6.4 Overall evaluation
Absence of excess body fatness prevents
6.2 Cancer-preventive effects in cancer in humans (Group A). Absence of excess
body fatness prevents cancers of the colon and
experimental animals rectum, oesophagus (adenocarcinoma), stomach
There is sufficient evidence in experimental (gastric cardia), liver (hepatocellular carcinoma),
animals for a cancer-preventive effect of limit- gall bladder, pancreas, breast in postmenopausal
ation of body weight gain by dietary restric- women, endometrium, ovary, kidney (renal cell
tion. Limitation of body weight gain by dietary carcinoma), and thyroid, as well as meningioma
restriction prevents cancer of the mammary and multiple myeloma.
gland, colon, liver, pancreas, skin, and pituitary
gland. In addition, inverse associations have
been observed for cancer of the prostate, and for
lymphoma and leukaemia.
645
A Working Group of 21 independent experts from 8 countries, convened by the
International Agency for Research on Cancer (IARC) in April 2016, reviewed the
scientific evidence and assessed the cancer-preventive effects of the absence of
excess body fatness.
The mean body mass index (BMI) in the adult population has increased dramatically
worldwide over the past 40 years, and IARC recently estimated that close to 4% of all
new cancer cases in adults were attributable to a high BMI; the number of cases is
highest in high-income countries and is expected to rise in low- and middle-income
countries.
This publication provides an important update of the 2002 IARC Handbook on Weight
Control and Physical Activity, with evidence-based evaluation of the association
between excess body fatness and cancer at more than 20 sites. In addition, the Working
Group reviewed the evidence on childhood obesity and cancer in later life, the impact
of excess body fatness in cancer patients on cancer survival and recurrence, and the
few intervention studies of weight control on cancer outcome.
© Tony Alter CC-BY-2.0

Absence of Excess Body Fatness

Uploaded by

Copyright:

Available Formats

Absence of Excess Body Fatness

Uploaded by

Document Information

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

Absence of Excess Body Fatness

Uploaded by

Copyright:

Available Formats

IARC HANDBOOKS

This publication represents the views and

LYON, FRANCE - 2018

How to cite this publication:

IARC Library Cataloguing in Publication Data

ISBN 978-92-832-3020-5 (NLM Classification: W1)

Cover image: Overweight people sitting on a bench. © Tony Alter CC-BY-2.0

NOTE TO THE READER. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.3 Risk factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

2. CANCER-PREVENTIVE EFFECTS IN HUMANS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

2.2.11 Cancer of the endometrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291

3. CANCER-PREVENTIVE EFFECTS IN EXPERIMENTAL ANIMALS. . . . . . . . . . . . . . . . . . . . . . 487

3.2.1 Obesity models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489

4. MECHANISTIC AND OTHER RELEVANT DATA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553

5. SUMMARY OF DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 633

5.2.9 Cancer of the breast in women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 636

Annie S. Anderson (Subgroup Chair, João Breda

Margot P. Cleary Ronald A. Herbert (Subgroup Chair, Cancer-

Susan M. Gapstur Department of Cancer Epidemiology

Marc Gunter University Hospital Regensburg

Andrew G. Renehan 2 Cornelia Ulrich

Henry J. Thompson (Subgroup Chair,

IARC/WHO Secretariat Administrative Assistance

Melina Arnold, Section of Cancer Marieke Dusenberg

A. GENERAL PRINCIPLES AND which are prepared by a Working Group of inter-

5. Meeting participants They may also contribute text on non-influential

1. Characteristics and occurrence of the descriptive epidemiological studies, case-series,

(c) Quantitative aspects study methodology. The results of studies judged

to characterize variation in the biotransforma- relationships, dose–response, and other quanti-

(a) Exposure data 6. Evaluation and rationale

Evidence suggesting lack of cancer-preventive The overall evaluation is described according

(iv) The intervention is unclassifiable as to References

This sixteenth Volume of the IARC Hand­ Previous evaluations

Table 1 Evaluations of IARC Handbooks Volume 6 (2002)

Intervention Humans Experimental animals Overall evaluation

Worldwide, it has been estimated that 481 000 Content of this Handbook

Weight loss References

AKT protein kinase B

GIANT Genetic Investigation of Anthropometric Traits

RCT randomized controlled trial

1.1 Background and definitions 1.1.2 Cultural definitions of obesity

Table 1.1 Overweight and obesity cut-off values

Population Measure Overweight Obese Obese I Obese II Obese III

IARC HANDBOOKS OF CANCER PREVENTION – 16

Fig. 1.2 Prevalence of obese population by age and sex in selected countries

China Colombia Ethiopia

France India Iran

Japan Malawi New Zealand

Russia United States Global

sex ● female male

Cambodia Colombia Iran Malawi South Korea USA

gain and obesity. Reviews and meta-analyses of (d) Fast food

Table 1.2 Endocrine-disrupting chemicals with obesogenic properties

Chemical Source Potential mechanism

and metabolic impairment (Heindel et al., 2015). (c) Tributyltin

and height obtained by telephone interviews have 1.4.2 Indexes of adiposity

(ii) Ethnic differences standardized metric than BMI. Correlation of

1.4.3 Other measures of adiposity or body

of the body and are detected by transducers. This

Table 1.5 Prediction of body compartments by anthropometric indices in multiple linear

This sixteenth Volume of the IARC Hand Previous evaluations