doi: 10.1111/ppe.

12447 1
Commentary

Optimal Birth Spacing: What Can We Measure and What Do We

Want to Know?
Katherine A. Ahrens,a Jennifer A. Hutcheonb
a
Office of Population Affairs, Office of the Assistant Secretary for Health, US Department of Health and Human Services, Rockville, MD
b
Department of Obstetrics and Gynaecology, University of British Columbia, Vancouver, Canada

The optimal spacing between births has long been a The ready-availability of interpregnancy interval on
topic of interest to epidemiologists and demogra- US birth certificate records has also influenced the
phers. In recent decades, numerous studies on the types of health outcomes studied. A large body of the
association between birth spacing and adverse research to date has focused on only a few neonatal
perinatal outcomes have been published from both health outcomes: preterm birth, low birthweight, and
high- and low-resource countries.1,2 Despite this small-for-gestational-age birth.1 There are large gaps
extensive body of work, there remains a profound dis- in our knowledge of how birth spacing affects many
connect between what these studies typically examine other important outcomes, such as maternal post-
and the questions that we would like to answer. We partum weight loss, and the risk of obesity and associ-
study the observed optimal association between birth ated conditions like gestational diabetes at the time of
spacing and infant health outcomes, but what we the subsequent pregnancy; infant and child develop-
want to know is if modifying when parous women ment; and non-health outcomes such as maternal
start trying to conceive again leads to better outcomes employment and lifetime earnings.
for the mother and child. The goal of future birth The dissonance between what we study and
spacing research should be to bring these 2 lines of what we would like to know becomes most appar-
research into closer alignment. ent when associations between interpregnancy
One reason for this disconnect is data availability. interval and adverse birth outcomes are interpreted
The timing of last livebirth or last pregnancy outcome as causal effects. Many authors suggest that their
is part of routinely collected hospital or birth records findings should be used by health care providers
in many countries.3 In the United States, for example, to counsel post-partum women on the optimal
this information was added to the US national birth spacing before the next pregnancy or to calculate
certificate in 1968.4 However, the interpregnancy the number of adverse health events that could be
interval itself (the time between delivery and the prevented by reducing the prevalence of short
beginning of a subsequent pregnancy) is not an interpregnancy intervals in a population.5,6 These
actionable intervention. Rather, it is the endpoint of types of leaps may be inappropriate because of the
numerous physiologic changes and behaviours after complexity in using observational data to disentan-
giving birth: women’s return to fecundability and gle the causal effect of interpregnancy interval
ovulation, pregnancy intentions, contraception use, from other closely intertwined determinants of
and sexual activity. Information on these factors is not adverse birth outcomes, particularly, low socio-
routinely collected. As a result, there is a paucity of economic status, and unintended pregnancy. Disen-
research examining how changes in factors that are tangling these factors is particularly challenging
more directly modifiable, and thus more relevant to when using administrative or vital records, which
counselling and informed decision making, affect often lack detailed information on socio-economic
birth outcomes. and life style factors. The extent to which unmea-
sured or uncontrolled confounding contributes to
Correspondence: the observed associations is not known, but adjust-
Katherine Ahrens, Office of Population Affairs, Office of the ment for maternal demographics nearly always
Assistant Secretary for Health, U.S. Department of Health and attenuates estimates, supporting an important role
Human Services, Rockville, MD, USA.
for these other factors.
E-mail: [email protected]

© 2018 John Wiley & Sons Ltd

Paediatric and Perinatal Epidemiology, 2018, , –
2 K. A. Ahrens and J. A. Hutcheon

The manuscripts by Zhang and colleagues7 and study is the findings on long interpregnancy interval
Conzuelo Rodriguez and Naimi8 in this issue of in women whose second child was born in 2016. The
Paediatric and Perinatal Epidemiology have furthered change from a 1-child to 2-child policy provides a
the conversation on this topic. Zhang and colleagues unique natural experiment on long interpregnancy
investigated the association between interpregnancy intervals, as the long interpregnancy intervals associ-
interval following livebirth on the risks of preterm ated with 2016 births may be less likely to be corre-
birth and small- and large-for-gestational-age birth lated with socio-economic and life style factors than
among a cohort of women in Guangzhou, China, similar intervals in other countries. The finding that
2000–2015. The study is timely, as guidance on birth interpregnancy interval ≥60 months is still associated
spacing for Chinese women is needed given a new with higher risk of preterm birth (but not small- or
universal 2-child policy issued in China in January large for gestational age birth) is therefore a novel
2016. The methodological study by Conzuelo contribution to the field—although it requires confir-
Rodriguez and Naimi examined changes in the mag- mation from analyses restricted to 2016 births con-
nitude of associations between short interpregnancy ceived post policy change. Further exploiting this
interval and preterm birth (as well as other factors) policy change to understand the consequences of long
based on what types of pregnancy outcomes (live- interpregnancy interval on other pregnancy outcomes,
births, miscarriages, stillbirths, induced abortions) such as preeclampsia or labour abnormalities, pro-
defined the interpregnancy interval. They argue that vides a valuable opportunity for reducing the discon-
previously reported associations between interpreg- nect between what is usually studied, and what we
nancy interval and birth outcomes may be biased want to know.
because interpregnancy interval has usually been The work of Conzeulo Rodriguez and Naimi found
calculated as the time between the delivery of one that the link between short interpregnancy interval
livebirth and the conception of the pregnancy (<18 months) and preterm birth was reasonably con-
leading to the next live birth, ignoring intervening sistent across different definitions of interpregnancy
pregnancies that did not result in a livebirth. These interval. These findings provide reassurance for pol-
authors use data from the National Survey of Family icy-makers seeking to interpret previous studies
Growth, a nationally representative sample of repro- based only on interpregnancy intervals between suc-
ductive-aged women in the United States that cessive livebirths. Nevertheless, this work could be
includes complete pregnancy history. Although refined to better answer the questions we most want
these 2 manuscripts examine interpregnancy interval to have answered. Combining livebirths with still-
in very different contexts and from different per- births, miscarriages, and induced abortions for
spectives, both help align what we can measure with women’s previous pregnancies is less useful because
what we would like to know. these are distinct clinical groups for whom separate
The study by Zhang and colleagues reports a spacing recommendations are likely needed. The evi-
J-shaped relationship for interpregnancy interval and dence for 18 months of recovery time needed for
preterm birth similar to that observed in the United women following miscarriages and induced abor-
States and other settings. They conclude that recom- tions, which are most likely to occur in the first trime-
mendations to wait at least 18 months, but not longer ster of pregnancy, is not compelling. Indeed, the only
than 60 months may be appropriate. However, as with recommendation for spacing following miscarriage or
administrative data studies from other settings, the induced abortion is the World Health Organization
extent to which these findings reflect causal associations recommendation to wait 6 months before attempting
—and thus the extent to which they should be used to the next pregnancy.3 This may itself be too long for
develop recommendations in China—is unclear. Their the United States given a recent cohort study reported
use of e-values to estimate the amount of unmeasured no increased risk of miscarriage/stillbirth, preterm
or uncontrolled confounding that would be needed to birth, gestational diabetes, or preeclampsia for cou-
explain the observed associations is useful in explor- ples who began trying to conceive within 3 months of
ing these concerns, but with results including mini- a pregnancy loss (although the study was not
mum unmeasured or uncontrolled confounding RRs designed to detect differences in these outcomes).9
<1.50 for some interpregnancy intervals, questions Providing the same guidance on birth spacing to
remain. Perhaps the most valuable contribution of the women whose first birth resulted in a miscarriage or

© 2018 John Wiley & Sons Ltd

Paediatric and Perinatal Epidemiology, 2018, , –
 Commentary 3

induced abortion as that given to women with a live- Jennifer Hutcheon is a perinatal epidemiologist and
birth could potentially cause psychological harm. An Associate Professor in the Department of Obstetrics
implicit risk in this scenario is the advancing maternal and Gynaecology at the University of British
age at next pregnancy attempt, which in itself confers Columbia in Vancouver, Canada. Her research uses
increased risk of adverse outcomes. It is even possible population databases and registries to improve our
that risks of adverse outcomes in subsequent preg- understanding of maternal and infant health. She
nancy may differ in women with multiple previous holds a Canadian Institutes of Health Research New
surgical (but not medical)-induced abortions.10 Thus, Investigator award and a Career Scholar award from
separating these groups would best inform guideline the Michael Smith Foundation for Health Research.
creation.
Even before the availability of modern contracep-
References
tives, couples controlled their fertility by increasing
spacing between births, presumably reflecting eco- 1 Conde-Agudelo A, Rosas-Bermudez A, Kafury-Goeta AC.
Birth spacing and risk of adverse perinatal outcomes: a
nomic concerns about increasing family size.11 Quali-
meta-analysis. Journal of the American Medical Association
tative work has shown that women’s birth spacing 2006; 295:1809-1823.
attitudes are complex, taking into account social and 2 Wendt A, Gibbs CM, Peters S, Hogue CJ. Impact of
financial resources, family composition, and partner’s increasing inter-pregnancy interval on maternal and infant
role, with less emphasis on the health effects associ- health. Paediatric and Perinatal Epidemiology 2012;
26(Suppl 1):239-258.
ated with birth spacing.12 An improved understand-
3 Report of a WHO Technical Consultation on Birth Spacing
ing of the complex decision making around having Geneva, Switzerland 13–15 June 2005. Department of
additional children can help women and couples Reproductive Health and Research (RHR).
achieve their desired reproductive goals, including 4 Lunde AS. Revisions of U.S. standard certificates on vital
how birth spacing can lead to better outcomes for the events. Public Health Reports 1967; 82:913-916.
mother and child. 5 Zhu BP, Rolfs RT, Nangle BE, Horan JM. Effect of the
interval between pregnancies on perinatal outcomes. New
England Journal of Medicine 1999; 340:589-594.
Acknowledgement 6 McKinney D, House M, Chen A, Muglia L, DeFranco E. The
influence of interpregnancy interval on infant mortality.
The work performed by Katherine Ahrens was as an American Journal of Obstetics and Gynecology 2017; 216:316.
employee of the US federal government. The findings e311-316.e319.
7 Zhang L, Shen S, He J, Chan F, Lu J, Li W, et al. Effect of
and conclusions in this article are those of the authors
interpregnancy interval on adverse perinatal outcomes in
and do not necessarily represent the official position
Southern China: a retrospective cohort study, 2000-2015.
of the Office of Population Affairs, Office of the Assis- Paediatric and Perinatal Epidemiology 2018; DOI: 10.1111/
tant Secretary for Health, Department of Health and ppe.12432.
Human Services. 8 Conzuelo Rodriguez G, Naimi A. The impact of computing
inter pregnancy intervals without accounting for intervening
pregnancy events. Paediatric and Perinatal Epidemiology 2018;
About the authors [Epub ahead of print].
9 Schliep KC, Mitchell EM, Mumford SL, Radin RG, Zarek SM,
Katherine Ahrens is a health scientist at the Office of Sjaarda L, et al. Trying to conceive after an early pregnancy
Population Affairs (OPA), Office of the Assistant Sec- loss: an assessment on how long couples should wait.
retary for Health, US Department of Health and Obstetrics and Gynecology 2016; 127:204-212.
10 Kc S, Gissler M, Virtanen SM, Klemetti R. Risks of adverse
Human Services. In this role, she conducts analyses
perinatal outcomes after repeat terminations of pregnancy by
and provides programme guidance related to family
their methods: a nationwide register-based cohort study in
planning and reproductive health services provided Finland 1996-2013. Paediatric and Perinatal Epidemiology 2017;
in the United States, including services provided 31:485-492.
through the Title X Family Planning Program. She 11 Van Bavel J. Deliberate birth spacing before the fertility
and Dr. Hutcheon recently served as co-chairs at an transition in Europe: evidence from nineteenth-century
Belgium. Population Studies 2004; 58:95-107.
expert work group meeting convened by OPA to
12 Bryant A, Fernandez-Lamothe A, Kuppermann M. Attitudes
address the methodological aspects of existing litera- toward birth spacing among low-income, postpartum
ture on short interpregnancy intervals and adverse women: a qualitative analysis. Maternal Child Health Journal
pregnancy outcomes. 2012; 16:1440-1446.

© 2018 John Wiley & Sons Ltd

Paediatric and Perinatal Epidemiology, 2018, , –