Association of Cesarean Delivery With Body Mass Index z Score at Age 5 Years

Cesarean delivery rates remain high and variable across hospitals, regions, and countries.^1,2 Cesarean delivery may be a risk factor for childhood obesity,^1,2 possibly because delivery route can influence the intestinal microbiomes,³ which may influence energy regulation.

Two meta-analyses, summarizing data from 24 studies, have reported an increased risk of obesity for individuals with cesarean birth (pooled odds ratio [OR], 1.22 [95% CI, 1.05-1.42] and 1.33 [1.19-1.48]).^1,2 Limitations of earlier studies include small sample size in several studies and lack of adjustment for maternal body mass index (BMI [calculated as weight in kilograms divided by height in meters squared]) and sociocultural factors. Even after adjusting for these measured maternal characteristics, residual confounding is likely. Within-family analysis is one way of controlling for such confounding. Because siblings grow up in similar social, economic, and cultural environments and share the same genetic predisposition to obesity, sibling studies minimize the variation in several of the noncausal factors that could explain why cesarean delivery could appear to be associated with a higher risk of obesity. The main objective of this study was to examine within-family and additional between-family associations of delivery route with BMI z score at 5 years of age in a large cohort of sibling pairs to separate most putative confounding effects from the effects of cesarean delivery.

We used data from the Linked CENTURY (Collecting Electronic Nutrition Trajectory Data Using Records of Youth) study, a longitudinal clinical database of well-child visits linked to each child’s birth certificate.⁴ The Linked CENTURY study includes singleton-birth children younger than 18 years who were seen for a well-child visit at Atrius Harvard Vanguard Medical Associates in eastern Massachusetts from 1980 through 2008 and whose data were linked to their Massachusetts birth certificate. We included 16 140 siblings born between 1987 and 2003 and their 8070 mothers. We obtained delivery route, which was the main exposure, from birth certificates. We extracted height and weight data for children at least 5 and less than 6 years of age from electronic health records and calculated BMI as weight in kilograms divided by height in meters squared. This study was approved and informed consent was waived by the Harvard Pilgrim Health Care Institutional Review Board.

Our main outcome was age-specific and sex-specific BMI z score at 5 years of age based on Centers for Disease Control and Prevention reference data.⁵ We used a single linear mixed model, adjusted for mother’s age, parity, and race/ethnicity and for child’s age, sex, and birth year, to decompose the total association of cesarean delivery with BMI z score into a within-family component and an additional between-family component.⁶ The model equation for this is

Y_ij = α + β_wX_ij + β_dX̅_i +v_i +e_ij,

where Y_ij = BMI z score for the ith individual from the jth family, X_ij = individual’s cesarean delivery status (0 = no and 1 = yes), X̅_i = family average cesarean deliveries (0 = both siblings vaginal delivery, 0.5 = one sibling vaginal delivery and one sibling cesarean, and 1 = both siblings cesarean), v_i = family random effect, and e_ij = random error; v_i and e_ij are assumed to be uncorrelated and normally distributed, each with a mean value of 0. The estimated within-family component β_w is mostly unconfounded by other factors. We also fit a model that did not decompose the effect of cesarean delivery; this model approximates the possibly confounded results we would have observed without accounting for sibling status. We fit all models using PROC GLIMMIX, SAS, version 9.4 (SAS Institute, Inc).

A total of 3804 of the 16 140 children (23.6%) had nonwhite mothers. A total of 7943 of the 16 140 children (49.2%) were female, 3204 children (19.9%) had cesarean delivery, and the mean (SD) BMI z score at 5 years of age was 0.48 (1.00). Mean BMI z score was 0.45 among siblings who both had vaginal delivery, 0.51 among siblings with 1 cesarean and 1 vaginal delivery, and 0.63 among siblings who both had cesarean delivery (Table 1).

In the covariate-adjusted model, the within-family association of cesarean vs vaginal delivery was 0.04 higher BMI z score at 5 years of age (95% CI, −0.04 to 0.11) and the additional between-family association was 0.13 (95% CI, 0.04 to 0.22) (Table 2). In a model adjusted for the same covariates, but without decomposing the within-family and between-family effects, children with cesarean delivery had 0.13 higher BMI z score (95% CI, 0.08 to 0.17).

We found that, within families, cesarean delivery was not associated with higher BMI z score at 5 years of age. This null finding suggests that confounding by unmeasured variables, such as maternal BMI and sociocultural factors, accounts for observed associations between cesarean delivery and BMI z score in some earlier studies. Such confounding is also reflected in the difference between our within-family estimate (0.04) and our additional between-family estimate (0.13).

This study had several limitations. All patients had health insurance; our results may not be generalizable to individuals without insurance. We did not have data on whether the cesarean delivery was an emergency procedure or data on maternal prepregnancy BMI. Maternal BMI, however, differs more between families than across pregnancies, which is another reason to use a sibling-pairs design. Strengths of this study include the large number of sibling pairs and our ability to link clinical data with birth certificates.

We did not observe a within-family effect of cesarean delivery. This finding suggests that reported associations between cesarean delivery and childhood obesity may be confounded by unmeasured variables.

Corresponding Author: Sheryl L. Rifas-Shiman, MPH, Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Health Care Institute, 401 Park Dr, Ste 401, Boston, MA 02215 (sheryl_rifas@hphc.org).

Accepted for Publication: February 27, 2018.

Published Online: June 11, 2018. doi:10.1001/jamapediatrics.2018.0674

Author Contributions: Ms Rifas-Shiman had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Rifas-Shiman, Gillman, Taveras, Kleinman.

Acquisition, analysis, or interpretation of data: Rifas-Shiman, Hawkins, Oken, Kleinman.

Drafting of the manuscript: Rifas-Shiman, Kleinman.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Rifas-Shiman, Kleinman.

Obtained funding: Hawkins, Taveras.

Administrative, technical, or material support: Rifas-Shiman, Gillman, Hawkins, Oken, Taveras.

Study supervision: Gillman, Oken.

Conflict of Interest Disclosures: None reported.

Funding/Support: This work was partially funded by grant R00 HD068506 from the National Institutes of Health (Dr Hawkins) and grant 200-2008-M-26882 from the National Center for Chronic Disease Prevention and Health Promotion (Dr Taveras).

Role of the Funder/Sponsor: The National Institutes of Health and the National Center for Chronic Disease Prevention and Health Promotion had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The views expressed in this article do not necessarily represent the views of the US Government, the Department of Health and Human Services, or the National Institutes of Health.