Mode of Delivery and Offspring Body Mass Index, Overweight and Obesity in Adult Life: A Systematic Review and Meta-Analysis

Background It has been suggested that mode of delivery, a potentially powerful influence upon long-term health, may affect later life body mass index (BMI). We conducted a systematic review and meta-analysis of the effect of Caesarean section (CS) and vaginal delivery (VD) on offspring BMI, overweight (BMI>25) and obesity (BMI>30) in adulthood. Secondary outcomes were subgroup analyses by gender and type of CS (in-labour/emergency, pre-labour/elective). Methods Using a predefined search strategy, Pubmed, Google Scholar and Web of Science were searched for any article published before 31st March 2012, along with references of any studies deemed relevant. Studies were selected if they reported birth characteristics and long-term offspring follow-up into adulthood. Aggregate data from relevant studies were extracted onto a pre-piloted data table. A random-effects meta-analysis was carried out in RevMan5. Results are illustrated using forest plots and funnel plots, and presented as mean differences or odds ratios (OR) and 95% confidence intervals. Results Thirty-five studies were identified through the search, and 15 studies with a combined population of 163,753 were suitable for inclusion in the meta-analysis. Comparing all CS to VD in pooled-gender unadjusted analyses, mean BMI difference was 0·44 kg·m-2 (0·17, 0·72; p = 0·002), OR for incidence of overweight was 1·26 (1·16, 1·38; p<0·00001) and OR for incidence of obesity was 1·22 (1·05, 1·42; p = 0·01). Heterogeneity was low in all primary analyses. Similar results were found in gender-specific subgroup analyses. Subgroup analyses comparing type of CS to VD showed no significant impact on any outcome. Conclusions There is a strong association between CS and increased offspring BMI, overweight and obesity in adulthood. Given the rising CS rate worldwide there is a need to determine whether this is causal, or reflective of confounding influences. Systematic review registration An a priori protocol was registered on PROSPERO (registration number: CRD42011001851)


Types of study to be included
There are no restrictions on the types of study design eligible for inclusion. All study types reporting Body Mass Index/obesity/overweight in adult life, categorised by method of delivery, will be included.

Condition or domain being studied
Mode of delivery has been implicated as an event that might influence health and the risk of disease in adult life (Hyde MJ et al (2011) Biological Reviews, DOI: 10.1111/j.1469-185X.2011. A few published reports support the hypothesis that the risk of overweight/obesity is affected (eg Goldani HS et al, Am J Clin Nutr 2011;93:1344) Delivery by Caesarean section and pre-labour Caesarean section is growing, as is obesity and overweight. Individual studies suggest an association between Caesarean section and the risk of obesity. There are plausible candidate mechanisms that might underpin such as association. We therefore propose to conduct a systematic review and meta-analysis of available data on adult Body Mass Index and weight in relation to mode of delivery.

Participants/ population
Outcomes are evaluated in adults (i.e. age 18 or above).

Context
The review will review all identified studies

Primary outcomes
Difference in Body Mass Index (kg/m^2) measured between control and intervention group in adult life (18years or older) Difference in prevalence of overweight (Body Mass Index greater than 25kg/m^2) /obese (Body Mass Index greater than 30kg/m^2) measured between control and intervention group in adult life (18years or older) Secondary outcomes None

Data extraction, (selection and coding)
Titles and abstracts (where available), of studies identified from the initial searches will be independently assessed by two reviewers for possible inclusion [KD and MJH]. The full text of all potentially eligible studies that have been identified will then be appraised by two assessors independently [KD and CG]. Where there is disagreement over eligibility for inclusion, this will be referred to a meeting of all authors. Data will be extracted independently onto a pre-piloted data collection form [KD and CG]. Quality of the included studies will be assessed using standard criteria, looking for potential bias [All authors independently].
The following data will be extracted for each study: type of study (cross sectional/longitudinal; retrospective/prospective); country of study; study population, setting and demographics of research subjects (offspring: gender, age at study, maternal Body Mass Index during pregnancy, maternal diabetes during pregnancy); details of intervention (type of delivery, year/period of deliveries being studied); inclusion and exclusion criteria; recruitment and study completion rates; outcome measurements; and adjusted analyses performed (yes/no and factors adjusted for). Missing data will be requested from study authors by MJH.

Risk of bias (quality) assessment
[1] Review level Bias: Bias from small study effects will be assessed visually using a funnel plot and Egger's test; where asymmetry is evident on the Funnel Plot a trim and fill analysis will be used. Possible causes for asymmetry other than publication bias (e.g. between-study heterogeneity) will also be considered. Where population based cohorts are identified by the search strategy but outcomes are not reported on the basis of mode of delivery, attempts will be made to extract this information for these studies by contacting the data custodian in order to reduce effects of publication bias.
[2] Study level bias: A modified Newcastle-Ottawa scale will be used to assess methodological quality of each individual study (carried out independently by SS and KD, discrepancies to be resolved by group discussion). A subgroup analysis of the studies scoring 7 out of 7 stars will be conducted. Ranked forest plots (on the basis of the score out of 7 that each study achieves) will be produced for primary outcomes, to allow assessment of study quality on effect size.

Strategy for data synthesis
Data collected will be systematised into a table. All studies identified in the systematic review will be included in the table whether or not they are included in the meta-analysis. A narrative description of these studies will be produced. Where studies are comparable on the basis of study population, intervention and outcome measure, the results will be pooled in a fixed-effects meta-analysis, with mean differences, 95% confidence intervals and two sided p-values. Heterogeneity will be assessed using the Chi-squared test for the Q statistic and calculation of Isquared, an estimate of the proportion of variance due to between-study heterogeneity. Where evidence of heterogeneity is present (p=value from Chi-squared test <0.05) a random effects meta-analysis will be used with meta-regression and subgroup analysis to explore sources of heterogeneity arising from study characteristics. Where between-study heterogeneity is very high (I-squared over 80%) and is not explained through differences in study characteristics we will re-consider whether quantitative data synthesis is appropriate. Data from studies adjusting for confounders will be synthesised in a separate meta-analysis to the unadjusted data.

Analysis of subgroups or subsets
[1] Subgroup analysis: For each subgroup analysis we will use meta-regression to investigate whether the effects are significantly different in the subgroups. Subgroup comparisons to be carried out: a) Type of Caesarean section (pre labour and in labour), and b) Gender.
[2] Meta regression: To investigate heterogeneity for continuous variables, we will use scatter plots of effect size against study characteristic and meta-regression. Meta regression analyses to be carried out: a) Age of offspring at assessment, b) Study quality, and c) Difference in Maternal Body Mass Index between Caesarean Section and Vaginal Delivery groups.
[3] Sensitivity analysis: Sensitivity analysis will be used to test whether conclusions are robust to alterations in the inclusion criteria. Sensitivity analyses to be carried out: a) Exclusion of studies including preterms (<37 gestational weeks), and b) Exclusion of studies including diabetic pregnancies

Dissemination plans
The results of this review will be presented at meetings of relevant societies and interest groups. It will be written up for peer-reviewed publication.

Funding sources/sponsors
Funding source: Imperial College project funding

Subject index terms status
Subject indexing assigned by CRD