Influence of maternal obesity on the association between common pregnancy complications and risk of childhood obesity: an individual participant data meta-analysis

Abstract

Background: Gestational diabetes and gestational hypertensive disorders are associated with offspring obesity, but the role of maternal adiposity in these associations remains unclear. We aimed to investigate whether these pregnancy complications affect the odds of offspring obesity independently of maternal obesity.

Methods: We did an individual participant data (IPD) meta-analysis of mother-offspring pairs from prospective birth cohort studies that had IPD on mothers with singleton liveborn children born from 1989 onwards and had information available about maternal gestational diabetes, gestational hypertension or pre-eclampsia, and childhood body-mass index (BMI). We applied multilevel mixed-effects models to assess associations of gestational diabetes, gestational hypertension, and pre-eclampsia with BMI SD scores and the odds of overweight and obesity throughout childhood, adjusting for lifestyle characteristics (offspring's sex, maternal age, educational level, ethnicity, parity, and smoking during pregnancy). We then explored the extent to which any association was explained by maternal pre-pregnancy or early-pregnancy BMI.

Findings: 160 757 mother-offspring pairs from 34 European or North American cohorts were analysed. Compared with uncomplicated pregnancies, gestational diabetes was associated with increased odds of overweight or obesity throughout childhood (odds ratio [OR] 1·59 [95% CI 1·36 to 1·86] for early childhood [age 2·0-4·9 years], 1·41 [1·26 to 1·57] for mid childhood [5·0-9·9 years], and 1·32 [0·97 to 1·78] for late childhood [10·0-17·9 years]); however, these associations attenuated towards the null following adjustment for maternal BMI (OR 1·35 [95% CI 1·15 to 1·58] for early childhood, 1·12 [1·00 to 1·25] for mid childhood, and 0·96 [0·71 to 1·31] for late childhood). Likewise, gestational hypertension was associated with increased odds of overweight throughout childhood (OR 1·19 [95% CI 1·01 to 1·39] for early childhood, 1·23 [1·15 to 1·32] for mid childhood, and 1·49 [1·30 to 1·70] for late childhood), but additional adjustment for maternal BMI largely explained these associations (1·01 [95% CI 0·86 to 1·19] for early childhood, 1·02 [0·95 to 1·10] for mid childhood, and 1·18 [1·03 to 1·36] for late childhood). Pre-eclampsia was associated with decreased BMI in early childhood only (difference in BMI SD score -0·05 SD score [95% CI -0·09 to -0·01]), and this association strengthened following additional adjustment for maternal BMI.

Interpretation: Although lowering maternal risk of gestational diabetes, gestational hypertension, and pre-eclampsia is important in relation to maternal and fetal pregnancy outcomes, such interventions are unlikely to have a direct impact on childhood obesity. Preventive strategies for reducing childhood obesity should focus on maternal BMI rather than on pregnancy complications.

Funding: EU's Horizon 2020 research and innovation programme (LifeCycle Project).

Figure 1

Percentages of childhood overweight according to maternal pre-pregnancy BMI category and presence or absence of A. gestational diabetes, B. gestational hypertension, and C. pre-eclampsia.

Figure 2

Associations of gestational diabetes with offspring BMI outcomes in early, mid, and late childhood.^{a a}Values are regression coefficients (95% confidence intervals) from multilevel linear mixed effects models and odds ratios (95% confidence intervals) from multilevel binary logistic models that reflect differences in early childhood (2·0 to 4·9 years), mid childhood (5·0 to 9·9 years) and late childhood (10·0 to 17·9 years) BMI SDS and risk of overweight and obesity, respectively, for children born to mothers with gestational diabetes, as compared with the reference group (children born to mothers with an uncomplicated pregnancy). Lifestyle characteristics models are adjusted for offspring’s sex, maternal age, educational level, ethnicity, parity, and smoking during pregnancy. Maternal BMI models are additionally adjusted for maternal pre/early-pregnancy BMI.

Figure 3

Associations of gestational hypertension with offspring BMI outcomes in early, mid, and late childhood.^{a a}Values are regression coefficients (95% confidence intervals) from multilevel linear mixed effects models and odds ratios (95% confidence intervals) from multilevel binary logistic models that reflect differences in early childhood (2·0 to 4·9 years), mid childhood (5·0 to 9·9 years) and late childhood (10·0 to 17·9 years) BMI SDS and risk of overweight and obesity, respectively, for children born to mothers with gestational hypertension, as compared with the reference group (children born to mothers with an uncomplicated pregnancy). Lifestyle characteristics models are adjusted for offspring’s sex, maternal age, educational level, ethnicity, parity, and smoking during pregnancy. Maternal BMI models are additionally adjusted for maternal pre/early-pregnancy BMI.

Figure 4

Associations of pre-eclampsia with offspring BMI outcomes in early, mid, and late childhood.^{a a}Values are regression coefficients (95% confidence intervals) from multilevel linear mixed effects models and odds ratios (95% confidence intervals) from multilevel binary logistic models that reflect differences in early childhood (2·0 to 4·9 years), mid childhood (5·0 to 9·9 years) and late childhood (10·0 to 17·9 years) BMI SDS and risk of overweight and obesity, respectively, for children born to mothers with pre-eclampsia, as compared with the reference group (children born to mothers with an uncomplicated pregnancy). Lifestyle characteristics models are adjusted for offspring’s sex, maternal age, educational level, ethnicity, parity, and smoking during pregnancy. Maternal BMI models are additionally adjusted for maternal pre/early-pregnancy BMI.