Keele Research Repository

Objective To synthesise the evidence on the overall and differential effects of interventions based on diet and physical activity during pregnancy, primarily on gestational weight gain and maternal and offspring composite outcomes, according to women’s body mass index, age, parity, ethnicity, and pre-existing medical condition; and secondarily on individual complications DESIGN Systematic review and meta-analysis of individual participant data (IPD).

DATA SOURCES Major electronic databases from inception to February 2017 without language restrictions.

ELIGIBILITY CRITERIA FOR SELECTING STUDIES Randomised trials on diet and physical activity based interventions in pregnancy. DATA SYNTHESIS Statistical models accounted for clustering of participants within trials and heterogeneity across trials leading to summary mean differences or odds ratios with 95% confidence intervals for the effects overall, and in subgroups (interactions).

RESULTS IPD were obtained from 36 randomised trials (12 526 women). Less weight gain occurred in the intervention group than control group (mean difference −0.70 kg, 95% confidence interval −0.92 to −0.48 kg, I2 =14.1%; 33 studies, 9320 women). Although summary effect estimates favoured the intervention, the reductions in maternal (odds ratio 0.90, 95% confidence interval 0.79 to 1.03, I2 =26.7%; 24 studies, 8852 women) and offspring (0.94, 0.83 to 1.08, I2 =0%; 18 studies, 7981 women) composite outcomes were not statistically significant. No evidence was found of differential intervention effects across subgroups, for either gestational weight gain or composite outcomes. There was strong evidence that interventions reduced the odds of caesarean section (0.91, 0.83 to 0.99, I 2 =0%; 32 studies, 11 410 women), but not for other individual complications in IPD meta-analysis. When IPD were supplemented with study level data from studies that did not provide IPD, the overall effect was similar, with stronger evidence of benefit for gestational diabetes (0.76, 0.65 to 0.89, I2 =36.8%; 59 studies, 16 885 women)