Physical activity–based interventions ranked highest in a network meta-analysis, and group delivery was associated with greater benefit, according to a recent study.
In an individual participant data (IPD) and network meta-analysis, investigators found that lifestyle interventions were associated with reductions in gestational diabetes, with physical activity–based programs consistently outperforming diet or mixed approaches.
The investigators included 104 randomized trials involving 35,993 pregnant women. IPD were available for 68% (n = 24,391) of the women across 54 trials. Lifestyle interventions were categorized as physical activity based, diet based, or mixed. The primary outcomes were gestational diabetes defined by any criteria and by National Institute for Health and Care Excellence criteria. Secondary outcomes included gestational diabetes defined by International Association of Diabetes in Pregnancy Study Groups criteria and key maternal and perinatal outcomes.
In IPD-only analyses, lifestyle interventions were associated with a 10% reduction in gestational diabetes defined by any criteria, corresponding to an absolute risk reduction of 1.3%. When aggregate data from non-IPD trials were incorporated, the reduction increased to 20%, with an absolute risk reduction of 2.6%. Physical activity–based interventions were associated with an absolute risk reduction of 5%. Diet-based interventions were associated with a modest reduction, whereas mixed interventions weren't associated with a statistically significant difference.
When gestational diabetes was defined by International Association of Diabetes in Pregnancy Study Groups criteria, lifestyle interventions were associated with a 14% reduction in IPD analyses, corresponding to an absolute risk reduction of 2.7%; this finding persisted after inclusion of aggregate data. No statistically significant reduction was observed when National Institute for Health and Care Excellence criteria were applied.
The investigators performed a two-stage IPD meta-analysis. Trial-level logistic regression models adjusted for maternal age and body mass index were fit in the first stage, with clustering accounted for where appropriate. In the second stage, effect estimates were pooled using random effects models with Hartung-Knapp correction. Absolute risk reductions were derived by applying pooled odds ratios to average baseline risks. A multivariate random effects network meta-analysis was used to rank interventions, identifying physical activity–based programs as most effective.
Treatment–covariate interaction analyses showed no differential effects by maternal body mass index, age, parity, or ethnicity. However, women with low educational levels experienced smaller relative benefits compared with those with middle or high educational attainment.
Study limitations included the incomplete availability of IPD, heterogeneity in intervention design and intensity, the limited reporting of ethnicity data, and the restriction of most trials to high-income settings. In addition, the structure of the network prevented formal statistical evaluation of consistency.
“Lifestyle interventions in pregnancy are likely to prevent gestational diabetes, with effects varying according to diagnostic criteria. Implementation strategies should address inequalities by maternal education and consider group formats, provider training, and physical activity based–interventions to prevent gestational diabetes,” noted lead study author John Allotey, Professor of the Institute of Life Course and Medical Sciences at the University of Liverpool in the United Kingdom, and colleagues.
Support for the research in this study was provided by the National Institute for Health Research (NIHR) Health Technology Assessment UK programme, NIHR Applied Research Collaboration North West Coast at the National Health Service (NHS) Cheshire and Merseyside ICB, University of Liverpool, NIHR Birmingham Biomedical Research Centre at University Hospitals Birmingham NHS Foundation Trust, and University of Birmingham. Full disclosures can be found in the study.
Source: The BMJ
