Effectiveness of pre-pregnancy lifestyle in preventing gestational diabetes mellitus—a systematic review and meta-analysis of 257,876 pregnancies

GDM is hyperglycaemia first detected during pregnancy and is typically diagnosed between 24–28 weeks’ gestation. It affects approximately 16% of live births globally, with over 90% of cases occurring in LMICs. Hyperglycaemia may be present earlier in gestation, and adverse fetal effects can occur by the time of GDM diagnosis. Maternal metabolic health influences offspring and potentially intergenerational metabolic risk, with GDM linked to adverse fetal programming. Current management reduces short-term complications but may not mitigate early programming effects, highlighting prevention as a priority. Lifestyle interventions prevent type 2 diabetes and may prevent GDM; however, antenatal lifestyle trials show mixed results, with early gestation interventions more promising. Preventing GDM via pre-pregnancy lifestyle changes could better address risks around conception. This study systematically reviews and meta-analyzes evidence on pre-pregnancy diet and physical activity in relation to GDM risk.

Prior research indicates lifestyle interventions can prevent type 2 diabetes and may reduce GDM risk when initiated in early pregnancy. Meta-analyses of antenatal diet and physical activity interventions have shown reduced gestational weight gain and some maternal/neonatal benefits, with ongoing IPD meta-analyses on GDM prevention. Evidence specifically on pre-pregnancy interventions has been limited and heterogeneous, with few studies and especially sparse data from LMICs. Observational cohorts suggest healthier dietary patterns, higher fiber and vitamin intake, and greater physical activity before pregnancy are associated with lower GDM risk, whereas higher intakes of fried foods, sugar-sweetened beverages, and pro-inflammatory diets are associated with higher risk.

Design: Systematic review and meta-analysis following a pre-registered protocol (PROSPERO CRD42020189574). Population: Pregnant women. Intervention/Exposure: Pre-pregnancy diet and/or physical activity lifestyle interventions or exposures. Comparator: No pre-pregnancy lifestyle intervention/exposure. Outcome: GDM incidence. Searches: MEDLINE, Web of Science, Embase, and Cochrane CENTRAL from inception to July 2022 using MeSH and free-text terms related to pregnancy, lifestyle intervention (diet, exercise, behavioral change, education), and GDM; limited to human studies without language restrictions. Study selection: RCTs, cohort, and case-control studies assessing pre-pregnancy lifestyle effects on GDM included. Exclusions: Pre-existing type 1 or type 2 diabetes, age <18 or >50 years, recent metformin for infertility, severe anaemia or hemoglobinopathies, serious medical illness, interventions starting only during pregnancy, drug-only interventions without a lifestyle arm, or studies not reporting GDM incidence. Two reviewers independently screened titles/abstracts and full texts using Rayyan; disagreements were resolved by discussion or third reviewer. Data extraction: Two authors independently extracted study characteristics (authors, year, country, sample size, demographics), intervention/exposure details and timing, assessment methods, and effect estimates (OR, RR, events). Risk of bias: RCTs assessed with Cochrane ROB-2; observational studies with Newcastle-Ottawa Scale (NOS). Publication bias and small-study effects assessed via funnel plots and Egger’s tests. Synthesis: Random-effects Mantel–Haenszel models produced pooled ORs with 95% CIs. Heterogeneity assessed with I² and Chi². For cohorts, comparable categories were meta-analyzed: pre-pregnancy physical activity, low carbohydrate/low sugar diet, and higher quality diet scores. Studies from the same cohort were included once per meta-analysis. RevMan 5.4.1 was used for analyses. Role of funder: No role in study design, conduct, analysis, or reporting.

- Inclusion: 30 studies totaling 257,876 pregnancies: 5 RCTs (n=2,471), 4 case-control (n=19,778), and 21 cohort studies (n=235,627). Of cohorts, 5 were physical-activity-focused (n=46,197) and 16 diet-focused (n=189,430). Most studies were from high-income countries (24/30), with 6 from LMICs. - RCTs (pre-pregnancy lifestyle interventions; 3 diet-only, 2 combined diet+PA): Pooled OR 0.76 (95% CI 0.50–1.17), p=0.21; moderate heterogeneity. Subgroup LMIC RCTs showed OR 0.61 (95% CI 0.43–0.85). - Cohort meta-analyses: • Pre-pregnancy physical activity (N=4; n=23,263): OR 0.66 (95% CI 0.44–0.99), p=0.04 (I²≈59%). • Low carbohydrate/low sugar diet (N=4; n=25,739): OR 0.86 (95% CI 0.68–1.09), p=0.22 (I²≈88%). • Higher quality diet scores (N=4): OR 0.72 (95% CI 0.60–0.87), p=0.0006 (I²≈65%). - Case-control highlights: Prudent/vegetable/vitamin-rich patterns associated with lower GDM risk (e.g., vitamin pattern OR per quartile 0.91, 95% CI 0.86–0.96; vegetable pattern OR per quartile 0.94, 95% CI 0.89–0.99); higher dietary inflammatory index associated with higher odds (tertile 3 vs 1 OR 2.1, 95% CI 1.02–4.34). - Risk of bias: Only ~20% of RCTs rated low risk; concerns mainly in randomization and missing data. Cohort and case-control studies generally low risk per NOS (score ~8). Funnel plots suggested low publication bias; sensitivity analyses supported result robustness.

The review addresses whether pre-pregnancy lifestyle can prevent GDM. While RCT evidence remains limited and heterogeneous, pooled RCTs suggest a potential 24% risk reduction that did not reach statistical significance, indicating feasibility but underscoring the need for more robust trials. Observational data consistently link higher pre-pregnancy physical activity and healthier dietary patterns (including higher-quality diet scores, greater vegetable/fruit intake, vitamins/folate, and fiber) with lower GDM risk, whereas pro-inflammatory diets, fried foods, and sugar-sweetened beverages relate to higher risk. These findings support the hypothesis that optimizing lifestyle before conception can reduce GDM incidence and potentially mitigate adverse fetal programming and intergenerational metabolic risk. The stronger and consistent associations in cohort studies, alongside signals from LMIC RCTs, highlight a meaningful preventive opportunity, yet causal inference requires well-designed, adequately powered RCTs, ideally combining diet and physical activity components and tailoring to cultural and individual contexts.

Pre-pregnancy lifestyle factors—particularly higher physical activity and healthier dietary patterns characterized by greater intake of vegetables, fruits, fiber, vitamins/folate, and overall high-quality diet scores—are associated with reduced GDM risk. Although current RCTs are few and heterogeneous, they suggest preconception interventions are feasible and potentially beneficial. Future research should prioritize rigorously designed, adequately powered, and personalized RCTs that combine multiple lifestyle components, with greater representation from LMICs and consideration of inter-pregnancy windows and scalable delivery (e.g., mobile health) to improve clinical and cost-effectiveness.

- High heterogeneity across studies due to variability in lifestyle components, populations, and measurements. - Heavy reliance on self-reported dietary and physical activity questionnaires in observational studies, introducing recall and reporting biases. - Limited number of pre-pregnancy RCTs, some with risk of bias in randomization and missing data. - Sparse evidence from LMICs, limiting generalizability to settings with the highest GDM burden. - Variability in GDM diagnostic criteria across studies.