Abstract
<title>Abstract</title> <p>Background: Gestational diabetes mellitus (GDM) is a major pregnancy complication associated with adverse maternal and neonatal outcomes. While established risk factors such as maternal age and pre-pregnancy body mass index (BMI) are well recognized, the role of parity remains controversial, with inconsistent findings across populations. These discrepancies may be explained by inadequate adjustment for confounders and limited exploration of effect modification. Methods: In this population-based retrospective cohort study, 423,273 pregnant women from the SINA electronic health record system (2018–2025) were analyzed. Parity was categorized as 0, 1, and ≥ 2. GDM was diagnosed based on IADPSG/WHO criteria. Multivariable logistic regression models were used to estimate adjusted odds ratios (ORs) and 95% confidence intervals (CIs). Stratified analyses and interaction tests were conducted to assess modification by maternal age and BMI. Results: The overall prevalence of GDM was 4.4%. Crude prevalence increased with parity (3.4%, 4.2%, and 5.7% for parity 0, 1, and ≥ 2, respectively; P < 0.001). In partially adjusted models, higher parity was associated with increased GDM risk (parity 1: OR 1.12, 95% CI 1.08–1.17; parity ≥ 2: OR 1.37, 95% CI 1.31–1.42). However, after full adjustment, higher parity was associated with lower odds of GDM (parity 1: OR 0.85, 95% CI 0.80–0.91; parity ≥ 2: OR 0.81, 95% CI 0.75–0.87). Significant interactions were observed for both maternal age and BMI (P for interaction < 0.001). The inverse association was limited to overweight/obese women (BMI ≥ 25 kg/m²) and was strongest among women aged ≥ 35 years. Conclusion: The apparent positive association between parity and GDM is largely explained by confounding, particularly by maternal age and BMI. After adjustment, higher parity is associated with a reduced risk of GDM in specific subgroups, highlighting the importance of considering effect modification in risk assessment and epidemiological analyses.</p>