Abstract
BACKGROUND: To investigate the relationship between body mass index (BMI) changes and large for gestational age (LGA) in women with gestational diabetes mellitus (GDM). METHODS: A retrospective cohort study including 10,486 women with GDM was conducted. A dose‒response analysis of BMI changes and the occurrence of LGA was performed. Binary logistic regressions were performed to assess crude and adjusted odds ratios (ORs) and 95% confidence intervals (CIs). Receiver operating characteristic (ROC) curves and areas under the curve (AUCs) were used to assess the ability of BMI changes to predict LGA. RESULTS: The probability of LGA increased with increasing BMI. The risk of LGA increased across the BMI change quartiles. The BMI change remained positively associated with the risk of LGAafter stratification analysis. The AUC was 0.570 (95% CI: 0.557 ~ 0.584)in the entire study population, and the best optimal predictive cut-off value was 4.922, with a sensitivity of 0.622 and a specificity of 0.486. The best optimal predictive cut-off value decreased from the underweight group to the overweight and obese group. CONCLUSIONS: BMI changes are related to the risk of LGA and may be a useful predictor of the incidence of LGA in singleton pregnant women with GDM.