Construction and validation of a risk prediction model for hypoglycemia in patients with gestational diabetes mellitus.

Objective: We explored the prevalence and determinants of hypoglycemia in patients with gestational diabetes mellitus (GDM), and we developed and validated a nomogram prediction model.

Methods: We extracted data from the clinical records of 475 patients with GDM attending the tertiary class A specialized hospital in Zhuhai City between December 2021 and June 2023 for a modeling group, and we used data of another cohort of 204 GDM cases for a validation group. We conducted a logistic regression analysis to identify factors associated with hypoglycemia in patients with GDM and generated a risk prediction model presented as a nomogram. The model was validated using data from the patients in the validation group.

Results: The prevalence of hypoglycemia in the study population was 25.5%. Our risk prediction model incorporated four predictors, including a fasting oral glucose tolerance test (OGTT) value, the number of fetuses, the presence or absence of intrahepatic cholestasis of pregnancy (ICP), and the blood glucose level self-monitoring frequency. The area under the receiver operating characteristic (ROC) curve was 0.786 for the modeling set and 0.742 for the validation set. The Brier score was 0.155, and the calibration slope was 0.750, demonstrating satisfactory clinical usefulness of the model. Moreover, a decision curve analysis further supported our model's clinical relevance.

Conclusion: The prevalence of hypoglycemia in patients with GDM is considerable. Our nomogram prediction model demonstrated good performance for identifying high-risk individuals. The model could serve as a valuable tool for screening and managing hypoglycemia among patients with GDM.