First-trimester Risk Score Model for Gestational Diabetes Mellitus based on a Systematic Review and Meta-analysis

Abstract

Objective: To construct and confirm a risk model for predicting gestational diabetes mellitus (GDM) in the early stage of pregnancy. Method: Meta-analysis was employed to determine the factors contributing to GDM. Based on the findings of this meta-analysis, the Rothman-Keller model was formulated. The model’s effectiveness was then evaluated using actual clinical data from real-world settings. Results: After screening, 1,548,515 women were ultimately incorporated into our meta-analysis. Nine risk factors linked to the occurrence of gestational diabetes, including maternal age ≥ 35, pre-pregnancy overweight or obese, family history of diabetes, history of GDM, polycystic ovary syndrome, parity ≥ 1, history of abortion, conception by assisted reproductive technology, and smoking before or during early pregnancy, were identified and incorporated into the model. In the external validation, the area under the receiver operating characteristic curve (AUC) of the model was 0.714 (95% CI 0.672–0.755). Conclusion: A first-trimester GDM prediction model was developed and validated through an extensive meta-analysis, identifying risk factors. Although the predictive accuracy of the model was promising, further refinement and validation were recommended to enhance its clinical utility.