A nomogram to predict gestational diabetes mellitus: a multicenter retrospective study.

Area of Science:

• Obstetrics and Gynecology
• Endocrinology
• Public Health

Background:

• Gestational diabetes mellitus (GDM) presents significant health risks to mothers and neonates.
• Current early prediction models for GDM lack standardization.
• There is a critical need for reliable tools for early GDM detection in pregnancy.

Purpose of the Study:

• To develop and validate a nomogram-based prediction model for identifying GDM risk in early pregnancy.
• To assess the clinical utility and stability of the developed prediction model.

Main Methods:

• A predictive model for GDM was developed using multivariate logistic regression analysis.
• A nomogram was employed for visualizing the prediction model.
• Data from 1824 pregnant women (training/internal validation) and 1604 (external validation) were utilized.
• Model performance was assessed using ROC curves, calibration curves, and DCA.

Main Results:

• The nomogram incorporated factors including maternal age, BMI, reproductive and family history, creatinine, triglyceride, LDL, neutrophil, and monocyte counts.
• The model demonstrated robust performance with AUC values of 0.804 (training), 0.800 (internal validation), and 0.829 (external validation).
• Calibration and decision curve analyses confirmed the model's accuracy and clinical usefulness.

Conclusions:

• The developed nomogram serves as a stable and effective tool for predicting GDM in early pregnancy.
• This prediction model can significantly aid in the early management of GDM, potentially improving health outcomes.
• The nomogram offers a standardized approach to early GDM risk assessment.