1239-P: Utilizing Continuous Glucose Monitoring for Early Detection of Gestational Diabetes Mellitus in a Multiethnic Asian Population

Introduction & Objective: Continuous Glucose Monitoring (CGM) proves reliable glycemic monitoring in subjects with gestational diabetes mellitus (GDM) during pregnancy. Therefore, we further explored the early GDM prediction potential of CGM in the first trimester.

Methods: We recruited 113 overweight or obese multi-ethnic Asian pregnant women (BMI ≥ 23 kg/m²) from a hospital-based, prospective cohort. Blinded CGM devices were worn for 10-14 days at gestation week 11-13 and universal GDM screening was performed at 24-28 week of gestation. EasyGV software was used to derive CGM parameters, including average glucose and glycemic variability parameters (e.g., liability index [LI], mean amplitude of glycemic excursions [MAGE], mean of daily differences [MODD], J-index, % in coefficient variability [% CV]) were assessed between GDM and non-GDM subjects. Forward stepwise logistic regression among 50 multiple imputation datasets identified parsimonious models for traditional risk factors and novel CGM parameters. We compared the predictive value of GDM using both sets of variables.

Results: Twenty-one GDM cases (18.6%) were ascertained using IADPSG criteria. Mothers with GDM had significantly higher values in mean glucose, time-above-range duration and all glycemic variability parameters mentioned above (all p<0.001). The novel CGM prediction Model (average glucose, LI & J-index) demonstrated better sensitivity (0.762 vs. 0.600), specificity (0.859 vs. 0.815), AUC (0.870 vs. 0.665), and superior R² (0.332 vs. 0.058, p=0.021) for incident GDM prediction, compared with the traditional risk model (maternal age, baseline BMI & baseline systolic blood pressure).

Conclusion: Our pilot data suggested CGM's potential clinical utility in early pregnancy for predicting GDM, particularly in individuals at risk due to BMI concerns.