Visual Abstract
A 4-12 week postpartum oral glucose tolerance test (GTT) is recommended for individuals with gestational diabetes mellitus (GDM) to identify those most likely to develop prediabetes and diabetes mellitus, but few people complete this testing. Thus we developed a clinical model at delivery to predict abnormal glucose metabolism at 1 year postpartum (HbA1c ≥ 5.7%) and compared it to the 4-12 week post-partum GTT in a prospectively recruited cohort of women with GDM who were followed for 1 year post-partum (n=203). Using multivariable Lasso regression, we investigated predictor variables available at or before the time of hospital discharge that had biologic plausibility for predicting subsequent abnormal glucose metabolism. We compared the receiver operating curve (AUC), sensitivity, specificity, positive and negative predictive values (PPV, NPV) between the Lasso multivariable model and the 4-12 week GTT to predict HbA1c ≥5.7% at 1 year postpartum. Nine variables were included in the Lasso model achieving an AUC and sensitivity that was better than the GTT at 4-12 weeks (Table). While additional work is needed to validate the model, it appears to better identify individuals at risk for abnormal glucose metabolism at 1 year postpartum than the 4-12 week postpartum GTT.
E. Werner: None. A. Espinosa dice: None. A. Bengtson: None.
American Diabetes Association (1-16-ICTS-118 to E.W.)
