Objective: To develop a predictive model to identify women with recent gestational diabetes (GDM) most likely to progress to impaired glucose tolerance postpartum. Design: Observational study. Setting: Academic medical center in the United States. Population: Postpartum women with recent GDM, defined by Carpenter-Coustan criteria & 1-year postpartum HbA1c assessment. Methods: We used lasso regression with k-fold cross validation to develop a multivariable model to predict progression to impaired glucose tolerance, defined as HbA1c ≥ 5.7%, by 1 year postpartum. Predictive ability was assessed by the area under the curve, sensitivity, specificity, positive and negative predictive values. Main Outcome Measures: Impaired glucose tolerance. Results: Of 203 women, 71(35%) had impaired glucose tolerance at 1 year postpartum. The final model had an AUC of 0.81 (95% CI 0.74, 0.87) and included eight indicators of weight, body mass index, Hispanic ethnicity, GDM in a prior pregnancy, GDM diagnosis < 24 weeks’ gestation, and fasting and 2-hour plasma glucose at 2 days postpartum. A cut-point of ≥ 0.24 predicted probability had sensitivity 80% (95% CI 69, 89), specificity 58% (95% CI 49, 66), PPV 57% (95% CI 46, 68) and NPV 83% (95% CI 74, 89) to identify women with impaired glucose tolerance at 1 year postpartum. Conclusions: Our predictive model had reasonable ability to predict impaired glucose tolerance around delivery for women with recent GDM. Funding: National Institute of Mental Health and American Diabetes Association. Keywords: gestational diabetes, impaired glucose tolerance, type 2 diabetes prevention; predictive model
