Gestational diabetes (GDM) is a frequent complication during pregnancy. We aimed to develop a score to predict future insulin dependency in gestational diabetes (GDM). Data from 1611 patients from Charite Berlins gestational diabetes clinic from 2015 to 2022 were utilized. A stepwise backwards regression, including patient characteristics obtained at the initial presentation, was performed. Predictors examined included age, fasting blood glucose level, blood glucose levels one and two hours after oral glucose tolerance test, pre-pregnancy BMI, number of previous pregnancies and births, and fetal sex. The ideal cutoff value between high and low risk for insulin dependency was assessed and the score was internally validated. There were 1249 (77.5%) women diagnosed with dietary GDM and 362 (22.5%) were diagnosed with insulin-dependent GDM. The CHarite AssessmeNt of GEstational Diabetes (CHANGED) Score achieved an area under the curve of 0.77 (95% confidence interval 0.75-0.80; 0.75 in internal validation). The optimal cutoff value was calculated at a score value of 9 (72% sensitivity, 69% specificity). We developed an easily applicable tool to accurately predict insulin dependency in gestational diabetes. The CHANGED Score is routinely available and can potentially improve maternal and fetal outcomes.
