Aims/hypothesis: Gestational diabetes mellitus (GDM) is generally defined based on glycaemia during an OGTT, but aetiologically includes women with defects in insulin secretion, insulin sensitivity or a combination of both. In this observational study, we aimed to determine if underlying pathophysiological defects evaluated as continuous variables predict the risk of important obstetric and neonatal outcomes better than the previously used dichotomised or categorical approaches.
Methods: Using data from blinded OGTTs at mean gestational week 28 from five Hyperglycemia and Adverse Pregnancy Outcome study centres, we estimated insulin secretion (Stumvoll first phase) and sensitivity (Matsuda index) and their product (oral disposition index [DI]) in 6337 untreated women (1090 [17.2%] with GDM as defined by the International Association of Diabetes and Pregnancy Study Groups). Rather than dichotomising these variables (i.e. GDM yes/no) or subtyping by insulin impairment, we related insulin secretion and sensitivity as continuous variables, along with other maternal characteristics, to obstetric and neonatal outcomes using multiple regression and receiver operating characteristic curve analysis.
Results: Stratifying by GDM subtype offered superior prediction to GDM yes/no only for neonatal hyperinsulinaemia and pregnancy-related hypertension. Including the DI and the Matsuda score significantly increased the area under the receiver operating characteristic curve (AUROC) and improved prediction for multiple outcomes (large for gestational age [AUROC 0.632], neonatal adiposity [AUROC 0.630], pregnancy-related hypertension [AUROC 0.669] and neonatal hyperinsulinaemia [AUROC 0.688]). Neonatal hypoglycaemia was poorly predicted by all models. Combining the DI and the Matsuda score with maternal characteristics substantially improved the predictive power of the model for large for gestational age, neonatal adiposity and pregnancy-related hypertension.
Conclusion/interpretation: Continuous measurement of insulin secretion and insulin sensitivity combined with basic clinical variables appeared to be superior to GDM (yes/no) or subtyping by insulin secretion and/or sensitivity impairment in predicting obstetric and neonatal outcomes in a multi-ethnic cohort. Graphical abstract.
Keywords: Caesarean delivery; Gestational diabetes; HAPO study; Insulin secretion; Insulin sensitivity; Large for gestational age; Neonatal outcome; Obstetric outcome; Oral disposition index; Preterm delivery.