353-OR: Comparison of Methods for Classifying Heterogeneity in Gestational Diabetes (GDM) and Association with Outcomes

Abstract

An underlying principle of the IADPSG GDM classification is that women at similar risk of adverse pregnancy outcomes related to hyperglycemia should be classified similarly. Powe et al have described heterogeneous GDM phenotypes related to impaired insulin secretion and/or insulin sensitivity and reported that pregnancy complications relate primarily to impaired insulin sensitivity. However, their Gen3G cohort was largely Caucasian and GDM women received treatment. Our study was designed to further examine this hypothesis in a diverse population of 6128 untreated women from 5 centers in the blinded epidemiologic HAPO study. Following similar methodology (Impairment = parameter value <25th centile for non GDMs), women with IADPSG GDM were classified as (1) GDMsecr - impaired insulin secretion - using either HOMA β (Hβ) or Insulinogenic index (II); (2) GDMsens - impaired insulin sensitivity - using Matsuda index; (3) Both 1 and 2; (4) GDMND - No Defect. Analyses compared neonatal outcome frequencies (LGA, excess neonatal (NN) adiposity, NN hypoglycemia, NN hyperinsulinemia) with non GDM women and between GDM groups using χ2 tests and logistic regression analyses adjusted for multiple confounders including maternal age, height, smoking, FH of diabetes, BMI at the OGTT visit, ethnicity and HAPO study center. Hβ and II gave divergent results for GDMsecr, with only 25% concordance. GDMND women (10% of all GDM by Hβ, 6% by II) showed outcome frequencies similar to non GDM women. NN hypoglycemia was unrelated to GDM group. In all adjusted analyses, using either Hβ or II, Groups 1-3 showed higher odds (aORs 1.6-3.1; p < 0.01 vs. non GDM) for LGA, NN adiposity and NN hyperinsulinemia. Although Hβ and II identify different women as having defects in insulin secretion, both may contribute to detection of a lower risk group of GDMND women. Beyond this, the division of women according to insulin secretion and insulin sensitivity status did not permit additional risk stratification. Disclosure D. McIntyre: None. L.R. Madsen: None. K.S. Gibbons: None. D.A. Sacks: None. R.C. Ma: Advisory Panel; Self; Boehringer Ingelheim International GmbH. Research Support; Self; AstraZeneca, Bayer AG, Pfizer Inc. Stock/Shareholder; Self; GemVCare. Other Relationship; Self; AstraZeneca. W.H. Tam: None. P. Catalano: None.