Prepregnancy Weight and Antepartum Insulin Secretion Predict Glucose Tolerance Five Years After Gestational Diabetes Mellitus

Abstract

OBJECTIVE: To identify phenotypic, genotypic, and metabolic parameters measured at the time of antepartum diagnosis of gestational diabetes mellitus that can indicate the risk of diabetes mellitus at early postpartum (≤6 mo after delivery) and at a 5-yr follow-up. RESEARCH DESIGN AND METHODS: The recommendations from the National Diabetes Data Group and International Workshop Conferences on Gestational Diabetes Mellitus were used for screening, diagnosing, and subclassifying gestational diabetes mellitus. National Diabetes Data Group criteria were also used for classification of glucose tolerance postpartum. Plasma glucose, insulin, and free fatty acids were measured after an overnight fast. Plasma glucose and insulin were measured 15, 30, 60, 120, and 180 min after the 100-g oral glucose load. Postpartum glucose tolerance was evaluated at 3–6 mo (early), 1 yr, and annually thereafter. RESULTS: The 5-yr cumulative incidence of diabetes during follow-up after gestational diabetes mellitus was nearly 50%. Among those who had diabetes within 5 yr, a history of diabetes in only the mother was nearly threefold more common than a history of diabetes in only the father (30 vs. 11%, P < 0.01). Those who displayed diabetes at early postpartum (≤6 mo) testing had significantly higher antepartum glucose levels at 60, 120, and 180 min compared with those whose early postpartum results were normal. They were also relatively insulinopenic at antepartum testing. Their fasting, acutely stimulated (15 and 30 min), and integrated 3-h response to oral glucose were all significantly lower relative to women who remained normal or had impaired glucose tolerance at early postpartum testing. Women who developed diabetes between 6 mo and 5 yr postpartum were more obese before the index pregnancy, and they had lower fasting, acutely stimulated (15 and 30 min), and integrated (1–3 h) insulin levels compared with women who remained normal or displayed impaired glucose tolerance at 5 yr postpartum. A multiple logistic regression model showed that diabetes present at early postpartum testing was independently associated with higher 2-h glucose and lower basal and total integrated insulin level. Later (≤6 mo-5 yr postpartum) development of diabetes was independently associated with prepregnancy weight and impaired insulin secretion at diagnosis of gestational diabetes mellitus. CONCLUSIONS: Impaired β-cell function and obesity at diagnosis of GDM were associated with the development of diabetes during a 5-yr, follow-up period. Studies designed to prevent diabetes in this high-risk group should examine strategies to maintain both optimal β-cell function and maximum insulin sensitivity.