452 Β-Cell UNLOADING AND THE RISK OF DIABETES IN THE PIOGLITAZONE IN PREVENTION OF DIABETES STUDY

Abstract

In the TRIPOD study, troglitazone` treatment of Hispanic women with prior gestational diabetes (GDM) reduced the incidence of type 2 diabetes by 55%, from 12.1% to 5.4% per year in a manner that was closely related to reduction of endogenous insulin requirements effected by amelioration of insulin resistance. Women who completed the TRIPOD study and post-trial drug washout were asked to participate in the Pioglitazone in Prevention of Diabetes (PIPOD) study if their HbA1C was ≤7%. Women were placed on pioglitazone, 45 mg/d, for a scheduled three years plus six months off-drug washout. Fasting glucose levels were measured at three month intervals during drug treatment and 6-month post-trial drug washout. Oral glucose tolerance tests were performed annually and at the end of the post-trial washout. Intravenous glucose tolerance tests (IVGTTs) performed at baseline and one year later were used to assess the initial impact of pioglitazone on insulin resistance and insulin responses. Data for this report were analyzed to assess the incidence of type 2 diabetes and its relationship to initial β-cell unloading in the 86 women who entered PIPOD without diabetes. The annual diabetes incidence rate in those women during a median follow-up of 3.4 years was 4.6%. Seventy of the women had an IVGTT at baseline and one year later. Mean initial β-cell unloading, calculated as the fractional change in total IVGTT insulin area between baseline and year 1, was 26%. Based on the relationship between initial β-cell unloading and diabetes incidence in TRIPOD, this initial loading in PIPOD predicts a 5.3% annual diabetes risk. The observed rate in the 70 women with baseline and 1-year IVGTTs was 5.4%/yr. Annual diabetes rates in tertiles of the cohort defined by initial β-cell unloading were 6.5%/yr, 3.8%/yr and 1.1%/yr for (smallest to largest unloading). Thus, treatment with pioglitazone in Hispanic women with prior GDM was associated with a relatively low annual diabetes rate that tracked closely with the degree of initial β-cell unloading induced by the drug.