African Americans (AA) have greater prevalence of type 2 diabetes mellitus (DM), and nondiabetic AA have demonstrated increased insulin resistance when compared with Caucasian Americans (CA). The objective of this study was to examine the impact of chronic use of an insulin sensitizer on glucose metabolism in normal glucose tolerant AA at risk for DM (previous gestational diabetes mellitus [GDM] or first-degree relative with DM). Forty-nine high-risk AA received 200 mg/d troglitazone (TRO) versus 81 age-, weight-, and body mass index (BMI)-matched high-risk AA who received placebo (PLA) for 24 months. Yearly anthropometric measurements, oral glucose tolerance test (OGTT) and frequently sampled intravenous glucose tolerance test (FSIVGTT) were performed. Biochemical parameters were monitored quarterly. There was no significant change in anthropometric measurements over 24 months in TRO versus PLA. There were no significant differences in serum glucose, insulin, or C-peptide incremental area under the curve (AUC) in TRO versus PLA at baseline or 24 months for OGTT and FSIVGTT. The insulin sensitivity (S(I)) for TRO and PLA increased from baseline to 24 months by 17% and 16%, respectively. The TRO demonstrated a 26% increase in insulin/glucose ratio versus 1% increase in the PLA at 24 months. The disposition index (DI) increased 33% from baseline in TRO versus 21% increase in PLA. Modest improvement in glucose metabolism was seen in TRO when compared with PLA. TRO was well tolerated without significant reported adverse events. Based on our current data, the treatment of normal glucose tolerant high-risk AA with thiazolidinedione (TZD) may be beneficial to "reset" and protect glucose metabolism by improving insulin responses. Because of the potential drug-related risks associated with use of TZD and the proven positive impact of diet and exercise in prevention of DM, studies of longer duration with examination of other potentially beneficial parameters, such as cardiovascular indices and inflammatory markers will be necessary to justify the cost in the nondiabetic population.