African-American women have been shown to be more insulin-resistant than age- and weight-matched Caucasian women, but the reasons for this difference are unclear. The purpose of the present study was to determine whether experimental manipulation of dietary fat intake has differential effects by race on insulin sensitivity (S(I)) in 20 African-American and 11 Caucasian women. Additionally, leptin levels before and after 3 weeks of an isocaloric high-fat ([HF] 50% fat, 35% carbohydrate, and 15% protein) or low-fat ([LF] 20% fat, 55% carbohydrate, and 15% protein) diet were compared. African-American and Caucasian women did not differ significantly in the body mass index (BMI) or percentage body fat at baseline. S(I) (adjusted for BMI) decreased on the HF diet and increased on the LF diet in both races combined relative to the baseline control (control, 2.42 +/- 0.22; HF, 2.29 +/- 0.22; LF, 2.75 +/- 0.21 x 10(-4) min(-1)/microU x mL; main effect of diet, P = .04). There was a 6% decrease in S(I) on the HF diet compared with the control in women of both races, while the LF diet increased S(I) by 6% in African-American and 20% in Caucasian women. Leptin levels increased by 14% on the HF versus control diet in African-Americans (35.2 +/- 3.0 v 30.8 +/- 3.0 ng/mL, P < .01), but did not change with diet in Caucasian women. Glucose and insulin administration had no effect on leptin levels. We conclude that a HF diet consumed over several weeks reduces S(I) in healthy women of both races; however, the magnitude of increase in S(I) on a LF diet is greater in Caucasian women. The HF diet significantly increased leptin levels in African-American women, although there were no other influences of diet, insulin, or race on serum leptin.