Troglitazone is effective in approximately 50% in patients with type 2 diabetes (NIDDM). In this study, we investigated the relations between serum leptin levels and clinical efficacy of troglitazone. Forty-five type 2 diabetic patients (23 men and 22 women) from our outpatient clinic were treated with troglitazone 400 mg daily for 12 weeks. Fasting plasma glucose (FPG), HbA1c, body weight, serum insulin and leptin concentrations were measured before and after troglitazone treatment. After 12 weeks of troglitazone treatment, FPG (before versus after, 179+/-33 vs. 138+/-26 mg/dl, mean+/-SD), HbA1c (7.8+/-1.3 vs. 6.9+/-1.0%), IRI (8.3+/-4.3 vs. 6.3+/-3.4 microU/ml) and HOMA-R index (homeostasis model assessment of insulin resistance) (3.8+/-2.4 vs. 2.2+/-1.3) decreased significantly, while body mass index (BMI) slightly increased (26.3+/-3.5 vs. 26.6+/- 3.6 kg/m(2)), and serum leptin remained unchanged (8.5+/-7.2 vs. 9.1+/-8.7 ng/ml). Reduction in FPG (DeltaFPG) after troglitazone treatment were correlated with reduction in HOMA-R (DeltaHOMA-R) (r=0.721, P<0.0001). DeltaFPG was correlated with serum leptin (r=0.441, P<0.01), HOMA-R (r=0.460, P<0.01) and FPG (r=-0.781, P<0.0001) at baseline, but not with BMI and serum IRI at baseline. Furthermore, serum leptin at baseline was significantly correlated with DeltaHOMA-R (r=0.634, P<0.01). Leptin concentration before treatment therefore, can be used as an predictor for clinical efficacy of troglitazone in patients with type 2 diabetes.