Five male non-obese newly diagnosed NIDDM and 5 age-, sex- and body mass index (BMI) matched healthy controls without a family history of diabetes were submitted to a frequently sampled intravenous (i.v.) glucose tolerance test modified by exogenous insulin administration for estimation of insulin sensitivity (SI) and glucose-mediated glucose disposal (SG) with Bergman's minimal model computer analysis of glucose kinetics. The tests were repeated after 3 months treatment with a second generation sulfonylurea, gliclazide, in the diabetics subjects. SI and SG were markedly reduced before gliclazide therapy in the diabetics in comparison to the paired controls. After gliclazide, despite significantly lower (almost normal) plasma glucose, normalization of glycosylated hemoglobin and increased fasting insulin levels, there was a slight but significant increase in SI while SG showed a further reduction, the improvement in glucose control being also associated to the significant increased first and 2nd phase insulin release for the first 20 min after glucose infusion.