An extrapancreatic effect of sulfonylureas has been postulated. However, in vivo results have been disputed because the amelioration of insulin action that follows sulfonylurea may represent the relief from glucose toxicity rather than a direct effect of the drug. Therefore, we studied the hypoglycemic action of gliclazide acutely and after 2 months of therapy in seven type 2 diabetic patients. All patients received a 240-minute i.v. glucose infusion with [3-3H]glucose. In a random order, 160 mg gliclazide (study 1) or placebo (study 2) was given orally before glucose infusion. Finally, the effect of 160 mg gliclazide was reassessed after a two-month treatment with the same sulfonylurea (80 mg t.i.d.). Basal plasma glucose, insulin, C-peptide and endogenous glucose production (EGP) were similar before the two initial studies. During glucose infusion, EGP was more suppressed after gliclazide in spite of comparable increase in plasma insulin and C-peptide. After the two-month therapy, basal plasma glucose levels and HbA1c were lower while plasma insulin and C-peptide were higher with respect to baseline (p < 0.05). Gliclazide further reduced plasma glucose, the incremental area above baseline, and EGP during glucose infusion, while plasma insulin and C-peptide achieved higher plateaus (p < 0.05). When data were pooled, plasma glucose concentration and EGP correlated both in the basal state (r = 0.71) and during the last hour of glucose infusion (r = 0.84; both p < 0.05). These data suggest that gliclazide enhances the suppression of EGP induced by insulin and that this effect is greater with chronic treatment because of concomitant improvement of insulin secretion.