A stepped intravenous metformin infusion was used in conjunction with the hyperglycaemic clamp technique to study the dose-response relationship of plasma metformin concentration with hepatic glucose production and peripheral glucose disposal in nine patients with Type 2 diabetes. The study was of double-blind crossover design, using NaCl infusion as control. Plasma metformin concentrations spanning the therapeutic range (1.64 +/- 0.13 mg l-1 and 6.57 +/- 0.61 mg l-1) were achieved. No differences in peripheral glucose disposal were demonstrated when compared with NaCl infusion (3.4 +/- 0.1 vs 3.6 +/- 0.2 (+/- SE) mg kg-1 min-1 and 3.4 +/- 0.2 vs 3.3 +/- 0.2 mg kg-1 min-1, respectively). There was also no difference in basal hepatic glucose production during metformin and NaCl infusion (2.7 +/- 0.3 vs 2.8 +/- 0.2 mg kg-1 min-1). No acute effect of metformin on hepatic glucose production or peripheral glucose disposal was observed, implying that a chronic persistent effect is more important in these respects than immediate effects consequent upon changes in plasma drug level.