OBJECTIVE To assess the reversibility of the defect in glycogen synthase (GS) activity in skeletal muscle from obese patients with NIDDM treated with a hypocaloric diet and metformin. METHODS Eighteen obese patients newly diagnosed with NIDDM were included in a randomized placebo-controlled double-blind parallel group trial and followed for 3 months. Euglycemic-hyperinsulinemic clamp including indirect calorimetry and biopsy of m. vastus lateralis was performed before and after treatment with a hypocaloric diet plus metformin or placebo. The patients were studied at basal, low, and high insulin concentrations. RESULTS The impaired GS activity in muscle biopsies was not reversed either by acute normalization of glycemia (for 8 h) or by chronic reduction of hyperglycemia by diet plus metformin. In both treatment groups, comparable effects on glycemic control and weight loss were found together with marked insulin suppression of nonesterified fatty acids and increased glucose oxidation. Total glucose disposal at euglycemic-hyperinsulinemic clamp increased significantly in the metformin group by 25% at high insulin level (259 +/- 31 vs. 207 +/- 21 mg x m(-2) x min(-1), P < 0.05). An insignificant increase by 13% was found in the placebo group. There were no significant changes in nonoxidative glucose metabolism. GS activity and glucose utilization showed no significant differences between the two treatment groups when regression coefficients, expressed as incremental changes by increments of insulin, were compared. CONCLUSIONS Defective GS activity in obese NIDDM patients is not secondary to hyperglycemia. Metformin and diet had no significant influence on GS activity. The added effect of metformin to that of a hypocaloric diet in improving insulin-stimulated glucose utilization is marginal when blood glucose reduction is obtained by weight loss.