Metformin has been demonstrated to lower blood glucose in vivo by a mechanism which increases peripheral glucose uptake. Furthermore, the therapeutic concentration of metformin has been estimated to be in the order of 0.01 mmol/l. We investigated the effect of metformin on insulin-stimulated 3-0-methylglucose transport in isolated skeletal muscle obtained from seven patients with non-insulin-dependent diabetes mellitus (NIDDM) and from eight healthy subjects. Whole body insulin-mediated glucose utilization was decreased by 45% (p < 0.05) in the diabetic subjects when studied at 8 mmol/l glucose, compared to the healthy subjects studied at 5 mmol/l glucose. Metformin, at concentrations of 0.1 and 0.01 mmol/l, had no effect on basal or insulin-stimulated (100 microU/ml) glucose transport in muscle strips from either of the groups. However, the two control subjects and three patients with NIDDM which displayed a low rate of insulin-mediated glucose utilization (< 20 mumol.kg-1.min-1), as well as in vitro insulin resistance, demonstrated increased insulin-stimulated glucose transport in the presence of metformin at 0.1 mmol/l (p < 0.05). In conclusion, the concentration of metformin resulting in a potentiating effect on insulin-stimulated glucose transport in insulin-resistant human skeletal muscle is 10-fold higher than the therapeutic concentrations administered to patients with NIDDM. Thus, it is conceivable that the hypoglycaemic effect of metformin in vivo may be due to an accumulation of the drug in the extracellular space of skeletal muscle, or to an effect of the drug distal to the glucose transport step.