Low doses of metformin (500 mg twice daily) were administered to 20 diabetic patients, combined with the original sulfonylurea treatment which had become ineffective even at full dosage. After 1 and 5 weeks, the effects of the drug on glycemic control, blood intermediate metabolites and monocyte insulin receptors were monitored. Metformin clearly improved glycemic control by reducing both fasting blood glucose from 189.88 +/- 21.11 mg/dl to 131.12 +/- 16.02 mg/dl after 1 week and to 130.11 +/- 13.29 mg/dl after 5 weeks (p less than 0.025 both after 1 and 5 weeks); the diurnal blood glucose average fell from 235.33 +/- 24.11 mg/dl to 174.66 +/- 23.45 mg/dl (p less than 0.0025) after 1 week and to 177.65 +/- 21.71 mg/dl (p less than 0.0005) after 5 weeks. Consequently both blood glycosylated hemoglobin (p = n.s. after 1 week, p less than 0.025 after 5 weeks) and serum fructosamine (p less than 0.0025 after both 1 and 5 weeks) also decreased after metformin treatment. No change in plasma insulin and C-peptide levels was reported and no modification in diurnal rhythms of blood lactate, pyruvate, alanine glycerol and beta-OH-butyrate was detected at any time during metformin treatment. All the changes documented in the binding values were already complete at the end of the first week; insulin binding to monocytes increased slightly but significantly (p less than 0.05) and the number of receptors per cell rose (p less than 0.05) but could not be correlated to any index of glycemic control. These data suggest that the antidiabetic action of metformin is neither related to its lactate-increasing activity nor does it depend upon its inducing an increase in insulin binding values. This metformin-related hypoglycemic effect might be the result, at least in part, of a reduced oxidative phosphorylation without inhibition of hepatic gluconeogenesis and/or of decreased hepatic glucose output. Moreover, our data are also consistent with the hypothesis that metformin might affect insulin action at a post-receptor level.