To analyze the effects of metformin (M) on the kinetics and pathways of glucose utilization after glucose ingestion, nine non-insulin-dependent diabetes mellitus (NIDDM) patients underwent two 5-hour oral glucose tolerance tests (OGTTs) preceded in random order by a 3-week treatment with either M (850 mg twice per day) or placebo. Each test included intravenous infusion of 3-3H-glucose and labeling of the oral dose (75 g) with 1-14C-glucose, with measurements of glucose kinetics, glycolytic flux (3H2O production), and glucose oxidation (indirect calorimetry and expired 14CO2). Basal glycemia was decreased by M (6.6 v 8.2 mmol/L, P < .01) with no changes in insulin levels, with the hypoglycemic effect correlating strongly (P < .001) with a decrease in glucose production. Mean 0- to 5-hour postprandial glycemia was also decreased by the drug (9.9 v 12.2 mmol/L, P < .04), lactate concentration was increased (1.79 v 1.44 mmol/L, P < .01), and absolute insulin levels were increased, but not to a significant extent. The rates of appearance (Ra) of exogenous and endogenous glucose were not modified, and the hypoglycemic effect of M in the postprandial state was entirely related to an increase in systemic glucose disposal (85.1 v 77.5 g/5 h, P < .001). Carbohydrate oxidation was unchanged, and glycolytic flux and nonoxidative glycolysis were increased by approximately 13 g/5 h (P < .01), with the excess lactate produced probably being converted to glycogen in the liver. Whole-body glycogen synthesis through the direct pathway tended to be reduced (-8 g/5 h, P > .05). Thus, M decreases postprandial glycemia by increasing glucose disposal and stimulates lactate production. The data also suggest that the drug increases the proportions of glycogen deposited through the indirect rather than the direct pathway.