Insulin and glucose responses to glibenclamide were studied in comparison to a novel non-sulphonylurea drug (AG) by means of the euglycaemic clamp technique. Nine fasting male subjects were connected to a Biostator and 1.75, 3.5 or 7.0 mg glibenclamide or 1.0, 2.0 or 4.0 mg AG were given and blood glucose concentrations were clamped at 10% below basal values. Glucose infusion rates were registered over 10 h after administration of the tablet. Maximal glucose infusion rates after glibenclamide were 40% higher compared to AG (1.75 vs 1.0 mg, 3.5 vs 2.0 mg, 7.0 vs 4.0 mg, respectively) and were reached after 3-3.5 h for all doses. After glibenclamide, area under the glucose infusion curves and maximal incremental serum insulin responses were higher by 25-40% and by 30% compared to AG when low, medium and high doses of each drug were tested. However, a linear dose relationship was obtained for both drugs when the glucose infusion rate was plotted against the area under the insulin curve. In fact, both drugs were equipotent on a molecular weight basis. The hypoglycaemic index of both drugs (integrated glucose infusion rate divided by integrated insulin release) expressed per mumol of drug revealed a dose-dependent and parallel inverse curvilinear relation to increasing doses. This methodological approach allowed us to quantify and compare the metabolic effects of oral hypoglycaemic agents under standardised experimental conditions.