The insulin-antagonistic effect of adrenaline was studied in seven healthy subjects with the euglycaemic clamp technique using two insulin infusion rates (40 and 1200 mU X (m2)-1 min-1). The adrenergic receptor mediating the adrenaline effect was characterized by concomitant infusion of propranolol (beta 1 + beta 2-antagonist) or metoprolol (beta 1-antagonist). Each subject was studied four times (placebo, adrenaline, adrenaline + propranolol, adrenaline + metoprolol). Glucose turnover was measured with D(3-3H)-glucose. Similar plasma insulin levels were reached in all studies with the two insulin infusion rates (mean; placebo 51 +/- 3 and 7421 +/- 337 mU/l respectively). Glucose production was completely inhibited by the low insulin level during placebo infusion. Adrenaline antagonized this effect so that a significant glucose production was seen at the low but not at the high insulin level. Propranolol, but not metoprolol, reversed this insulin-antagonistic effect of adrenaline. Glucose utilization increased from 2.53 +/- 0.17 to 7.28 +/- 0.88 mg X kg-1 X min-1 during placebo when the insulin levels were increased from 4 +/- 0.3 to 51 +/- 3 mU/l. Increasing the insulin levels 150-fold to approximately 7500 mU/l only doubled the glucose utilization (14.68 +/- 1.14 mg X kg-1 X min-1). Adrenaline induced a pronounced inhibition of glucose utilization at both insulin levels (78% and 37% inhibition respectively). Propranolol, but not metoprolol, prevented this effect of adrenaline. Thus, physiological adrenaline levels exert a pronounced insulin-antagonistic effect which is mediated by beta 2-receptor stimulation. The inhibitory effect on glucose uptake is maintained even at high insulin levels when hepatic glucose production is completely abolished.