Glucagon-like peptide 1(7-36) amide (GLP-1) is postulated to be the major physiological incretin in humans, but evidence is indirect. We report the first studies examining the physiological role of GLP-1 in the postprandial state in humans using the GLP-1 antagonist exendin 9-39. Exendin 9-39 completely blocked GLP-1-induced glucose-stimulated insulin release from perifused human islets of Langerhans. In healthy fasted volunteers, intravenous infusion of exendin 9-39 at 500 pmol x kg(-1) x min(-1) in the hyperglycemic state abolished the insulinotropic effect of a physiological dose of GLP-1 and fully reversed the glucose-lowering effect of GLP-1. Nine healthy subjects consumed a 150-g oral glucose tolerance test and were infused with 500 pmol x kg(-1) x min(-1) exendin 9-39 or saline. Exendin 9-39 increased the peak postprandial glucose level (exendin 9-39, 8.67 +/- 0.35 vs. saline, 7.67 +/- 0.35 mmol/l, P < or = 0.005) and increased postprandial plasma glucose incremental area under the curve by 35% (exendin 9-39, 152 +/- 19 vs. saline, 113 +/- 16 mmol x min x l(-1), P < or = 0.05). This could be explained as partly secondary to the blockade of glucose-induced suppression of glucagon and maybe also to an increased rate of gastric emptying. Thus, in humans exendin 9-39 acts as an antagonist of GLP-1 both in vitro and in vivo. When infused alone, exendin 9-39 causes a deterioration in postprandial glycemic control, suggesting that GLP-1 may be important for maintenance of normal postprandial glucose homeostasis in humans.