The factors responsible for prevention of duodenal mucosal injury are not known. This series of experiments was performed to determine whether the human duodenum secretes bicarbonate that could prevent mucosal damage. To isolate a 4-cm segment of proximal (i.e., the duodenal bulb) or distal duodenum free of contamination from either gastric or pancreaticobiliary secretion, or both, methods were developed using occlusive balloons. The test segment was perfused with NaCl (2 ml/min, 37 degrees C) containing [14C]PEG as a nonabsorbable marker, and bicarbonate output was quantitated. Mean (+/- SE) basal proximal duodenal bicarbonate output was 143 +/- 17 mumol/cm X h. A 5-min infusion of 25, 50, and 100 mM HCl directly into the isolated proximal duodenal test segment increased bicarbonate output to 167 +/- 29, 199 +/- 19, and 278 +/- 49 mumol/cm X h, respectively, during the hour after acidification. Distal duodenal acidification (25, 50, and 100 mM) also increased bicarbonate output from the isolated proximal duodenal test segment. A synthetic prostaglandin E1 analogue, misoprostol (1.67-13.3 micrograms/min), infused directly into proximal or distal test segments significantly stimulated bicarbonate outbreak; peak responses were 644 +/- 35 mumol/cm X h and 171 +/- 20 mumol/cm X h (p less than 0.001), respectively. Thus, in humans, the proximal and distal duodenal mucosa secretes bicarbonate at rest; direct acidification of the proximal duodenum stimulates bicarbonate output; acidification of the distal duodenum beyond the isolated test segment also increased proximal duodenal bicarbonate output; and a synthetic prostaglandin E1 analogue stimulated both proximal and distal bicarbonate output; however, distal duodenal bicarbonate output was significantly less, indicating a proximal-to-distal gradient in bicarbonate secretion.