The effects of plain and enteric-coated fenoprofen calcium (Nalfon, Dista, Indianapolis, Ind.) on gastrointestinal microbleeding were studied in 32 normal male volunteers in a randomized, open-label, parallel trial at two inpatient research facilities. A 1-week placebo (baseline) period preceded 2 weeks of fenoprofen therapy (enteric coated or plain, 600 mg q.i.d.). Fecal blood loss was measured by 51Cr-tagged erythrocyte assay and averaged over days 4 to 7 (baseline) and 11 to 14 and 18 to 21 (active therapy). At one center gastrointestinal irritation was evaluated endoscopically before and after active therapy. Endoscopy showed both formulations to cause mucosal damage not evident by subject-reported symptoms. Four of the 16 subjects developed asymptomatic duodenal ulcers. Mean daily fecal blood loss was significantly lower (P = 0.03) with enteric-coated (mean +/- SD, 1.104 +/- 0.961 ml/day) than with plain fenoprofen calcium (mean +/- SD, 1.686 +/- 0.858 ml/day), suggesting that tolerance of fenoprofen can be improved with administration in an enteric-coated form.