A prospective endoscopic study was carried out in 65 patients with rheumatoid arthritis to assess the prevalence of gastroduodenal lesions on aspirin alone or aspirin plus another anti-inflammatory drug (n = 26). All patients were taking greater than or equal to 8 tablets aspirin/day for greater than or equal to 3 months. Drug therapy consisted of regular, buffered or enteric-coated aspirin +/- one other non-steroidal anti-inflammatory drug or less than or equal to 10 mg prednisone/day. Endoscopic findings were graded as normal, or gastric and/or duodenal erythema, erosions, or ulcer; only the most serious lesion was tabulated. Fifty-three percent of patients on aspirin alone has a gastric ulcer or erosions, and 13% duodenal ulcer or erosions compared to 35% with gastric lesions and 4% with duodenal lesions on aspirin plus a second drug. Twenty-two percent of patients taking regular aspirin had gastric ulcers compared to only 6% taking enteric coated. Patients on enteric coated aspirin and prednisone had only a 10% prevalence of severe gastro-duodenal lesions. In conclusion, the combination of a second anti-inflammatory drug and aspirin therapy did not result in a higher prevalence of gastro-duodenal damage over that produced by aspirin therapy alone. Enteric coated aspirin produced significantly fewer serious lesions than regular or buffered aspirin. The combination of enteric-coated aspirin plus low dose prednisone caused a low prevalence of severe gastro-duodenal lesions.