One hundred and fifty patients with various rheumatic complaints who were taking nonsteroidal anti-inflammatory drugs (NSAIDs) and required upper gastrointestinal endoscopy were compared with 150 patients, matched for age and sex, who were also referred for gastroscopy but who did not have rheumatic complaints and were not taking these drugs. A comparison was made between the indications for endoscopy and the endoscopic findings in the two groups. Significant differences were found. Fewer patients with rheumatic complaints had normal findings, more had chronic gastric ulcer, duodenal ulcer, gastritis or mucosal erosions. Patients with rheumatic complaints were referred more frequently for anaemia and less frequently for dysphagia or abdominal symptoms. The increased severity of gastric morbidity in patients with rheumatism is probably due to damage caused by NSAIDs. The frequency and severity of upper gastrointestinal lesions in patients with rheumatic complaints taking NSAIDs cannot be estimated from the patient's complaint rate, anaemia rather than pain being the most frequent finding. Upper gastrointestinal tract endoscopy is required to make a definitive diagnosis. This has important implications for the licensing policies of regulatory authorities when considering the licensing of new nonsteroidal anti-inflammatory drugs.