Non-steroidal anti-inflammatory drugs (NSAIDs) are widely used and some 24 million prescriptions are issued for them each year in the United Kingdom. Although animal and human experimental studies can yield confusing and disparate findings, complaints of gastro-intestinal side effects are common, and spontaneous reports of serious gastro-intestinal adverse effects of NSAIDs form by far the most commonly reported variety to the United Kingdom's Committee on the Safety of Medicines. Experimental suggestions that non-aspirin non-steroidal anti-inflammatory drugs (NSAIDs) cause gastritis and erosions have been overshadowed by evidence that takers of NSAIDs tend to develop serious complications with acute bleeding and perforations of duodenal and gastric ulcers. Findings in clinical case series that a third or more of patients with bleeding or perforated ulcers have been taking non-aspirin NSAIDs contrast with the outcome of surveillance studies where, in sets of 10,000 takers of such drugs, there appeared to be no material differences in ulcer complication rates in takers and non-takers. Data obtained in formal case control studies however, have suggested a relative risk which is increased by between two- and four-fold in non-aspirin NSAID takers for both gastric and duodenal ulcer in elderly people. The findings in case-control and surveillance studies can be reconciled by taking account of the very large numbers of prescriptions for non-aspirin NSAIDs issued each year, now approaching 24 million in the United Kingdom.(ABSTRACT TRUNCATED AT 250 WORDS)