Nonsteroidal anti-inflammatory drugs (NSAIDs) have been associated with an increase in upper gastrointestinal complications. There is no agreement, however, on whether all conventional NSAIDs have a similar relative risk (RR), and epidemiologic data are limited on acetaminophen. We studied the association between these medications and the risk of upper gastrointestinal bleed/perforation in a population-based cohort of 958,397 persons in the United Kingdom between 1993 and 1998. Our nested case-control analysis included 2,105 cases and 11,500 controls. RR estimates were adjusted for several factors known to be associated with upper gastrointestinal bleed/perforation. Compared with non-users, users of acetaminophen at doses less than 2 gm did not have an increased risk of upper gastrointestinal complications. The adjusted RR for acetaminophen at doses greater than 2 gm was 3.6 [95% confidence interval (95% CI) = 2.6-5.1]. The corresponding RRs for low/medium and high doses of NSAIDs were 2.4 (95% CI = 1.9-3.1) and 4.9 (95% CI = 4.1-5.8). The RR was 3.1 (95% CI = 2.5, 3.8) for short plasma half-life, 4.5 (95% CI = 3.5-5.9) for long half-life, and 5.4 (95% CI = 4.0-7.1) for slow-release formulations of NSAIDs. After adjusting for daily dose, the differences in RR between individual NSAIDs tended to diminish except for apazone. Users of H2 receptor antagonists, omeprazole, and misoprostol had RRs of 1.4 (95% CI = 1.2-1.8), 0.6 (95% CI = 0.4-0.9), and 0.6 (95% CI = 0.4-1.0), respectively. Among NSAID users, use of nitrates was associated with an RR of 0.6 (95% CI = 0.4-1.0).