Evidence has accumulated that nonsteroidal anti-inflammatory drugs (NSAIDs) cause clinically important gastroduodenal ulcers. The pathogenesis, which involves the impairment of mucosal resistance to injury in an acid-peptic environment, is multifactorial and controversial. Ulcers caused by NSAIDs can occur either in mucosa inflamed because of infection with Helicobacter pylori or in histologically normal mucosa. The use of these drugs has been linked to an unexpectedly high incidence of ulcer complications, and a history of peptic ulcer disease is common in such cases. Nonsteroidal anti-inflammatory drugs thus appear both to exacerbate an underlying peptic diathesis and to cause de novo ulcers. The association between the use of these drugs and ulcer complications is supported by ulcer prevalence data from cross-sectional studies, and by data from case-controlled and cohort studies, and from randomized, experimental trials. Drug-induced gastric ulcers have been prevented by misoprostol, but not by H2 blocker therapy. Several therapies have been reported to promote ulcer healing despite continued use of NSAIDs, but adequate controlled trials have not been done. Small gastric and duodenal ulcers readily heal, whereas larger gastric ulcers require vigorous and prolonged therapy. The relative efficacies of various therapies in preventing ulcers, healing ulcers, or preventing complications remain to be established.