We used a large computerized data base to study the relationship between exposure to nonsteroidal anti-inflammatory drugs (NSAIDs), anticoagulants, and corticosteroids and development of complications of peptic ulcer disease. In addition to contributing to the information on the etiology of ulcer disease, these studies offer some clues regarding the phenomenon of mucosal protection. Our series of studies indicate that mucosal adaptation to NSAIDs is of little clinical importance, that NSAIDs are associated with a strong dose-response effect that is probably more important than any between-drug differences, and that there are no "safe" NSAIDs. In addition, our data suggest that the effects of NSAIDs are mediated through mechanisms other than merely their antiplatelet effect, as they cause a different spectrum of disease than anticoagulants, and that the ulcerogenic effects of NSAIDs and corticosteroids are probably mediated through separate mechanisms as well. The effects of these drugs are synergistic, such that the combination of NSAIDs with anticoagulants or corticosteroids puts patients at a very high risk for serious ulcer disease.