ABSTRACT
A number of new nonsteroidal antiinflammatory drugs (NSAIDs) have been marketed in the past. The major motivation for the use of these drugs is the known gastrointestinal toxicity of aspirin. From pre-marketing studies it was known that these drugs, as well, could cause subclinical gastrointestinal bleeding, but there were no controlled studies of the association between the NSAIDs and clinically apparent upper gastrointestinal (UGI) bleeding. After excluding patients known to have been prescribed aspirin, generally part of combination cold preparations or combination analgesics, there were a total of 100,124 patients exposed to NSAIDs. In order to examine duration- and dose-response relationships, a subsampled cohort study was performed. Cases were all subjects with a diagnosis of UGI bleeding during a day at risk. Univariate analyses were performed to determine if each potential confounding variable was associated with the UGI bleeding.
