Nonsteroidal anti-inflammatory drugs (NSAIDs), including aspirin, appear to have clinically significant anticarcinogenic effects in the gastrointestinal tract. Epidemiological data indicate that use of these drugs is inversely associated with the risk of sporadic colorectal cancer, and clinical trials among patients with familial polyposis coli show that NSAIDs can lead to the regression of large bowel adenomas. Animal studies have reported a similar efficacy of NSAIDs against experimental carcinogenesis. A consistent pattern in this research is that continued long-term use of NSAIDs is required for an anticancer effect--up to 15 or 20 years before a reduced risk of colorectal cancer appears. Epidemiological data also suggest possible protective effects in the stomach and esophagus. The mechanisms underlying any chemopreventive effect of NSAIDs are not clear. Inhibition of cyclooxygenase is one possibility, but pathways independent of cyclooxygenase and prostaglandins are also possible.