Use of nonsteroidal antiinflammatory drugs (NSAIDs) has been associated with a reduced risk of colorectal cancer, but limited information is available on the effect of individual nonaspirin NSAIDs. In addition, the dose-response relation of aspirin in reducing the risk of colorectal cancer has not been described. We carried out a population-based cohort study with secondary case-control analysis to determine the association between the risk of colorectal cancer and use of aspirin and individual NSAIDs, including the role of dose and duration. The General Practice Research Database in the U.K. was the source population. We traced 943,903 persons 40-79 years of age and free of cancer and colorectal adenoma between January 1994 and September 1997. A total of 2,002 incident cases of colorectal cancer were ascertained. The incidence rate of colorectal cancer per 10,000 person-years was 7.3. The risk of colorectal cancer was reduced in users of nonaspirin NSAIDs and became evident after 6 months of continuous treatment. The adjusted relative risk was 0.5 (95% confidence interval = 0.4-0.7). The reduction in risk disappeared completely 1 year after stopping NSAID treatment. The risk of developing colorectal cancer was reduced in long-term users of aspirin at doses of 300 mg daily (relative risk = 0.6; 95% confidence interval = 0.4-0.9). Daily doses of 75 and 150 mg aspirin were not associated with a reduced risk of colorectal cancer. Our data support the existence of an important protective effect of nonaspirin NSAID continuous intake against colorectal cancer and point to a similar reduction in risk for aspirin at doses of at least 300 mg daily. One-year treatment with NSAIDs would prevent one case of colorectal cancer in a population of 1,000 persons 70-79 years of age.