Self-reported NSAID use at entry (January 1995-May 1997) was updated through June 2006, using a nationwide prescription database. CRC incidence was ascertained from nationwide registers. Cox proportional hazards regression was used to compute confounder-adjusted incidence rate ratios (RRs) and 95% confidence intervals (CIs).
From 51,053 cohort subjects, we identified 615 CRC cases during 1995-2006. Daily aspirin use at entry was associated with a decreased risk of CRC (RR, 0.73; 95% CI, 0.49-1.10). A similar risk reduction was seen among subjects with 10 or more prescriptions for aspirin or non-aspirin NSAIDs and five or more years of follow-up. Most aspirin prescriptions were for 75-150 mg aspirin tablets. Among non-aspirin NSAID users, only those with body mass index (BMI) above 25 showed risk reductions [RR, 0.69 (0.47-1.03) for 10 or more prescriptions].
Long-term consistent use of aspirin or non-aspirin NSAIDs appears necessary to achieve a protective effect against CRC. Further studies of the effective dose of aspirin and the potential interaction between NSAID use and BMI are warranted.