Whether non-aspirin non-steroidal antiinflammatory drugs (NSAIDs) affect the risk of myocardial infarction is unclear. Also, it is unknown whether the effect varies by individual NSAIDs. To summarize the evidence from published observational studies on the risk of myocardial infarction associated with both traditional NSAIDs (tNSAIDs) and selective inhibitors of cyclooxygenase-2 (Coxibs), the authors conducted a systematic review of cohort and case-control studies on NSAIDs and myocardial infarction published between 2000 and 2005. Sixteen original studies were selected according to predefined criteria. Two researchers independently extracted the data on individual study characteristics and results. The authors calculated pooled relative risk (RR) estimates of myocardial infarction for specific NSAIDs compared with no NSAID use, tested the heterogeneity of effects, and evaluated potential reasons for heterogeneity. The pooled RR of myocardial infarction was 0.98 (95% confidence interval (CI): 0.92-1.05) for naproxen, 1.07 (95% CI: 1.02-1.12) for ibuprofen, 1.44 (95% CI: 1.32-1.56) for diclofenac, 0.96 (95% CI: 0.90-1.02) for celecoxib, and 1.26 (95% CI: 1.17-1.36) for rofecoxib (all doses). The pooled RR for rofecoxib at doses >25 mg/day was 1.78 (95% CI: 1.36-2.34), and 1.18 (95% CI: 1.07-1.31) for doses < or =25 mg/day. The RR associated with naproxen was 0.83 (95% CI: 0.72-0.90) among non-users of low-dose aspirin. The RR associated with rofecoxib (all doses) was 1.39 (95% CI: 1.25-1.54) among subjects without a history of myocardial infarction. The risk of myocardial infarction varies with individual NSAIDs. An increased risk was observed for diclofenac and rofecoxib, the latter one with a clear dose-response trend. There was a suggestion of a small increased risk with ibuprofen. Also, data suggest a small reduced risk for naproxen present only in non-users of aspirin, mainly people free of clinically apparent vascular disease.