Rofecoxib and valdecoxib have been withdrawn from the market because they increase the risk of cardiovascular events. To determine whether these drugs and other types of selective cyclooxygenase 2 inhibitors (coxibs) and noncoxib NSAIDs can also increase the risk of stroke, Roumie et al. conducted a retrospective study of 336,906 individuals (aged 50–84 years) enrolled in the Tennessee Medicaid program. Over a follow-up of 989,826 person-years, 4,354 hospitalizations for stroke were recorded, 89.3% of which were ischemic. Among nonusers of NSAIDS (the reference group) there were 4.51 strokes per 1,000 personyears. By comparison, users of rofecoxib had an increased risk of incident stroke (adjusted hazard ratio 1.28, 95% CI 1.06–1.53), as did users of valdecoxib (adjusted hazard ratio 1.41, 95%CI 1.04–1.91). None of the other NSAIDS examined (celecoxib, naproxen, ibuprofen, diclofenac and indomethacin) were associated with an increased risk of incident stroke. Restriction of the analyses to new users of NSAIDs produced similar results. Compared with the reference group, individuals who began rofecoxib having not received any other NSAID in the 365 days before filling their first rofecoxib prescription had a 1.46-fold increased risk of incident stroke, and new users of valdecoxib had a 1.39-fold increased risk. These results indicate that the risk of stroke is increased with the use of either rofecoxib or valdecoxib. By contrast, none of the other NSAIDs examined seem to affect stroke risk. The authors advise caution when prescribing NSAIDs with rofecoxib-like or valdecoxib-like coxib selectivity.