Background: Cyclooxygenase-2 (COX-2) inhibitors have recently been implicated in the occurrence of adverse cardiovascular events. Rofecoxib has been withdrawn from the market, but celecoxib (Celebrex) continues to be used. Other nonsteroidal anti-inflammatory drugs (NSAIDs) also may increase the risk of cardiovascular events. McGettigan and Henry performed a systematic review and meta-analysis of controlled observational studies to compare the cardiovascular risks associated with COX-2 inhibitors and other NSAIDs.

The Study: The study included case- or cohort-controlled population studies that looked at the cardiovascular effects of COX-2 inhibitors and conventional NSAIDs.

Twenty-three studies were eligible for inclusion. Almost all of the reported cardiovascular events were myocardial infarction or sudden cardiovascular death. Prescriptions were considered markers for drug use in most studies. Most studies collected data through hospital records, or, in some cases, direct interviews. Many of the studies adjusted for comorbidities. The reference points of the studies were no use or remote use of the medications.

Results: The relative risk (RR) summary estimates were separated for the case-control studies and the cohort studies. For rofecoxib, these were 1.31 and 1.53, respectively, or 1.35 when both groups were combined. The data showed that higher doses produced higher cardiovascular risk. For celecoxib, the RRs were 1.01 and 1.22 for the case-control and cohort studies, respectively, and 1.06 for the combined studies. The summary RRs for otherstudydrugswere:naproxen(Naprosyn), 0.97; ibuprofen (Motrin), 1.07; piroxicam (Feldene), 1.06; diclofenac (Voltaren), 1.40; and indomethacin (Indocin), 1.30.

When the analysis was reformulated to include only studies that compared the COX-2 inhibitors with conventional NSAIDs, rofecoxib had RRs of 1.37, 1.32, 1.61, and 1.32. Diclofenac had a summary RR from three studies of 1.80. Celecoxib had RRs of 1.07, 1.00, 1.28, and 0.98. Indomethacin had a summary RR from two studies of 1.48. Ibuprofen and naproxen showed no or modest increases in RR. Early risk for new users of rofecoxib showed elevated RR (summary RR of 1.66) but not for celecoxib (RR of 1.32).

Conclusion: The authors conclude that, regarding conventional NSAIDs, there is no increased cardiovascular risk found in this study with naproxen or ibuprofen, and the combination of aspirin and ibuprofen did not alter this effect. No cardioprotective effect was found with naproxen, as suggested by an earlier trial. Indomethacin was associated with increased cardiovascular risk.

Regarding selective COX-2 inhibitors, rofecoxib appears to have an early and late risk. Celecoxib appears to be safer, but only in dosages of 200 mg per day, although it was safer than rofecoxib at higher doses. Diclofenac has a risk profile similar to that of rofecoxib. Overall, COX-2 inhibitors have a lower safety profile than naproxen and ibuprofen; and celecoxib, if used, should be given in low dosages.