Selective COX-2 Inhibitors as Risk Factors for Cardiovascular Events
August 2001
The August 22/29, 2001 issue of the Journal of the American Medical Association features a meta-analysis entitled "Risk of cardiovascular events associated with selective COX-2 inhibitors" which has stirred considerable and immediate media attention and raises concerns about possible thromboembolic risks facing patients taking COX-2 inhibitors and physicians who prescribed them.
This meta-analysis studied two large randomized post-marketing trials of rofecoxib and celecoxib (VIGOR, CLASS respectively), as well as other published and unpublished data submitted to the relevant FDA committees. The VIGOR study with 8076 patients was designed to evaluate gastrointestinal toxicity of rofecoxib vs. naproxen in a large rheumatoid arthritis cohort2. Low dose aspirin use was not permitted. There were 46 cases of serious thrombotic cardiovascular events (e.g., MI, TIA and stroke) in the rofecoxib group, and 20 events in the naproxen group (RR 2.38; p <0.001). The CLASS study with 8059 patients evaluated GI toxicity of celecoxib vs. ibuprofen vs. diclofenac3. Aspirin use was permitted. There were no significant differences in cardiovascular events between groups in the CLASS study. The cardiovascular risk seen in the rofecoxib and celecoxib studies was then compared to a meta-analysis of two previous unrelated studies using aspirin vs. placebo for prevention of cardiovascular events (the US Physicians Health Study and the UK Doctors Study). This analysis suggested that the annualized MI rate for the placebo group from the US/UK studies was 0.52%, statistically significantly lower that for both rofecoxib (0.74%; p=0.04) and celecoxib (0.80%; p = 0.02).
The meta-analysis of Mukherjee et al has important limitations. It is well appreciated that selective COX-2 inhibitors do not significantly inhibit platelet aggregation, one of several factors influencing thrombotic cardiovascular disease. Conventional NSAIDs inhibit both thromboxane and PGI2. It has been suggested that by decreasing antithrombotic PGI2 production and not affecting thromboxane production, selective COX-2 inhibitors could cause an increase in thrombotic cardiovascular events. This hypothesis was not directly studied by Mukherjee et al. At the same time, all NSAIDs, including selective COX-2 inhibitors, may have important effects in reducing the inflammatory component of atherogenesis and actually reduce cardiovascular risk. Patients in the COX-2 trials were heterogeneous; the CLASS study included patients with both OA and RA; patients with RA have a higher risk for MI, a fact that likely had important, but unknown meaning for the study conclusions. Finally, none of the studies examined in the meta-analysis was actually designed to evaluate differences in the rates of cardiovascular events.
Current evidence indicates that available selective COX -2 inhibitors are associated with a significant reduction in major NSAID related GI toxicity, especially in high-risk patients. These GI toxicities have been and continue to be a major source of morbidity and mortality in patients with arthritis. The decision to use selective COX-2 inhibitors is multidimensional, and must balance possible risks for major cardiovascular events against the benefits of these agents. For the individual patient, it is important to consider factors that affect the risk/benefit ratio of the therapeutic choice and the patient's willingness to accept therapy recommendations. These factors include risks for major GI toxicity, cardiovascular events, renal failure, cost and others. These are balanced against measures that can ameliorate these risks, including where appropriate, concomitant low dose aspirin, gastric mucosal protection, and use of effective analgesics and antiinflammatories with minimal GI toxicities. A well-designed study focusing on cardiovascular risk and benefit of selective COX-2 inhibitors will be needed to properly examine this issue.
August 2001
Eric Matteson, MD, MPH; John J. Cush, MD
Hotline Editors
- Mukherjee et al, JAMA 2001;286:954-959
- Bombardier et al, N Engl J Med 2000;343:1520-1528
- Silverstein et al, JAMA 2000;284:1247-1255
