1. Citation
Mukherjee D, Nissen SE, Topol EJ. Risk of Cardiovascular Events Associated With Selective COX-2 Inhibitors. JAMA. 2001;286(8):954-959.
2. Background
Although the selective COX-2 inhibitors rofecoxib and celecoxib are now widely prescribed, largely because of relative gastrointestinal safety, little is known about their possible cardiovascular (CV) effects. Biochemical data suggest that COX-2 inhibition may have either anti-atherogenic or prothrombotic effects and are difficult to extrapolate to the clinical arena.
3. Methods
This is a meta-analysis of four randomized, double-blind trials of COX-2 inhibitors: two major studies, the Vioxx Gastrointestinal Outcomes Research Study (VIGOR) and the Celecoxib Long-term Arthritis Safety Study (CLASS), and two smaller studies (one each for rofecoxib and celecoxib) submitted to the U.S. Food and Drug Administration (FDA). Gastrointestinal effects were the primary outcome of all of these studies. Patients with recent CV or cerebrovascular events were excluded. However, the meta-analysis looked at CV events, including myocardial infarctions (MI), unstable angina, cardiac thrombus, resuscitated cardiac arrest, sudden or unexplained death, ischemic stroke, and transient ischemic attacks.
4. Results
In VIGOR, 8,076 elderly patients with rheumatoid arthritis (RA) were randomized to rofecoxib 50 mg qd or to naproxen 500 mg bid. Concurrent use of aspirin or other anticoagulants was an exclusionary criterion. During median follow-up of 9 months, 1.1 CV events per 100 patients were observed in the rofecoxib arm, representing a relative risk of 2.38 as compared to the naproxen. There was no difference in overall mortality.
In CLASS, 8,059 patients with osteoarthritis (73%) or RA (27%) were randomized to celecoxib 400 mg bid, ibuprofen 800 mg td, or diclofenac 75 mg bid. Patients were allowed to receive daily low-dose aspirin for cardiac protection. There was no difference in the frequency of CV events or overall mortality among the three groups.
In the two small FDA studies, patients with osteoarthritis were randomized to a COX-2 inhibitor (rofecoxib or celecoxib), a traditional anti-inflammatory agent (nabumetone), or placebo. Again, low-dose aspirin was permitted. There was no observed difference in the frequency of CV events or overall mortality between the three groups.
Because there was no placebo control in the VIGOR and CLASS studies, the authors compared annualized MI rates of a historical aspirin-free control cohort of patients culled from primary coronary artery disease prevention trials to those of the COX-2 inhibitor groups from VIGOR and CLASS. The annualized MI rate for the control group was 0.52%, while those for the rofecoxib and celecoxib groups were 0.74% (P = 0.04) and 0.80% (P = 0.02), respectively.
5. Commentary
This study raises legitimate concerns over a possible association between cardiovascular disease (CVD) and use of COX-2 inhibitors. The authors of the VIGOR study have conjectured that naproxen might have some cardioprotective effects not possessed by rofecoxib, thereby accounting for the observed difference in CV risks. However, there is limited clinical evidence to support or refute this contention.
Accordingly, the conclusion that rofecoxib causes CVD is equally premature. A prospective study comparing the CV effects of COX-2 inhibitors versus placebo as a primary outcome is lacking and warranted. Moreover, the 50 mg dose of rofecoxib used in the VIGOR study is two to four times the standard prescribing dose, and it is certainly reasonable to ask whether there is a dose-dependent effect.
The comparison that Mukherjee, et al, make between COX-2 inhibitor users in their analysis and historical aspirin-free controls from primary CVD prevention trials has inherent problems. The cohorts are culled from different patient populations and, therefore, are questionable comparison groups. For example, there are almost certainly a disproportionately higher number of patients with RA in the COX-2 inhibitor user group, and it is known that patients with RA are at increased risk for vascular disease.
Nonetheless, given the overwhelming popularity in the use of COX-2 inhibitors, it is important to establish whether non-gastrointestinal (in particular, cardiovascular) adverse effects do, in fact, present significant concern. This is especially true because many candidate patients for COX-2 inhibitors may be at risk for multiple comorbidities (e.g. elderly patients).
