A Modern Day Faustus: COX-2 Inhibitors — Are We Trading Gastric Protection for Cardiac Death?

Abstract & Commentary

Synopsis: This study suggests that high-dose rofecoxib may increase the risk of cardiovascular events.

Source: Ray WA, et al. Lancet. 2002;360:1071-1073.

This retrospective cohort trial assessed the risk of low- and high-dose COX-2 inhibitors (celecoxib and rofecoxib) for serious coronary heart disease defined as a new myocardial infarction or cardiovascular death. The Tennessee Medicaid program enrolled 202,916 subjects who were assessed for a 2-year period. Subjects currently taking COX-2 inhibitors (10,298 person years) were compared to nonusers (237,975 person-years) and to those taking nonsteroidal medications (ibuprofen, naproxen; 37,423 person-years). Eligible subjects were non-nursing home patients without a life-limiting illness, aged 50-84 years.

When compared to nonusers, patients taking high-dose rofecoxib (> 25 mg/d) had an adjusted risk ratio of 1.93 (95% CI, 1.1-3.4; P = 0.02). There was no statistical difference between nonusers and those taking other nonsteroidals, celecoxib or low-dose rofecoxib (< 25 mg/d). The event rate in patients not taking COX-2 inhibitors or NSAIDs was 1.3%; the event rate in the high-dose rofecoxib population was 1.1%. The absolute risk of harm was 0.2%; 500 patients would have to be treated with high-dose rofecoxib instead of no therapy before 1 additional cardiovascular event could be expected (number needed to harm = 500).

Comment by Jeff Wiese, MD

The VIGOR trial found that subjects who took rofecoxib 50 mg/d were 5 times as likely to have a myocardial event as those who took 1000 mg of naproxen per day.1 The initial explanation was that naproxen was cardioprotective, but this was discounted in subsequent studies that found no benefit of naproxen in preventing cardiac disease.2

COX-2 inhibitors selectively inhibit the cyclo-oxygenase 2 enzyme, and in doing so, decrease inflammation without inhibiting the prostaglandins important for gastrointestinal protection. One potential complication of this selective inhibition may be an increase in arachidonic acid metabolites shunting down the thromboxane pathway, thereby increasing platelet aggregation. Bombardier and colleagues have suggested that COX-2 inhibition may affect nitric oxide synthetase, thereby increasing vascular tone and myocardial work. The observation of this study and the VIGOR trial may be the clinical manifestation of these physiologic postulates.1

The results of this study should be interpreted with caution. There are many potential confounders that may explain the results of this study. It is likely that most patients in this study began with a nonsteroidal such as ibuprofen, advancing to a COX-2 inhibitor if this failed. The patients in the rofecoxib group may therefore represent a patient population with a greater severity of inflammation and a greater risk of cardiovascular events. Furthermore, patients with comorbid diseases prohibiting nonsteroidal therapy such as renal insufficiency or gastrointestinal bleeding would have been excluded from the ibuprofen group but may have been included in the COX-2 group. These patients may have been disproportionately represented in the rofecoxib group.

This study suggests that high-dose rofecoxib may increase the risk of cardiovascular events. Further randomized trials will be required to eliminate the numerous potential confounders from interpretation of the results. Until that time, high-dose rofecoxib should be used with caution, especially in those with underlying cardiovascular disease.

Dr. Wiese, Assistant Professor of Family Medicine, Medical College of Ohio, Toledo, OH, is Associate Editor of Internal Medicine Alert.


1. Bombardier C, et al. N Engl J Med. 2000;343:1520-1528.

2. Ray WA, et al. Lancet. 2002;359:118-123.

3. Solomon DH, et al. Arch Intern Med. 2002;162: 1099-1104.