Am Fam Physician. 2005;72(01):158
Clinical Question: Does celecoxib (Celebrex) increase the risk of cardiovascular events?
Setting: Outpatient (specialty)
Study Design: Randomized controlled trial (double-blinded)
Allocation: Concealed
Synopsis: This study was an attempt to broaden the indication of celecoxib by determining whether it can prevent recurrent colon polyps. When the results of a similar study of rofecoxib (Vioxx) were released to the media, the authors of this study decided to similarly analyze their existing data to look for cardiovascular complications.
In this study, 2,035 patients who were 32 to 88 years of age with a history of a single large colon adenoma or multiple adenomas were randomized to treatment with placebo, celecoxib in a dosage of 200 mg twice daily, or celecoxib in a dosage of 400 mg twice daily. Groups were balanced at the start of the study, analysis was by intention to treat, and patients were followed for three years.
Cardiovascular endpoints were assessed blindly, and mortality data were available for all patients. Although 41 percent of patients had hypertension, only 4 percent had a history of myocardial infarction (MI); 2 percent had a history of stroke; 7 percent, angina; and 2 percent, heart failure. After 2.8 to 3.1 years of follow-up, the likelihood of cardiovascular death, MI, stroke, or heart failure was 1.0 percent in the placebo group, 2.3 percent in the 200-mg group, and 3.4 percent in the 400-mg group. The increase in this group of outcomes between the 400-mg and placebo groups was statistically significant (95 percent confidence interval [CI], 1.4 to 7.8). The difference between the 200-mg and placebo groups just missed being statistically higher with celecoxib (95 percent CI, 0.9 to 5.5). The number needed to treat to harm with celecoxib in a dosage of 400 mg per day for three years is 42, or 126 per year of use.
Bottom Line: One additional cardiovascular event or cardiovascular death occurs for every 126 patients treated for one year with celecoxib. There appears to be a dose-response relationship. It is difficult to justify continued use of this and other coxibs in all but the most exceptional circumstances. (Level of Evidence: 1b)