Risk of cardiovascular events associated with selective COX-2 inhibitors.
Level 1 - systematic review of randomized trials
Systematic review of randomized controlled trials and regulatory submissions
PubMed 11509060 · doi:10.1001/jama.286.8.954
What was done
Authors conducted a MEDLINE search (1998 to February 2001) and reviewed FDA submissions for trials evaluating selective COX-2 inhibitors for arthritis and musculoskeletal pain in patients without coronary artery disease. The search identified two large randomized trials (VIGOR with 8,076 patients and CLASS with 8,059 patients) and two smaller randomized trials of approximately 1,000 patients each.
What was found
In VIGOR, rofecoxib showed a significantly higher relative risk of confirmed adjudicated thrombotic cardiovascular events compared with naproxen (relative risk 2.38, 95% CI 1.39-4.00, P = .002). In CLASS, no significant difference in cardiovascular event rates (myocardial infarction, stroke, death) was observed between celecoxib and nonsteroidal anti-inflammatory drugs. Annualized myocardial infarction rates for rofecoxib (0.74%, P = .04) and celecoxib (0.80%, P = .02) were significantly higher than the 0.52% rate reported in the placebo group of an external meta-analysis of 23,407 primary prevention patients.
Why it matters
This analysis provided early, high-impact evidence that selective COX-2 inhibitors are associated with an increased risk of serious thrombotic cardiovascular events.
Limits
The review relies on only four trials, and the primary comparisons for absolute myocardial infarction risk were made against an external, historical placebo cohort from a separate meta-analysis rather than within-trial placebo controls. Comparators also differed between the major trials (naproxen in VIGOR versus other NSAIDs in CLASS).
Cited by
- supports Vioxx was removed from the market after research by Dr. Steven Nissen at the Cleveland Clinic demonstrated an increased risk of heart attacks.