Valdecoxib (Bextra), unlike rofecoxib, has not been subjected to long-term randomized trials to assess either its safety or efficacy; however, two short-term trials that evaluated valdecoxib for treating acute pain following coronary artery bypass graft (CABG) surgery are available.

A double-blind, controlled clinical trial was performed to compare a combination of parecoxib (Dynastat, Pharmacia/ Pfizer) and valdecoxib to placebo. Parecoxib is the pro-drug of valdecoxib, and it is administered only parenterally.

Sequential parecoxib/valdecoxib was given to 462 patients for 14 days with a 30-day follow-up period after therapy was completed. One MI (0.7%) occurred in the control group, and five MIs (1.6%) occurred in the treatment group. A similar trend was seen with cerebrovascular disorders; the treated patients experienced a higher incidence of cerebrovascular events (2.9%) than the controls (0.7%) did. The sample size might have been inadequate to demonstrate statistical significance.

A second randomized, double-blind trial involved 10 days of treatment and 30 days of follow-up in 1,500 patients. Three groups of patients were assigned to receive one of these therapies:

  • intravenous (IV) parecoxib for three days, followed by oral valdecoxib through 10 days
  • IV placebo, followed by oral valdecoxib for 10 days
  • placebo for 10 days

The incidence of CV events was significantly greater with parecoxib and valdecoxib (2%) than with placebo (0.5%) (RR, 3.7; 95% CI, 1.0-13.5; P = .03). All CV events occurred equally during the treatment and follow-up periods. In each trial, patients began ASA therapy immediately following surgery.

The CABG trials, which involved very-high-risk patients, led to an FDA-mandated boxed warning cautioning against the use of valdecoxib in patients after CABG surgery. Thereafter, in April 2005, after a review of an excessive number of case reports of severe skin reactions, valdecoxib was withdrawn from the market.


Only one long-term trial has been published with lumira-coxib (Prexige), and it is the only coxib trial adequately powered to measure the incidence of CV events. Over one year, the Therapeutic Arthritis Research and Gastrointestinal Event Trial (TARGET) compared the safety and efficacy of lumira-coxib with naproxen canadian and ibuprofen for the treatment of OA. No significant differences between lumiracoxib and naproxen (RR, 1.77; 95% CI, 0.82-3.84; P = .1471) or ibuprofen (RR, 0.66; 95% CI, 0.21-2.09; P = .4833) were noted in the incidence of MIs. In addition, no differences were observed in the patients taking ASA.


Limited data exist for the evaluation of the long-term safety effects of etoricoxib as well. The Etoricoxib Diclo-fenac Gastrointestinal Evaluation (EDGE) study compared once daily with three times daily in more than 7,000 patients over a nine-month period. There were 26 confirmed CV events (0.72%) in the etoricoxib patients and 19 events (0.54%) in the diclofenac group. Thirty percent of patients were receiving concurrent ASA. No differences were found among patients taking ASA, but the non-ASA subgroup had more MIs in the patients taking etoricoxib canadian.

As noted in these randomized clinical trials, the risk of CV events was increased with rofecoxib, celecoxib, and valdecoxib but was not clearly evident with lumiracoxib and etori-coxib.

Observational, Case-Control, Retrospective Cohort Studies

Global intrigue has surrounded the randomized clinical trials just described. For the many analyses that have been forth­coming, most have compared rofecoxib and celecoxib with other nonselective NSAIDs (e.g., ibuprofen). For rofecoxib, many studies have demonstrated an increased risk of CV or cerebrovascular events, when compared with either placebo or nonselective NSAIDs, but other studies have not shown such find-ings. Fewer studies have demonstrated an increased risk with celecoxib, and several head-to-head studies found that rofecoxib carried a higher risk in direct comparison with celecoxib.

It is difficult to draw conclusions from these data, considering the multiple variables that might have affected the results, such as the dose, the duration, and comparator agents. It appears that rofecoxib may be associated with a greater risk of CV events than celecoxib.

Meta-analyses Rofecoxib

Meta-analyses have been performed to evaluate both rofe-coxib and canadian celecoxib independently. The first pooled analysis of patient data from participants in phase IIb-V rofe-coxib clinical trials of four weeks or more in duration compared rofecoxib with placebo, naproxen, and non-naproxen NSAIDs. Data from 23 studies, involving multiple disease states and more than 28,000 patients, were analyzed.

The Antiplatelet Trialists’ Collaboration (APTC) endpoint consists of the combined incidence of the following:

  • CV events, hemorrhage, and death from unknown causes
  • nonfatal MI
  • nonfatal stroke Results were as follows:
  • rofecoxib versus placebo (RR, 0.84; 95% CI, 0.51-1.38)
  • rofecoxib versus non-naproxen NSAIDs (RR, 0.79; 95% CI, 0.40-1.55)
  • rofecoxib versus naproxen (RR, 1.69; 95% CI, 1.07-2.69)

The differing results were explained by the perceived cardio-protective effect, similar to the conclusion of the VIGOR trial. No differences were found in the duration of the study or in the rofecoxib dose.

A second analysis was conducted to evaluate 5,435 patients with OA from eight studies that compared rofecoxib with placebo and non-naproxen NSAIDs. With findings similar to those of the aforementioned analysis, no differences in the combined endpoints from the Antiplatelet Trialists’ Collaboration were found between any of the groups receiving placebo, non-naproxen NSAIDs, or rofecoxib.

A third and more recent meta-analysis evaluated 18 randomized, controlled trials and 11 observational studies. More than 25,000 patients were evaluated, and the combined relative risk for the rofecoxib patients was more than twice that of controls (RR, 2.24; 95% CI, 1.24-4.02; P = .007). In this analysis, the authors reported that by the year 2000, the increased risk of MI with rofecoxib had become evident, and it has remained constant since then.

These findings could not be explained by the naproxen-cardioprotection theory. Comparator agents, doses, and duration of the trials did not appear to affect the trial results.