Valdecoxib (Bextra) meta-analysis signals significant cardiovascular risk

November 10, 2004

Nov 10, 2004

New Orleans, LA - A new meta-analysis of trials with Pfizer's COX-2 inhibitor valdecoxib (Bextra) has suggested that this agent is also associated with increased risk of major cardiovascular events.

However, coxib experts have strongly criticized the way in which this meta-analysis was carried out, say that the conclusions are "problematic," and are concerned that these reports are "scaring away patients".

The meta-analysis, presented by Dr Garret FitzGerald (University of Pennsylvania, Philadelphia) during an invited lecture at the American Heart Association Scientific Sessions 2004 this week, included two placebo-controlled trials of valdecoxib used as pain control in CABG patients along with several other placebo-controlled studies of the drug in arthritis patients. The total number of patients included was around 7500. Drs Curt Furberg (Wake Forest University, Winston-Salem, NC) and Bruce Psaty (University of Washington, Seattle) also helped to put together the meta-analysis.

Results showed that valdecoxib was associated with more than twice as many MI or stroke events than placebo (relative risk 2.19). FitzGerald reported that the confidence intervals did not embrace 1, so it was a significant result. He also pointed out that this was actually higher than the 1.96 relative risk shown with rofecoxib (Vioxx, Merck) in the APPROVE trial that led to withdrawal of the drug. "This is a time bomb waiting to go off," he commented.

Talking to the media about his findings, FitzGerald explained that both the CABG trials tested a combination of valdecoxib and its prodrug, the prodrug being given initially for fast onset of action, followed by valdecoxib orally. One of the CABG trials was published last year and showed a cluster of cardiovascular events in the valdecoxib group [ 1 ]. Limited information is available on the other CABG trial, which has not been published, but Pfizer did release some data on this trial recently that also showed a cluster of cardiovascular events in the valdecoxib group. FitzGerald used this data in the meta-analysis. The other trials included in the meta-analysis were placebo-controlled trials of valdecoxib in arthritis patients, which were included in a previous published meta-analysis by White et al [ 2 ].

Number of subjects in studies included in meta-analysis

Ott (initial CABG study)
2nd CABG
White (placebo-controlled studies only)

Number of cardiovascular events (MI and stroke)

95% CI
2nd CABG
Meta-analysis, p=0.11

To download tables as slides, click here

FitzGerald said that of the total patients in his meta-analysis, about 2000 were in the CABG trials and the remainder were in the arthritis trials. When the trials were tested for heterogeneity, this was not significant. "Insofar as you can ask the question 'Is the risk different in the two populations—CABG and arthritis patients?' " FitzGerald said, "the answer is no."

Canary in the coal mine

FitzGerald said the CABG patient trials were probably "the canary in the coal mine," as the likelihood of detecting a signal of increased cardiovascular risk with C0X-2 inhibitors is expected to be linked to the degree in which the hemostatic system is activated. "If the thrombosis risk is elevated, as it is in CABG patients, we are much more likely to see an effect if it is there. The clustering of these events in CABG patients is exactly the sort of signal we would be concerned about."

He added that if this meta-analysis had been conducted before the APPROVE results and the Vioxx withdrawal, then it would have probably just been taken as reason to do more studies, but now the situation has changed. "Okay, this is not a prospective randomized trial like APPROVE was, even though there are twice as many patients as were in APPROVE, but in the wake of APPROVE and the Vioxx withdrawal and the fact that there is proposed clear mechanism that suggests a class effect, the context has now changed. The burden of proof has now shifted, and there is good evidence that cardiovascular risk will be increased with all drugs in this class."

We need to find a way to salvage the utility of these drugs. I believe that can be done. We need to find a way to develop these drugs safely so the baby is not washed out with the bathwater.

FitzGerald said he did not believe all the COX-2 inhibitors should be withdrawn, rather that ways should be explored to make sure they are used safely. "I would recommend that the FDA take a careful look at the entire class, which I'm sure they will do, and consider whether they should carry a warning against use in patients at high cardiovascular risk. We also need to advise people at low cardiovascular risk how to use these drugs."

Pointing out that the increased risk in the APPROVE trial did not become evident until 18 months, he suggested that patients may be advised to avoid prolonged drug exposure. Reminding the audience that two controlled trials have suggested that COX-2 inhibitors do reduce serious gastrointestinal events compared with regular NSAIDS, he added: "We need to find a way to salvage the utility of these drugs. I believe that can be done. We need to find a way to develop these drugs safely so the baby is not washed out with the bathwater."

Too late for Vioxx

But it seems to be too late for Vioxx. Having already been withdrawn, it would be very unlikely that it would be returned to the market. However, Merck has a second COX-2 inhibitor — etoricoxib — which it could still market. FitzGerald commented: "Vioxx is probably not going to come back, but the question on Merck's table is now 'How do we develop etoricoxib?' "

Pfizer responds

Responding to a New York Times article on the meta-analysis published earlier today [ 3 ], Pfizer said the analysis had drawn "unsubstantiated conclusions" about the cardiovascular safety of Bextra and that the analysis "is based on information that has not been published in a medical journal or subject to independent scientific review [ 4 ]."

Pfizer claims that the previous meta-analysis just in arthritis patients by White et al demonstrated that the drug "was not associated with an increased incidence of thrombotic events relative to nonselective NSAIDS or placebo."

Internet Explorer users: to download the PowerPoint presentation, click here.


  1. Ott E, Nussmeier NA, Duke PC, et al. Efficacy and safety of the cyclooxygenase 2 inhibitors parecoxib and valdecoxib in patients undergoing coronary artery bypass surgery. J Thorac Cardiovasc Surg 2003 Jun; 125(6):1481-92 Abstract

  2. White WB et al. "Effects of the cyclooxygenase-2 specific inhibitor valdecoxib versus nonsteroidal antiinflammatory agents and placebo on cardiovascular thrombotic events in patients with arthritis." Am J Ther 2004; 11: 244-250 Abstract

  3. Harris G. New study links Pfizer's Bextra, similar to Vioxx, to heart attacks. New York Times November 10, 2004. Available at:

  4. Pfizer statement on Bextra (valdecoxib). November 10, 2004. Available at:


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.