12:00 AM
 | 
May 28, 2007
 |  BioCentury  |  Politics, Policy & Law

Deconstructing data

Political Defibrillator

Deconstructing data

Although Steven Nissen's meta-analysis of safety data for diabetes drug Avandia rosiglitazone is flawed, that doesn't mean the drug has no cardiovascular risk. But quantifying the level of risk requires large-scale controlled studies, such as the ongoing RECORD study being run by GlaxoSmithKline plc.

RECORD, a 4,400-patient study designed to assess cardiovascular side effects of the drug, should provide a definitive answer in 2009 if patients are willing to remain in the study given the heightened awareness of possible safety problems.

In the meantime, FDA says the accumulated data are contradictory and an interim analysis of RECORD hasn't raised any red flags. And last week, EMEA said it had taken care of the issue in September 2006 by adding a warning to the product label about the risk of cardiac ischemic events (see "Avandia Labeling History," A2).

Indeed, EMEA advised patients "not to stop treatment with rosiglitazone and to discuss the medication with their doctor at their next regular visit."

Even the scientists who conducted the meta-analysis agree the retrospective study published in last week's New England Journal of Medicinesuffers from methodological flaws that limit the strength of its conclusions.

But Nissen, chairman of the Cleveland Clinic's department of cardiovascular medicine, told BioCentury that despite the caveats, he is convinced Avandia should have been withdrawn from the market a long time ago.

"Caution is appropriate in any manuscript, but I'm convinced the drug increases adverse cardiac outcomes," he told BioCentury. "FDA has allowed a drug to remain on the market that has resulted in a huge number of adverse cardiac outcomes."

The meta-analysis

Nissen and Kathy Wolski, a biostatistician at the Cleveland Clinic, analyzed 42 trials with about 27,000 patients in which Avandia was compared to placebo or a variety of active controls.

They reported that the odds ratio for myocardial infarction was 1.43 (p=0.03), meaning patients who received Avandia experienced a 43% increase in the chance of having an MI compared to those who received placebo or another drug, and the odds ratio for death from cardiovascular causes was 1.64 (p=0.006). The 95% confidence interval around the MI odds ratio ranged from 1.03 to 1.98,...

Read the full 1814 word article

User Sign in

Trial Subscription

Get a 4-week free trial subscription to BioCentury

Article Purchase

$150 USD
More Info >