HOME > STATS > ARTICLES

A recent meta-analysis (a study of existing studies) linking the diabetes pill Avandia (rosiglitazone) with an increased risk of heart attacks led to more patient worry over whether their medications were doing more harm than help. Business Week and many other publications note that the market share for Avandia plummeted from 10 percent to zero within two days of the publication of the meta-analysis in the New England Journal of Medicine (NEJM), while the rate of patients switching to a rival drug, Actos, tripled. Doctors treating diabetes have been flooded with calls from panicked patients, largely in response to the way the media covered the news.

“Millions of Americans taking the highly promoted diabetes drug Avandia could be putting themselves at risk for deadly heart problems, according to new research,” began ABC News in its segment on the drug.

“Another study has found that diabetes drugs intended to help patients live longer, healthier lives may in fact increase the chances they will die -- this time of heart attacks and other causes,” reported Reuters in the lede to its story.

“A widely prescribed drug to treat Type 2 diabetes substantially increases the risk of heart attacks and death from cardiovascular disease, according to a study released today that critics say questions the government's ability to monitor drug safety,” reported the Los Angeles Times in its lede.

But the rush to damn and ditch Avandia is far from justified by the actual data compiled and analyzed in the study. The Washington Post’s Health Section contacted experts in the field and found that despite the appearance of a scandal (claims that the FDA knew about the risk and failed to take appropriate steps to issue a warning), doctors aren’t rushing to recommend that patients switch to other medication. This follows on a remarkable editorial in the Lancet that rebuked the New England Journal of Medicine for scaremongering.

“To avoid unnecessary panic among patients, a calmer and more considered approach to the safety of rosiglitazone is needed. Alarmist headlines and confident declarations help nobody.”

To be fair, the authors of the study Dr. Steven Nissen et al do note that their study has “important limitations”

“We pooled the results of a group of trials that were not originally intended to explore cardiovascular outcomes. Most trials did not centrally adjudicate cardiovascular outcomes, and the definitions of myocardial infarction were not available. Many of these trials were small and short-term, resulting in few adverse cardiovascular events or deaths. Accordingly, the confidence intervals for the odds ratios for myocardial infarction and death from cardiovascular causes are wide, resulting in considerable uncertainty about the magnitude of the observed hazard. Furthermore, we did not have access to original source data for any of these trials. Thus, we based the analysis on available data from publicly disclosed summaries of events. The lack of availability of source data did not allow the use of more statistically powerful time-to-event analysis. A meta-analysis is always considered less convincing than a large prospective trial designed to assess the outcome of interest.”

In other words, given the small number of cases of non-fatal heart attack overall when all trials were combined (86/14,371 for those taking Avandia versus 72/11,634 in the control groups), small changes either way could dramatically alter the statistical perception of risk. This is the problem with meta-analyses: the result can often be skewed by the way the data was added together. Until there is supporting data from randomized controlled trials, a meta analysis should be treated as a potential signal of an effect.

There is no question that the media should have been much more circumspect in reporting such a study given that it was a meta analysis and given that it contradicted other more rigorous studies showing that Avandia had not increased heart risk. And reporters should have noted the statistical limitations in the data.

The tougher question is whether these limitations should have led to a more guarded release of the study, and more circumspection from Nissen. He told WebMD, for instance, that there was “not any reason to panic;” but he also told the Associated Press that he “was physically sick to my stomach and I couldn't sleep," after he saw the results of the statistical analysis. Such mixed messages, coupled with a highly critical editorial in NEJM that urged the FDA to pull the drug completely guaranteed alarmism in the media. And so everybody else panicked.

And, now, in one of the most critical responses to date, A Wall Street Journal op-ed by American Enterprise Institute resident fellow, Scott Gottlieb M.D. says that NEJM “rushed onto its Web site a limited and flawed analysis” in order to advance a political agenda:

“The publication was timed to get ahead of the Food and Drug Administration's more careful evaluation of the same issues. The journal seemed bent on beating the FDA to the punch. The goal? Painting the FDA as impotent, in order to argue for legislation winding through Congress that would increase regulatory hurdles for drug approvals. The journal's motives were made bare by its own editorial on the matter.”

As evidence for collusion, Gottlieb points out that while the FDA, which is in the process of conducting an extensive long-term safety study of Avandia, was not given any notice by NEJM of the study’s pre-emptive publication online, Rep Henry Waxman (D., California) issued a “substantive” statement as soon as the study was posted, which included a date and location for oversight hearings on the FDA’s handling of the drug. Gottlieb also noted that both Waxman and NEJM are supporters of tighter regulation of FDA drug approvals.

In a further twist, the 10Q detective argues that the media should have noted a potential conflict of interest in that Nissen’s clinic receives financial support from the manufacturer of Avandia’s rival Actos, made by Takeda, and that Nissen has been a critic of Avendia for years.

View the Technorati Link Cosmos for this entry