Thursday, May 10, 2012

Story of Biased Chantix Study by Pfizer-Funded Author Gets Worse; Article Violates FDA Recommendations, Excludes Heart Attack and Stroke Events, and Doesn't Adequately Consider Low Power of Studies

Monday, I discussed the results of a re-analysis of a previous study that showed an increased risk of adverse cardiovascular events among patients receiving Chantix. The new study concludes that Chantix is perfectly safe for use in patients in terms of cardiovascular risk. One of the authors of the new study is a recipient of grant funding from Pfizer, the manufacturer of Chantix, and therefore has a significant financial conflict of interest. I have previously argued that: "The bias present in the study has the appearance of being a result of this conflict of interest."

Today, I reveal several more dimensions of this story.

First, as Dr. Sonal Singh points out in his rapid response to the re-analysis, it appears that the authors excluded from their analysis several severe adverse cardiovascular events that have already been reported, including by the FDA itself. As Dr. Singh notes: "For example, in the largest clinical trial, the Rigotti study, the authors recorded 10 events in each treatment arm. In contrast, the updated FDA label, which contains warnings about an increase in cardiovascular risk associated with varenicline in patients with cardiovascular disease, notes several more treatment-emergent cardiovascular events during 30 days in varenicline than placebo [in this same study]: “These include treatment emergent events (on treatment or 30 days after treatment of angina pectoris (13 patients in varenicline arm vs 7 in placebo arm), and the serious cardiovascular events of non-fatal MI (4 vs 1) and nonfatal stroke (2 vs 0).”"

These exclusions presumably occurred because the authors did not count any adverse events that happened more than 30 days after discontinuation of the drug. However, the FDA-mandated clinical trial to assess the cardiovascular risks associated with Chantix required that follow-up of treated patients for adverse cardiovascular outcomes must continue out to at least one year. Thus, the study violates the FDA's recommendation that a one-year follow-up period is necessary. Apparently, the FDA is concerned that adverse effects of Chantix could occur well after discontinuation of the drug.

And well they should be. We all know about the severe cardiovascular adverse events associated with Vioxx use. Vioxx was in fact pulled from the market because of these effects. It is critical to note that these adverse cardiovascular effects of Vioxx persisted up to one year after discontinuation of the drug. How does it make any sound scientific sense, then, to exclude events occurring more than 30 days, but less than one year, after discontinuation of Pfizer.

This dimension of the story seems to make the apparent bias of the study even more striking.

Second, I do not believe that the article appropriately considers the lack of power in the clinical trials to detect a significant increase in a rare, serious adverse event such as heart attacks. These studies were powered to detect differences in smoking cessation between treatment arms, not to detect differences in rare adverse events. Pfizer itself notes that these clinical trials should not be used for conclusive evidence on whether the drug poses cardiovascular risks. Speaking about one clinical trial, Pfizer notes: "The study was powered for assessing efficacy (ie quit rates) but not for assessing differences in the occurrence of serious CV events between CHAMPIX and placebo. Therefore, the study was not large enough to allow conclusions regarding the difference in the incidence of CV events reported in the two arms."

In this light, it is difficult to understand how the paper could conclude not only that there was no statistically significant increase in risk, but also that there was no clinically significant increase. As I pointed out yesterday, even a 0.27% absolute increased risk would translate into thousands of adverse events because of the large number of patients being treated with the drug. And if the drug is not particularly effective in achieving smoking cessation, then the risk-benefit analysis could easily tip towards the risk side.

Even if one accepts the results of the article at face value, the upper end of the confidence interval for the risk difference was 0.63%. This means that because of the small sample sizes in the clinical trials, this meta-analysis cannot rule out an effect as large as a risk difference of 0.63%. With 1.8 million annual prescriptions for Chantix, however, this excess risk of 63 events per 10,000 treated patients would be quite significant. To conclude definitively that there is no clinically significant cardiovascular risk associated with Chantix based on the meta-analysis thus appears to be quite biased.

Finally, while the article does not provide details of the purpose for which the conflicted author is receiving funding from Pfizer, it turns out that this funding is for a clinical trial of Chantix among hospitalized smokers. In my opinion, this makes the conflict of interest even stronger.

Clearly, the verdict is not yet in and an investigator who has a conflict of interest by virtue of taking money from Pfizer for a clinical trial of Chantix should not draw a conclusion at this point in time, based on the existing under-powered clinical trials, that the drug has no clinically significant cardiovascular risks.

No comments: