Vytorin Is Beaten Up Some More

The Vytorin-cancer scare resurfaces with the simultaneous presentation of data from the Simvastatin and Ezetimibe in Aortic Stenosis (SEAS) trial at the annual congress of the European Society of Cardiology (ESC) and the online publication of study results in the NEJM. Accompanying the NEJM article are an analysis of cancer data from 3 Vytorin studies (including the SEAS trial) led by biostatistician (Sir) Richard Peto and a cautionary editorial headed by the journal's Jeffrey Drazen.

Peer-reviewed results from the prospective, randomized, double-blind, placebo-controlled SEAS trial (N = 1873) are consistent with those in a Merck/Schering-Plough press release from July (previous post here). Namely Vytorin did not reduce the relative risk of the primary endpoint, major cardiovascular events associated with aortic valve or atherosclerotic disease, at a median follow-up of 52.2 months. Moreover, cancer (but not cancer deaths) occurred significantly more often in Vytorin-treated patients (105 vs 70; P = .01). The latter finding was driven by a significantly higher incidence of (unspecified) skin cancer with Vytorin treatment (18 vs 8; P = .08).

In their analysis of the cancer data, Peto et al compared the rate of cancer in the SEAS trial (mean follow-up, 4.1 years) with unblinded cancer data from 2 larger, ongoing Vytorin studies, SHARP (N = 9264) and IMPROVE-IT (N = 11,353). Data from the SEAS trial revealed an increase in new cancer from several sites with Vytorin treatment in this analysis (105 vs 65), but there was no increase in the risk of cancer overall or at a particular site in SHARP and IMPROVE-IT (313 vs 326). When combining data from these 2 studies, cancer deaths were nonsignificantly increased with Vytorin treatment (97 vs 72; P = .08). But Peto et al argue that any increase in the risk of cancer death should be accompanied by an increased risk of cancer, which was not observed.

In their editorial, Drazen et al advise caution when interpreting the cancer analysis of Peto et al given the relatively short follow-up times of SHARP and IMPROVE-IT (2.7 and 1.0 years, respectively), despite that fact that these studies are much larger than the SEAS trial. The editorial also notes that Peto et al have received research funding from Merck and Schering-Plough (intimating some kind of conflict of interest), although the cancer analysis was performed independently of the companies.

Forbes's Matthew Herper, an avid follower of the Vytorin story, covers the clinical divide created by these recently presented data (a divide that has arguably been fomented by media outlets like Forbes). At the center of the cancer controversy is how Vytorin, and specifically the ezetimibe component, might increase the risk of cancer. Some argue that the drug may block the absorption of anticancer plant sterols, reports Herper, a touchy-feely explanation that MD Anderson's biostatistician Donald Berry* finds "pretty weak."

Another debate is whether it's statistically sound to pool cancer data from all 3 studies, SEAS, SHARP, and IMPROVE-IT, in which case there are significantly more cancer deaths with ezetimibe treatment than placebo or simvastatin (Zocor) treatment. According to Forbes, cardiologist Steven Nissen, who's made a national name for himself by performing inherently problematic meta-analyses, advocates the act; statisticians Peto and Berry advise against it—possibly because of substantial differences in the designs of these studies.

Debate will undoubtedly continue, given the results of the SEAS trial and a fervent interest in anything remotely anti-Vytorin. In the meantime, Merck and Schering-Plough are left with the very difficult task of proving a negative—namely that ezetimibe is not associated with an increased risk of cancer or cancer deaths. Even if clinical doubt could be removed by positive, final data from SHARP or IMPROVE-IT, these data won't be available for at least 2 years.

IMPROVE-IT = Improved Reduction of Outcomes: Vytorin Efficacy International Trial; SHARP = Study of Heart and Renal Protection.

* Berry was cited in a previous post on drug-testing stats.