Participants in the ongoing safety debate over GlaxoSmithKline's Avandia (rosiglitazone) are refining their arguments ahead of the July 13-14 joint FDA advisory committee meeting on the troubled thiazolidinedione's future.

Recent publications and public forums have served as a preview of the rhetoric expected at the Endocrinologic and Metabolic Drugs and Drug Safety and Risk Management advisory committee meeting, which promises to be the most anticipated drug review event of the year and one that could have ramifications for years to come on FDA's drug safety oversight.

In late June, Avandia critics took to the medical literature to bolster their arguments against keeping rosiglitazone on the market, with publication of studies by FDA's David Graham and Cleveland Clinic's Steven Nissen that will be showcased at the advisory committee meeting. Regardless of how much credence the studies are given by the committees, the extensive pre-meeting publicity spawned by their publication has escalated the pressure on the agency and its experts.

In a study published online in the Journal of the American Medical Association June 28, Graham, FDA's Office of Surveillance and Epidemiology associate director for science and medicine, and co-authors from CMS, Stanford University and Acumen LLC reported findings from an observational, retrospective study of more than 227,000 Medicare Part D beneficiaries who started taking Avandia or Takeda's competing thiazolidinedione Actos (pioglitazone) and were followed for up to three years.

Avandia was associated with significantly increased risks of stroke (27 percent), heart failure (25 percent), death (14 percent) and a composite of acute myocardial infarction, stroke, heart failure or death (18 percent). But rosiglitazone was not associated with an increased risk of acute MI compared to pioglitazone.

The number needed to harm for the composite endpoint was 60 persons treated for one year to generate one excess event.

The study, which Graham is expected to present at the advisory committee meeting, gives the OSE associate director another jumping off point from which to argue that Avandia should be withdrawn from the market.

At the 2007 advisory committee review of Avandia's safety, Graham presented data from a head-to-head study that showed a non-significant 3.5-fold increase in cardiovascular risk with Avandia compared to Actos. He also said a Takeda meta-analysis of clinical data showing a 25 percent overall decrease in cardiovascular risk with pioglitazone bolstered his argument that Avandia should be withdrawn because it conferred no unique benefit over Actos while posing a greater risk.

Presentation of the data, which had not yet been analyzed by the Office of New Drugs, reflected what was described as a "fundamental disagreement" between OND and OSE on conclusions that could be drawn on the two drugs' relative safety (¹ Also see "CDER Offices Go Public With “Fundamental Disagreement” Over Avandia" - Pink Sheet, 6 August, 2007.). The advisory committee took a dim view of the unvetted data.

Interactions With Fibrates

At the July re-review, the advisory committee will not hear the results from an OSE epidemiological study, using the same Medicare Part D cohort examined in the JAMA report, that investigated the cardiovascular safety of Avandia and Actos when taken concomitantly with fibrates. Although the study's existence was mentioned in a June update to the agency's PDUFA IV drug safety five-year plan, analysis of the data is not expected to be completed until the fall and will not be presented at the July 13-14 meeting.

Graham's JAMA study was accompanied by an editorial from David Juurlink, Sunnybrook Research Institute, Toronto. "Whether rosiglitazone and pioglitazone really do have different cardiovascular safety profiles is an intriguing question but one with a misplaced focus," he wrote. "Accumulating concerns about rosiglitazone make it difficult to advance a cogent argument regarding why, exactly, a patient might want to receive the drug or why a physician would choose to prescribe it when there is an available and quite possibly safer alternative."

Juurlink has previously established himself firmly in the anti-Avandia camp. In a May letter coauthored with Public Citizen Health Research Group Director Sidney Wolfe, Juurlink asked FDA Commissioner Margaret Hamburg to halt GSK's ongoing TIDE (Thiazolidinedione Intervention in Vitamin D Evaluation) trial comparing Avandia and Actos due to "accumulating evidence of the increased dangers" of rosiglitazone relative to pioglitazone and other older antidiabetic agents (² Also see "GSK Settles Hundreds Of Avandia Suits Before FDA Advisory Cmte. Meeting" - Pink Sheet, 17 May, 2010.).

Nissen Strikes Again

Also back for the second round of advisory committee debate is the Cleveland Clinic's Nissen and Kathy Wolski, who are largely credited with drawing public scrutiny to rosiglitazone's safety with a 2007 meta-analysis of Avandia studies which linked the drug to a significant 43 percent increased risk of MI and a non-significant 64 percent increase in CV mortality.

Nissen and Wolski published an update to their 2007 meta-analysis online in the Archives of Internal Medicine on June 28. The updated report includes results with and without the RECORD (Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycemia in Diabetes) trial, which were announced in 2009.

The authors found a significantly increased risk of MI for rosiglitazone ranging from 28-39 percent, with a number needed to harm of 52 or 37 (with and without the RECORD data included). No significant increase was seen in CV death or all-cause mortality.

The authors' 2007 study was attacked by some as biased because it excluded clinical trials in which there were no events. Anticipating these criticisms again, Nissen and Wolski conducted an alternative analysis in which smaller studies were pooled by randomization ratios, which allowed for inclusion of a total of 56 studies regardless of the presence or absence of CV adverse events. This analysis however, yielded similar results and showed a significant increased risk of MI.

GSK's statement in response to the Graham and Nissen studies also serves as a preview of the evidence upon which it will rely in defending Avandia's safety to the advisory committees.

The company points out that six controlled clinical trials, including RECORD, reported since the 2007 advisory committee review collectively show rosiglitazone does not increase the overall risk of heart attack, stroke or death. In RECORD, an open-label study evaluating Avandia compared to metformin or sulfonylurea as add-on therapy for patients who failed monotherapy, there was no increase in the MACE composite endpoint of death, MI or stroke.

An updated company meta-analysis encompassing 52 trials does not show an increase in myocardial ischemia, GSK said. "There have also been several new observational studies published since 2007. Review of these studies with tighter confidence intervals have risk ratios that are very close to one, indicating no difference in the risk of MI between rosiglitazone and pioglitazone."

GSK will face pressure in defending RECORD at the advisory committee meeting. Critics have said the open-label study, designed in conjunction with European regulators, was underpowered, a point that panel members are sure to raise (³ Also see "GSK Stands By Avandia: Seeks Label Change, Asks Docs To Give It Another Try" - Pink Sheet, 22 June, 2009.).

Nissen reiterated his criticism of RECORD during a debate with David Orloff, former director of FDA's Division of Metabolism and Endocrinology Products, at the American Diabetes Association meeting in Orlando, Florida in late June.

"If you try to do everything wrong in a clinical trial you'd get RECORD," Nissen said, according to a transcript of the question-and-answer session provided by the consulting group Close Concerns. "It was open-label and unblinded. Then the company, when the concern comes about, does an unplanned interim analysis and publishes it - now every patient and doctor involved knows the results. Half the patients randomized to therapy aren't even taking the drug by the end of the study, yet it's intent-to-treat. They didn't report the crossover. It's going to come up in a couple of weeks in the panel meeting. Anybody who looked deep into the trial knows you couldn't do the trial worse."

It remains to be seen how much stock the advisory committee members will put in the Graham and Nissen studies, which may carry little weight since they are not randomized controlled trials and present mixed findings on MI risk. Nevertheless, their publication has heightened the pressure that both the agency and its advisors will feel in their deliberations.

Nowhere was this more evident than in a statement by Rep. Rosa DeLauro, D-Conn., who chairs the Appropriations Subcommittee on Agriculture, which controls FDA's purse strings. Commenting on the JAMA study, DeLauro made clear what she thinks FDA should do.

"Coupled with the previous findings, these new studies appear to confirm that Avandia is dangerous and should be pulled from the market. The evidence now seems to be overwhelming that taking Avandia significantly increases a patient's risk of a heart attack, even when compared to a similar prescription drug. This information will make it even more difficult for the FDA to justify keeping this dangerous drug on the market."

- Sue Sutter ( ⁴ s.sutter@elsevier.com )

Avandia Safety Rhetoric Ramps Up Ahead Of Advisory Committee Review

More from Archive

More from Pink Sheet