WebMD Newsroom

Wow. The usual FDA advisory panel meeting is a sleep-inducing sequence of dreary data and appreciative nods from one insider to another.

Today’s meeting was not your usual FDA advisory meeting. Comparing it to previous meetings is like comparing a three-ring circus (elephants and all) to a too-long sermon.

The two-day meeting will decide whether the diabetes drug Avandia is pulled from the market for safety reasons — or whether, as GlaxoSmithKline would like, the drug should be shorn of its black-box warning label and left on drug-store shelves.

Kicking off the meeting was FDA commissioner Margaret Hamburg, MD. FDA commissioners almost never attend these things, but Hamburg was there with what would have been a surprise announcement, had not briefing documents been released last week: The FDA is split over the question. It’s not just asking the panel for help; it’s pleading with it to heal the federal agency’s rift between overcautious safety experts and overzealous drug approvers.

Did I mention elephants? The elephant in the room is the RECORD study, just finished last year. Like the fabled elephant surveyed by the fabled blind men, RECORD feels very different to people standing in different spots.

GSK says the study proves not only that Avandia is safe, but that it’s safer than the longstanding diabetes drugs metformin and sulfonylurea. Top GSK scientists made their case to the panel with statistics and tables and talking points.

“I find it hard to see how one could conclude that [Avandia] increases cardiac events,” said Philip Home, DPhil, MD, chair of the RECORD trial and a diabetes researcher at the U.K.’s Newcastle University.

But wait! Next to the podium is Steven Nissen, MD, chair of cardiology at Cleveland Clinic and the guy whose searing looks at GSK data forced the FDA to hold this meeting. Nissen, too, had statistics and tables and talking points. He also told of data long kept under wraps by GSK and the FDA, of a New York settlement that forced GSK to post its data on a website, and of consistent data pointing to an increased risk of heart attacks in people taking Avandia.

“What is the cost that if these estimates are correct, if there is a 28% to 39% increase in heart attacks?” Nissen asked the panel. He calculates that if one accepts the RECORD data, there will be one heart attack for every 52 people treated with Avandia. If one does not accept the trial, there may be a heart attack for every 37 people treated.

But wait, there’s more. Nissen was followed by FDA researcher Thomas A. Marciniak, MD, who spent six months painstakingly going over about an eighth of the actual case reports on patients in the RECORD trial. Marciniak pointed to heart attacks that were not recorded, patients listed as on treatment who had left the trial, and suspicious possible heart attacks that never were investigated. The biggest problem: The study was never blinded, so investigators — and the sponsor — knew which patients were taking Avandia and which weren’t before they analyzed the data.

“The RECORD trial was inadequately designed. You cannot rely on it for reassurance about safety,” Marciniak said. “And it does suggest [Avandia] increases the risk for heart attacks.”

Marciniak says he thinks the problem was with RECORD’s extremely poor design, and not with dishonesty on the part of the researchers. Even so, his presentation was electric. Before he finished, he was leaning toward the table where the GSK researchers sat, lecturing to them.

Ellis Unger, MD, deputy director of the FDA office of drug evaluation, followed Marciniak and, while praising him as a careful researcher, noted that his work cannot be taken as the final word on the subject, as much more study would have to be done to examine the entire RECORD database.  Unger said he felt much of the RECORD data is useful, even reassuring.  That drew a sharp remark from panel member Sanjay  Kaul, MD, of  Cedars-Sinai Medical Center, Los Angeles.

“The key question is whether the FDA review uncovered  deliberate attempts to cover up harm from [Avandia]. Is it malfeasance?” Kaul asked.

Presentations will continue today; the panel discussion and vote is slated for tomorrow.

Are you concerned about the FDA process for drug approval? Do you have questions about Avandia? Share your comments and questions with the Diabetes Community.