Accusations of Bias Among FDA Rosiglitazone Panel

September 14, 2010

September 14, 2010 (Silver Spring, Maryland) — Accusations of bias among the FDA advisory committee that voted on the fate of rosiglitazone (Avandia, GlaxoSmithKline) earlier this summer have been made by two members of the FDA's Office of Surveillance and Epidemiology [1].

In a letter to BMJ published online September 7, 2010, Drs David Graham and Kate Gelperin point out that half the members of the July 2010 advisory panel had previously voted in favor of allowing rosiglitazone to stay on the market at an earlier FDA advisory committee in 2007. They write: "In an unprecedented move, the FDA's Center for Drug Evaluation and Research (CDER), which originally approved rosiglitazone and has defended its continued marketing, invited not only the current members of these committees [the Endocrinologic and Metabolic Drugs Advisory Committee and the Drug Safety and Risk Management Advisory Committee] but also all members from the 2007 meeting, even though they were no longer active members of either committee. Of the 32 advisers who voted at the 2010 meeting, 16 (50%) attended the 2007 meeting, and 15 of them had voted that rosiglitazone remain on the market (one attendee was a temporary nonvoting invitee)."

Graham and Gelperin say this biased the vote at the 2010 meeting. They calculate that members voting for the first time were 4.4 times more likely to vote that rosiglitazone be withdrawn from the market than were members who had voted previously to keep it on the market. "Had these former members not been included, the vote would have been 10 out of 17 (59%) in favor of rosiglitazone withdrawal, with an additional three in favor of severely restricted distribution," they add.

Panel "Loaded"?

Commenting on this for heartwire , Dr Steve Nissen (Cleveland Clinic, OH), the coauthor of the meta-analysis at the center of the rosiglitazone safety concerns, who has been leading the campaign to get rosiglitazone withdrawn, said it was obvious that the 2010 panel was "loaded."

They have rigorously defended rosiglitazone again and again, and obviously they feel they cannot be seen to have made such an error.

He added: "The outcome of FDA advisory panels is significantly influenced by who is on the panel. When I saw the roster, I knew it was going to be an uphill struggle. I was very surprised that they had invited back people who had already voted to keep the drug on the market in 2007. Once you have opined publicly, it is very difficult to turn around and admit you made a mistake." Nissen said it would have been the leadership at the CDER that decided who to invite to this panel meeting, and this is another demonstration they do not want to act against this drug. "They should have acted much earlier on, when the first negative data were reported. But they have rigorously defended rosiglitazone again and again, and obviously they feel they cannot be seen to have made such an error."

Another who has a similar view is Dr David Juurlink (University of Toronto, ON), who has also published research suggesting rosiglitazone is harmful. "Of course the inclusion of these older panel members biased the outcome. This is consistent with the widely held perception that the powers that be at the FDA do not want to take rosiglitazone off the market, because they were responsible for allowing it onto the market in the first place, and then not doing anything when the reports of cardiovascular adverse effects first came to light. There was probably enough information to remove it in 2003, but now it's 2010 and the drug's still here."

This is another element in this whole debacle that should make people stand back and ask what exactly is going on at the FDA.

Juurlink suggested that the number of voting options given to the FDA panel at the recent meeting also raised suspicions of bias. "Why didn't they just ask whether the drug should be withdrawn--yes or no--as they did at the 2007 meeting? This time they were given five different options to choose from. This means it is more difficult to get a straightforward withdrawal vote and allows the result to be spun. So now the sponsor is saying that the committee voted to keep the drug on the market, when in actual fact 22 of the 32 members voted for withdrawal or strong restrictions," he commented to heartwire .

He added: "This is another element in this whole debacle that should make people stand back and ask what exactly is going on at the FDA. Decisions by the FDA should be made on scientific grounds and for the public safety, not for the best interests of the people who work there. The FDA is really dragging its feet on this. Even from the July meeting, it is now two months later, and nothing has happened except that more people have been injured and undoubtedly some will have died because the drug is still in use."

Kaul Defends FDA Position

But Dr Sanjay Kaul (Cedars Sinai Medical Center, Los Angeles, CA), who sat on the 2010 FDA panel and has questioned Nissen's meta-analysis, defended the FDA's position. He suggests the claim that the former panel members did not come to the meeting with an open mind, ready to be persuaded by the data, is itself a "biased assumption." He cites the quotation: "I think the signature feature of bias is that it is only detectable when it is present in a flavor different from one's own."

Kaul adds: "The authors should remain open to the possibility that the evidence remains inconclusive to adjudicate the cardiovascular risk associated with rosiglitazone" and that "it is not unusual for the FDA to invite experts even though they are no longer active members of either committee."