Avandia, once the best-selling diabetes pill on the planet, is now given only to a few thousand Americans. In much of the rest of the world it has been pulled from the market because of analyses that showed it might increase the risk of heart attacks. The controversy over the drug changed the way the Food and Drug Administration regulates diabetes medicines, resulting in steep requirements that new drugs prove they are unlikely to cause heart side effects.
Tomorrow, three years after an FDA advisory committee voted that Avandia's use should be sharply curtailed or that the drug should be pulled, a new FDA advisory committee will take two days to reconsider that opinion. It's possible that Avandia could go back into wider use, or that it could be withdrawn from the market entirely. Whatever happens could have far-reaching consequences for diabetes-focussed companies like
Why is the FDA re-considering a decision that to all the world seemed over and done with? Doing so is politically risky, and exposes the agency's internal battles. But the FDA says in briefing documents released yesterday that it committed to look at the Avandia matter when more data became available, and so it will.
Avandia lost last time around in large degree because of one of FDA drug safety reviewers, a doctor named Thomas Marciniak. The questions about Avandia had originally been raised by Steven Nissen, a Cleveland Clinic cardiologist who also crusaded against Vioxx, who also presented at the last Avandia meeting. But Nissen's alarm was bolstered by Marciniak's analysis on the day of the panel, which eviscerated a large clinical trial conducted by
RECORD was meant to be the refutation of Nissen's analysis, which pooled small trials comparing Avandia and placebo, and of a similar analysis by the FDA that also showed a heart risk for Avandia. But the trial had issues that everyone agrees on. The largest of them is that doctors and patients knew which drugs each patients were getting, which can allow bias to creep in.
But Marciniak argued the flaws went far deeper than anyone thought. He went back to original case reports from RECORD and argued that cases that should have been seen as cardiac deaths or adverse events in the Avandia group had not been recorded properly. This led the FDA panel to nearly completely disregard the RECORD trial.
The 2010 panel was asked to vote on a scale from A to E what should happen to the drug, with A meaning removal of safety warnings and E meaning complete withdrawal from the market. Nobody voted for A. Three voted for option B (do nothing), 7 for option C (add warnings), 10 for option D (restrict use), and 12 for option E (yank it), with one abstention. The largest single vote was for withdrawal, and 80% of panelists were for dramatically restricting or removing the drug.
After the meeting, the FDA asked GlaxoSmithKline to fund an independent group to look at the RECORD data again, patient record by patient record, and to try to determine if the study could be salvaged. Glaxo chose the Duke Clinical Research Institute at Duke University, one of the most respected institutes when it comes to the conduct of clinical trials. The Duke re-review found, once again, that Avandia did not increase the overall rate of death in RECORD.
Now the FDA's two-day meeting is shaping up as something like an action movie, with Marciniak playing the lone wolf and the FDA and Glaxo playing the army he must fight. In briefing documents made public yesterday, another FDA reviewer accused Marciniak of not being truly “blinded” in his review – that is, of looking for problems with Avandia patients but not with others. That's a serious charge, and if it were true it would invalidate Marciniak's analysis.
Marciniak came out swinging in his own memo. He called Duke's integrity as an institution into question, noting the scandel around Anil Potti, a genetics scientist whose work was allegedly fabricated. He said others at the FDA had been unwilling to recognize serious problems with Duke. Marciniak and the rest of the agency seem to be disagreeing frequently; last week, the Wall Street Journal published an account that says that Marciniak has been arguing that commonly used blood pressure medicines called angiotensin receptor blockers (ARBs) raise the risk of cancer, again putting him in conflict with others at the regulatory agency.
The big question is how the new 27-member panel of doctors and other experts will view the RECORD data. Fourteen of the experts on the panel were on the panel last time around, and these people were more favorable toward Avandia than other experts who will not be part of the deliberations this time. This time, Nissen will not be there to help make the argument; the FDA decided that he did not need to be part of the day's program.
Two of the three B votes (Drs. Felner and Moss) are returning; along with five of the seven C votes (Drs. Burman, Oakes, Kaul, Van Belle, and Geller) and four of the 10 D votes (Drs. Day, Hammerschmidt, and Moratta, and patient representative Killian.) Although option E, to withdraw Avandia, was the most popular one last time around, only three of those experts are returning (Drs. Konstam, Heckbert, and Flegal.)
“We've confirmed that all current Special Government Employees (SGEs) that were voting members at the July 2010 advisory committee meeting were invited to participate at the June 2013 advisory committee meeting,” says Morgan Liscinsky, an FDA press officer, in a statement. “We cannot comment specifically on why individuals are not able to participate at advisory committee meetings.”
Thomas Fleming of the University of Washington, a statistician who has been important in several drug safety controversies who voted E, says that he was invited to present at the meeting but then told that the FDA would not give him a waiver for a conflict of interest. “I would think that when you had people that are involved in a process that is being reiterated that there is particular value to having some of those key players retained,” Fleming says.
Panel chair Clifford Rosen and Richard Platt, both of whom voted negatively with regard to Avandia, both say they were invited to the meeting but could not attend. Platt initially said he had not been invited, but then emailed back that he actually had been but could not attend. Another panelist, Lewis Nelson, a drug safety expert at New York University, initially said that he had not been invited, but then said he could not remember for sure. The FDA says it has documentation that the panelists were invited.
Curt Furberg of Wake Forest University, a fiery drug safety advocate, says he was not invited by the FDA. Furberg is not currently a special government employee, according to the FDA.
Nelson says that he does not believe there is any reason for a “conspiracy theory” about who was invited to the panel. “The FDA really seems to be on the up and up,” he says. “They really seem very nonconflicted and unbiased,” Nelson says.
