The Harvard prof who wrote the book on FDA (literally) deconstructs the decision on Avandia—and its future implications.
GlaxoSmithKline’s diabetes drug Avandia became mired in such controversy regarding its safety in recent years that it was dubbed “another Vioxx”—exactly what the pharmaceutical industry had vowed never to produce. Two weeks ago, FDA’s decided to restrict access to Avandia, earning it “drug of last resort” status. This long-awaited decision was expected, but there were many unexpected aspects to the way the agency made its decision public.
Pharmaceutical Executive asked Daniel Carpenter, a professor of government at Harvard University and the author of a big new book about FDA called Reputation and Power (Princeton University Press) that is getting a lot of attention in the press, for his take on how the Avandia story has played out—and what it may mean for both FDA and pharma in the future. Walter Armstrong
Pharm Exec: FDA decided to leave Avandia on the market and to impose a REMS that essentially turns it into a diabetes drug of last resort. (The drug already carried a black box warning about cardiovascular risks.) GSK has said it will no longer spend any money marketing Avandia, and most analysts agree that new prescriptions will likely come to a halt. So why didn’t the FDA, like the EMA, simply yank the drug?
Daniel Carpenter: I see the Avandia decision as a kind of withdrawal, with a very strong, long-term message for the pharmaceutical industry. Essentially 95 percent or more of the drug’s once-robust market is gone, and the decline is attributable almost entirely to post-marketing studies by third parties and a series of FDA regulatory decisions.
GSK’s decision to discontinue marketing for Avandia was not unrelated to FDA’s decision. My guess is that GSK had an implicit understanding that FDA would allow sales for existing Avandia patients without other alternatives, but only if marketing stopped to new patients.?The FDA can now use a REMS to carve out virtually all of a drug’s market while also satisfying medical libertarians by leaving it to be prescribed for the few who have no alternative. Some consumer advocates have decried the decision to leave the drug on the market. I see it differently: Avandia has been all but taken off the market, and a critical precedent has been established.
If I had told you four years ago that, in the absence of a single randomized clinical trial demonstrating greater cardiovascular risk for the drug, 95 percent of that drug’s market was going to be taken away by virtue of a meta-analysis and post-market epidemiology studies—plus some unsavory evidence revealed about a company’s clinical trial practices and failure to disclosure risk information—I think you would have bet against me.
Vioxx was pulled only when a randomized clinical trial demonstrated greater myocardial infarction risk. In some ways, given the evidence base, FDA’s decision on Avandia is a more aggressive regulatory action. It may well be the right one, but regardless, it’s momentous.
If I had to wager one other thought—based on this case’s circumstantial evidence—it’s that FDA’s leadership might be using Avandia as a try-out opportunity for a tougher REMS. Again, there is a real signal being sent to the drug industry that these REMS are not mere “management tools” but can be used to reduce a drug’s market to a fraction of the sponsor’s original ambitions.
Three aspects about the way the decision was communicated seemed surprising to me: First, FDA and the EMA made their announcements in a coordinated fashion.
Second, FDA posted on its website the memos written by John Jenkins at the Center for New Drugs, David Graham at the Center for Drug Safety, and other key officials, showing the dramatically conflicting views of their risk/benefit analysis of the drug.
Third, Commissioner Margaret Hamburg, Deputy Commissioner Josh Sharfstein, and CDER head Janet Woodcock co-wrote a piece in the New England Journal of Medicine explaining the difficulty of coming to a conclusive decision. Each of these actions was a remarkable departure from FDA tradition. They appear to speak to the new transparency at the agency, which seems to be a very high value of the Margaret Hamburg.
DC: I think you’re right about transparency being cherished by the new leadership. It’s being pursued for a number of reasons, not least because the new FDA leadership understands, in ways the past few commissioners did not, that the FDA’s credibility has been severely compromised in the last two decades.
The current FDA leadership also understands the symbolic importance of their actions, and I can only surmise that that understanding, plus the possible confusion caused by separate EU and US decisions, was the reason for the coordination with the EMA. ?There are, of course, risks to transparency. In this case, by not presenting a united front to the public and the scientific community, people will perhaps begin to view FDA as characterized by massive infighting. That said, in an age when anyone can tweet, blog, talk to a reporter, and so on, I think Commissioner Hamburg understands that there is a greater risk from trying to present a united front when there isn’t one.
Her policy of transparency serves other purposes, too. It now forces people like Jenkins and Graham to come clean and public with their strong views. Graham had been doing that already, but this policy requires him and other drug safety officials to articulate their views on the same platform (scientific and administrative memoranda) shared by Jenkins and Woodcock, the people whom they regularly oppose. Jenkins’ view that Avandia needed essentially no further regulation had been articulated internally in the past, but from now on it has to be communicated openly and immediately.
So this policy may make it more difficult for people like Jenkins and Graham to maintain rigid stances. To keep credibility, they may need to pick and choose their battles.
GSK attempted to refute the charge that Avandia caused more heart attacks and other CV complications than other diabetes treatments by saying that the two most damning studies were both meta-analyses, while their own RECORD trial, which didn’t show an increased risk, was gold-standard randomized controlled trial (RCT). (Of course, their defense was demolished when an FDA investigator revealed that some of the outcome data in the trial had been dropped or faked.)Do you agree that the two different kinds of studies are too different to be usefully compared?
DC: On the memo from FDA reviewer Thomas Marciniak severely criticizing the RECORD trial, I have claimed elsewhere (Pharmalot, July 2010) that its effect was to undermine GSK’s credibility in running trials. That could have long-term damaging consequences for the company, especially when combined with the New York Times reports and the evidence coming from the Senate Finance Committee. The issue is not so much whether New York Times readers trust the company’s trials, but whether advisory committees and FDA and EU medical reviewers trust the company’s trials.
This question—the comparability and relative value of RCTs versus epidemiologic studies—is one of the critical issues facing 21st century governance and science of pharmaceuticals. It’s clear that RCTs are still preferable for many questions, but it’s also the case that other kinds of evidence (such as pharmacoepidemiology) are becoming more rigorous.
The question of internal and external validity is also very important. I tell my students that perhaps the most important medical and public health finding of the 20th century—that cigarette smoking causes lung cancer and heart disease—came entirely from observational studies.
One other thing. If we are truly entering a world where we care not just about a single average treatment effect, but instead a battery of treatment effects that differ by phenotype or genotype, then we are likely entering a world where observational studies are more—not less—important to the scientific inferences we make.
How would you characterize GSK’s approach to the entire question of its drug’s association with CV events, especially heart attacks. The coverage in the New York Times, in particular, of the documents GSK gave to the Senate Finance Committee paints the picture of a company almost trying its hardest not to get to the bottom of the problem—in fact, using tactics like intimidation of critics, hiding negative data from the public and even the FDA, and possibly paying doctors to fudge clinical data. Is that the grim reality of pharma’s methods?
DC: Large pharmaceutical companies have a severe trust deficit right now—in the public, in academic medicine and science, and among regulators and public health officials. And I daresay that trust deficit has been richly earned.
Every time another memo is leaked suggesting that risk information has been suppressed or shrouded, every time a critic has been bought off or intimated, every time a clinical trial has been oversold or sloppily analyzed, the reputation of the entire industry suffers. It really is a case of one bad apple ruining the whole bunch.
In my scholarship, I’ve noticed this kind of associative property of reputations and trust, where people who had nothing to do with success or malfeasance are nonetheless implicated symbolically in the success or malfeasance by virtue of their common membership in an organization or an industry.
Part of the problem is that we all inhabit a world where institutions and organizations generally have lost legitimacy. My own “sector”—higher education and places like Harvard—are not immune from this critique. But the reputation of the pharmaceutical industry seems to have fallen in a different way, by a different magnitude. If I had to name the top two or three issues facing pharmaceutical managers and executives over the coming generation, restoring public trust through circumspection and transparency would have to be one of them. This issue isn’t going away.
(Pharm Exec’s November issue will contain an in-depth interview with Daniel Carpenter about his 10-year study of FDA for his book, Reputation and Power, as well as some of the intense responses the book has already received in the media.)