Avandia and Rezulin: Parallels that Should Make GSK Nervous

History doesn’t repeat itself but it does rhyme. That old Mark Twain saying must be making GlaxoSmithKline sweat as Avandia is starting to look more and more like another Rezulin. By our reading of the tea leaves, Avandia is in much more peril than anyone seems to realize.

GSK is hunkering down for continuing assaults on its number two drug, battered initially by Cleveland Clinic’s Steve Nissen whose meta-analysis showed a 43% increase in heart attack risk for Avandia patients compared to control.

During the company's second quarter conference call, CEO JP Garnier clearly was using the "If you sound like a winner, you are a winner" strategy when it came to discussing Avandia with investors, analysts and media.

"We are still encouraged [about Avandia] because we have seen...a lot of evidence recently," Garnier said of the data GSK has submitted to the agency in advance of a Monday advisory committee fact-gathering meeting. "The evidence is supportive of Avandia's risk/benefit ratio, and of its effect on cardiovascular safety."

GSK is hyping a 400,000-patient epidemiology study of patients on Avandia and Takeda's Actos among other treatments that apparently bodes well for the diabetes drugs.

Glaxo has said that they simply have been unsuccesful in boiling down their message on Avandia to a "7-second soundbyte" which is the reason for the more than 45% decline in new Avandia scripts. " In the US, the media has ... had more of an impact on physician and patient impressions than the data itself," GSK's pharma operations chief David Stout said on the call.

Clearly, the message from GSK is: We stand behind Avandia. Unfortunately for the company, there are some discouraging parallels between their diabetes drug and Warner-Lambert's Rezulin.

Warner-Lambert pioneered the glitazone class, but Rezulin caused liver toxicity that ultimately led to its withdrawal. A recap of the regulatory history suggests some uncomfortable parallels with Avandia and the concerns about cardiovascular safety.

Two months after it got to market in 1997, FDA slapped Rezulin with a stricter warning on its packaging (thanks to 35 post-marketing reports of liver injury). At that point, 500,000 patients were already on the drug. Several "Dear Doctor" letters later, FDA's Endocrine and Metabolic Drugs Advisory Committee reviewed the liver tox issues, and recommended keeping Rezulin on the market, but only for patients not well-controlled on other diabetes drugs. One year later, the drug was taken off the market when reports kept coming in.

For Avandia, the toxicity is different--cardiovascular rather than liver--but the slow motion, repeated regulatory reactions are similar.

Avandia labeling was rewritten to strengthen cardiovascular safety warnings in 2001, and the company issued a "Dear Doctor" letter on the topic at that time. The concerns were raised more directly in the context of the review of Avandia for an indication for use with insulin; that use was ultimately approved in 2003. The Nissen paper now has put the regulatory machinery into fast forward, and an advisory committee will discuss Avandia's fate on Monday.

All of that is uncomfortable enough, but there is another parallel to the end of Rezulin emerging at the worst time for GSK: a Senate Committee is raising concerns that FDA reassigned a medical officer who wanted to put stronger warnings on Avandia.

If that sounds familiar, it should. In early March 2000, FDA senior medical officer Robert Misbin wrote a letter to Rep. Henry Waxman (D-Calif.) expressing frustration over FDA's handling of Misbin's attempts during the previous two months to convince the agency's Center for Drug Evaluation & Research that Rezulin had to be withdrawn from the market. Misbin asserted that FDA officials had stopped him from releasing information related to deaths of Rezulin patients.

One other thing: Misbin was the primary reviewer on Avandia and was taken off of the review several years ago. (Apparently, he's not the whistleblower this time around -- for more speculation on who the whistleblower might be, see the next post.)

All in all, Monday's advisory committee meeting doesn’t look good for Avandia. Even if the medical officers keep quiet, FDA will not be presenting a united front to the committee. That's because the agency is once again going to let its most prominent whistleblower, director for science and medicine in OSE David Graham, make a formal presentation. Graham most recently helped ensure that Merck's Arcoxia died a painful public death before an FDA advisory committee. FDA has apparently concluded that they have to let Graham speak at these meetings rather than wait for him to go to Congress to make his presentations. (Here is our coverage of the Arcoxia debacle.)

On Avandia, Graham has already made his position clear in FDA briefing documents. He argues that the current postmarketing studies (in particular the key RECORD study) can't, statistically, demonstrate a heart attack risk related to Avandia: they're underpowered. In other words, the current scientific evidence is all FDA is going to get to make their decision on the future of GSK's drug. Anyone want to venture a guess at where Graham will stand on Avandia?

And its not like FDA won't let the discussion go into whether the drug needs to be pulled. Quite the opposite. Here is one question posed to the committee: "Does the overall risk-benefit profile of Avandia support its continued marketing in the US (VOTE requested)? If yes, please comment on what FDA should do to maximize the risk-benefit considerations (e.g., limit to certain patients, incorporate a boxed warning….)"

That question means FDA is thinking awfully hard about whether this drug should stay on the market. You could argue that they added the question for political cover in order to leave it on the market, but I'm not buying it. I think they really want to know the experts' opinion.
And what will that opinion be? Nissen himself has said Avandia should remain on pharmacy shelves. NIH's Malozowski told us that he didn’t think FDA would pull it. “They will probably add a warning for a subpopulation of patients and a contraindication for its use with insulin."

I also asked Tom Garvey, a former FDA reviewer who runs his own drug development consulting business, what he thought. He concurs with Malozowski. Sort of.

Rezulin could be pulled off the market with less risk, he argued, because there were two other marketed drugs without Rezulin’s liabilities. Moreover, "the absolute risk found by Nissen is small (if, indeed, it exists) and the benefit conferred by Avandia is not inconsequential, especially in certain types of type II diabetics.”

But then he added, surprisingly: “All of this having been said, I too get the sense that the drug is probably doomed."

In short, as with Rezulin, an FDA advisory committee could recommend keeping Avandia on the market, in a limited way—while, in parallel, the political and historical momentum builds to yank it off. On the scientific front, the data isn’t clear. Nissen's meta-analysis has come under intense fire, but his results were confirmed by FDA's own meta-analysis, and they had access to a much larger data set. Meanwhile, GSK's RECORD study has been inconclusive on the heart attack risk question.

But the political front will evolve in its own way. And if history really is rhyming, if not precisely repeating itself, FDA will have a hard time keeping Avandia on the market.