A Guide to the Competing Storylines at the FDA's Hearing on Glaxo's Diabetes Drug

It almost doesn't matter whether the FDA will withdraw GlaxoSmithKline (GSK)'s Avandia diabetes drug from the market or not. The brand has received so much negative publicity it's hard to imagine that many doctors will be prescribing it for new patients, and plenty of existing Avandia patients will now want off the drug.

An FDA advisory committee meets July 13 and 14 to decide the drug's fate. As the New York Times reported Saturday, Thomas A. Marciniak, an FDA cardiovascular expert, has filed a devastating slideshow in advance of the meeting that describes "a dozen instances in which patients taking Avandia appeared to suffer serious heart problems that were not counted in the study's tally of adverse events."

That's one of many storylines that will be debated at the meeting. Those storylines are important: Although the decision ought to be made purely on the basis of whether the statistics say Avandia is safe or not, the fact is that anyone who has propped their eyelids open with matchsticks to endure a conference room PowerPoint session knows that the "winning" presentation is the one that tells the best story.

Steven Nissen of the Cleveland Clinic, GSK's nemesis in the Avandia fight, knows this. His slideshow suggests that the hearing will be a Zapruder-esque experience in which the panel will hear several competing versions of reality. One of his slides is titled: "Seeing the Forest Plots Through the Trees." It states, "GSK will present its own meta-analysis, which will differ from the FDA and our independent analysis":

He's right, of course, but then that was an easy guess. The first issue the FDA asks the panel to address is:

- the potential for publication bias

But there's something missing. There's no chart showing only cardiovascular event hospitalization rates for Avandia vs. the control drugs. There are charts for "CV hospitalization or CV death," "Time to all-cause death," and "Time to CV death," but no single graph isolates from the database the people who had a heart-attack, went to hospital, and lived to tell the tale.

Unfortunately for GSK, Marciniak has such a slide, titled "Time to First HF Hospitalization," where "HF" stands for heart failure:

Mushing the two together (in a statistically inappropriate fashion), it seems that while people may live just as long on Avandia as on a competing regimen, the Avandia folks will have extra heart attacks (that they survive) along the way.

The other Matlock moment you're likely to see at the panel is FDA reviewer David Graham's slideshow. You'll remember that Graham, like Nissen, is an Avandia dissident and both believe that the Avandia experience highlights the need for reform at FDA to split up drug approval from drug marketing. Graham's slides compare Avandia (rosiglitazone) to Takeda's Actos (pioglitazone):

Those diverging lines indicate a slightly higher risk of CV events for Avandia. Graham, Nissen and Marciniak will likely argue that even though the statistical difference may appear small on their charts we're talking about millions of patients and therefore thousands of "differences."

At that point, a fat lady will sing and Avandia will be removed from pharmacy shelves.

Related:

• The Diabetes War: Avandia's Goose May Be Cooked, But Are the Chefs' Kitchens Clean?
• Heart Association Finds Drug Risks "Inconclusive" After Taking $3.6M From Its Maker Glaxo
• Glaxo Strikes Back at Drug Critics, But It's a Big Swing and a Miss
• Secret Tapes Lay Bare Glaxo's Scorched-Earth Corporate Culture