The Pink Sheet, the venerable trade publication, conducted an interesting question-and-answer interview with the controversial Cleveland Clinic cardiologist after the recent FDA advisory panel meeting, in which new, but unspecified restrictions were recommended for the Glaxo diabetes drug. This is an excerpt of what he had to say about clinical trial design, advisory panels and safety issues….

The Pink Sheet: What is the proper role of meta-analyses in drug safety oversight, and is that role changing?
Steven Nissen: A meta-analysis is a fine way to assess the safety of a drug - it’s not necessarily the best way to assess the efficacy of a drug. I actually thought the committee maybe didn’t understand that. [To] put it into perspective, what the FDA did in their meta-analysis is they essentially took every randomized controlled trial done with the drug and so it’s an integrated analysis of everything we know about the drug. And if there was ever a way to do safety, that’s the way to do safety.”

“Because if you take everything together about the drug and lump it together and you see, as they did here, a 40 percent increase in the risk of heart attack, then you can be pretty sure there is an increase in the risk of heart attack. And so in general my advice would be to the FDA, don’t give a label claim for efficacy on the basis of a meta-analysis, but for safety purposes consider it to be very powerful and very useful. And of course the FDA, when they want to, agrees with that. They pulled [Novartis' irritable bowel syndrome drug] Zelnorm off the market … because there were 13 myocardial infarctions in a meta-analysis of Zelnorm compared to one or two I think in the control groups. So it was a significantly significant excess risk, but the number of cases was very, very small and it was considered sufficient to actually pull the drug. So I think that it’s the right thing to do, to use a meta-analysis for that purpose. It is also important obviously to have some confirmatory evidence when you can. The problem of course in the Avandia case is that the company never did a definitive outcomes trial. And so there wasn’t any way to refute or support the results of the meta-analysis, and that was what the committee was struggling with.

The Pink Sheet: Curt Furberg has suggested that advisory committees have too many clinicians represented, while there are too few drug safety experts, and you’ve made similar comments in the past. What would you recommend for establishing an appropriate committee roster?
Steve Nissen: The Office of New Drugs, which just really didn’t want to go there, were the ones that created the panel (for the Avandia meeting). They did, in fact, have relatively few people that I would consider to be drug safety oriented on the panel. There certainly were outstanding physician-scientists on the panel, but it is very tough if you’re an endocrinologist and you’ve been using this drug for years to look at it and say ‘OK, well, I’m going to vote to remove the drug from the market.’ … At the COX-2 inhibitor panel in 2005, the panel members voted 15-13 that the benefits of Vioxx exceeded its hazards - after the drug had actually been withdrawn from the market. So to my knowledge, no advisory panel ever has actually voted to withdraw a drug from the market. And that tells me something about what we may need to do to get panels that are perhaps a little more balanced towards the safety side in the future.

* a subscription is required to read the full interview and the rest of the publication