Dr. Steven Nissen, head of cardiovascular medicine at the Cleveland Clinic and a central figure in the debate over the diabetes drug rosiglitazone (Avandia), has a hard-hitting commentary at the Forbes blog:
Dr. Nissen published a meta-analysis in the New England Journal in 2007 claiming that rosiglitazone increased the risk of having a heart attack by 43%. It later turned out that the drug maker, GlaxoSmithKline, had done its own in-house analysis that more or less confirmed these figures. At first GSK sat on the data but eventually they did notify the FDA's Center for Drug Evaluation and Research. As FDA critics have long noted, the CDER has a sort of structural conflict of interest, being responsible both for approving new drugs and then for assessing the safety of drugs on the market. If they later find a drug unsafe and having to be restricted or removed from the market, they are basically admitting that they blew it the first time around when they approved the drug.
Dr. Nissen says: "Initially, GSK withheld the internal analysis from the FDA, but in 2006, the company informed CDER of the findings. FDA statisticians confirmed the risks, but, incredibly, CDER and GSK agreed privately to conceal this hazard from patients and practitioners....[After Nissen published his analysis in NEJM] The leadership of CDER was intensely embarrassed by these revelations and furious with us for publicly challenging the safety of Avandia (and indirectly the competence and integrity of CDER). The FDA appeared incredibly insensitive to the welfare of patients. GSK knew, FDA knew, but patients and physician were not warned."
Later, GSK came forth with data from a new trial called RECORD, which they claimed absolved rosiglitazone from charges of causing excess heart disease. CDER latched onto RECORD and hoped to use these findings to claim that they had not done anything wrong in concealing the risks of the drug. As Dr. Nissen recounts the story, this all backfired when FDA's own internal analysis of RECORD showed that the trial was so flawed as to be unreliable. As I commented in an earlier post on the Nissen case:
--it appeared that GSK itself invalidated RECORD by inappropriately breaking the blind early, trying to mine the RECORD data for early signs that Nissen's meta-analysis was off base. In the end the FDA's own statisticians agreed with Nissen and dismissed RECORD.
Despite all this, CDER has announced a new round of advisory hearings on rosiglitazone, despite the fact that after all the negative publicity only a handful of patients are still taking the drug. The hearings basically exclude anyone who might say critical things about the RECORD trial or the drug. Dr. Nissen's charge: none of this has anything to do with protecting the public health, but it has everything to do with a small, entrenched group of careerists within CDER leadership protecting their supposed reputations by insisting they did nothing wrong back in 2006.
He concludes: "We have endured a series of drug and device safety disasters (Vioxx, Avandia, Ketek, defective defibrillator leads, and recently, compounding pharmacy oversight, and the list goes on) that have harmed many of our fellow citizens. In each case, the permanent FDA bureaucracy has sought to deny responsibility and often failed to learn the appropriate lessons necessary to prevent future catastrophes. In the middle of the sequester, CDER is willing to spend a large sum of taxpayer dollars to conduct a 2-day advisory meeting on a drug nobody uses, for the sole purpose of absolving its own bureaucracy of responsibility for a terrible drug safety tragedy. The current Director of CDER, Janet Woodcock, has directed this division for nearly 20 years, increasingly insulated from the Agency’s true constituency, the American public. If CDER is allowed to re-write the history of Avandia, this vital FDA Center will continue to function as an unsupervised, self-regulating bureaucracy, accountable to no one. That’s just unacceptable when the public health is at stake."
Hat tip to Marilyn Mann for sending along this link.