Waxman: Why did FDA Approve a Drug with no Health Advantages that May Worsen Heart Problems?

On June 6, Henry Waxman will hold a hearing to "assess" the FDA’s approval of Avandia in 1999.  Avandia is a drug marketed by GlaxoSmithKline to control elevated blood sugar in diabetics.  One of the company’s top sellers, it produced over $3 billion in revenue last year.  

The immediate reason for the hearing is a May 21, The New England Journal of Medicine (NEJM) (PDF, requires free registration) report that Avandia might increase the risk of heart attack by 43 per cent and the risk of death from cardiovascular causes by 64 per cent.  Cardiovascular disease is a major result of diabetes, so if a diabetes treatment increases the risk of dying from such disease, the treatment can only be evaluated as a danger to the millions of diabetics who have used it since 1999.

 A NEJM editorial (PDF, registration required) accompanying the article suggested that there was no evidence of countervailing cardiac benefits from using the drug.  In fact, the editorial points out that the drug was never subjected to trials that would evaluate its overall health benefits or lack thereof.  It was only tested to evaluate its effect on lowering blood sugar.  The underlying assumption was a simple one:  elevated blood sugar increases health risks, so a reduction in blood sugar will improve health outcomes for patients.

But Avandia has many effects on the body, and not all of them are even known.  One or more of those unknown effects might cause changes in the cardiovascular or other body systems that are more detrimental to overall health than the reduction of blood sugar is beneficial to it.  The only scientifically valid means to evaluate the drug’s effect on overall health is to carry out large scale, long duration trials to evaluate the health effects of any drug, especially one like Avandia that is to be taken everyday for a lifetime.

But such trials were not done before approval, and according to the NEJM editorial are still not available to physicians.  Moreover, such trials were not done before the approval of Vioxx, withdrawn because of adverse cardiovascular effects in 2004 after being used by millions of patients to treat pain.  In fact, the 1992 enactment of The Prescription Drug User Fee Act (PDUFA) institutionalized a bargain between the industry and the FDA:  The industry would pay the Agency several hundred million dollars each year to support only its new drug approval activities; the FDA would do everything possible to reduce the time it took to approve new drugs.

PDUFA had to be reauthorized every five years.  Each reauthorization contained provisions requiring closer cooperation between the industry and the Agency.  By the time the law reached its third iteration in 2002, the FDA had agreed to work with the drug companies to specify appropriate clinical trial designs even before drugs were submitted for approval and to be bound by the results of these pre-specified trials.  Since the goal was speedy approvals, trials were designed to produce relatively quick results.

The kinds of large scale, long term trials described above do not produce quick results; they require several years of effort.  Consequently, industry and the government agreed to create trial designs relying on "surrogate end points".  Rather than design studies to evaluate the effects of new drugs on aspects of health evident to the physician and important to the patient, such as morbidity and mortality, they designed studies that looked at effects assumed to contribute to such health outcomes.  Thus, Avandia was evaluated solely on its effect on blood sugar levels and not on its effects on the patient’s broader health status.  As the NEJM editorial stated,

The current approach to drug approval involves an intensive, high-quality evaluation in the preapproval setting. For many sponsors, approval marks the transition from research to marketing.

The 2002 PDUFA reauthorization contained provisions allowing the FDA to require post-approval trials to evaluate some health effects of long term use of the drug.  But, as the NEJM editorial points out, such studies have severe limitations:  They only address questions about efficacy or safety evident at the time of approval, and there is no provision to examine potential problems that become evident only after approval.  Sponsors themselves propose the trial designs and can easily manipulate them to make the drug in question look good.  And, of most importance, few required studies are completed.  Between 1998 and 2003, only about a quarter of requested post-marketing trials were completed, and in September 2006, 899 such studies remained pending.  Avandia was one of those.

PDUFA is again in the process of reauthorization.  The Senate has already completed action on its version of the bill, The Food and Drug Administration Revitalization Act (S.1082).  But, according to the NEJM, there is nothing in it that would have identified the cardiovascular risks of Vioxx or Avandia in a timely way.

The House has yet to consider the bill.  Waxman’s hearing, along with the publicity generated by Avandia’s problems, may lead to a strengthened bill that requires rigorous post market testing to identify meaningful health outcomes resulting from the use of new drugs.  It would also be helpful to call your congresscritter and let them know that it is time to require the FDA to give more consideration to the safety of American inhabitants than to the bottom line of pharmaceutical companies.