Reply to comment

This post is genuinely excellent. You are asking exactly the right questions, and the PLX4032 trial is the perfect example. The drug clearly works, and its efficacy is solidly supported by the science. It is being tested against a 35 year old drug (dacarbazine) that has never, even in the opinion of the FDA (the agency that approved it in 1975) been shown to provide clinical benefit, and in many clinical trials, it has never been shown to confer a survival benefit. This trial is being run solely to meet the requirements of a policy unilaterally established by Dr. Richard Pazdur (Director of the Office of Oncology Drug Products at the FDA) in 2003. Concerned at the time that such trials were difficult to enroll as Phase IV (post-approval) trials when a drug this promising received what is called Accelerated Approval based on highly promising Phase I and II results, he decided he would no longer grant Accelerated Approval until the unethical trial was run as a Phase III, pre-approval trial. We call it his Decelerated Approval Initiative. The law and the agency's regulation actually intend that a drug like PLX4032 receive Accelerated Approval after showing to a substantial degree of certainty that it is safe and effective. The idea was accelerating the delivery of progress to patients who can't wait, and while it has saved the lives of tens of thousands of HIV/AIDs sufferers, and before 2003, thousands of cancer patients, Pazdur all but eliminated it as an available approval pathway for cancer drugs in 2003. PLX4032 more than meets the standard for Accelerated Approval right now (and has for many months). But Pazdur has rigidly applied his policy to all cancer drugs whether the standard for Accelerated Approval is met or not, so the most fundamental test of ethics, that a trial that puts patents at risk must be asking a question we still need the answer to, is not met. We already have the answer. PLX4032 is clearly and obviously better than dacarbazine, and so different from the older drug that a comparison is meaningless to medical practice and progress against the disease. It is highly likely that the trial will be stopped early by the trial Data Safety Monitoring Board due to the major imbalance in outcomes that will emerge between the two arms, at that point causing it to become "unethical" under current FDA policies, but the imbalance will emerge because of quick progression and deaths in the control arm, a trial result that could be predicted with almost complete certainty before starting the trial. The details with PLX4032, and the FDA's role (which is only partialy explained here) in causing this trial to be run, are very troubling. An even more troubling fact is that these situations have been common over the last 7 years, now coming in to the light because of the NYTs coverage, and the startling gap between the safety and efficacy of the two drugs being tested in the trial. The trial is designed to measure overall survival (Pazdur's "endpoint" for this specific medical indication), which means that even when patients quickly progress on dacarbazine, they are not allowed to cross over to PLX4032, which poses an 80 percent chance of helping them. The reason? To calculate the difference in survival, the control arm patients must be denied PLX4032, and die on the schedule of patients getting only dacarbazine, the drug the FDA actually beleives based on a mountain of data, to be little more than a toxic placebo.

How is this ethical? It clearly isn't, but it is FDA's policy. Your discussion should also include questions about the responsibilities of the FDA and IRBs to make sure unethical human research is not being conducted. In fact, IRBs are not empowered to stop a trial like this one based on it being unethical on its face, as long as the patients - who have no where else to go but the grave - sign informed consents. But what about the FDA? By mandating trials like these through their policies and practices, they declare them not only ethical, but required. The clear message to drug companiesover the last seven years has been, run these trials and produce perfect statistics from the trial (as measured by an arcane, hypertechical and medically meaningless statistical metric called a p-value) or your drug won't get approved.

Is the FDA requiring unethical human clinical research for approval of new cancer drugs? The answer is yes, and that is a huge problem for thefield of bioethics, which has largely spent the last seven years ignoring, or trying to justify and rationalize, this problem.

Steven Walker, Co-Founder
Abigail Alliance for Better Access to Developmental Drugs
www.abigail-alliance.org