Drug Regulation in all the Wrong Places
Our Clinical Trial Process is Misguided
Can you trust the drug in your medicine cabinet or the medical device in your chest or knee? Hardly a month goes by without a drug or device being pulled off the market, lawyers popping up on TV and the web scrounging for victims, and befuddled patients left to talk with their equally confused doctors about whether they should throw out their pills or get their device replaced given the newly detected slew of deaths, risks, structural failures, or nasty side effects.
Lots of folks seem to think we have too much regulation of drugs and devices already—among them Paul Howard at the Manhattan Institute and Scott Gottlieb at the American Enterprise Institute—so much so that it is choking innovation to death. But, if that is so, then why are there so many scandals?
One possible answer is that the companies know they have problems but sit on that knowledge. If that’s sometimes (or oftentimes) the case, then we need a regulatory system that can get around that kind of immoral behavior. We don’t have that system.
What we have is a regulatory system that is too skewed toward looking at the earliest stages of research. Moreover, the way it is designed makes recalls almost inevitable. The diabetes drug Avandia is the latest in a long parade of failures of our current post-clinical trial drug approval process.
Avandia went through the usual approval process with the U.S. Food and Drug Administration. The drug was a blockbuster. But sales began to fall after a 2007 study of people taking the drug suggested that Avandia could cause heart attacks and strokes. I first learned about this while serving on a bioethics advisory board for GlaxoSmithKline, the developer of the drug—a panel that was looking at research ethics issues in poor nations. The panel, on which Science Progress Editor-in-chief Jonathan Moreno also served, came to an abrupt halt.
In response, the FDA put a black-box warning on the drug telling doctors of the heart attack risk. GlaxoSmithKline was not happy. There was a lot of back and forth about the safety of the drug. Over the past few months more evidence become public that shed doubt on Avandia’s safety. Worse, it appears the company withheld data about serious side-effects. The FDA appointed an advisory panel this July to consider these allegations, but the panel itself quickly got caught up in charges of conflict of interest among its members. It is likely that more black-box warnings to doctors will follow, should GlaxoSmithKline choose to keep Avandia on the market.
Avandia is not alone. The drug’s problems in the marketplace follow hard on the heels of Prempro, a hormone replacement therapy made by Wyeth Pharmaceuticals, now part of Pfizer Inc., which became caught up in lawsuits alleging it caused breast cancer. More recently, the FDA released a warning about the Afluria flu vaccine, made by CSL Ltd. of Australia, concerning high fever and seizures. Prior to that was Merck & Co.’s widely publicized recall of Vioxx, which came after problems with Astra Zeneca’s Seroquel, Abbott Laboratory’s Meridia, Pfizer’s Rezulin, C.R. Bard Inc.’s G2 filter, Bayer’s Baycoll, Boston Scientific Inc.’s Express Stent, and on and on.
So is there a real phenomenon here or just more PR associated with recalls? And if there are more recalls going on then what is wrong with the oversight of new drugs and devices? It is not clear from the literature whether there are more recalls taking place in recent years—there is no real database that would show such a trend. There are certainly more stories about recalls and more people studying the objectivity of the marketplace surveillance being done by pharmaceutical, biotech, and device companies.
No one seems to have reliable numbers on recall trends, yet the Institute of Medicine and other groups still warn that the existing system of drug protections after the FDA approval process is complete does not seem adequate to handle the products that are reaching the market. Fortunately, there is a way to fix the system.
The major problem today is that too many lousy or dangerous drugs and devices get to you without adequate safety review because drug and device regulation is heavily weighted in the United States toward early stages of research. Every drug has to be tried in animals to roughly determine safety. Then drugs are introduced into a small number of humans to further check safety—so-called phase one trials. Then dose and mode of administration are checked for safety, biological activity, and signs of effectiveness—phase two. Only after all this safety testing is a drug or device ready to go to phase three clinical trials. In these studies hundreds or sometimes thousands of subjects are recruited to receive the drug or product for periods of time that range in nearly every case from a few months to a year. Phase three trials are almost always placebo controlled randomized, blinded studies.
So there is a lot of effort to try and make sure that subjects are not hurt in phase three trials. The deaths of subjects in phase one clinical trials, among them Ellen Roche and Jesse Gelsinger in early stage studies over a decade ago seem to have reinforced regulatory anxiety about the risk of deaths in first in-human studies.
Meanwhile, the weakest link is the fourth and final step in the research process—phase four—in which drugs are to be monitored when out in actual use in the world for adverse events and problems. Drug companies sometimes promise to do these trials to get final product approval but don’t. These studies are heavily weighted to support the funders of these studies, Big Pharma, which results in much more rosy reporting then studies done by independent groups.
Reporting of problems in phase four is left to doctors and patients who rarely do so. And there is no systematic tracking of a subpopulation taking new drugs or other medical products to see what is going on with real patients in real world conditions. Deaths have to mount rapidly and obviously to get regulatory or physician attention before phase four studies are ever seriously undertaken.
But a lack of independent, well designed phase four trials is not the only problem. Approving drugs based on current standards for phase three testing has its own built-in limits. Testing drugs and devices in randomized, blinded, placebo control trials is great, but it means that approval is given on the basis of highly controlled studies on highly selective populations—often subjects who are not that old, not that sick and are highly compliant. That’s not the real world, where patients take lots of drugs, some legal some not, are poorly compliant, have multiple diseases, and can be very old or very young.
So what looks safe in a phase two or phase three study can prove lethal when given to real people in uncontrolled, unsupervised environments. What’s more, phase three studies are also relatively short. What looks safe after three months exposure or six may not be after three or six years.
The seemingly endless parade of horrors of FDA approved drugs gone bad merits a reexamination of a regulatory system that is not keeping us safe. The issue is not too much bureaucracy and too much red tape, but a strategy of safety that puts the emphasis in the wrong place—early not late—and then uses techniques that by themselves cannot ensure safety for real people in the long run.
Update August 17 2010: CNN Money reported that the number of drug recalls in the U.S. surged to 1,742 in 2009, up 309 percent from the 2008 level. Recalls so far in 2010 are also on pace greatly exceed previous levels. However, not all of these recent recalls were due to the drugs themselves being unsafe—some were due to problems with the manufacturing process of generic, over-the-counter drugs.
Arthur Caplan, PhD, is the Director of the Center for Bioethics and the Sidney D Caplan Professor of Bioethics at the University of Pennsylvania
Comments on this article