Congress Seeks to Balance Drug Safety, Quick Approval

By David Brown
Washington Post Staff Writer
Thursday, July 5, 2007

Three drugs -- Vioxx, Ketek and Avandia -- are casting long shadows over the congressional debate on how to update and revise the 15-year-old system that has dramatically sped up the process of bringing prescription medicines to American consumers.

All three ran into problems after the Food and Drug Administration approved them for use. The problems ranged from thousands of heart attacks partly attributable to the painkiller Vioxx to a few cases of liver failure caused by the antibiotic Ketek. And the hazards of the diabetes drug Avandia are only now emerging.

Together, the three drugs epitomize what can happen when a new pharmaceutical moves from the small and orderly world of scientific testing to the big and messy world of everyday use. Many experts now hope that efforts to do a better job of identifying the "unexpected toxicities" in drugs that have reached medicine cabinets will get a boost from an unexpected source -- the Prescription Drug User Fee Act.

The act's chief purpose when it was passed in 1992 was to help drugs get through FDA review quicker and out to patients sooner. It did so by creating a "user fee" that a company must pay when it submits the mass of research and paperwork that the FDA reviews before approving or rejecting a drug. As Congress debates reauthorizing the user-fee law, it is considering several proposals that would raise the fees and devote a larger portion of them to "post-market surveillance" of drugs -- the regulatory equivalent of chasing the horses after they are out of the barn.

The debate touches not only on how much to spend on post-market surveillance but also on the user-fee system itself. Those fees now provide more than half of the money the FDA spends, mostly on salaries, to review new drugs. The agency wants to collect $393 million next year, $87 million more than this year.

In April, the Senate passed a huge FDA-related bill that included a reauthorization of the user fees for five more years. The House will consider numerous bills on the same subjects -- not only user fees for drugs and medical devices, but also incentives for testing drugs in children and long-range strategies for enhancing drug safety -- probably later this month.

Differences would have to be resolved in conference. But it seems likely that heavy reliance on user fees -- rather than appropriations by Congress -- will continue to be a key feature of pharmaceutical regulation.

"I am deeply disappointed in the way things are headed. This is our last chance to get it right before five more years go by," said Jerry Avorn, a Harvard Medical School physician and expert on drug safety.

Avorn says that not enough money is being spent on "pharmacovigilance" -- the ongoing study of drug safety -- and that the FDA's dependence on industry money may be corrupting it in subtle ways.

The user-fee act, which goes by the awkward acronym PDUFA, was born of attacks from two quarters.

Industry wanted faster action by the FDA so that its review of new drugs would consume as little as possible of the period in which a product is under patent. Companies were willing to pay the FDA to hire more people because quicker reviews might ultimately increase profits. At the same time that companies were complaining of the FDA's slowness, advocates for patients were lambasting the agency for insufficient urgency in making drugs -- particularly AIDS drugs -- available to sick or dying patients.

As a strategy for addressing those complaints, the law has been a huge success.

From 1993 to 2003, the average approval time for standard drugs fell from 22 to 14 months. For fast-tracked "priority" drugs it declined more steeply, from 13 months to six. This dramatically changed the United States from a country where new drugs came late to one where they came early. In the decade before the PDUFA, only about 8 percent of new prescription drugs became available first in the United States; now that number is about 50 percent.

The user-fee system nearly doubled the number of FDA workers reviewing new drugs, from about 1,300 in 1992 to 2,500 in 2004. Equally important, it changed the work culture at the agency.

The act laid out guidelines requiring that at least 90 percent of drug reviews be done within a specified time after an application was filed. It specified how quickly the agency had to respond to company requests for meetings with FDA reviewers. It allowed manufacturers to ask the FDA to look over their drug-testing protocols before the studies were done -- an option that by 2004 had added 346 meetings a year to the 2,132 regularly scheduled ones between the agency and the companies.

Although drug reviewers suffered no penalties if they missed these goals, they took them very seriously -- and consistently met them, FDA annual reports show.

The payoffs and the hazards of this more company-friendly FDA were revealed in two unusual studies published recently.

Tomas J. Philipson of the University of Chicago and four colleagues analyzed the economic and human effects of the first decade of the user-fee system.

They estimated that the speedier approvals yielded at least $11 billion in new profits for the industry -- an average of about $39 million for each new drug. Faster action also saved 180,000 to 310,000 years of life for patients -- a gain partially offset by up to 56,000 years lost to the harmful effects of drugs approved but later withdrawn for safety reasons.

The FDA touts the consumer payoff.

"Early access to safe and effective medicine is good for public health," Theresa Mullin, FDA assistant commissioner for planning, said in a recent interview. "For people with Parkinson's, with early Alzheimer's, with cancer, time is a luxury you just don't have."

On the other hand, Daniel Carpenter, a government professor at Harvard, has studied the potential effects of the time pressure that the PDUFA placed on the FDA.

He found that drugs approved in the two months before the deadlines were three to five times as likely to be taken off the market for safety reasons as drugs approved earlier in the cycle or after, when the agency missed the deadline.

"Our analyses suggest the [act's] clocks have dramatically changed the behavioral structure of the FDA review cycle," he wrote.

Few doubt, though, that even the most carefully tested drugs can produce surprises. That is true both when the problems they cause are extremely rare (such as spontaneous liver failure with Ketek) or very common (such as heart attacks with Vioxx, and possibly with Avandia). Either way, the problems are hard to detect.

"Many of the adverse reactions that we care about are not recognized as adverse reactions most of the time," said Richard Platt, a researcher at Brigham and Women's Hospital in Boston. Traditionally, the FDA has relied on spontaneous reports from many doctors, or on the sharp eyes of a few, to catch the bad actors. The latter occurred 10 years ago when several physicians at the Mayo Clinic noticed rare heart-valve lesions in women taking the diet-drug combination fen-phen.

Now, many experts -- Platt is one of the leaders -- believe large databases maintained by Medicare and HMOs can be "interrogated" by sophisticated computer programs to identify problems far earlier than if regulators just wait for reports to trickle in. A report last year by the National Academies' Institute of Medicine pushed for more such studies, among other drug-safety strategies.

Such database analyses helped reveal the hazards of Vioxx, which got fast-track approval in 1999 because it caused less intestinal bleeding -- and thus appeared safer -- than other nonsteroidal painkillers.

Five months before Merck voluntarily pulled the drug from the market in September 2004, a research team that included Avorn examined the experience of 54,000 elderly enrollees in two large pharmacy benefits programs. They found that people taking Vioxx had more heart attacks. A later study by Swiss researchers concluded that if another strategy called cumulative meta-analysis had been used to study the totality of research on Vioxx, the drug's risks could have been recognized by early 2001.

Today, nearly everyone agrees that more attention should be paid to tracking the safety of drugs already in use.

Originally, user-fee money could not legally be spent for that purpose. Now it can be, but only to study a drug's performance during its first three years on the market. Even that restriction would be lifted under the current proposals.

Meanwhile, the FDA increased the number of employees doing post-market surveillance from 62 to 96, with at least 80 more to be added next year. Bills in both the House and the Senate would more than double -- from $30 million to about $75 million -- the amount of user-fee money for post-market review of drugs.

But many people think that is not nearly enough for a country where the average 65-year-old takes six prescription medicines, $180 billion is spent on those drugs and manufacturers spend $12 billion advertising them.