Study Shows Faster Approval Does Not Affect Safety

Faster approval times have not lead to increased withdrawals due to safety problems, according a new report by the Tufts Center for the Study of Drug Development. Reforms in the 1990s instituted user fees and sped up the approval process.

    Tufts researchers found that of the 20 drugs that have been withdrawn since 1980, nine came off the market before the 1993 user fee changes. According to the report, seven of 217 drugs approved in the 1980s were withdrawn and 11 of 311 drugs approved in the 1990s were withdrawn. Two of these 11, however, were withdrawn prior to the reforms that led to user fees. So far only two of the 120 drugs approved since January 2000 have been withdrawn.

    The report also indicates that quickly approved drugs are no more likely to face safety withdrawals than other drugs. For example, data indicate that drug approvals took an average of 2.3 years in the 1980s. But approval times for withdrawn drugs averaged 2.9 years. In the 1990s, approvals for withdrawn drugs took 1.1 years as compared with 1.3 years overall.

    But Tufts CSDD director Kenneth Kaitin said that this difference is small enough that there is “no discernible pattern.” He also indicated that the time differential for drugs approved since 2000 is not reliable because of the small number of withdrawals.

     “FDA’s process was inordinately lengthy in previous years,” Kaitin said. “Now we have an application process that does not limit FDA’s ability to detect safety problems.”

    Further research, Kaitin said, should examine how to identify risks after a drug is on the market.

    Charles Bennett, professor at Northwestern University’s Feinberg School of Medicine, agreed that optimizing post-marketing risks is more important than changing the speed of approval.

    “A little faster or a little slower, I don’t think it’s short-cutting safety steps,” he said. “It’s not the approval that’s the issue; it’s after approval.”

    Bennett also criticized the Tufts study for not differentiating between drugs approved under the regular process and those with accelerated approval. He stressed that it is especially important to understand the assessment of post-marketing risk for accelerated drugs, which have smaller clinical trials.

    The Tufts study also found that 70 percent of the drugs withdrawn since 1980 were in three classes --- anesthetic/analgesic, cardiovascular, and anti-infective. Kaitin speculated that the first two classes of drugs might have lead to more safety issues because they are generally taken over a longer period of time.

    “Increasing the length of treatment increases the likelihood that you will have safety problems due to long-term toxicity,” he said.

    He also thought that anti-inflammatory drugs might lead to safety problems because the processes targeted by these drugs are often part of a variety of physiological functions.

    He indicated that almost all of the anti-infective drugs pulled off the market were in one class, quinolones, which were associated with liver toxicity.

    On the other hand, Brian Strom, professor of public health and preventive medicine at the University of Pennsylvania, said that these three classes probably have some of the highest numbers of drugs in them. So, the larger number of withdrawals for these classes is most likely proportional.

    Bennett suggested that instead of focusing on classes of drugs, future research looks at the types of toxicities that lead to the withdrawals. He suggested looking at liver and bone marrow registries or at data from plasmapheresis centers. Information on toxicities from these samples or registries could be compared with knowledge about what drugs the people with these symptoms were taking.