Skip to main content Back to Top


Safety of Newly Approved Drugs Questioned

Kate Traynor

New research urges physicians to think twice before prescribing a newly marketed drug product instead of an older but effective medication.

The study, reported in the Journal of the American Medical Association (JAMA),1 found that 10% of the 548 drugs initially approved by FDA from 1975 through 1999 had been pulled from the market by 2000 or required a new black-box warning for safety reasons.

What Is Behind the Box?

A 1998 study identified six criteria that seem to influence FDA's decisions to have black-box warnings added to the labeling of drug products.1

Based on the wording of 375 black-box warnings for 206 drugs listed in the 1995 edition of the Physicians' Desk Reference, the research team found that a warning may

  • Identify a drug-associated adverse event that can be prevented through monitoring and intervention,    
  • Identify specific patients for whom the treatment is particularly dangerous,    
  • Advise that the risks of treatment may outweigh its benefits,    
  • Identify a potentially harmful drug interaction or describe critical dosing information,    
  • State that the drug should be administered only by a specially trained health care practitioner or in a special setting, or    
  • Caution that the method of drug administration requires exceptional care.

According to the report, 25% of the black-box warnings studied had identified patients for whom use of the drug was particularly risky. The communication of harmful drug interactions or dosing considerations, the purpose of 20% of the black-box warnings, was the next most common reason for the warning. Eighteen percent of the warnings advised that the product should be administered in a special setting or by a health care provider who has been specially trained in the product's use.

1. Beach JE, Faich GA, Bormel FG et al. Black box warnings in prescription drug labeling: results of a survey of 206 drugs. Food Drug Law J. 1998; 53:403-11.
Journal%20Online/53_3/art2.pdf (accessed 2002 May 14).

A probability analysis showed that each new drug had a 4% chance of being withdrawn from the market during the study period. Half of the withdrawals occurred within two years after FDA had initially approved the product.

In all, 81 additions of black-box warning to the labeling of the new drugs were documented in various editions of the Physicians' Desk Reference (PDR), the research team's source of data on product labeling. Half of these changes first appeared in PDR within seven years after the drug was first approved for marketing in the United States.

Based on the timing of PDR-documented labeling changes and drug withdrawals, the research team predicted that one in five new drugs would, within 25 years, acquire a new black-box warning or be withdrawn from the U.S. market because of adverse events in patients.

The research team—Sidney M. Wolfe of the Public Citizen Health Research Group and five physicians affiliated with the Harvard Medical School—concluded that physicians should let several years elapse before prescribing a newly approved drug if safe, effective, older alternatives to the new product exist. The physicians also encouraged FDA to "consider raising its threshold for approving new drugs when safe, effective therapies also exist."

In an editorial in the same issue of JAMA, two FDA physicians, Robert J. Temple and Martin M. Himmel, questioned the usefulness of the study's findings and attempted to put the labeling changes into perspective.2 Temple and Himmel noted that many black-box warnings "reflected refinements of monitoring that were surely useful, but did not affect the basic use of the drug."

According to the editorialists, eight of the labeling changes described in the research article affected antineoplastic or antiviral drugs that were already known to cause significant toxicity before the new warnings were added. Temple and Himmel noted that none of these drugs were withdrawn from the market.

Some of the labeling changes, Temple and Himmel noted, applied to entire classes of drugs, with each class member counted once by the researchers. Seven labeling changes applied to antiarrhythmic agents, which received a classwide relabeling based on the results of a single study of an unapproved use of the drugs. Classwide labeling changes advising against the use of angiotensin-converting-enzyme inhibitors during pregnancy affected six drugs.

Another issue raised in the editorial was the reliability of PDR as a data source for calculating the timing of black-box warnings. PDR is published once each year, but black-box warnings can appear at any time. According to the editorial, the researchers’ reliance on PDR to date a black-box warning "will make the warnings added early in the life of the drug seem more delayed than they were but will not greatly affect the warnings added late."

Temple and Himmel noted that adverse events, such as torsades de pointes, that are discovered after the initial marketing of a product alert FDA to watch for similar problems when other drugs are evaluated for marketing approval. Five drugs described in the JAMA article were given a black-box warning related to torsades de pointes, an adverse event that the editorialists said is unlikely to catch FDA by surprise in the future.

1. Lasser KE, Allen PD, Woolhandler SJ et al. Timing of new black box warnings and withdrawals for prescription medications. JAMA. 2002; 287:2215-20.

2. Temple RJ, Himmel MH. Safety of newly approved drugs: implications for prescribing. JAMA. 2002; 287:2273-5. Editorial.