From Our 2008 Archives

Picture of Antidepressants Too Rosy?

Review Shows "Selective" Reporting of Studies Inflates Effectiveness of Antidepressants

By Kathleen Doheny
WebMD Health News

Reviewed By Louise Chang, MD

Jan. 16, 2008 -- Commonly prescribed antidepressants may not work as well as published studies suggest, according to new research. "Selective" reporting of study results inflates the effectiveness of antidepressants, researchers found.

"I don't want people to think these drugs are not effective, that's not the take-away," Erick Turner, MD, the study's lead author, tells WebMD. "The drugs are still effective, but less so than what would be apparent from the medical literature."

His team found that some negative studies didn't get published, and some study results on effectiveness are inflated.

"Selective publication exaggerated the efficacy of the drugs by 32% overall," says Turner, an assistant professor of psychiatry and pharmacology at Oregon Health & Science University in Portland and medical director of the Portland VA Mood Disorders Program. "But for individual drugs, the efficacy [reported in published studies] increased from between 11% to 69%." The study is published this week in The New England Journal of Medicine.

Study Details: Effectiveness of Antidepressants

Turner decided to do the study after his three-year stint at the FDA, working as a reviewer on drug studies from 1998 to 2001.

For the study, he and his colleagues zeroed in on 12 antidepressants that had been approved by the FDA between 1987 and 2004. They looked at 74 FDA-registered studies involving 12,564 patients, comparing the FDA reviews of the studies with study results published in medical journals.

The published results on the effectiveness of antidepressants didn't always jibe with the FDA review of the study. For instance:

  • The FDA viewed 38 studies as positive, and all but one were published in medical journals.
  • Of the 36 other studies viewed by the FDA as having negative or questionable results, 22 were not published and 11 were published in a way that Turner's team believed reported a positive outcome. (That might be accomplished, Turner says, by looking at statistics differently or using a different rating system for depression, for instance.)
  • While the FDA reviews concluded that 51% of the studies had positive results for the antidepressants, the published literature reported that 94% of the studies had positive results.
  • Among the drugs looked at were Prozac, Cymbalta, Paxil, and Zoloft.

Effectiveness of Antidepressants: The Rules and the Disconnect

In the U.S., drug companies must register with the FDA all trials they intend to use to support their request to market a drug, Turner says, and then report those results. The FDA analyzes their results and goes back over the raw data to see if they come to the same conclusions.

Exactly why the FDA reviews and the published results don't agree isn't known, Turner says.

Besides looking at statistics differently, he says, it might be a failure by the pharmaceutical companies or researcher to submit the manuscript, a decision by medical journal editors or reviewers not to publish a study result, or a combination of those factors.

Bioethicist Weighs in

The study results point to the need for more oversight, says Arthur Caplan, PhD, director of the Center for Bioethics at the University of Pennsylvania in Philadelphia, who believes the system needs improvement.

"So there can be no doubt, no institute or institutional review board [which approves studies] should approve any clinical trial without documentation that it has been registered and all results will be published -- positive or negative or incomplete," he says.

Pharmaceutical Companies Respond

The situation with reporting of clinical trials has improved, pharmaceutical companies contend. "We have a policy that we publish the data of all randomized clinical trials regardless of outcome," says Tammy Hull, a spokeswoman for Eli Lilly and Co., which makes Prozac and Cymbalta.

She notes that two study results on Cymbalta that Turner says he could not find published actually were published in peer-reviewed journals. Lilly also posts study results on its site.

Mary Anne Rhyne, a spokeswoman for GlaxoSmithKline, which makes Paxil, says the company also posts clinical trial results on its site, which now includes results of more than 3,000 trials of 82 medicines and vaccines.

The Pharmaceutical Research and Manufacturers of America (PhRMA) took issue with the study, too. The industry has taken steps recently to make study results information more accessible, according to a PhRMA statement issued in response to the study.

In 2004, PhRMA launched an Internet database, www.clinicalstudyresults.org, which now includes summaries for 497 prescription drugs. And the recently passed FDA Amendments Act provides for a clinical trial database to be set up and run by the government, PhRMA notes.

Message for Depressed Patients

Despite Turner's observation that the effectiveness of antidepressants is sometimes inflated, he tells depressed patients: "Keep taking the drugs if they are needed. The drugs are effective but perhaps less consistently effective than we might have thought.

"Don't be too disappointed if you don't have an excellent response to the first drug you try," he says. "Be patient."

SOURCES: Erick Turner, MD, assistant professor of psychiatry and pharmacology, Oregon Health & Science University, Portland; medical director, Portland VA Mood Disorders Program. Turner, E. The New England Journal of Medicine, Jan. 17, 2008, vol 358: pp 252-260. Arthur Caplan, PhD, director, Center for Bioethics, University of Pennsylvania, Philadelphia. Lydia Stuckey, spokeswoman, Pharmaceutical Research and Manufacturers of America (PhRMA), Washington, D.C. Mary Anne Rhyne, spokeswoman, GlaxoSmithKline, Research Triangle Park, N.C. Tammy Hull, spokeswoman, Eli Lilly & Co., Indianapolis.

© 2008 WebMD Inc. All rights reserved.





STAY INFORMED

Get the Latest health and medical information delivered direct to your inbox!