A systematic review of studies on antidepressants concludes that the positive effects of these drugs are probably overstated in the medical literature.

But it's not clear if the bias comes from a reluctance to submit negative manuscripts or decisions by journals not to publish them, or a combination of both, according to Oregon Health and Science University researchers, whose report is published in the January 17 issue of the New England Journal of Medicine.

The researchers compared drug efficacy inferred from published studies with drug efficacy reported to a mandatory US government registry of clinical trials, in which all results, including raw data, must be included.

Only 51% of studies in the US Food and Drug Administration registry were considered by the agency to have positive results. In the published medical literature, however, 94% of studies appeared positive.

The increase in the effectiveness of the drug ranged from 11% to 69% for individual drugs, and was 32% overall. The antidepressant bupropion appeared to show a high level of bias. The authors also noted studies that were not positive were often published with a slant that made them seem positive.

"For the average patient, yes, these drugs are effective, just less effective than they appear," said study author Dr Erick Turner, an assistant professor of psychiatry and pharmacology at Oregon Health Sciences University. "Don't be overly disappointed if you don't have an excellent response to the first antidepressant you try. You might have to try two or three, and you might need to add psychotherapy."

One expert hailed the finding. "This is, in my opinion, an excellent paper and what it shows is really consistent with the past five years," said Adil Shamoo, founder of Citizens for Responsible Care and Research and a professor of biochemistry and molecular biology at the University of Maryland School of Medicine in Baltimore. "Publications have indicated that [the] industry basically has a tendency to bias publications towards positive results and to not publish negative results."

"Research is not regulated by anyone, so therefore they don't have to submit to anyone, and that's really the key," Shamoo added. "It's a systemic failure."

Other experts presented slightly differing points of view. "It's not that the drugs are ineffective, but that the public's perception is that they are more effective than they are [something the authors also point out]," said Dr Julio Licinio, chairman of psychiatry and behavioural sciences at the University of Miami Miller School of Medicine. "The data was good enough for the drugs to get approval."

"The concern that selective publication of clinical studies routinely overestimates the benefit of medical services is not new," added Dr A. Mark Fendrick, a professor of internal medicine and health management and policy at the University of Michigan. "While the creation of clinical trial registries will enhance access to unpublished data, the availability of information alone will likely not be very useful to patients and clinicians unless the results can be analysed in a careful manner, as performed in this important study."

The buzzword in healthcare these days is "evidence-based medicine," meaning therapies and preventive strategies that are based on careful, scientific evidence as reported in clinical trials. However, the study authors pointed out, such evidence-based practices are helpful only insofar as the data available is complete and unbiased. And, they said, it is difficult to study "selective publication" primarily because data from unpublished trials is hard to locate.

"Pharmaceutical companies have responded to the heat," Turner said. "[But] it's hidden under our very noses." For example, as a result of a 2004 lawsuit, GlaxoSmithKline initiated a clinical trial registry. "I have yet to speak to a colleague who knew about that," Turner said, adding that physicians generally look to peer-reviewed journals for guidance.

Licinio believes the reasons for the bias are a combination of people not submitting papers and editors/publishers not printing them. As the editor of the journal Molecular Psychiatry, Licinio receives more than 700 submissions a year and publishes an average of 144. "It becomes very difficult to select," he said.

"There should be a national repository for these kinds of reports [that don't get published in journals]," Licinio added. "Those papers don't see the light of day."

More health research news

Originally posted February 2008.

Disclaimer

This is a summary article from HealthScout. Knowledgeable New Zealand health consumers may also find this article useful. This information is intended solely for New Zealand residents and is of a general nature only and no person should act in reliance on any statement contained in the information provided and at all times should obtain specific advice from a health professional. All rights reserved. © UBM Medica (NZ) Ltd. This publication is copyright.