Study Causes A Stir

Controversial Findings Cast Doubt On Anti-Depressants

By: Matt Lin

Posted: 3/25/08

People who think anti-depressants are a cure-all for depression might be surprised to learn about the results of a recent study.

The study, led by emeritus UConn Clinical Psychology Professor Dr. Irving Kirsch and supported by meta-analysis application from Social Psychology Professor Blair T. Johnson and Post-Doctoral Associate Tania B. Huedo-Medina, has generated great interest in the scientific community and beyond. This study, published on Feb. 26, revealed from previous clinical research data - both published and unpublished - that new-generation anti-depressants lack clinical significant efficacy for all patients except those who are very severely depressed, when compared against placebo trials.

This study utilizes a new approach of pooling data that was previously unreleased by the federal Food and Drug Administration (FDA). Due to the introduction of the Freedom of Information Act (FOIA), researchers have gained access to clinical research data that were not previously readily available.

Kirsch approached Johnson during the publication process to enhance the statistical portions of the study. Johnson invited his associate, Huedo-Medina, to join the team as well.

"[Kirsch] was doing some meta-analysis before, using the same database, but they couldn't test whether or not the efficacy of anti-depressants is dependent on the severity of depression," Huedo-Medina said. She explained that Kirsch needed someone who has been trained in state of the art of meta-analysis techniques.

"He and his collaborators thought that this pattern was there but they were surprised at just how clear the picture was after we used meta-analytic techniques, so bringing us onto the team helped make the picture clearer," Johnson said.

Data sets from 47 separate clinical trials - from both the U.S. and UK - were included in this meta-analysis, which is a statistical method that pools the results of independent studies on a topic. Of the 47 trials, nine were not previously published by the FDA for reasons unknown.

Other evidence has shown that these unpublished trials tend to show anti-depressants to be less effective than previously thought.

The trials that the Kirsch team examined include six types of commonly prescribed anti-depressants, a class of drugs usually labeled as Selective Serotonin Reuptake Inhibitors or SSRIs, which includes fluoxetine (Prozac), venlafaxine (Effexor), nefazodone (Serzone), paroxetine (Paxil), citalopram (Celexa) and sertraline (Zoloft). However, the latter two were excluded from the main analyses due to missing parts of a complete data set that left Kirsch and his colleagues with 35 complete trial data sets.

With the help of Johnson and Huedo-Medina, the team's work resulted in a controversial study that has been one of the most viewed research articles on the Public Library of Science (PLoS) Medicine Journal Web site.

Important scientific publications such as Nature have also taken notice. A recent Nature editorial titled "No more scavenger hunt," published on March 6, cited the study for its use of FOIA to obtain less-biased data through the FDA. "I think Nature is on to something here, because if we can get rid of the biases that hide the truth, it will provide more reliable results that hold up better," Johnson said.

Another recent British Journal of Medicine (BMJ) editorial quotes the results from Kirsch's study. According to the editorial, current clinical study trials conducted by researchers contracted by pharmaceuticals have not been required to register or publish their results prior to the marketing of any drug. The FDA has made amends by enforcing registration processes for registered researchers and companies to post their research proposal within the first 21 days of the first patient's enrollment and "the results within one year - with extensions up to two years - of the time that the FDA approves the drug or some other action is taken or the trial is concluded." However, there are now current laws that require such cooperation for drugs that have already been on the market.

Although the results from the study are compelling, Dr. Johnson also points out that the data set does not cover the entire range of depression. Due to the ongoing debates pertaining to interpretations of the Hamilton Rating Scale of Depression (HRSD) - a standard used by most clinicians and researchers working on depression - it may not be possible to get a large enough sample beyond the ranges defined by individual studies.

Moreover, some researchers claim that anti-depressants take a longer time period - as long as three months - to achieve maximum affect. There is much to be explored in this field. Yet the implication is that this meta-analysis study has improved the basic understanding of anti-depressants' efficacy.

Contact Matt Lin at Kuan-hung.Lin@UConn.edu.