Do antidepressants really work?

Perhaps only modestly in severely depressed patients, say the results of a recent meta-analysis.

The findings are based on data obtained from all published and unpublished clinical trials submitted to the Food and Drug Administration (FDA) for the licensing of four selective serotonin or serotonin-noradrenaline reuptake inhibitors: fluoxetine (Prozac), venlafaxine (Effexor), nefazodone (Serzone), and paroxetine (Seroxat, Paxil). The study focused on the degree to which patients improved compared with placebo and relative to the initial severity of their disease. Including unpublished data helped to eliminate bias that can occur when unpublished negative trials are excluded from analysis.

The researchers found almost no difference between the effects of the drugs and placebo in patients with moderate or mild depression at baseline. And while efficacy reached clinical significance in the most extremely depressed patients, it was due to a decrease in response to placebo and not to an increase in response to medication.

Because the data submitted to the FDA reached statistical significance for the effects of the drugs compared with placebo, but not clinical significance and because more than 40 million people worldwide have been treated with fluoxetine alone, the authors of the meta-analysis concluded by suggesting that the criteria for licensing these drugs be changed.

Kirsch I, Deacon BJ, Huedo-Medina TB, et al. Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration. PLoS Med. 2008;5:260-268.