Special review article Is drug–placebo difference in short-term antidepressant drug trials on unipolar major depression much greater than previously believed? ☆ Zoltán Rihmer a,b, ⁎ , Xénia Gonda a,c a Department of Clinical and Theoretical Mental Health, Semmelweis University, Faculty of Medicine, Budapest, Hungary b Department of Psychiatry and Psychotherapy, Semmelweis University, Faculty of Medicine, Budapest, Hungary c Department of Pharmacology and Pharmacotherapy, Semmelweis University, Faculty of Medicine, Budapest, Hungary Received 13 December 2007; received in revised form 15 January 2008; accepted 23 January 2008 Available online 15 February 2008 Abstract A pooled analysis of randomized placebo-controlled short-term antidepressant trials on unipolar major depression shows that the rate of antidepressant and placebo responders is 50% and 30% respectively. The traditional calculation of antidepressant– placebo difference (50–30 = 20%) in these drug trials is based on the assumption that all placebo responders should be antidepressant responders, a postulation that has been never investigated and proved. Further studies are needed investigating directly the relationship of placebo response in relation to antidepressant response/nonresponse. If a substantial part of placebo responders were antidepressant nonresponders, the drug–placebo differences in all short-term antidepressant drug trials on unipolar major depression are much more than 20%, and the previously published data on antidepressant–placebo difference should be re- calculated. © 2008 Elsevier B.V. All rights reserved. Keywords: Antidepressants; Placebo; Antidepressant–placebo difference; Short-term antidepressant drug trials; Unipolar major depression Contents Role of funding source ................................................... 198 Conflict of interest ..................................................... 198 References ......................................................... 198 The role of placebos in evaluating the efficacy of antidepressants is generally well accepted, yet continues to be a frequently debated topic (Laporte and Figueras, 1994; Quitkin, 1999; Walsh et al., 2002). In randomized placebo-controlled clinical trials (RCTs), the differences in drug –placebo response rates (drug –placebo Journal of Affective Disorders 108 (2008) 195 – 198 www.elsevier.com/locate/jad ☆ An earlier version of this paper has been published by Z. Rihmer, in Neuropsychopharmacologia Hungarica, 2007; 9: 35–37. ⁎ Corresponding author. Semmelweis University, Faculty of Med- icine, Kutvolgyi Clinical Center, 1125 Budapest Kútvölgyi út 4, Hungary. E-mail address: rihmerz@kut.sote.hu (Z. Rihmer). 0165-0327/$ - see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2008.01.020