Are washout periods useful in antidepressant trials?

Abstract

Single-blind placebo washout periods before randomisation enable the elimination of the psychotropic agents previously received. They subdue carryover effects which could be achieved without using a placebo. Washout periods also purport to identify and eliminate the placebo responders. Trivedi et al.performed a meta-analysis which included 101 studies. They demonstrated that placebo washout periods do not reduce the response rate in the placebo group and do not increase the difference between the placebo and the treated group. This held true for all the different antidepressant classes. In another study, Greenberg et al.analysed 28 antidepressant controlled trials published between 1983 and 1992 and found no difference between trials with or without a placebo washout period in terms of response rate in either the placebo or the treated group. Therefore, placebo washout periods, although appealing and widely used, may not reduce the number of patients who respond to placebo. Besides, the patients who respond during the washout period have very diverse outcomes after three months. This subgroup is likely to be heterogeneous and should be better studied. Some authors have stated that washout periods may bring in confounding effects such as lowering the observed difference between the treated and placebo group. Their explanation was that response to placebo is not a stable characteristic and that responding to placebo during the washout period may subsequently lower the level of placebo-induced improvement. It would also be cumbersome if washout periods covered the problems related to the placebo and blindness issues, which are often neglected. Finally, it appears necessary to further assess the usefulness of single-blind washout periods.