To the Editor: Ward T. Smith, M.D., and his colleagues’ (1) recent randomized controlled trial of fluoxetine and fluoxetine plus clonazepam represents an important contribution to the problem of antidepressant-benzodiazepine combination therapy for major depression, which is widely practiced in the real world but may not have a solid evidence base. The editor of the Journal, Nancy C. Andreasen, M.D., Ph.D., called the rapidity of the response to this combination therapy striking.

Dr. Smith et al.’s review of the extant literature on this topic, however, is slanted. The authors cite four randomized controlled trials in which they claim to have demonstrated positive results for a combination therapy of tricyclics and benzodiazepine. There are many ways that the experimental therapy can be better than the control therapy, but when we limit ourselves to the primary end point measure of depression, at least one (2) and possibly two more (3, 4) of these four did not report any positive results. More important, a systematic search of MEDLINE revealed at least five more relevant randomized controlled trials, three of which were negative (5–7), and another randomized controlled trial that reported significant worsening of depression upon discontinuation of benzodiazepine after successful combination therapy.

The effectiveness of a treatment should be judged on the basis of the totality of the evidence available to humankind, because any one randomized controlled trial, even when not subject to publication bias, is always subject to random variation, and if you conduct enough trials, there will almost always be those with positive results. The researchers need to conduct their own study and allow readers to view it under the appropriate perspectives.