Effectiveness of paroxetine in the treatment of acute major depression in adults: a systematic re-examination of published and unpublished data from randomized trials

As a family physician, I see many patients with mood disorders, and the SSRIs are an important therapeutic option. I have the clinical impression that they can be very effective and I prescribe them often. I was thus interested to read what a systematic overview revealed about their effectiveness (Barbui et al, 2008).

I was disappointed and puzzled by the bizarre primary outcome measure, the proportion of patients who left a study early for any reason, selected by the authors. Consider the ideal situation in which no one in either arm drops out; it would be impossible for the active treatment to be better than placebo, even if all treated subjects went into remission. How can this be a measure of effectiveness?

In my practice, the biggest challenge is persuading patients to persist with therapy through the first few days of unpleasant side effects until the beneficial effects become manifest. I consider early dropout to be a failure of my persuasive powers, and not an indication that the SSRIs are ineffective. I believe that it is important to distinguish between dropout in the first days of treatment, as a manifestation of predictable and often transitory unpleasant side effects, and delayed dropout which may reflect treatment failure. The authors have failed to stratify their analysis on the time of dropout, and their analysis is thus not informative with respect to dropout for the important endpoints of treatment failure or persistent side effects.

As a clinician I am primarily interested in the effectiveness of the drug in those who actually take it. It is thus disappointing that the authors gave short shrift to their secondary outcome measures, all of which show significant benefit of active treatment.

Sincerely,

Murray M. Finkelstein PhD MDCM CCFP

Department of Family and Community Medicine

Mt Sinai Hospital

Toronto

Conflict of Interest:

None declared