McGrath PJ, Stewart JW, Quitkin FM, et al
Am J Psychiatry. 2006;163:1542-1548
The goal of this study was to determine whether patterns of response to antidepressant medications might predict future relapse to a depressive episode. Treatment for 570 patients diagnosed with a major depressive disorder was initiated by prescribing increasing dosages of fluoxetine for a duration of 12 weeks. Responders (n = 292) were then randomized to either continue on fluoxetine or be placed on placebo for 52 weeks or until time of relapse. Responders were differentiated as to being "true drug" responders (a delayed and persistent response to treatment) vs "placebo" responders (early and/or intermittent response to treatment). Other factors that were evaluated in terms of predicting relapse included:
Presence or absence of atypical depressive neurovegetative symptoms (weight gain and hypersomnia)
Chronicity of illness
Symptom severity
Despite the fact that maintenance on fluoxetine was predictive of a lower relapse rate, "true drug" response vs "placebo" response was not predictive of relapse. What was associated with increased rates of relapse, with no difference between fluoxetine or placebo treatment, included:
Chronicity of illness
Severity of depressive symptoms at the beginning of the maintenance phase
Atypical neurovegetative depressive symptoms
Female gender
Patients who appear to be successfully treated with an antidepressant medication will relapse up to 40% of the time. It is often difficult to predict which of our patients will become victim to relapse when adherent to medications. This study tried to evaluate response pattern as a possible predictor of future relapse. While the active medication fluoxetine was superior to placebo in maintaining treatment response and decreasing relapse rates compared with placebo, the premise that "placebo" responders would relapse at a higher rate than "true drug" responders was not borne out in this study. This may be due, in part, to the extreme difficulty in differentiating a true pharmacologic response from a placebo response by looking at initial time course of response. The investigators defined placebo response as occurring before week 3 of treatment. This might be a psychopharmacology folklore, since recent studies indicate that initial response to an antidepressant medication might occur within the first 2 weeks of treatment.
As confirmed in this trial, factors such as illness chronicity and reverse neurovegetative symptoms were associated with increased rates of relapse, both in the placebo and the fluoxetine cohorts. It was surprising, however, that females had a higher relapse rate; previous research has shown that they may be higher responders to antidepressant treatment. Overall, this study highlights how little we currently know about predicting response and relapse for patients on antidepressant medications.
Abstract