Relationship of prior antidepressant exposure to long-term prospective outcome in bipolar I disorder outpatients.

OBJECTIVE The long-term impact of prior antidepressant exposure on the subsequent course of bipolar illness remains controversial. METHOD 139 outpatients (mean age, 42 years) with bipolar I disorder diagnosed by DSM-IV criteria had a detailed retrospective examination of their prior course of illness on the National Institute of Mental Health Life Chart Method. Number of prior antidepressant trials and total duration of antidepressant exposure were assessed. Prospective long-term response (for at least 6 months) to naturalistic treatment in the network from 1996 through 2002 was the primary outcome measure as it related to prior antidepressant exposure (and other illness variables) by logistic regression, with P < .05 used for statistical significance in this post hoc analysis. RESULTS Greater number of antidepressant trials, but not duration of antidepressant exposure, was related to prospective nonresponse (P = .0051) whether or not antidepressants were covered by concurrent treatment with a mood stabilizer or atypical antipsychotic. Poor prospective response was also independently related to having had an anxiety disorder and 20 or more prior affective episodes. CONCLUSIONS That the number of antidepressant trials, but not duration of antidepressant treatment, was associated with prospective nonresponse suggests that it is the repeated use of antidepressants to treat episodes of depression that is related to poor prospective response to naturalistic treatment. The direction of causality is unclear as to whether more antidepressant trials led to this increased treatment resistance or whether a difficult course of illness with more episodes and anxiety comorbidity engendered more attempts at antidepressant treatment.