• treatment response
• recurrence
• major depression
• bipolar spectrum
• bipolar disorder

Objective: The presence of unrecognized bipolar disorder or “bipolar spectrum” features has been suggested to contribute to poor treatment response in major depressive disorder. We aimed to investigate the association between putative bipolar spectrum features and clinical outcomes in a cohort of fluoxetine-treated patients with MDD.
Method: N= 602 outpatients aged 18-65 years with major depressive disorder recruited at 2 academic medical centers first entered a 12-week phase of open-label treatment with fluoxetine titrated up to 60mg/day. Patients who met the response criteria by week 12 entered the second phase of the study during which they were double-blindly randomized either to continue the same fluoxetine doses to which they had responded or to take placebo, for 52 weeks or until the occurrence of a relapse. The following clinical features suggestive of bipolar illness were selected for analysis: a history of early onset and recurrent depression, baseline atypical depressive features, irritability, psychoticism, suicidality, interpersonal sensitivity, comorbid anxiety disorders, and substance abuse/dependence. These measures were condensed into a summary score of bipolarity ranging from 0 to 10 points. We considered as primary outcomes time to response, remission, and discontinuation during acute treatment and time to relapse in the second phase of the study, utilizing survival analyses.
Results: Higher scores on the summary measure of bipolarity were not associated with differential acute treatment outcomes. They were significantly associated with a shorter time to relapse (p = 0.016). The median time until first recurrence was 52.0 weeks in the group with lower scores vs. 12.0 wk in the group with higher scores.
Conclusion: Bipolar spectrum features may be associated with shorter time to recurrence in MDD patients after recovery, suggesting some predictive validity for this measure.