Personality disorders predict relapse after remission from an episode of major depressive disorder: a 6-year prospective study.

OBJECTIVE To examine prospectively the course of major depressive disorder (MDD) and to test for the moderating effects of personality disorder (PD) comorbidity on relapse after remission from an episode of MDD. METHOD Participants were 303 patients (196 women and 107 men) with current DSM-IV-diagnosed MDD at baseline enrollment in the Collaborative Longitudinal Personality Disorders Study. Major depressive disorder and Axis I psychiatric disorders were assessed with the Structured Clinical Interview for DSM-IV, and Axis II PDs were assessed with the Diagnostic Interview for DSM-IV Personality Disorders. The course of MDD was assessed with the Longitudinal Interval Follow-up Evaluation at 6 and 12 months and then yearly through 6 years. Survival analyses were used to analyze time to remission and time to relapse. The study was conducted from July 1996 to June 2005. RESULTS Of 303 patients, 260 (86%) remitted from MDD; life table survival analyses revealed that patients with MDD who had PDs at baseline had significantly longer time to remission from MDD than patients without PDs. Among the 260 patients whose MDD remitted, 183 (70%) relapsed. Patients with MDD with PDs-specifically those with borderline and obsessive-compulsive PDs-at baseline had significantly shorter time to relapse than patients with MDD without PDs. Cox proportional hazards regression analyses revealed that the presence of PDs at baseline (hazard ratio = 1.5) and recurrent-type MDD (hazard ratio = 2.2), but not sex (hazard ratio = 1.03) or dysthymic disorder (hazard ratio = 0.97), significantly predicted time to relapse. CONCLUSIONS Personality disorders at baseline were robust predictors prospectively of accelerated relapse after remission from an episode of MDD. Personality disorders at baseline significantly moderated eventual time to relapse in MDD among patients who remitted from an episode of MDD, even when controlling for other potential negative prognostic predictors.