Major depressive disorder: A prospective study of residual subthreshold depressive symptoms as predictor of rapid relapse

Background: The study tested whether level of recovery from major depressiv e episodes (MDEs) predicts duration of recovery in unipolar major depressiv e disorder (MDD) patients. Methods: MDD patients seeking treatment at five academic centers were followed naturalistically for 10 years or longer. Pat ients were divided on the basis of intake MDE recovery into residual depres sive symptoms (SSD; N = 82) and asymptomatic (N = 155) recovery groups. The y were compared on time to first episode relapse/recurrence, antidepressant medication, and comorbid mental disorders. Recovery level was also compare d to prior history of recurrent MDEs ( > 4 lifetime episodes) as a predicto r of relapse/recurrence. Results: Residual SSD compared to asymptomatic rec overy patients relapsed to their next MDE > 3 times faster (median = 68 vs. 23 weeks) and to any depressive episode > 5 times faster (median = 33 vs. 184 weeks). Residual SSD recovery status was significantly associated with early episode relapse (OR = 3.65) and was stronger than history of recurren t MDEs (OR = 1.64). Rapid relapse in the SSD group could not be attributed to higher comorbidity or lower antidepressant treatment. Limitations: Altho ugh inter-rater agreement on weekly depressive symptom ratings was very hig h (ICC > 0.88), some error may exist in assigning recovery levels. Antidepr essant treatments were recorded, but were not controlled. Conclusions: NIDE recovery is a powerful predictor of time to episode relapse/recurrence. Re sidual SSD recovery is associated with very rapid episode relapse which sup ports the idea that SSD is an active state of illness. Asymptomatic recover y is associated with prolonged delay in episode recurrence. These findings of this present study have important implications for the goals of treatmen t of MDD and for defining true MDE recovery. (C) 1998 Elsevier Science B.V. All rights reserved.