Outcome of a three-phase treatment algorithm for inpatients with melancholic depression

In patients suffering from major depressive disorder, non-response to initial antidepressant monotherapy is relatively common. The use of treatment algorithms may optimize and enhance treatment outcome. A single-center 3-phase treatment algorithm was evaluated for inpatients with major depressive disorder, i.e. phase I (n = 85): 7 weeks optimal antidepressant monotherapy (imipramine or venlafaxine); phase II (n = 39): 4 weeks subsequent plasma level-targeted dose lithium addition in case of insufficient improvement of antidepressant monotherapy; and phase III (n = 8): subsequent electroconvulsive therapy in case of insufficient improvement of antidepressant‑lithium treatment. Overall feasibility of the 3-phase algorithm was determined by the number of dropouts, and overall efficacy was evaluated using weekly scores on the 17-item Hamilton Rating Scale for Depression (HAM-D) during the treatment phases of the algorithm. This paper is based on an RCT comparing the two antidepressants in phase I and adding lithium in phase II. Of the 85 patients analyzed, overall dropout during the 3-phase treatment algorithm was 24 (28%) patients. When analyzing the 3-phase treatment algorithm on a modified intention-to-treat basis, 39 (46%) patients achieved complete remission (HAM-D score ≤ 7) by the end of the algorithm. Regarding response (HAM-D score reduction ≥50%): of the 85 patients, 60 (71%) were responders by the end of the algorithm. The favorable outcome of the 3-phase treatment algorithm emphasizes the importance of pursuing stepwise antidepressant treatment in patients who are nonresponsive to the first antidepressant. This study protocol is registered at http://www.controlled-trials.com, “Pharmacological Treatment of Depression” (identifier: ISRCTN73221288). •In the completers, the favorable outcome of the algorithm is shown by a 93% response rate and a 64% remission rate.•Feasibility of the algorithm was reasonable, with an overall dropout rate of 28%.•Both an index episode less than one year, and not having received previous adequate treatment predicted a greater chance of achieving remission.