The article you requested is

Early Improvement of Specific Symptoms Predicts Subsequent Recovery in Bipolar Depression: Reanalysis of the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) Data

J Clin Psychiatry 2017;78(2):e146–e151

Objective: The aim of this post hoc analysis was to evaluate which specific depressive items could predict subsequent durable recovery in patients with bipolar depression.

Methods: The study population was at least 18 years old and met DSM-IV criteria for a major depressive episode associated with either bipolar I or II disorder. The data were derived from the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD), in which patients with bipolar depression were randomly assigned to treatment for acute depression with a mood stabilizer plus an adjunctive antidepressant drug or placebo. The primary and secondary outcomes were durable recovery (ie, 8 consecutive weeks of euthymia) and treatment-emergent affective switch (ie, transition to mania or hypomania), respectively. Binary logistic regression analysis was performed to identify specific symptoms whose improvement during the first 2 weeks predicted those outcomes; the score change of each individual symptom in the continuous symptom subscales for depression (SUM-D) from week 0 to week 2 was used as an independent variable.

Results: In the evaluable 188 participants who took placebo and active drugs, the improvement in loss of self-esteem (P = .037) or loss of energy (P = .040) at week 2 was significantly associated with higher chances of subsequent durable recovery. For participants taking active drugs (n = 91), solely the improvement in loss of energy at week 2 was significantly associated with subsequent durable recovery (P = .027). There was a significant association between the improvement of psychomotor retardation at week 2 and subsequent affective switch (P = .008).

Conclusions: These findings imply that focusing on individual symptoms is important in bipolar depression, rather than relying solely on a summed score in rating scales.

Trial Registration: The original STEP-BD dataset is registered on (identifier: NCT00012558).