Early symptom change and prediction of subsequent remission with olanzapine augmentation in divalproex-resistant bipolar mixed episodes

Potential predictors of remission in mixed bipolar I disorder were identified using early Clinical Global Impression-Severity (CGI-S) improvement criteria in divalproex-resistant patients randomized to olanzapine augmentation (olanzapine + divalproex; N = 101) in a 6-week, double-blind, placebo-controlled trial. In a post-hoc analysis, receiver operating characteristics of 1-point decreases in the CGI-S total score after 2, 4, 7, and 14 days were examined as predictors of endpoint (Week 6 or last observation) remission of depression and/or mania as defined by 21-item Hamilton Depression Rating Scale (HDRS-21) and Young Mania Rating Scale (YMRS) total score ≤8. Based on a 1-point improvement in CGI-S as a predictor of remission, all odds ratios (ORs) and 95% confidence intervals (CIs) were statistically significant for depression or mania remission criteria. ORs for mixed symptom remission with a decrease ≥1 in CGI-S scores at Day 2 for olanzapine augmentation were (6.727; CI: 2.382, 18.997; p < .001) with negative predictive value = 89.5% and positive predictive value = 44.2%. Changes in HDRS-21 and YMRS individual item scores after 2 days of augmentation as predictors of endpoint remission identified that decreases in HDRS-21 symptom item scores (early, middle, and/or late insomnia; paranoid; agitation; and somatic/gastrointestinal) predicted depressive symptom remission at endpoint, and decreases in YMRS item scores (language-thought disorder and irritability) were associated with manic symptom remission at endpoint. Because remission with augmentation therapy may occur in as few as one in ten individuals who lack very early symptom reduction, lack of early improvement may indicate a need to expediently reassess treatment strategy.