A diagnostic test to examine early improvement as a predictor of later response to lurasidone in bipolar depression

Introduction: Kato et al. reported results of a 6-week, double-blind, randomized, placebo-controlled trial of lurasidone in adults with bipolar depression (BDep). Aim: We performed a post hoc analysis using data from the lurasidone trial to predict later responses from early improvements. Methods: An early improvement was defined as a ≥20% reduction in Montgomery–Åsberg Depression Rating Scale (MADRS) total score at Week 2; response was defined as a ≥50% reduction in MADRS total score at Week 6; symptomatic remission were defined as a score of ≤8 on MADRS total score at Week 6. Results: Both sensitivity and negative predictive value (NPV) were higher for the remission outcome than for the response outcome. The interpretation of sensitivity and NPV in the lurasidone group when remission is an outcome is as follows. It means (1) that, from all remitters at Week 6, 80.6% was identified as such at Week 2 on the basis of their early improvement and (2) that a patient showing non-improvement at 2 weeks had 93.5% probability of being a non-remitters at Week 6. However, the values of specificity for both response and remission in the lurasidone group were not high. Conclusion: Patients who did not show an early response at Week 2 cannot be predicted with a high probability to also show poor improvement at Week 6. In fact, some patients who did not show early response at 2 weeks might have marked improvement at 6 weeks.