Early Nonresponse in the Antipsychotic Treatment of Acute Mania : A Criterion for Reconsidering Treatment? Results From an Individual Patient Data Meta-Analysis

OBJECTIVE: To investigate whether early nonresponse to antipsychotic treatment of acute mania predicts treatment failure and, if so, to establish the best definition or criterion of an early nonresponse. DATA SOURCES: Short-term efficacy studies assessing antipsychotics that were submitted to the Dutch Medicines Evaluation Board during an 11-year period as part of the marketing authorization application for the indication of acute manic episode of bipolar disorder. Pharmaceutical companies provided their raw patient data, which enabled us to perform an individual patient data meta-analysis. STUDY SELECTION: All double-blind, randomized, placebo-controlled trials assessing the efficacy of antipsychotics for acute manic episode of bipolar disorder were included (10 trials). DATA EXTRACTION: All patients with data available for completer analysis (N = 1,243), symptom severity scores on the Young Mania Rating Scale (YMRS) at weeks 0, 1, and 2 and at study end point (week 3 or 4). RESULTS: The a priori chances of nonresponse and nonremission at study end point were 40.9% (95% CI, 38.2%-43.6%) and 65.3% (95% CI, 62.0%-68.6%), respectively. Early nonresponse in weeks 1 and 2, defined by cutoff scores ranging from a ≤ 10% to a ≤ 50% reduction in symptoms compared to baseline on the YMRS, significantly predicted nonresponse (≤ 0% symptom reduction) and nonremission (YMRS score higher than 8) in week 3. The predictive value of early nonresponse (PVnr_se) at week 1 for both nonresponse and nonremission at study end point declined linearly with increasing cutoff scores of early nonresponse; nonresponse: 76.0% (95% CI, 69.7%-82.3%) for a ≤ 10% response to 48.7% (95% CI, 45.5%-51.9%) for a ≤ 50% response; nonremission: 92.2% (95% CI, 88.3%-96.1%) for a ≤ 10% response to 76.8% (95% CI, 74.4%-79.5%) for a ≤ 50% response. A similar linear decline was observed for increasing cutoff scores of early nonresponse at week 2 for nonresponse, but not for nonremission at end point: nonresponse 90.3% (95% CI, 84.6%-96.0%) for a ≤ 10% response to 65.0% (95% CI, 61.4%-68.6%) for a ≤ 50% response; nonremission: 94.2% (95% CI, 89.7%-98.7%) for a ≤ 10% response and 93.2% (95% CI, 93.1%-95.1%) for a ≤ 50% response. Specific antipsychotic characteristics did not modify these findings at either time point (week 1: P = .127; week 2: P = .213). CONCLUSIONS: When patients fail to respond early (1-2 weeks) after the initiation of antipsychotic treatment for acute mania, clinicians should reconsider their treatment choice using a 2-stage strategy