LAUSR

OBJECTIVE In youth with first-episode psychosis, we evaluated 1) whether early-non-response to antipsychotics predicts non-response and non-remission; 2) patient and illness characteristics as outcome predictors; and 3) the response prediction of the Positive and Negative Syndrome Scale (PANSS)-30-items compared to the PANSS-6-items and Clinical Global Impression-Improvement scale (CGI-I). METHOD Post-hoc-analysis from a 12-week, double-blinded, randomized trial of aripiprazole versus quetiapine-extended-release in adolescents with first-episode psychosis, age=12-17 years. We defined 1) Early-non-response (week-2 or week-4) as <20% symptom reduction (PANSS-30) [or <20% symptom reduction (PANSS-6) or CGI-I score=4-7 (less than "minimally improved")]; 2) non-response (week-12) as <50% symptom reduction (PANSS-30); 3) non-remission (week-12) as a score of >4 on eight selected PANSS-items. We analyzed 1) Positive/negative predictive values (PPV/NPV) and Receiver Operating Characteristics; 2) Binary logistic regression models; 3) PPV/NPV using PANSS-6 and CGI-I. RESULTS 84 of 113 randomized patients were included in this post-hoc-analysis (age=15.7±1.3 years; male=28.6%). The 12-week symptom decrease was 31.9±27.9%, most pronounced within the first two weeks (61.1% of total PANSS reduction). Response (27.4%) and remission (22.6%) rates were low. Results indicated that 1) early-non-response reliably predicted 12-week non-response (PPV: week-2=82.2%; week-4=90.0%) and non-remission (PPV: week-2=80.0%; week-4=90.0%); 2) early-non-response at week-4 was a statistically significant baseline predictor for 12-week non-response; 3) PANSS-6 had a similar predictive significance as PANSS-30, however outcomes were heterogeneous using CGI-I. CONCLUSION In youth with first-episode psychosis showing early-non-response to aripiprazole or quetiapine-extended-release, switching antipsychotic drug should be considered. PANSS-6 is a feasible and clinically relevant alternative to PANSS-30 to predict 12-week non-response/non-remission.