Anne Katrine Pagsberg1, Pia Jeppesen1, Dea Gowers Klauber1, Karsten Gjessing Jensen1, Ditte Rudå1, Jens Richardt M. Jepsen1, Birgitte Fagerlund1, Amanda Krogmann2, Laura von Hardenberg2, Anders Fink-Jensen1, Christoph U Correll2,3, Britta Galling2; 1Copenhagen University Hospitals, 2Charité Universitätsmedizin, 3Hofstra University
Objective: To evaluate whether early response/non-response (ER/ENR) to antipsychotics can predict ultimate response/non-response (UR/UNR) and remission/non-remission, and to compare the Positive and Negative Syndrome Scale (PANSS)-30-items to the PANSS-6-items and Clinical Global Impressions Scale (CGI) regarding prediction of response/non-response and remission/non-remission in youth with psychosis. Methods: Data from 12-week, double-blinded, randomized trial of aripiprazole (2.5-20 mg/day) versus quetiapine (50-600 mg/day) in adolescents with first-episode psychosis (age=12-17 years). Therapeutic antipsychotic doses were reached at day 9. ER definition: ≥20% symptom reduction (PANSS-30) at week-2 or week-4 (ER2/4) (or: “≥minimally improved” (CGI-I) and ≥20% symptom reduction (PANSS-6)). UR definition: ≥50% symptom reduction. Remission definition: “Andreasen criteria”. Analyses: positive/negative predictive values (PPV/NPV), and logistic regression to identify moderators/mediators of UR/UNR and remission/non-remission. Results: Patients (n=87; age=15.2±1.4 years; male=35.6%) had a mean 32.0±28.2% symptom decrease, most pronounced within the first two weeks. UR and remission rates were low (28.7%; 19.5%). ER2/ER4 could not predict UR (ER2: PPV=40.0%, ER4: PPV=38.2%) and remission (ER2: PPV=27.5%, ER4: PPV=27.3%), while ENR reliably predicted UNR (ENR2: NPV=80.0%, ENR4: NPV=87.5%) and non-remission (ENR2: NPV=84.4%, ENR4: NPV=90.6%). Besides ER/ENR-status, no significant demographic, illness or treatment variables were significant moderators/mediators of endpoint outcomes. PANSS-6 had similar predictive significance as PANSS-30 for non-response/non-remission. Outcomes were heterogeneous using CGI. Conclusion: Clinical treatment decisions should be informed by standardized assessments. A <20% reduction of PANSS-6-items at week-2 and, even more so, at week-4 is a feasible and clinically relevant alternative to PANSS-30 assessments to reliably predict 12-week non-response/non-remission. In patients with ENR4, consider antipsychotic switch.
