Early Prediction of Clinical Response in Schizophrenia Patients Receiving the Atypical Antipsychotic Zotepine

Objective: Prior early prediction models for antipsychotic treatment response demonstrate good specificity but poor sensitivity (i.e., high false-negative rates). The purpose of this study was to refine the early prediction model in schizophrenia patients taking an atypical antipsychotic agent, zotepine.

Method: 135 acutely ill inpatients with DSM-IV-defined schizophrenia received 4 weeks of 150 mg/day zotepine treatment. Psychopathology severity was assessed weekly with the Brief Psychiatric Rating Scale (BPRS) and subscales for positive, negative, and general symptoms. Clinical response was defined as a reduction of 20% or more in the BPRS total score at week 4. A logistic regression model was used to obtain early predictors. The receiver operating characteristic curve was employed to determine the optimal cutoff points of the variables for predicting response. The study was conducted from June 2004 to April 2005.

Results: The most significant early predictors for ultimate response at week 4 were BPRS positive subscale score changes at week 1 and, better, at week 2 (p < .001 at both timepoints). At week 1, a BPRS positive score reduction of 4 appeared to be the optimal cutoff point for predicting eventual response, providing a sensitivity of 0.77 and specificity of 0.77. At week 2, a BPRS positive score reduction of 6 was the best for prediction, with a sensitivity of 0.83 and specificity of 0.91.

Conclusions: These findings suggest that using the first 2 weeks’ improvement in positive symptoms to predict the fourth week’s treatment response is favorable in terms of both specificity and sensitivity. Further studies are needed. Moreover, whether this model could be applied to establish a prediction system for other antipsychotics or other psychotropics also deserves research.