Early improvement on antipsychotic treatment as a predictor of subsequent response in schizophrenia: analyses from ziprasidone clinical studies

Abstract

We examine data from short-term placebo-controlled and comparator-controlled clinical trials of ziprasidone in schizophrenia to confirm the predictive capacity of early symptom changes for response. We pose the question of how early is too early to consider "stay or switch" and evaluate the predictive capability of a clinical measure in this regard. We presented two separate pooled analyses of (i) two placebo-controlled and (ii) two active comparator (risperidone and olanzapine) randomized trials of ziprasidone in schizophrenia. Relationship between early changes in Positive and Negative Syndrome Scale (PANSS) total, Brief Psychiatric Rating Scale (BPRS), and Clinical Global Impression-Improvement (CGI-I) scores and treatment outcome was evaluated. Week 2 improvement was more reliably predictive of subsequent outcome than week 1 improvement using PANSS and BPRS scores with high sensitivity and specificity, whereas CGI-I had much lower specificity. Overall, non-improvement at week 1 or week 2 was highly predictive of non-response using BPRS scores and PANSS but not CGI-I. These data, independent of antipsychotic used, confirm prior research showing that early improvement in symptoms is predictive of response. There appears to be an important window of time, beyond week 1, during which important clinical decisions to stay or switch medication may be made.