This work may not be copied, distributed, displayed, published, reproduced, transmitted, modified, posted, sold, licensed, or used for commercial purposes. By downloading this file, you are agreeing to the publisher’s Terms & Conditions.

Original Research

Determinants of Antipsychotic Response in Schizophrenia: Implications for Practice and Future Clinical Trials

Jonathan Rabinowitz, PhD; Nomi Werbeloff, PhD; Ivo Caers, PhD; Francine S. Mandel, PhD; Virginia Stauffer PharmD; François Ménard, MD; Bruce J. Kinon, MD; and Shitij Kapur, MBBS, PhD

Published: April 15, 2014

Article Abstract

Background: Response to antipsychotics in schizophrenia is highly variable, and determinants are not well understood or used to design clinical trials.

Objective: We aimed to understand determinants of response to antipsychotic treatment.

Method: Supported by the Innovative Medicines Initiative, as part of a large public-private collaboration (NEWMEDS), we assembled the largest dataset of individual patient level information from randomized placebo-controlled trials of second-generation antipsychotics conducted in adult schizophrenia patients by 5 large pharmaceutical companies. The dataset included all placebo-controlled trials of risperidone, paliperidone, ziprasidone, sertindole, olanzapine, and quetiapine. We examined patient- and trial-design-related determinants of outcome as measured by change on the Positive and Negative Syndrome Scale in 29 placebo-controlled trials (drug, n = 6,971; placebo, n = 2,200) and initial findings confirmed in additional data from 5 separate trials (drug, n = 1,699; placebo, n = 580).

Results: While it is conventional for trials to be 6 weeks long, drug-placebo differences were observable at week 4 with nearly the same sensitivity, and dropout rates were lower. Having any of these attributes was associated with significantly greater drug versus placebo differences in symptom improvement and rates of study completion: being female (P ≤ .04), being a young adult patient who is a few years beyond the first episode (P ≤ .03), having prominent positive and negative symptoms (P ≤ .03), and living in Eastern Europe versus North America (P ≤ .04). Contrary to prevalent clinical opinion, age at onset and use of benzodiazepines did not show a differential treatment response, and patients just above PANSS inclusion threshold were not overrepresented.

Conclusions: Proof-of-concept trials can be shorter and efficiency improved by including an even distribution of sexes and of patients with prominent symptomatology, thus reducing patient exposure to placebo and experimental treatments.

J Clin Psychiatry 2014;75(4):e308-e316

Submitted: October 22, 2013; accepted January 28, 2014 (doi:10.4088/JCP.13m08853).

Corresponding author: Jonathan Rabinowitz, PhD, Bar Ilan University, Ramat Gan, Israel 91000 (

Volume: 75

Quick Links:

Continue Reading…

Subscribe to read the entire article


Buy this Article as a PDF