Trajectories of response to treatment with atypical antipsychotic medication in patients with schizophrenia pooled from 6 double-blind, randomized clinical trials

Abstract

Research has identified distinct trajectories of antipsychotic response in patients with chronic schizophrenia in short-duration trials (~ 12 weeks). This post-hoc analysis identified trajectories in patients with chronic schizophrenia treated for ≤ 24 weeks. We pooled data from 1990 patients with chronic schizophrenia from 6 randomized, double-blind, olanzapine-comparator trials of atypical antipsychotics. Trajectory analysis identified homogeneous subpopulations within the larger heterogeneous population. Baseline demographics were compared between the identified latent classes. Five distinct response trajectories based on Positive and Negative Syndrome Scale (PANSS) Total score were identified: Dramatic Responders (n = 47/1990, 2.4%), severely-ill patients (PANSS = 124) with rapid and sustained improvement (51%) by Week 3; Partial Responders (n = 1802/1990, 90.6%), moderately-ill (PANSS = 90) with minimal improvement (21%) by Week 4, and little further improvement; Partial Responders-Unsustained (Late) (n = 32/1990, 1.6%), markedly-ill (PANSS = 95) with minimal initial improvement followed by worsening after Week 12; Partial Responders-Unsustained (Early) (n = 28/1990, 1.4%), markedly-ill (PANSS = 102) with minimal initial improvement followed by worsening after Week 8; and Delayed Responders (n = 81/1990, 4.1%), markedly-to-severely-ill (PANSS = 113) with minimal (11%) improvement at Week 8, but noticeable improvement thereafter (49%). Significant differences were noted for several baseline characteristics (p < .05) and discontinuation rates (46%–72%). Dramatic Responders were younger and more likely to be female and Hispanic with higher baseline illness severity. Analysis of antipsychotic response over 24 weeks in a large, pooled, heterogeneous population treated for schizophrenia revealed 5 distinct trajectories. Most patients had modest and sustained improvements during atypical antipsychotic treatment, regardless of their baseline illness severity, representing a partial response to currently available treatments.

Introduction

Schizophrenia is a devastating mental illness affecting approximately 24 million people worldwide (World Health Organization Statistics on Schizophrenia, 2010). Though some patients experience dramatic symptom improvement in response to currently available antipsychotic treatment, 70% of patients fail to show even minimal response early in treatment (Ascher-Svanum et al., 2008, Kinon et al., 1993, Kinon et al., 2008) and over the longer term, medication will effectively treat only about half of patients (Lieberman et al., 2003). Identifying the earliest possible point at which a given degree of response is indicative of how a patient will respond over the longer term if that treatment is continued could minimize patient exposure to ineffective, potentially even harmful medication, and allow for earlier assessment and adjustment of medication to achieve optimal outcome. Until recently, these analyses have identified patients with early response to treatment by using a prespecified threshold for measures assessed early in treatment. For instance, patients with ≥ 20% improvement in Positive and Negative Syndrome Scale (PANSS) (Kay et al., 1987) Total score at Week 2 were identified as Early Responders and those with less response at Week 2 were identified as Early Non-Responders. Such strategies produced dichotomous categories, possibly excluding a patient from a clinically relevant response category. As an example, a patient with a 19% improvement in PANSS Total score at Week 2 may not have differed qualitatively from a patient with a 20% improvement at that time point, though the former would have been identified as a Non-Responder, and the latter as a Responder. Even when change over multiple time points was considered, mean values for improvement over time were usually presented by treatment group as smooth lines which did not provide a sense of the heterogeneity within each group, nor the likely overlap between them.

Using varied statistical methods, others have investigated these longitudinal profiles to reveal latent classes of response within the heterogeneity of treatment response trajectories over time. Trajectory analysis characterizes and identifies subgroups of patients who differ in response to treatment and identifies baseline characteristics that help distinguish to which subgroup a given individual may belong (Muthén et al., 2002, Muthén and Shedden, 1999). These methodologies have been applied to datasets from clinical trials of treatments for schizophrenia (Case et al., 2010, Levine and Leucht, 2010, Levine and Rabinowitz, 2010, Levine et al., 2010a, Levine et al., 2010b, Marques et al., 2010; Table 1).

In this post hoc analysis of data pooled from 6 clinical trials, our objective was to describe trajectories of response for a large population of patients with chronic schizophrenia, and to assess whether patient characteristics at baseline are associated with specific response trajectories. This analysis differs from those previously published because it examines response trajectories in a heterogeneous population, primarily patients in the chronic phase of schizophrenia who were treated with antipsychotics for up to 24 weeks.