Innovations In Clinical Neuroscience

NOV-DEC 2017

A peer-reviewed, evidence-based journal for clinicians in the field of neuroscience

Issue link:

Contents of this Issue


Page 56 of 83

57 ICNS INNOVATIONS IN CLINICAL NEUROSCIENCE November-December 2017 • Volume 14 • Number 11–12 R E V I E W domains in schizophrenia; 2) they have high orthogonality, and, thus, because of the low level of between-factor correlation, they are measuring the effect of a drug on a symptom domain with a high degree of specificity; 3) they exhibit minimal loss of information (i.e., after the UPSM transformation, >90% of information variance was retained, with high R-squared values between sums of the transformed PANSS factors and PANSS total score); and 4) the weighted UPSM coefficients have generalizable utility, and, thus, can be applied to PANSS data in a wide range of clinical trials to generate transformed PANSS factors with the same clinimetric properties as the original transformed PANSS factors. Effect size estimates for endpoint change (lurasidone vs. placebo) were notably different when calculated using the standard (Marder) PANSS factors and the transformed PANSS factors. The uniformity of effect sizes calculated using the Marder PANSS factors reflects the high level of between-factor correlation, particularly for the PANSS positive factor (range: r, 0.52–0.74). This level of between- factor correlation contributes to the likely overestimate of the effect of treatment on many of the PANSS factors. The effect size profile of lurasidone, when estimated using transformed PANSS factors, would appear to provide a more valid measure of the efficacy of lurasidone in treating the key clinical symptoms domains of schizophrenia. An important next step will be to replicate these Results for other antipsychotic agents, both first and second generation. The results of the UPSM analysis of baseline and Week 6 (on treatment) data confirm and extend our previously reported 14 UPSM change score analysis. At baseline, standard (Marder) PANSS factor severity scores and transformed PANSS factor severity scores both exhibited low levels of between-factor correlation, suggesting that each factor is measuring a separate clinical symptom domain. However, when measuring the effect of treatment, standard (Marder) PANSS factor change scores exhibited moderate between-factor correlations, indicative of pseudospecificity, while transformed PANSS changes scores exhibited ergodicity, since they continue to show the same domain- independent factor structure that was observed at baseline. Thus, the UPSM transform yields results that support the notion that there is a consistent underlying schizophrenia symptom structure present in both exacerbated and stable (treated) states. Previously,we showed that the transformed PANSS factor change scores were highly correlated with their respective standard (Marder) PANSS factor change scores, 14 indicating that the transformed PANSS factors are measuring the same symptom domains as the Marder factors. In the new analyses reported here, we have shown that transformed PANSS factor scores were highly correlated with their respective standard (Marder) PANSS factor scores at baseline. In addition, we have summarized analyses from a single study where MADRS and NSA assessments were available that provide preliminary evidence of concurrent validity between these measures and the transformed PANSS depression and negative symptom subscores, respectively. CLINICAL IMPLICATIONS For more than a quarter of a century, the PANSS has been the gold standard measure of efficacy for clinical trials in schizophrenia. However, the high degree of correlation among Marder PANSS factors has proved to be its Achilles heel, preventing clinicians, researchers, and regulatory agencies alike from clearly determining whether improvement in (for example) the PANSS negative factor is a targeted drug effect or whether (and to what extent) negative symptom improvement might reflect improvement in a (correlated) PANSS factors. The term pseudospecificity was coined to describe improvement in a symptom domain that is highly correlated (overlapping) with a separate outcome domain in terms of phenomenology and measured treatment response. We view the current UPSM procedure for generating transformed PANSS factors as a means to generate reliable signals of drug effect across symptom domains in acute treatment trials of patients with schizophrenia. In an acute trial involving exacerbated patients, demonstration of improvement in, for example, the transformed negative symptom factor might be viewed as a more reliable indicator of specific treatment effect than change in the standard PANSS negative symptom subscale. Such signal-generating results may then be used to support confirmatory studies or could, in some cases, be potentially used for labelling without the need for additional studies. In addition, the availability of transformed PANSS factors with a high degree of orthogonality/specificity should help clinicians to have a better understanding of the structure of symptom change in schizophrenia and the profile of treatment effects across schizophrenia symptom domains. Thus, reanalysis of existing clinical trials data sets using transformed PANSS factors could provide important insights into the effectiveness of existing antipsychotic agents for the treatment of specific symptom domains in patients with schizophrenia. TABLE 2. Pearson's correlations among PANSS factor severity scores at Week 6 Endpoint (Study 233 22 data) MPOS: Marder positive symptoms; MDIS: Marder disorganized thinking; MNEG: Marder negative symptoms; MHOS: Marder hostility/ excitement; MDEP: Marder depression/anxiety; UPSM: uncorrelated PANSS score matrix; UPSM-POS: positive symptoms; UPSM-DIS: disorganized thinking; UPSM-NAA: negative-apathy/ avolition; UPSM-NDE: negative-deficit of expression; UPSM-DEP: depression symptoms; UPSM-ANX: anxiety symptoms; NSA: Negative Syndrome Assessment scale; MADRS: Montgomery-Åsberg Depression Rating Scale

Articles in this issue

Archives of this issue

view archives of Innovations In Clinical Neuroscience - NOV-DEC 2017