T26. CLINICAL OUTCOME ASSESSMENTS FOR TRIALS IN COGNITIVE IMPAIRMENT ASSOCIATED WITH SCHIZOPHRENIA

Abstract

AbstractBackgroundFor clinical development programs in cognitive impairment associated with schizophrenia (CIAS), it has been stated that significant improvement on a cognitive performance endpoint would be necessary, but not sufficient, for drug approval. In addition, regulators require improvement on a functionally meaningful co-primary. A co-primary approach may require a more conservative criterion be set for statistical significance: first a modest impact of correlation between the endpoints is observed on required sample size for a given effect size and power; second a more important impact of effect size is observed. If the standardized effect size for one endpoint is smaller, the sample size is determined by this. In the recent draft guidance for treatment of early AD, the FDA acknowledge the entity of cognition as meaningful and they outline circumstances in which integrated cognition-function, or cognition only outcomes would be acceptable as single primary endpoints. In addition, it is stated that when measured using conventional assessments, the meaningfulness of cognition may not be apparent.MethodsA review was conducted of Phase 2 and 3 industry clinical trials for schizophrenia listed on clinicaltrials.gov, and which studied potential for cognitive improvement. Fifty-two trials were identified and the most commonly reported outcomes for cognition and function identified.ResultsThe most commonly used and reported outcomes have been the MCCB for cognition and the UPSA for function. The other cognitive outcomes used most frequently (≥7 trials) have been the Cogstate, and BACS batteries, whilst the other most frequently used functional outcomes (≥5 trials) have been the SCoRS and PSP. Furthermore, several versions of the UPSA have been employed (e.g. UPSA-B, UPA-2, UPSA-VIM). These clinical outcome assessments (COAs) fall into groups of cognitive performance-based outcome (PerfO), functional capacity assessments, and clinician-reported outcome (ClinRO) assessments. Whilst no treatment is yet approved for CIAS, several trials have shown encouraging data in early clinical development and certain antipsychotic trials have also shown evidence for improvement.DiscussionCognition as measured by conventional cognitive PerfO assessments is most commonly an indirect measure of meaningful health outcomes (i.e. the test activities are not a part of a person’s usual normal life). Consequently, the meaning of any change in a cognitive score is not intrinsically known. This lack of direct measurement of meaningful health outcomes drives the requirement for the use of co-primary endpoints. There may be alternate approaches such as establishing cognitive only outcomes as intermediate or surrogate endpoints or creating integrated cognition-function endpoints with greater statistical power. Sufficient data likely exist across completed industry trials to facilitate the development and validation of such endpoints. Furthermore, novel cognitive PerfO may be developed that are more direct measures of health outcomes and could also form single primary endpoints.