Abstract
BACKGROUND: Cognitive impairment is common in people with schizophrenia and is among the strongest predictors of functional decline in this patient group. Despite considerable efforts, there are however no pharmacological treatments for cognitive impairment associated with schizophrenia (CIAS) that have successfully reached efficacy criterion in phase III trials to receive regulatory market approval. The high rate of phase III failures calls into question the reasons why compounds keep failing, despite promising clinical evidence for cognitive improvement in phase II trials. This presentation will provide a brief history on these trials before moving on to discuss the extent to which objective cognitive performance has been used as an eligibility and/or stratification criterion in CIAS intervention trials. The potential implication of these findings for future CIAS research and development will be discussed. METHODS: A systematic search was carried out using ClinicalTrials.gov for all protocols associated with schizophrenia listed between January 2000 and October 2018. Eligible studies were randomized, double-blind, placebo-controlled pharmacotherapy trials conducted in patients with a diagnosis of schizophrenia, in which a cognitive endpoint was the primary outcome measure. For all eligible trials, information was collated (where provided) on: (1) study characteristics (sponsor, year of publication, phase, country where the work was performed); (2) inclusion criteria; (3) any objective cognitive tasks used to assess patient eligibility; (4) randomization procedures; (5) primary and secondary study objectives; (6) pharmacological agent under study. RESULTS: Of the trials that used cognition as an endpoint, only a small minority employed inclusion criteria requiring the presence of an objectively measured cognitive deficit at baseline. In contrast, a much larger number of trials included exclusion criteria to eliminate subjects who had severe cognitive deficits or dementia. The only consistent inclusion criteria across clinical trials were confirmation of a diagnosis of schizophrenia, as determined using the Diagnostic and Statistical Manual of Mental Disorders and/or the Positive and Negative Syndrome Scale. CONCLUSIONS: Even when cognition is the primary outcome and an intervention is intended to ameliorate cognitive dysfunction, the majority of studies did not include formal eligibility criteria to ensure study participants had a cognitive deficit on entry into the trial. While this is consistent with consensus guidelines that have previously recommended such an approach, the increase in CIAS trial failures call this view into question, particularly as neither existing diagnostic criteria nor psychotic symptom severity is indicative of cognitive ability. An evidence-base is building which suggests that not all patients with schizophrenia may benefit from a pro-cognitive agent if they have relatively intact cognition at baseline, relative to normative performance thresholds. Exclusion of these individuals, or at least ensuring equal stratification of these ‘normal’ cognitive performers across trial arms during randomization, may provide additional power to observe pro-cognitive treatment effects in CIAS trials.