Ultrarare Coding Variants and Cognitive Function in Schizophrenia

IMPORTANCE: Impaired cognitive function in schizophrenia is associated with poor functional outcomes, but the role of rare coding variants is unclear. OBJECTIVE: To determine whether ultrarare constrained variants (URCVs) are associated with cognition in patients with schizophrenia. DESIGN, SETTING, AND PARTICIPANTS: Linear regression was used to perform a within-case genetic association study of URCVs and current cognition and premorbid cognitive ability. A multivariable linear regression analysis of the outcomes associated with URCVs, schizophrenia polygenic risk score, polygenic risk score for intelligence and schizophrenia associated copy number variants on cognitive ability was performed. Exome sequencing data from 802 participants with schizophrenia were assessed for current cognition using the Measurement and Treatment Research to Improve Cognition in Schizophrenia Consensus Cognitive Battery and for estimated premorbid IQ using the National Adult Reading Test. Individuals were recruited from clinical and voluntary mental health services in the UK. Those with a diagnosis of intellectual disability or a neurological disorder known to affect cognition were excluded. Data collection occurred between 2007 and 2015. Data were analyzed between April 2020 and March 2022. MAIN OUTCOMES AND MEASURES: Association between URCVs, current cognition, and current cognition adjusted for premorbid IQ. RESULTS: Of the 802 participants, 499 (62%) were men and 303 (38%) were women; mean (SD) age at interview was 43.36 (11.87) years. Ultrarare constrained variants (n = 400) were associated with lower current cognition scores (β = −0.18; SE = 0.07; P = .005). In the univariable analysis, premorbid IQ was associated with URCVs (β = −0.12; SE = 0.05; P = .02) and partly attenuated the association with current cognition (β = −0.09; SE = 0.05; P = .08). Multivariable analysis showed that measured genetic factors combined accounted for 6.2% of variance in current cognition, 10.3% of variance in premorbid IQ, and supported outcomes of URCVs associated with current cognition independent of premorbid IQ (β = −0.10; SE = 0.05; P = .03). CONCLUSIONS AND RELEVANCE: The findings of this study suggest that URCVs contribute to variance in cognitive function in schizophrenia, with partly independent associations before and after onset of the disorder. Although the estimated effect sizes were small, future studies may show that the effect sizes will be greater with better annotation of pathogenic variants. Genomic data may contribute to identifying those at particularly high risk of cognitive impairment in whom early remedial or preventive measures can be implemented.