W73. GENETIC AND SOCIOENVIROMENTAL CORRELATES OF SYMPTOM DIMENSIONS ACROSS THE PSYCHOSIS SPECTRUM AT BASELINE AND 10-YEAR FOLLOW-UP

Abstract

Background: Psychosis primarily manifests in the popula- tion as a spectrum influenced by genetic susceptibility to mental health conditions, neurodevelopmental impairment, and exposure to socio-environmental risk factors. The interplay of these determinants makes it challenging to use discrete operationalised diagnoses to their fullest potential in clinical practice and to validate them with biological markers. Consequently, it has been suggested that patient characterisation should encompass transdiagnostic symptom dimensions and presumed aetiology. In line with this approach, this study aims to explore the biological and environmental correlates within the transdiagnostic continuum of psychosis in samples from the general population, individuals suffering from a psychotic disorder, and their siblings. Methods: We examined baseline and follow-up data from the GAP, EU-GEI, and UK Biobank study samples. Using multidimensional item response modelling, we estimated bifactor models of psychopathology to generate dimension scores for positive, negative, disorganization, manic, and depressive symptoms of psychosis. These bifactor models were based on 1) Subclinical psychosis experiences in the general population and patients diagnosed with a psychotic disorder (UK Biobank, GAP, EU-GEI), as well as their siblings (EU-GEI), using the Community Assessment of Psychic Experiences (CAPE); 2) full psychosis symptoms in first episode psychosis patients and their 5- and 10-year follow-ups (GAP and EU-GEI), using the OPerational CRITeria (OPCRIT) system. Results: The bifactor measurement demonstrated good strength and reliability indices across psychosis experiences and symptoms in the general population, patients with a psychotic disorder and their siblings, and at different stages of the condition. We found that symptom dimension presentation at baseline was associated with specific polygenic and environmental exposures. Individuals presenting with higher scores on the negative experience/symptom dimension had greater genetic susceptibility for schizophrenia and aggression, accompanied by differentially methylated regions. Conversely, those with higher scores on the positive experience/symptom dimension had increased exposure to environmental risk factors, particularly cannabis use. The follow-up data suggested that schizophrenia polygenic scores, baseline symptom dimensions, and exposure to multiple environmental risk factors were associated with different symptom trajectories. Additionally, cannabis use at follow-up was associated with violence and aggression outcomes. Discussion: Our study shows that the bifactor model of psychopathology is robust across categories of subclinical and clinical psychosis, diagnosis, and stages of the condition. Furthermore, our findings support the evidence that specific dimensions of psychosis experiences and symptoms are influenced by different genetic and environmental factors across the spectrum. These findings advocate for characterising the expression of psychosis by symptom dimensions and presumed aetiology.