SA9. Associations of Subclinical Psychosis Spectrum Characteristics, Childhood Depressive Symptoms, and Schizophrenia Polygenic Risk Score: Analyses From the Philadelphia Neurodevelopmental Cohort

Abstract

Background: Despite heterogeneity in adult clinical presentation of psychotic disorders, the childhood precursors and genetic underpinnings of this heterogeneity are unclear. Risk profile scores (RPS) summarize an individual’s risk for developing schizophrenia based on common genetic variants associated with the disease. The Philadelphia Neurodevelopmental Cohort (PNC) is a publicly available data set including genetic and psychopathology information for a large sample of adolescents/young adults. In light of findings of increased affective symptoms in youth at risk of schizophrenia, we examined the association of childhood psychosis spectrum (PS) characteristics and childhood depressive symptoms in relation to RPS. Methods: Complete neuropsychiatric and RPS data were available for 3217 participants of European-American ancestry (ages 11–21) from the PNC. From a large psychopathology assessment, items related to prodromal symptoms and psychosis-like experience were combined into a composite PS score. Using other items tracking DSM-IV criteria, we created binary variables representing “quasi-diagnoses” (QDx) for childhood psychiatric conditions including depression. RPS were calculated using the most strongly illness-associated genetic variants identified by the multinational Psychiatric Genetics Consortium. We tested for associations of PS and depression QDx with RPS using logistic regression, then tested for differences in RPS across QDx and levels of PS with mixed model analyses. Analyses in the full sample were repeated for “11–15” and “16 and older” subgroups. Results: We divided the sample into top 20%/bottom 80% PS groups for the analyses. Prevalence of the QDx for at least 1 depressive episode was 12.1%. PS group alone was not associated with RPS. Depression episode QDx was associated with RPS at the more restrictive RPS P value thresholds (e.g., RPS_5e-08 P = .003; R2 = .009). In mixed model analyses for the full sample (ages 11–20), we observed a small main effect of depression QDx on RPS for schizophrenia (e.g., RPS_5e-08 P = .003; η2 = .003). We observed no effect of depression QDx or psychotic experience on RPS in the younger subgroup (11–15) and no interaction. The older group (16–20) again showed a main effect of depression (RPS_5e-08 P =.0003; η2 = .009) and, also, a small but significant interaction of depression with PS(RPS_5e-08 P = .012; η2 = .004). Findings were similar for males and females. Conclusion: Affective symptoms are frequently observed in youth at risk of schizophrenia. These analyses suggest that, especially in the years leading up to the typical age of onset, common genetic risk of schizophrenia may be more strongly associated with depressive symptomatology than with psychotic-like experience. The associations of RPS with psychopathology were evident for the RPS based on the most strongly associated (“GWAS significant”) markers and not at more inclusive levels of RPS.