65. HIGHER POLY-METHYLATION SCORE FOR SCHIZOPHRENIA IS ASSOCIATED WITH GREATER NEGATIVE SYMPTOM SEVERITY AT 10-YEAR FOLLOW-UP

Background Negative symptoms are among the most disabling features of psychotic disorders and frequently persist over time. Baseline clinical measures (e.g. symptom severity at illness onset) explain only part of the variability in long-term outcomes, highlighting the need for improved prognostic markers. DNA methylation reflects the combined influence of genetic liability and environmental exposures and has been associated with psychotic disorders. Poly-methylation scores, which aggregate methylation signals across multiple loci into composite indices, may therefore serve as biomarkers of treatment outcomes. This study investigated whether a schizophrenia-derived poly-methylation score measured at inclusion (i.e., baseline bSCZ-PMS) is associated with negative symptom severity at a 10-year follow-up, beyond established clinical measures including the Positive and Negative Syndrome Scale-Negative subscale (PANSS-N) and the Global Assessment of Functioning-Functioning subscale (GAF-F). Methods We calculated a bSCZ-PMS using discovery data from 3,181 schizophrenia cases and controls and tested the score in 160 individuals with psychosis enrolled in the TOP longitudinal sub-study (University of Oslo, Norway). We first examined whether bSCZ-PMS measured at inclusion was associated with three clinically defined outcome groups (sustained, transient, or minimal/moderate). We further evaluated whether bSCZ-PMS was associated with PANSS-N at 10 years outcome and if adding bSCZ-PMS to models containing baseline predictors (PANSS-N and GAF-F) improved explanatory value. Associations with 10-year PANSS-N were assessed using linear regression, with a residualized-change model conducted as a sensitivity analysis. Results Higher bSCZ-PMS was observed in individuals with sustained negative symptoms at 10-year follow-up compared with the transient or minimal/moderate symptoms groups (p = 0.009). bSCZ-PMS provided additional explanatory value beyond baseline clinical predictors (β = 0.061, 95 % CI = 0.018–0.105, p = 0.006). A one-standard deviation increase in bSCZ-PMS corresponded to a 6.3% increase in PANSS-N score at follow-up. Consistent findings were observed in the residualized-change analysis. Discussion bSCZ-PMS was positively associated with long-term negative symptom severity independently of baseline symptoms and functioning. These findings indicate that methylation-based scores may capture biologically encoded information not reflected in standard clinical assessments. Replication in independent cohorts will be necessary to determine the potential value of bSCZ-PMS alongside clinical measures and to evaluate whether it may ultimately serve as a biomarker of long-term outcome in psychotic disorders.