Sex differences in the independent and combined effects of genomic and exposomic risks for schizophrenia on distressing psychotic experiences: insights from the ABCD study

Abstract

Purpose: To investigate sex-dependent effects of polygenic risk (PRS-SCZ) and exposome score (ES-SCZ) for schizophrenia, both independently and jointly, on distressing psychotic experiences (PEs) in early adolescents. Method: Baseline to 3-year follow-up data of the Adolescent Brain and Cognitive Development Study (ABCD) were used. PRS-SCZ and ES-SCZ were calculated to assess cumulative genetic and environmental (childhood adversity, cannabis use, hearing impairment, and winter births) risk for schizophrenia, respectively. The primary outcome was past-month distressing PEs at the 3-year follow-up. Secondary outcomes included distressing PEs across four yearly assessments: lifetime (≥ 1 wave), repeated (≥ 2 or ≥ 3 waves), and persisting (≥ 4 waves). Sex-stratified multilevel logistic regression models were used to test the independent and joint associations of binary modes (> 75th percentile) of PRS-SCZ (PRS-SCZ<inf>75</inf>) and ES-SCZ (ES-SCZ<inf>75</inf>) on the outcomes. As sensitivity analysis, the sex-stratified analyses were repeated on a randomly selected unrelated sample, and the coefficients of males and females were compared. Results: PRS-SCZ<inf>75</inf> was not associated with past-month distressing PEs in either sex but significantly associated with lifetime and repeated (≥ 2 waves) distressing PEs only in females. In both sexes, ES-SCZ<inf>75</inf> was significantly associated with all PE outcomes but did not additively interact with PRS-SCZ<inf>75</inf> in predicting them. Sensitivity analysis confirmed the findings and revealed a significant sex difference in the association between PRS-SCZ<inf>75</inf> and lifetime distressing PEs. Conclusion: The influence of genomic risk for schizophrenia on distressing PEs might be sex-dependent, whereas that of the exposomic risk was universal in early adolescence. Further studies in larger samples are needed.