Sex differences in polygenic risk for sex-biased brain disorders: Associations with diagnostic status and brain age gap

Biological sex influences the prevalence and progression of several brain disorders, yet it remains unclear whether genetic risk interacts with sex in relation to case-control status or brain health. Using data from up to 220 836 females and 187 651 males in the UK Biobank (aged 39–81), we assessed how sex and polygenic risk scores (PRSs) for major depressive disorder (PRS_MDD), Alzheimer’s disease (PRS_AD), schizophrenia (PRS_SCZ), and Parkinson’s disease (PRS_PD) relate to case-control status and MRI-derived brain age gap (BAG) as a proxy of brain health. We tested for sex differences in PRSs and performed regression analyses examining associations between sex, PRSs, and both outcomes. We then compared models with sex-pooled or sex-specific PRSs to explore whether sex-specific PRSs improve predictive accuracy. PRS_SCZ was higher in females, while no other PRSs differed by sex. A significant PRS_AD-by-sex interaction revealed weaker associations with AD case-control status in males than females. Sex-pooled models outperformed sex-specific models for case-control status, apart from PRS_AD, where the female-only PRS_AD model outperformed the sex-pooled model. No sex-by-PRS interactions were significantly associated with BAG. However, male sex, higher PRS_MDD, higher PRS_AD, and lower PRS_PD were associated with higher BAG, indicative of an older brain age. BAG model performance did not differ between sex-pooled and sex-specific PRS models. In summary, while sex-specific PRSs currently offer limited value over sex-pooled scores, the superior performance of the female-specific PRS_AD model suggests that the value of sex-specific PRSs will likely grow with increased statistical power.