Sex-specific early cognitive changes are linked to global and pathway-specific genetic risk for Alzheimer's disease in at-risk individuals

Biol Sex Differ. 2026 Feb 17. doi: 10.1186/s13293-025-00800-w. Online ahead of print.

ABSTRACT

Alzheimer's disease (AD) is a progressive neurodegenerative condition in which genetic predisposition plays a key role, yet the sex-specific mechanisms linking genetic risk to early cognitive changes remain unclear. This study examined the impact of polygenic risk scores (PRS) on early cognitive changes in 318 cognitively unimpaired participants from the ALFA+ cohort, a nested longitudinal cohort from the ALFA study (see details in Study Participants Section, Methods). Participants were followed for three years, with assessments across five cognitive domains and a preclinical composite (PACC). Global AD PRS, including and excluding the apolipoprotein E (APOE) gene, alongside five biologically informed pathway-specific PRS (amyloid, immune, external stimuli signaling, cholesterol efflux, lipoprotein metabolism) were computed. Generalized linear models including interaction by sex and stratified by sex and amyloid status (CSF Aβ42/40 < 0.071) assessed associations between PRS and cognitive change. In women, APOE-independent AD genetic risk predicted worse executive function, particularly via cholesterol efflux and external stimuli signaling pathways. Among Aβ + women, PRS also predicted lower memory performance, partially modulated by reproductive span. In Aβ - women, worse executive functioning performance was linked to amyloid, immune, and signaling pathways. In contrast, men showed associations between AD PRS and worse visual (Aβ-) and attentional (Aβ+) performance, independent of pathway-specific mechanisms. These findings reveal distinct, domain-specific cognitive vulnerabilities to AD genetic risk by sex and amyloid status, highlighting APOE-independent and mechanistic contributions to early and subtle cognitive changes. Results support the need for sex-aware, biologically informed genetic models in preclinical AD for risk stratification and early intervention.

PMID:41703636 | DOI:10.1186/s13293-025-00800-w