Domains of Frailty as Early Risk Factors for Alzheimer Dementia - Genetic and Causal Evidence

BackgroundAlzheimers dementia (AD) is a leading cause of disability and dependency in older adults. Frailty is a phenotypic risk factor for AD, but its causal role remains unclear - partly due to reliance on composite scores that may obscure distinct mechanisms. MethodsWe applied a Mendelian randomisation (MR) framework using two large GWAS of clinically diagnosed AD. Frailty exposures were derived from a multivariate genomic SEM model comprising one general factor and six domain-specific domains. Bidirectional univariable MR (UVMR) estimated total effects on AD, while multivariable MR (MVMR) adjusted for educational attainment, household income, and longevity. ResultsThe general frailty factor showed no evidence of a causal effect on AD. The unhealthy lifestyle frailty domain increased AD risk in UVMR (OR = 6.60, 95% CI: 2.36-18.50, q = 0.021 and OR = 3.12, 95% CI: 1.57-6.19, q = 0.001), but these effects attenuated substantially in MVMR, with betas reduced by [~]66-94% and ORs falling to [~]1.1-1.5 (non-significant), consistent with overlap with socioeconomic pathways. The disability frailty domain was nominally protective in UVMR, but positively associated with AD in MVMR (OR = 1.69, 95% CI: 1.21-2.36, q = 0.007). The poorer cognition frailty domain increased AD risk in MVMR when adjusted for education and income (OR = 1.55, 95% CI: 1.12-2.15, q = 0.025), though this effect attenuated by [~]37% with further adjustment for longevity. Sensitivity analyses indicated that this cognition-AD signal was partly driven by SNPs in SPI1, a well-established AD risk locus, consistent with pleiotropic overlap rather than an independent domain. No reverse effects were observed. ConclusionsFrailty comprises genetically distinct domains with differential causal relevance for AD. These findings highlight disability-related frailty as a potential causal contributor to AD, whereas cognitive frailty is likely predominantly driven by shared aetiology, -underscoring the need to move beyond composite frailty indices when evaluating dementia risk.