BACKGROUND: Frailty, which is considered a potential modifiable risk factor for dementia, continues to generate debate when it comes to Alzheimer's disease (AD). Furthermore, the underlying pathological mechanisms linking frailty to AD remain uncertain. OBJECTIVE: We aimed to investigate the relationship between frailty and the risk of AD while elucidating the connections between frailty, AD biomarkers, and cognitive function. METHODS: Total of 829 non-frail (261 robust, 568 pre-frail) and 94 frail individuals from the Alzheimer's Disease Neuroimaging Initiative database were recruited. Kaplan-Meier analysis and Cox regression assessed AD risk across diverse frail statuses in 923 non-demented individuals. Multiple linear regression, mixed effects models and causal mediation analyses bootstrapped 10,000 iterations were conducted to examined underlying associations. RESULTS: The frail group had a 67.7% increased risk of AD than non-frail group (HR = 1.677
95%CI, 1.179-2.385
p = 0.004), a 61.8% increased risk of AD than pre-frail group (HR = 1.618
95%CI, 1.131-2.316
p = 0.009) and a far higher risk of AD than robust group (HR = 2.011
95%CI, 1.263-3.202
p = 0.003). Frailty was associated with cognitive decline (global cognition, memory and executive function), whole brain and hippocampus atrophy, and ventricle dilation. Higher frail degree predicted faster cognitive decline, brain atrophy and ventricle dilation. Frailty's association with cognition was partially mediated by volume of whole brain (29.54%-30.17% of total effect), hippocampus (18.21%-24.55% of total effect), and ventricle (baseline, 7.62%-10.87% of total effect
change rate, 13.30%-24.33% of total effect). CONCLUSIONS: Frailty as a potential risk factor for AD, further mechanisms investigation is warranted
mitigating frailty could potentially contribute to AD prevention.