Investigation of frailty as a moderator of the relationship between neuropathology and dementia in Alzheimer's disease: a cross-sectional analysis of data from the Rush Memory and Aging Project

Background: Some people with substantial Alzheimer's disease pathology at autopsy had shown few characteristic clinical symptoms or signs of the disease, whereas others with little Alzheimer's disease pathology have been diagnosed with Alzheimer's dementia. We aimed to examine whether frailty, which is associated with both age and dementia, moderates the relationship between Alzheimer's disease pathology and Alzheimer's dementia.

Methods: We did a cross-sectional analysis of data from participants of the Rush Memory and Aging Project, a clinical-pathological cohort study of older adults (older than 59 years) without known dementia at baseline, living in Illinois, USA. Participants in the cohort study underwent annual neuropsychological and clinical evaluations. In the present cross-sectional analysis, we included those participants who did not have any form of dementia or who had Alzheimer's dementia at the time of their last clinical assessment and who had died and for whom complete autopsy data were available. Alzheimer's disease pathology was quantified by a summary measure of neurofibrillary tangles and neuritic and diffuse plaques. Clinical diagnosis of Alzheimer's dementia was based on clinician consensus. Frailty was operationalised retrospectively using health variable information obtained at each clincial evaluation using the deficit accumulation approach (41-item frailty index). Logistic regression and moderation modelling were used to assess relationships between Alzheimer's disease pathology, frailty, and Alzheimer's dementia. All analyses were adjusted for age, sex, and education.

Findings: Up to data cutoff (Jan 20, 2017), we included 456 participants (mean age at death 89·7 years [SD 6·1]; 316 [69%] women). 242 (53%) had a diagnosis of possible or probable Alzheimer's dementia at their last clinical assessment. Frailty (odds ratio 1·76, 95% CI 1·54-2·02; p<0·0001) and Alzheimer's disease pathology (4·81, 3·31-7·01; p<0·0001) were independently associated with Alzheimer's dementia, after adjusting for age, sex, and education. When frailty was added to the model for the relationship between Alzheimer's disease pathology and Alzheimer's dementia, model fit improved (p<0·0001). There was a significant interaction between frailty and Alzheimer's disease pathology (odds ratio 0·73, 95% CI 0·57-0·94; p_interaction=0·015). People with an increased frailty score had a weakened direct link between Alzheimer's disease pathology and Alzheimer's dementia; that is, people with a low amount of frailty were better able to tolerate Alzheimer's disease pathology, whereas those with higher amounts of frailty were more likely both to have more Alzheimer's disease pathology and for it to be expressed as dementia.

Interpretation: The degree of frailty among people of the same age modifies the association between Alzheimer's disease pathology and Alzheimer's dementia. That frailty is related to both odds of Alzheimer's dementia and disease expression has implications for clinical management, since individuals with even a low level of Alzheimer's disease pathology might be at risk for dementia if they have high amounts of frailty. Further research should assess how frailty and cognition change over time to better elucidate this complex relationship.

Funding: None.