INTERACTIVE ASSOCIATION OF LIVING ALONE AND AMYLOID-Β BURDEN WITH LONGITUDINAL COGNITION IN COGNITIVELY UNIMPAIRED OLDER ADULTS

Living alone has previously been shown to be a risk factor for cognitive decline and incident dementia in older adults, however this risk has not been evaluated in the context of early Alzheimer's disease. We investigated whether living alone and brain amyloid-β, the Alzheimer's disease pathologic p... Full description

Living alone has previously been shown to be a risk factor for cognitive decline and incident dementia in older adults, however this risk has not been evaluated in the context of early Alzheimer's disease. We investigated whether living alone and brain amyloid-β, the Alzheimer's disease pathologic protein, were additively or interactively associated with longitudinal cognition in a sample of cognitively unimpaired older adults. 272 men and women from the Harvard Aging Brain Study, ages 62-90, 59% female, with a Clinical Dementia Rating score of 0, were included. At baseline, 173 participants reported living with a spouse or partner, family member(s), or with a friend, and 99 reported living alone. Neocortical amyloid-β was measured at baseline using Pittsburgh Compound B-Positron Emission Tomography (PiB). Cognitive performance was assessed at baseline and annually using the Preclinical Alzheimer Cognitive Composite (PACC) (number of assessments: mean 3.93?years, range 1-9). The association of living situation (reference group: living with others) and cognitive performance over time was estimated using linear mixed-effects models with random intercepts and slopes, adjusting for age, sex, socioeconomic status, depressive symptoms, PiB, and their effects by time. A second model examined the multiplicative interaction of marital status and continuous PiB to estimate cognition over time, adjusting for the same covariates. The main model was repeated using dichotomous, high and low PiB groups (PiB+, PiB-) based on a PiB Distribution Volume Ratio threshold of 1.32. Compared to participants who lived with others, participants who lived alone demonstrated worsening PACC performance in the adjusted model (β =-0.06, p=0.005; Figure 1). Further, living alone and higher PiB were interactively associated with longitudinal PACC beyond their independent effects such that those who lived alone exhibited steeper PiB-related PACC decline (β=-0.18, p=0.0007). Findings were consistent in the dichotomous PiB models (Living Alone/ PiB+ × Time: β=-0.11, p=0.01; reference Living with others/PiB-; Figure 2). In a sample of cognitively unimpaired older adults, living alone was associated with accelerated amyloid-β-related cognitive decline prior to overt cognitive impairment. Future research on evidence-based interventions targeting specific, high risk older adults who live alone may help attenuate cognitive decline and mitigate the rising prevalence of mild cognitive impairment an