Five-year changes in 24-hour sleep-wake activity and dementia risk in oldest old women.

Abstract

Sleep disruptions are associated with cognitive aging in older adults. However, it is unclear whether longitudinal changes in 24-hour multidimensional sleep-wake activity are linked to cognitive impairment in the oldest old. We studied 733 cognitively unimpaired women (mean age=82.5±2.9 years) who completed two actigraphy assessments over five years. We performed hierarchical clustering on principal components in nine sleep, napping, and circadian rest-activity rhythm parameters to identify multidimensional sleep-wake change profiles and multinomial logistic regression to evaluate the associations between sleep-wake changes and risk of cognitive impairment at follow-up. We identified three sleep-wake change profiles: Stable Sleep (43.8%), Declining Nighttime Sleep (34.9%), and Increasing Sleepiness (21.3%). After adjustment for demographics and comorbidities, women with Increasing Sleepiness had approximately doubled (odds ratio=2.21, p=0.018) risk of dementia compared to those with Stable Sleep. Increasing sleepiness may be an independent marker or risk factor for dementia in oldest old women.