BACKGROUND AND OBJECTIVES: Aging is associated with changes in circadian rhythms. Rest-activity rhythms (RARs) measured using accelerometers are markers of circadian rhythms. Altered circadian rhythms may be risk factors of neurocognitive outcomes; however, results are mixed, and previous studies were often conducted in homogeneous populations. We aimed to assess the association between circadian RARs and incident dementia in a racially diverse sample of older adults.
METHODS: This was a retrospective examination of data from the Atherosclerosis Risk in Communities (ARIC) study, a community-based cohort study conducted at 4 US centers. ARIC participants who wore the Zio XT® long-term continuous monitoring patch in 2016-17 for ≥3 days and were free of prevalent dementia were included. RARs were derived from investigational accelerometer data from the patch. Nonparametric RARs included relative amplitude (rhythm strength), intradaily variability (rhythm fragmentation), and interdaily stability (rhythm consistency). Cosinor RARs included measures of rhythm strength (amplitude, mesor) and circadian timing (acrophase). Dementia cases were adjudicated through 2020 using in-person and phone assessments, hospitalization codes, and death certifications. Cox proportional hazards models were used.
RESULTS: Of the 2,183 participants (mean ± SD age 79 ± 4.5 years, 58% female, 24% Black), 176 (8%) developed dementia. The median follow-up time was 3 years, and the mean Zio XT Patch wear time was 12 days. After multivariable adjustment, each 1-SD decrement in relative amplitude and 1-SD increment in intradaily variability were associated with 54% (95% CI 32%-78%) and 19% (95% CI 2%-38%) greater risk of dementia, respectively. Amplitude and mesor were associated with elevated dementia risk after multivariable adjustment (hazard ratios per 1-SD decrement: 1.43 [95% CI 1.15-1.78] and 1.33 [95% CI 1.08-1.63], respectively). A later acrophase was associated with 1.45 times (95% CI 1.01-2.07) greater risk of dementia compared with a normal acrophase.
DISCUSSION: Using accelerometer data from a commonly used ambulatory ECG monitor, weaker and more fragmented circadian RARs and later peak activity time were prospectively associated with elevated dementia risk in older Black and White adults. Limitations of our study include the lack of dementia subtype data and objective sleep disorder measurements. Further research to determine whether circadian rhythm interventions can reduce dementia risk is warranted.
