24‐hour Activity Rhythms, Sleep, and Amyloid Pathology in the population‐based Rotterdam Study

Abstract

AbstractBackgroundPrevious research found associations between activity rhythms, sleep, and amyloid pathology (Winer et.al, JAMA Neurology, 2021; Leng et.al, Lancet Neurology, 2019). Shorter self‐reported sleep duration and more napping were robust predictors in the amyloid prediction model we recently developed (A4 Study, n = 4,119) and validated (Rotterdam Study, n = 500). Studies with more population‐representative samples, longer follow‐up time and objective activity/sleep measures are needed to understand whether disturbances of the activity rhythm and sleep increase the risk of developing amyloid pathology.MethodNondemented participants (n = 319) from the prospective, population‐based Rotterdam Study (mean age 69.2±5.3, 47.0% women) were asked to wear an actigraph watch for 7 days before undergoing amyloid 18F‐florbetaben PET on average 7.8 years [range 4‐16] later (Fig.1). 24‐hour activity rhythms (intradaily stability and variability, L5 start time) and sleep (sleep efficiency, sleep onset latency, total sleep time, and wake after sleep onset) were associated with continuous PET readouts (log‐transformed SUVr values) using linear regression models, adjusted for age, sex, education, APOE4 risk allele count, time between actigraphy and PET assessment, difference in type of actigraphy watches, employment status, depressive symptoms, and vascular risk factors (BMI, diabetes, hypertension, physical activity, smoking).ResultHigher intradaily variability, i.e. a more fragmented activity rhythm, was associated with higher amyloid‐PET SUVr (b(SE) = 0.149(0.055), pFDR = 0.049). This association was significantly stronger in APOE4 carriers than non‐carriers (interaction: pFDR = 0.014, Fig.2). Our results remained after excluding participants who probably were already on the AD pathway (excluded lowest 10% of hippocampal volume and highest 10% atrophy on two consecutive MRIs) at the time of measuring activity rhythms. This may suggest that a fragmented activity rhythm is a risk factor rather than a result of amyloid pathology. There were no significant associations of other 24‐hour activity rhythm and sleep variables with amyloid.ConclusionHigher fragmentation of the 24‐hour activity rhythm was associated with more cerebral amyloid pathology eight years later, even after accounting for potential reverse causality. Since activity disturbances are potentially modifiable, early risk factors of AD, further research should be conducted to examine the preventative effect on the development of AD neuropathology.