Amyloid on autopsy in the oldest‐old may be associated with sleep duration thirty years earlier: The 90+ Study

Abstract

AbstractBackgroundAmyloid production and clearance from the brain are thought to be related to sleep duration. The association between brain amyloid accumulation and self‐reported sleep duration has been mostly established in cross‐sectional studies using PET. Our aim is to examine the longitudinal association between amyloid in the oldest‐old (90+ years), measured on autopsy and PET, with sleep duration self‐reported ∼30 years earlier.MethodWe analyzed data on 289 individuals with one‐time report of their usual nocturnal sleep duration as part of Leisure World Cohort Study (LWCS) in the early 1980s, and brain autopsy or 18F‐florbetapir (amyloid) PET between 2004 and 2021 with The 90+ Study. Sleep duration was analyzed as continuous and categorical variables (<7(short), 7‐8(optimal, reference) and >8 hours(long)). National Institute on Aging Alzheimer Association (NIA‐AA) amyloid (A) and neuritic (C) plaque scores were dichotomized as negative (0‐1, reference) and positive (2‐3). Standardized uptake value ratio (SUVr) for the posterior cingulate/precuneuswas analyzed as continuous and binary (SUVr <0.76 vs. ≥0.76) variables. We analyzed sleep duration, as predictor, in relation to amyloid on autopsy (A and C scores) and PET (SUVr), as outcomes, using multiple logistic and multiple linear regressions, controlling for age, sex, education, positive items from the Zung depression scale, and time between sleep report and PET or autopsy.ResultSleep was reported on average at 68 years, autopsy was done ∼29 years and PET was done ∼32 years later (Table 1). Long sleep, compared to optimal sleep, was associated with 67% decrease in the odds of being NIA‐AA A positive (Table 2). Although PET amyloid results were not significant, they suggest that every hour increase in sleep duration may be associated with 0.01 unit decrese in SUVr and with 11% lower odds of being PET amyloid positive (Table 3). PET amyloid by sleep group analysis was underpowered and not done.ConclusionIn this group of oldest‐old individuals, long, compared to optimal, sleep duration self‐reported 29 years earlier is associated with lower odds of amyloid deposition on autopsy. One of the plausible mechanisms is that longer sleep duration may support greater amyloid clearance and decreased production.