Worse objective and subjective sleep‐wake metrics are associated with locus coeruleus hypopigmentation: a retrospective post‐mortem study

Abstract

AbstractBackgroundThe brainstem locus coeruleus (LC) was recently put forward as a key nucleus to connect early sleep‐wake disruption and Alzheimer’s disease (AD) pathophysiological processes. Here, we sought to investigate the relationships between post‐mortem LC hypopigmentation, reflecting LC neurodegeneration, and retrospective subjective/objective sleep‐wake regulation metrics in two longitudinal datasets.MethodLC pigmentation ratings were extracted from autopsy data across 517 Rush Memory and Aging Project (MAP) and 637 National Alzheimer’s Coordinating Center (NACC) cases. Actigraphy‐derived intradaily variability, an objective measure of rest‐activity rhythm fragmentation, was obtained in the MAP dataset, and subjective evaluations of sleep quality were retrieved from the NACC cohort. For both studies, sleep‐wake measures were extracted at the furthest and closest available time points, on average 6.8 y. and 2.5 y. before death, respectively. Logistic regressions adjusted for age at death, sex, race, and post‐mortem interval were used to assess whether sleep‐wake variables could predict post‐mortem LC pigmentation ratings. In the next step, additional models including AD‐related variables of APOE genotype and cortical neuropathology (amyloid‐beta and tau burden) were further performed to test for independence in the relationships between sleep‐wake measures and LC pigmentation.ResultsIn the MAP cohort, higher rest‐activity rhythm fragmentation at both time points was significantly associated with greater odds of LC hypopigmentation at autopsy (Table 1). In models considering additional AD‐related variables, these relationships remained significant independently of the effects of APOE and cortical AD neuropathology (Table 1). In the NACC dataset, worse subjective sleep quality at the closest time point, but not at the furthest, was significantly linked to greater odds of LC hypopigmentation at autopsy (Table 2). Likewise, additional models indicated that this relationship is independent of APOE and cortical AD neuropathology (Table 2).ConclusionThese findings highlight the utility of evaluating sleep and wakefulness, especially with objective measurements, to predict LC neurodegenerative processes in older populations, as far as 6.8 years before death and beyond the effects of APOE genotype and cortical AD neuropathology. Our results thus have implications for improved detection of at‐risk individuals.