Sleep Disturbances, Cognitive Status, and Biomarkers of Dementia.

BACKGROUND While sleep disturbances appear to be risk factors in Alzheimer's disease (AD) progression, information such as the prevalence across dementia severity and the influence on the trajectory of cognitive decline is unclear. OBJECTIVE We evaluate the hypotheses that the prevalence of insomnia differs by cognitive impairment, that sleep disturbances track with AD biomarkers, and that longitudinal changes in sleep disorders affect cognition. METHODS We used the National Alzheimer's Coordinating Center Database to determine the prevalence of clinician-identified insomnia and nighttime behaviors in normal, mild cognitive impairment (MCI), and demented individuals. We evaluated mean Montreal Cognitive Assessment (MoCA) scores, hippocampal volumes (HV), and CSF phosphorylated tau:amyloid-β ratios at first visit using analysis of variance with age as a covariate. In longitudinal evaluations, we assessed changes in MoCA scores and HV in insomnia and nighttime behaviors between the first and last visits. RESULTS Prevalence of insomnia was 14%, 16%, and 11% for normal, MCI, and dementia groups. Prevalence of nighttime behaviors was 14%, 21%, and 29% respectively. Insomnia patients had higher MoCA scores, larger HV, and lower pTauBeta than individuals without insomnia, indicating less neurodegeneration. In contrast, nighttime behaviors were associated with worse cognition, smaller HV, and higher pTauBeta. Similar findings were seen between longitudinal associations of sleep disorders and cognition and HV. CONCLUSION Our findings suggest that insomnia is unreliably recognized in patients with cognitive impairment. Nighttime behaviors may better indicate the presence of sleep disturbances and have diagnostic specificity in AD over insomnia.