Cross‐sectional associations between sleep quality reports and core Alzheimer’s disease biomarkers in cognitively unimpaired adults from the European Prevention of Alzheimer’s Dementia Longitudinal Cohort Study (EPAD LCS)

Abstract

AbstractBackgroundSleep disturbances are prevalent in Alzheimer’s disease (AD), with sleep quality having been reported to be already impaired at the preclinical stage of the disease. However, most results originate from studies with rather limited sample sizes. This research aimed to evaluate the association of subjective sleep measures with cerebrospinal fluid (CSF) AD biomarkers in cognitively unimpaired adults from the European Prevention of Alzheimer’s Dementia Longitudinal Cohort Study (EPAD‐LCS).MethodThis cross‐sectional study included 1257 cognitively unimpaired adults [CDR(clinical dementia rating scale)=0] aged over 50 years who underwent CSF sampling and filled out the Pittsburgh sleep quality index (PSQI) questionnaire. The PSQI includes 7 components (scored 0‐3), which summed yield a total score ranging from 0 to 21. We used multivariate linear regressions to analyse associations between CSF biomarkers (i.e. Aβ42, p‐tau and t‐tau) and the following PSQI measures: total score, binarized score (poor sleep categorized as PSQI>5), sleep latency, duration, efficiency and disturbance. For sleep duration and disturbance, the two highest categories were collapsed into one due to an insufficient number of responses. We used separated models, with each biomarker as the dependent variable and each sleep measure as the predictor, and adjusted them by known potential confounders.ResultPoor sleep quality (PSQI total>5) was associated with higher CSF levels of p‐tau (std. β=0.060, p=0.040) and t‐tau (std. β=0.066, p=0.019). Shorter sleep duration (6‐7 hours vs >7 hours) was associated with higher CSF levels of p‐tau (std. β=0.063, p=0.017) and t‐tau (std. β=0.054, p=0.037). A higher degree of sleep disturbance (1‐9 vs 0 and >9 vs 0) was associated with lower CSF levels of Aβ42 (std. β=‐0.106, p=0.016; std. β=‐0.091, p=0.046).ConclusionOur results show that self‐reported sleep quality is associated with AD biomarkers in people without cognitive impairment. Specifically, overall poor sleep quality and shorter sleep duration are related to higher p‐tau and t‐tau levels, whilst more profound sleep disturbances are associated with lower Aβ42 levels. These results potentially support sleep impairment before cognitive symptom onset in AD. Future longitudinal studies using objective sleep quantification will more precisely illuminate the role of sleep in the early stages of the AD continuum.