Poor Sleep Quality May Predict Preclinical Alzheimer’s Disease
Abstract & Commentary
By Alan Z. Segal, MD
Associate Professor of Clinical Neurology, Weill Cornell Medical College
Dr. Segal reports no financial relationships relevant to this field of study. This article originally appeared in the July 2013 issue of Neurology Alert.
Synopsis: Poor sleep quality may predict future
Alzheimer’s disease and may play a pathogenic role
in its etiology.
Source: Ju YE, et al. Sleep quality and preclinical Alzheimer disease. JAMA Neurol 2013;70:587-593.
Insomnia is a common complaint among adults and it is well recognized that mental performance can be impaired by poor sleep. Both difficulty falling asleep and trouble staying asleep have been associated with poor cognitive function in cross-sectional elderly cohorts. Sleep impairments leading to excessive daytime sleepiness also have been shown to contribute to cognitive decline. Getting more sleep, however, clearly is not the solution. The quantity of sleep has been shown to have variable effects, with some studies showing a U-shaped relationship between sleep duration and cognition. The Nurse’s Health Study, for instance, suggested that subjects getting 7 hours of sleep had optimal functioning compared to those getting 5 or 9 hours of sleep. Sleep may be further compromised by obstructive sleep apnea, which not only causes oxidative stress, but also disrupts sleep architecture and has been shown to increase the risk of dementia. Snoring itself, in the absence of documented sleep apnea, has shown similar results.
Most prior studies have used subject self-report surveys or sleep diaries to assess variables such as sleep efficiency — defined as the actual time asleep divided by the total time in bed. Formal polysomnography eliminates the imprecision introduced by such patient estimates, but is an artificial environment which itself limits sleep efficiency. Sleep may be measured at home using movement detection (actigraphy) as in the current study or with increasingly available technologies, such as smartphone apps that capitalize on embedded accelerometers and movement detectors to create modestly accurate sleep maps.
In the current study, cognitively normal subjects wore actigraphs for 2 weeks to measure total sleep time and efficiency. Sleep was then correlated with levels of cerebrospinal fluid (CSF) ß-amyloid (Aß) 42. It has been shown that low soluble Aß42 levels in the CSF are suggestive of insoluble amyloid deposition in brain senile plaques. This decrease in CSF Aß42 may predate clinical Alzheimer’s disease by 15 years. Low Aß42 levels have also been correlated with positive findings on amyloid PET imaging.
Among subjects with Aß42 levels < 500 pg/mL, sleep efficiency was 80.4% compared with those having Aß42 levels > 500 pg/mL, who had a sleep efficiency of 84.7% (P = 0.04). Frequent napping (3 or more days per week) was also associated with amyloid deposition (31.2% among low Aß42 subjects compared with 14.7% in the high Aß42 group, P = 0.03). Total sleep time did not differ between the two groups; however, low Aß42 subjects spent a longer total time in bed in order to obtain adequate sleep.
Because the relationship between sleep efficiency and Aß42 could be bidirectional, the investigators also used Aß42 as an outcome variable and dichotomized sleep into poor (< 75% efficiency) vs good (> 89% efficiency) groups. Subjects with poor sleep had a 5.6-fold increased risk of having low Aß42 levels compared to subjects with good sleep quality.
As the authors nicely demonstrate in a cyclical flow diagram, these data clearly suggest a two-way street. Aß accumulation may be disruptive of sleep and conversely impaired sleep may promote Aß deposition. There are multiple mechanisms by which amyloid plaques may disrupt sleep circuits in the basal forebrain. This pathogenesis is supported by transgenic mouse models prone to amyloid plaques. These animals have markedly impaired sleep-wake cycles. Interestingly, when such mice are immunized with Aß, preventing plaque formation, sleep architecture is restored. Because Aß is secreted during active neuronal activity, the investigators further hypothesize that prolonged wakefulness results in a higher concentration of amyloid and therefore an increased plaque burden.
Multiple other behavioral factors that impair sleep in the elderly may further amplify the relationship between poor sleep and Aß. Depression, schedule changes related to retirement (due to cognitive decline or merely due to age), lack of exercise, and institutionalization (impairing regular light exposure) all may potentially contribute to Aß deposition. As the authors note, however, Aß secretion also may have a strong circadian influence. This would vary independent of the subject’s sleep-wake state.
These data confirm that poor sleep efficiency may play a pathogenic role in Alzheimer’s disease. This study has important strengths in its use of actigraphy to quantify sleep and also its use of Aß42 as a biomarker of Alzheimer’s disease risk in place of psychometric data. Of note, these subjects all underwent exhaustive neuropsychiatric testing, confirming them to all be normal. Aß42 in a cognitively normal person, however, is not equivalent to a future diagnosis of Alzheimer’s disease. Low CSF levels of Aß42 may not consistently correlate with pathologically confirmed amyloid plaques, and these plaques may not be uniformly predictive of Alzheimer’s disease. As the authors observe, the use of amyloid imaging and the inclusion of individuals with elevated CSF tau, planned in future studies, could further strengthen the connection between sleep disruption and preclinical Alzheimer’s disease.
Despite these limitations, it is clear that addressing reversible sleep disorders, such as obstructive sleep apnea, may be an important step toward addressing Alzheimer’s disease in its earliest stages. Sleep disorders, including sundowning, excess daytime napping, and insomnia occur in up to 40% of patients with Alzheimer’s disease. Although these phenomena may be the consequence of Alzheimer’s disease, they might also mark a population of people for whom Alzheimer’s disease is in the future and might be prevented.