ABSTRACT Alzheimer's disease (AD) is an important, non‐curable disease today. Several possible risk factors have been discussed, among others sleep. Evolutionary hypotheses were proposed to explain sleep variation and AD risk, such as a potential advantage of short or interrupted sleep in ancient insecure environments, the evolution of increased plasticity of the human brain, or antagonistic pleiotropy with increased AD risk. The aim of this systematic review was to investigate in which way sleep is associated with AD risk, in the light of evolutionary hypotheses. Following PRISMA guidelines, the databases PubMed and Embase were searched for longitudinal observational studies on the association between sleep and incident AD in cognitively healthy people at baseline. Potential confounders were assessed (e.g., age, sex, country, etc.). Search results were deduplicated and assessed for inclusion or exclusion by two independent reviewers. We summarized the results in tables and performed meta‐analyses in risk factor subgroups where appropriate. Out of 5800 studies, 39 were suitable for this review and 35 were meta‐analyzed. Long sleep duration showed a positive association with AD (HR = 1.35, 95% CI = 1.12–1.63) that was not significant after accounting for heterogeneity using prediction intervals (95% CI = 0.74–2.49). Short sleep duration showed a weak, non‐significant association with AD risk (HR = 1.07, 95% CI = 0.98–1.18). Several measures of sleep quality (hypnotics use, daytime sleepiness, bad overall sleep quality, and early bedtime) showed an increased risk for AD, while others (e.g., late bedtime) showed a protective effect. This systematic review showed possible associations between sleep characteristics and the risk for AD. However, several results showed heterogeneity that we could not explain with the information given in the publications. More work is therefore needed to assess the risk factors for AD connected to sleep.
Spirig et al. (Sun,) studied this question.