Sleep disturbances in anorexia nervosa

IntroductionSleep is crucial for physical and psychological health, and disturbances are closely linked to psychiatric disorders, making their management essential for improving treatment outcomes and preventing relapse. Although sleep disturbances are implicated in psychopathology of eating disorders, its role in anorexia nervosa (AN) remains unclear. This review aimed to characterise sleep in AN.MethodsA systematic search was conducted in four scientific databases, including papers from inception to 25 May 2024. A machine learning algorithm (ASReview) was utilised to screen titles and abstracts for eligibility. Sleep quantity, architecture, and quality were investigated. Meta-analyses were conducted to investigate the difference between patients with AN and healthy controls (HC) in total sleep time (TST), wake after sleep onset (WASO), sleep onset latency (SOL), sleep efficiency, Sleep Stage 1, 2, slow wave sleep, rapid eye movement (REM) sleep and REM latency. Certainty of evidence was assessed using the GRADE approach.ResultsOut of 67 potentially eligible papers, 31 were included in this review, with 15 in the meta-analyses. Statistically significant average mean differences were found for TST (-32.1 min [95% CI: -50.9, -13.4]), WASO (19.0 min [95% CI: -2.4, 40.3]), and sleep efficiency (-4.4% [95% CI: -7.9, -0.9]). Additionally, Sleep stage 1 was significantly increased (2.4% [95%-CI: 0.05, 4.7]), while REM sleep was reduced (-2.1% [95%-CI: -4.2, -0.02]). Subgroup analysis showed that TST and WASO did not improve significantly after weight restoration.ConclusionSleep in patients with AN is impaired, with lower TST and sleep efficiency, higher WASO, more time in stage 1 sleep, and less in REM. Weight restoration alone may not improve sleep. While more research is needed, substantial accompanying sleep disturbances in AN justifies addressing these in current treatment practice, also because of the chronic character of AN and importance of sleep for long term (mental) health.