Different circadian rest-active rhythms in Kleine-Levin syndrome: a prospective and case-control study

Study Objectives: Kleine-Levin syndrome (KLS) is a rare recurrent hypersomnia. Our study aimed at monitoring the movements of patients with KLS using actigraphy and evaluating their circadian rhythm. Methods: Twenty young patients with KLS and 14 age-matched controls were recruited. Each individual wore an actigraphy for more than 6 months to monitor at least two attacks. Controls kept wearing the device for at least 7 days. The activity counts were averaged in hourly basis and the day-to-night amplitude was quantified by the differences of the averaged activity counts during daytime and nighttime. The hourly activities of different days were aligned and averaged to construct the circadian profile. Parametric and nonparametric estimation of circadian rhythm was calculated. We applied detrended fluctuation analysis to evaluate the temporal correlations beneath the activity fluctuations at multiple time scales. Results: Circadian rhythm in asymptomatic period showed no significant difference compared to the controls. During hypersomnia attack, the amplitude of the circadian rest-active rhythms drastically decreased and decreased interdaily stability (IS) was found, as well as significant decreased M10 and short-time fractal correlation (alpha 1). Drastically decreased mean and standard deviation of activity were noted, compared to the pre-attack phase and recovery phase. alpha 1 and M10 increased during the late attack phase, and overcompensated IS was noted in the recovery phase. Conclusions: This study confirmed that circadian rest-active rhythms was affected when KLS hypersomnia attack. Several parameters including M10, IS, and alpha 1 may be physiological markers of KLS, which can help to predict the end of hypersomnia episodes. Statement of Significance Kleine-Levin syndrome is a rare recurrent hypersomnia disorder. Our study monitored the movements of these patients using actigraphy. We found their circadian active/rest rhythms in asymptomatic periods showed no significant difference compared to the controls, but during hypersomnia attacks, the amplitude of the circadian rhythm drastically decreased. Further analysis identified several physiological markers which could help to predict the end of hypersomnia episodes.