Vergleich der automatischen Auswertung von Polysomnographien von Patienten mit REM Schlaf-Verhaltensstörung und Polysomnographien von Patienten mit differenzial-diagnostisch relevanten Diagnosen (Schlafwandeln, Restless Legs Syndrom, obstruktive Schlafapnoe) zur Beurteilung der Sensitivität und Spezifität der Methode

Human states of being basically consists of three states: Awake state, NREM sleep and REM (rapid eye movement) sleep. These three states are regulated by a complex system of generators in the central nervous system. REM Sleep Behavior Disorder (RBD) is a parasomnia described among 80 other sleep disorders in the International Classification of Sleep Disorders. (ICSD-3, 2014) It is characterized by the occurrence of abnormal nocturnal behaviors, as well as by an in-crease of skeletal muscle activity during REM sleep. Currently the disorder is defined by acting out dreams. The content of the patient’s dream is often aggressive, with at-tacks, threats or persecution of the individual by animals or unknown persons. Also vocalizations such as talking, shouting, laughing, etc. are possible. Serious injury and sleep interruption may result. Current epidemiological data on the prevalence and incidence of RBD are inadequate. The aetiology distinguishes between idiopathic RBD and secondary forms in context of neurodegenerative diseases, drug associated and alcohol consumption. In particular, the association with neurodegenerative diseases i.e. mainly alpha-synucleinopathies such as PD, DLB or MSA, have led to a strong increase of research interest in recent years. The fact that RBD may be a prodromal stage of alpha-synucleinopathies leads to the need developing neuroprotective or disease modifying therapies. Results from these studies are not yet available. Therefore it is important to find a reliable and efficient way to diagnose patients. RBD is characterized by an abnormal intermittent increased muscle tone, measurable in the electromyographic (EMG) recordings of polysomnography (PSG). The diagnosis of RBD requires loss of muscle atonia in REM sleep (RWA) in combination with documented motor activity in the video-audio-PSG. The quantification of EMG tonus is complex and not uniform yet. In addition, the hand scoring is extremely time-consuming. Therefore, there is great interest in the development of a valid computerized EMG analysis method. In recent years, various quantification methods of EMG tonus have been proposed. (Consens et al., 2005), (Bliwise et al., 2006), (Frauscher et al., 2012) Unfortunately there are no current uniform standards of cut-off values defined in order to differentiate between normal and abnormal EMG-tones. The ICSD-2 only mentioned an “excessive amount of phasic or tonic EMG activity.“ (American Academy of Sleep Medicine, 2005) In this study the automatic computerized EMG analysis was further investigated. It was developed by Mayer and Kesper and uses indices to describe muscle activity during sleep. (Mayer et al., 2008) Polysomnographies of RBD (n=20), sleepwalking/night terrors (n=10), restless legs syndrome (n=10) and obstructive sleep apnea syndrome patients (n=10) were studied to validate the computerized analysis and to determine the sensitivity and specificity of this method (SMI & LMI). For this purpose SMI and LMI values were calculated by using the software EDFTrace (sleep laboratory Marburg, Germany), based on the EMG recordings of the mentalis muscle, flexor digitorum superficialis (FDS) and tibialis anterior muscle (TA). The mentalis muscle has a sensitivity of 72.5% and a specificity of 86.7%. Cut-off values for the mentalis muscle for SMI were 90.1 /h REM and LMI 43.1 /h REM. (AUC 0.819, p 0,002) TA showed a low sensitivity of 65% and a specificity of 63.3%. For the sum of the two TA EMG channels, cut-off values were 323.0 /h REM for SMI and 133.5 /h REM for LMI. (AUC 0.628, p 0.051) The FDS channels had a sensitivity of 65% and a specificity of 95%. The corresponding cut-off values for the sum of right and left FDS EMG channel were for the SMI 124.3 /h REM and LMI 50.1 /hREM (AUC 0.771, p 0.017). To provide additional validation of the algorithm, consistency of the automatic analysis and of the manually scored video-PSG was performed. For the RBD patients a percent agreement of 77% was determined. This seems to be a good value, as e.g. movements of the torso can theoretically occur without activation of the five analysed EMG channels. In terms of sensitivity and specificity of the method, promising results were especially found for EMG recordings of the mentalis and FDS muscle. In summary the computerized EMG analysis is a feasible, non-invasive method to as-sess RBD diagnosis with appropriate reliability. With this method (SMI & LMI) it is possible to distinguish between RBD, sleepwalking, OSA and RLS. The method can be performed in routine vPSG (including recording of flexor digitorum superficialis) and scored automatically fast and properly. However, to establish this method, cut-off values should be reproduced in larger study populations.