Novel sleep scoring method for neurodegenerative diseases: feasibility and reliability in advanced Parkinson disease

In neurodegenerative diseases, damage to sleep-wake regulating brain regions alters sleep architecture (e.g., loss of spindles and K-complexes, EEG slowing during wakefulness and REM sleep), challenging the current American Academy of Sleep Medicine (AASM) rules for scoring sleep. In Parkinson disease (PD) patients, we assessed the feasibility and interrater reliability of a previously described novel sleep scoring methodology.

Video-polysomnography of 15 PD patients and five age-matched controls were scored blinded by five trained raters (two neurologists and three technicians) with different levels of expertise. Recordings were first scored empirically following the AASM methodology and, when necessary, scored again using the following additional abnormal stages: abnormal Wake, Subwake, Undifferentiated NREM (UNREM) sleep, Poorly Structured N2 sleep, and abnormal REM sleep. Sleep microstructure alterations (e.g., encapsulated REM sleep behavior disorder) were also evaluated.

The novel scoring method showed substantial interrater agreement (kappa = 0.75), comparable to the AASM method (kappa = 0.73). UNREM sleep was identified in 12 patients, Subwake in nine, and abnormal Wake, encapsulated REM sleep behavior disorder and abnormal REM sleep in seven. No abnormal stages were observed in three early PD patients or in any control. The presence of the abnormal sleep stages was associated with ageing and advanced PD with cognitive impairment and motor fluctuations.

The novel sleep scoring methodology, which includes abnormal wakefulness as well as NREM and REM sleep stages, is feasible, reliable, and applicable in clinical practice for advanced PD patients undergoing video-polysomnography.