Guevarra, C

Abstract Purpose Compare agreements between polysomnography-based (PSG) diagnosis of isolated REM-sleep-behavior-disorder (iRBD) and Non-REM-Hypertonia (NRH), a novel biomarker independently associated with synucleinopathy-related neurodegenerative diseases. Methods Sixteen patients with histories of dream-enactment-behavior (DEB)(women=38%; age:64.6±13.0) underwent PSG with simultaneously-recorded Sleep Profiler (SP). Two boarded sleep neurologists independently characterized iRBD. Physician1 combined abnormal qualitative REM-sleep-without-atonia (RSWA) by submental electromyography, with video-confirmation of probably DEB. Physician2 relied solely on qualitative RSWA. SP was auto-staged, technically reviewed, and reprocessed for automated abnormal NRH detection. Kappa scores measured physician and NRH agreements. Results In the 14 records with REM sleep, iRBD was characterized in: Physician1=64%, Physician2=79%, NRH=71% of the records. Across the three methods, unanimous iRBD agreement occurred in 57% of the records (positive=7, negative=1). The between-physician agreement in iRBD classifications was fair (kappa=0.32). The agreement between NRH and Physician1 was moderate (kappa=0.52) versus slight with Physician2 (kappa=0.05). NRH comparisons to consensus physician agreement yielded one false-positive and one false-negative iRBD finding. Physician2 classified: a) iRBD in two cases that were negative by Physician1 and NRH, and b) one negative case that Physician1 and NRH characterized as iRBD. Physician1 identified one negative case that was classified iRBD by Physician2 and NRH. Additionally, NRH was abnormal in one of the two records with no REM sleep. Discussion NRH may assist in iRBD risk assessment, given it agreed with at least one physician in 86% of the cases and the between-physician iRBD agreement was only fair. NRH also characterized iRBD-risk in patients with insufficient REM sleep for RSWA assessment.

Abstract Purpose Evaluate the sleep staging agreement between polysomnography (PSG) and Sleep Profiler (SP) in patients with suspected isolated REM-sleep-behavior-disorder. Methods Twenty-six patients with reported dream-enactment-behavior (Site1=16, Site2=10; 27% women; age 64±13 years) underwent a diagnostic PSG with simultaneously recorded SP. A registered sleep-technologist at each site performed PSG-staging, while SP was auto-staged and technically reviewed/edited. Across technicians, the initial staging was blinded. Site1 then performed unblinded restaging of PSG=N3(N2) vs. SP=N2(N3) epochs, while Site2 conducted a blinded, carefully-targeted restaging of N3. Statistics included Cohen’s kappa and Chi-square analyses. Results Agreement between SP and Site1 vs. Site2 were significantly different for Wake (kappa:Site1=0.816;Site2=0.650;combined=0.736), stage N1 (kappa:Site1=0.149;Site2=0.228;combined=0.188), stage N2 (kappa:Site1=0.632;Site2=0.718;combined=0.659), stage N3 (kappa:Site1=0.715;Site2=0.368;combined=0.525) and REM (kappa:Site1=0.827;Site2=0.719;combined=0.766)(all P<0.001). After restaging of N3, the kappa values improved at Site1 (unblinded:N2=0.659/N3=0.883) and Site2 (blinded:N2=0.775/N3=0.736)(combined:N2=0.735/N3=0.851). The proportion of PSG-epochs restaged from N3 to N2 was 17% at Sites1 and 38% at Site2 (P<0.001), while Site1 had fewer remaining PSG=N3 vs. SP=N2 conflicts (5.6% vs. 20.8%, P<0.001). Compared to Site2, Site1 had a superior: REM kappa due to fewer SP=N2 disagreements (8.5% vs. 16.8%, P<0.001), and Wake kappa resulting from fewer SP=N1 (6.6 vs. 15.6%, P<0.001) and SP=N2 conflicts (5.9 vs. 12.0%, P<0.001). Conversely, the Site1 N1 kappa was inferior due to greater SP=wake disagreement (41.6% vs. 19.8%, P<0.001). Discussion N3 was excessively stage by both PSG technicians before restaging. At Site1, Wake, N3, and REM had almost-perfect-agreement with SP, while N2 had substantial-agreement. At Site2 and across-site, substantial-agreement was observed for Wake, N2, N3, and REM.