Diagnostic cutoffs of REM sleep without atonia for identifying isolated REM sleep behavior disorder in a Korean population

Abstract

Study objectives: To derive diagnostic cutoffs for REM sleep without atonia (RWA) in isolated REM sleep behavior disorder (iRBD) in a Korean cohort using a multichannel, multi-method approach, and to externally evaluate the stability of specificity in an independent cohort.

Methods: EMG activity was automatically detected in submentalis (SM) and bilateral flexor digitorum superficialis (FDS) muscles using a validated algorithm. RWA was quantified by four methods: SINBAR (3s mini-epochs), AASM (30s epochs), RAI, and CRWA%. The variables were labeled using a 'channel-event-method' format. Cutoffs were derived in 68 patients with iRBD and 46 age- and sex-matched controls. The specificity of SM-based cutoffs was externally evaluated in 68 matched controls from a population-based cohort.

Results: With specificity constrained to 1.00 for diagnostic purposes, the cutoffs were 36.2 % (95 % CIs, 29.2-42.7 %; AUC = 0.98) for combined SM-any and FDS-phasic-3s; 21.2 % (7.6-31.9 %; 0.98) for combined SM-any and FDS-phasic-30s; 6.5 % (2.8-6.9 %; 0.82) for SM-tonic-30s; 0.78 (0.74-0.92; 0.93) for RAI; and 7.1 % (5.5-7.5 %; 0.94) for SM-CRWA%. In external evaluation, specificities for the diagnostic cutoffs were ≥0.90. For the optimal cutoffs, the specificities lay within the derivation cohort's 95 % confidence intervals, except for SM-any-3s. In external controls, the 95th percentile RWA values were 24.6 % for any-3s, 21.9 % for any-30s, 9.7 % for CRWA%, and 0.91 for RAI.

Conclusion: Diagnostic and physiologic RWA cutoffs in Koreans broadly align with Western findings. The discrepancy with earlier Asian cohort data suggests that multinational, multi-ethnic studies are warranted to assess generalizability across Asian populations.

Keywords: Cutoff; Diagnosis; REM sleep behavior disorder; REM sleep without atonia.