Rapid Eye Movement Sleep-Related Indicators with Promising Performance in Diagnosing Narcolepsy

The dysregulation of rapid eye movement (REM) sleep is an intrinsic nature of narcolepsy. This study aimed to evaluate the performances of REM sleep-related clinical features and the parallel objective indicators for distinguishing narcolepsy type 1 (NT1) and, more specifically, narcolepsy type 2 (NT2) from patients with hypersomnias.

Drug-free patients with excessive daytime sleepiness (EDS) were consecutively recruited between January 2020 and March 2023. All the patients were categorized into three groups, including NT1, NT2, and EDS controls. The REM sleep-related clinical features and REM sleep parameters were compared across these three groups. Receiver operating characteristic (ROC) analyses were applied to quantify the diagnostic performances of REM sleep-related indicators in differentiating NT1 and NT2 from patients with hypersomnias, with bootstrap resampling to stabilize cut-offs and related metrics.

A total of 206 patients were enrolled in this study (NT1 = 87, NT2 = 27, EDS controls = 92). In the multiple sleep latency test (MSLT), a daytime REM sleep duration of ≥ 11.91 (7.90-18.75) minutes for NT1 (AUC: 0.943 [0.908-0.974], sensitivity: 0.87 [0.76-0.95], specificity: 0.93 [0.84-1.00]) and ≥ 8.41 (3.00-12.50) minutes for NT2 (AUC: 0.922 [0.852-0.973], sensitivity: 0.90 [0.77-1.00], specificity: 0.88 [0.73-0.96]) yielded optimal differentiation from other hypersomnias. Stratified subgroup ROC analyses focusing on NT1 vs IH (AUC: 0.948 [0.897-0.986], cut-off: 12.67 [0.50-18.50] min, sensitivity: 0.85 [0.72-0.99], specificity: 0.97 [0.80-1.00]) and NT2 vs IH (AUC: 0.918 [0.819-0.986], cut-off: 7.18 [1.75-17.50] min, sensitivity: 0.89 [0.60-1.00], specificity: 0.86 [0.67-1.00]) further support the promising discriminatory potential of this marker. Additionally, ROC analyses showed high AUC values for any REM sleep without atonia (RSWA) index during MSLT in discriminating NT1 (AUC: 0.882 [0.826-0.930], cut-off value: 36.61% [3.71-71.81%], sensitivity: 0.91 [0.81-0.98], specificity: 0.80 [0.69-0.89]) and NT2 (AUC: 0.865 [0.796-0.924], cut-off value: 18.83% [7.41-43.85%], sensitivity: 0.99 [0.90-1.00], specificity: 0.75 [0.65-0.85]) from other hypersomnias. Comparable separation performance was also seen in separate NT1 vs IH (AUC: 0.885 [0.772-0.968], cut-off: 42.81% [3.71-82.58%], sensitivity: 0.86 [0.67-0.99], specificity: 0.86 [0.65-1.00]) and NT2 vs IH (AUC: 0.863 [0.701-0.974], cut-off: 20.09% [7.41-82.29%], sensitivity: 0.97 [0.71-1.00], specificity: 0.77 [0.54-1.00]) subgroup analyses. The co-occurrence of REM sleep-related symptoms of sleep paralysis, hypnagogic hallucinations, and REM sleep behavior disorder exhibited a high specificity in diagnosing patients with NT1 (AUC: 0.711 [0.647-0.773], specificity: 0.87 [0.80-0.94]).

The daytime REM sleep duration, any RSWA in MSLT, and REM sleep-related symptoms represent promising candidate biomarkers in distinguishing between narcolepsy subtypes and other hypersomnias, with particular utility for identifying NT2.