Increase in Heart Rate as a Surrogate EEG Arousal for Type 3 Ambulatory Recordings

Abstract

SESSION TITLE: Sleep Disorders Posters I: Diagnosis SESSION TYPE: Original Investigation Poster PRESENTED ON: Wednesday, October 28, 2015 at 01:30 PM - 02:30 PM PURPOSE: It is known that type 1 sleep recordings are costly and an alternative to EEG recording is needed. The aim of this study is to assess the heart rate (beat per min/bpm) cut-off that can be a sensitive and a specific surrogate to EEG arousals. METHODS: 72 subjects suspected of obstructive sleep apnea (OSA) underwent one night sleep laboratory recording. The following respiratory outcomes were scored: 3% oxygen desaturation index (ODI), apnea-hypopnea index -AASM 2012- (AHI), respiratory disturbance index with EEG arousals (RDIe), and RDI with autonomic arousals (RDIa). Among the 72 subjects, data from 24 subjects were randomly chosen for preliminary analysis of cut-off assessments. A total of 2669 respiratory events were selected and 1762 hypopneas and 561 RERAs were included. Analyses performed for each event were: 1) EEG arousal and desaturation classification; 2) receiver operating characteristic (ROC) curves with EEG arousal as gold standard to assess different autonomic arousal cut-offs based on sensitivity and specificity. RESULTS: Out of all the 1762 hypopneas, 714 were positive for EEG arousals and had no desaturations. Out of the 561 total RERAs, 371 RERAs had EEG arousals. A total of 1085 events were scored with EEG arousals. For a 5 bpm autonomic arousal cut-off, the sensitivity was 97.5% and the specificity 1.5%; with a cut-off of 10 bpm the sensitivity was 74.6% and the specificity 41.4%; with a cut-off of 15 bpm the sensitivity was 45.3% and the specificity 75.9%; and with a cut-off of 12 bpm the sensitivity was 61.8% and the specificity 57.2%. CONCLUSIONS: An autonomic arousal cut-off value of 12 bpm identified 1020 over 1085 RDIe events. CLINICAL IMPLICATIONS: The use of autonomic arousals is a good alternative to identify EEG arousals respiratory events. DISCLOSURE: The following authors have nothing to disclose: Serguei Marshansky, Pierre Rompré, Zeina Kaddaha, Jobin Vincent, Gilles Lavigne, Pierre Mayer No Product/Research Disclosure Information