P006 Adding EMG to combined EEG/EOG improves limited channel sleep scoring

Abstract Introduction Attended polysomnography (PSG) remains the gold standard for OSA diagnosis in children, however home based limited channel sleep monitoring is more accessible. We previously demonstrated that a single combined electro-encephalogram/electro-oculogram (EEG/EOG) often misclassified N1 to awake or REM sleep. We hypothesised addition of EMG would correct this misclassification and improve agreement with PSG scoring. Methods Sample size calculations indicated 1000 epochs would be sufficient to demonstrate a kappa statistic of at least 0.7 for sleep stage comparison. Four children underwent in-laboratory PSG for suspected OSA with additional EEG/EOG channel. Sleep was scored by a single senior scientist using EEG/EOG and the EEG/EOG plus EMG montage (EEG/EOG + EMG) and compared to standard PSG scoring. Results Epoch-by-epoch comparisons were made for 3930 epochs. Agreement with PSG scoring for awake vs sleep only was kappa = 0.82(p<.001) using EEG/EOG, and kappa = 0.88 (p<.001) using EEG/EOG + EMG. Agreement with PSG scoring for awake, NREM and REM sleep was kappa = 0.79 (p<.001) using EEG/EOG, and improved to kappa = 0.88 (p<.001) using EEG/EOG + EMG. Agreement with PSG scoring for awake, N1, N2, N3 and REM sleep was kappa = 0.77 (p<.001) using EEG/EOG, and improved to kappa = 0.84 (p<.001) using EEG/EOG + EMG. N1 misclassification as REM sleep improved when using EEG/EOG + EMG (4%) compared with EEG/EOG (32%). N1 misclassification as awake remained unchanged (28%) between EEG/EOG + EMG and EEG/EOG montages. Discussion Accurate total sleep time is important for accurate respiratory index calculation. The addition of EMG to a combined EEG/EOG sleep scoring montage improved agreement with PSG scoring and is recommended for future abbreviated sleep scoring montages.