P042 Comparing NOX auto-analysis of sleep staging and quality parameters in a paediatric cohort

Abstract The comparison of NOX auto-analysis (NOXAA) and manual score is well documented in adult populations. This is not well defined in paediatric populations due to nuances in staging criteria, as children <9years have lower posterior dominant rhythm (3.5-9.5Hz) compared to children >9years (8-13Hz). This study will compare the accuracy of the NOXAA with a highly trained manual reference score, and asses the reliability of NOXAA in a paediatric setting. A retrospective analysis compared NOXAA (A) (V6.3.1.34324) and manually scored (J) paediatric polysomnography data. 52 studies were assessed with 16 analysed post exclusion. Exclusion criteria: ages <1, oxygen studies, moderate-severe OSA and EEG impedance >200Hz. Two data sets were separated based on age: data-set 1 = <9years and data-set 2= >9years. Parameters compared: sleep stage (N1, N2, N3 and REM)%, total sleep time (TST)(hours), sleep efficiency (SE)%, sleep latency (SL)(min), REM latency (RL)(min), arousal index (AI) and apnoea hypopnea index (AHI). Data set 1 (J)/(A): N1=1.525%/0.25%*, N2=26.275%/25.425%*, N3=34.9%/64.475%*, REM=27.3%/9.925%*, SL=49.2625/7.1375*, RL=143.45/98.8875, TST=8.16/8.48, SE=80.625%/82.225%, AI=7.8375/5.775*AHI=2.7375/5.325* Data set 2(J)/(A): N1=5.4375%/21.3%*, N2=45.1125%/44.7375%, N3=26.7125%/34.1875%*, REM=22.725%/18.4%*, SL=15.5125/9.5875, RL=143.325/82.025, TST=8.05/8.16, SE=90.4%/89.05%, AI=7.6625/7.7, AHI=4.775/7.675* *indicates statistical difference between J and A There was a significant difference between J and A in N1, N3 and REM for both data sets. TST and SE showed no difference. The data suggests NOXAA is not reliable for paediatric sleep staging, however there may be clinical utility for quality parameters such as TST and SE. Further statistical analysis with larger populations is needed to further validate NOXAA’s accuracy in a paediatric setting.