Validation Of An Automated Algorithm For Detecting Apneas And Hypopneas By Acoustic Analysis Of Breath Sounds
Alshaer H, Fernie GR, Maki E, Bradley TD, Sleep Medicine. 2013; 14:562.
The objective of this study was to test the validity of a single-channel portable monitoring system that captures breath sounds (BSs) during sleep on a microphone that is embedded in a face-frame and written on to a microprocessor. BSs were recorded from 50 patients undergoing simultaneous polysomnography (PSG) for suspicion of sleep apnea. Using novel software, BSs were analyzed to identify apneas and hypopneas from which the acoustic apnea-hypopnea index (AHI-a) was calculated. The AHIs from PSG (AHI-p) were scored by three technicians blinded to the scoring of the other two technicians according to two criteria: 1) a tidal volume (TV) based criterion that required at least a 50% reduction in the TV signal for at least 10 seconds but not requiring an oxygen desaturation or an arousal from sleep to score a hypopnea (TV50), and 2) AASM criteria that included at least a 30% reduction in the TV signal for at least 10 seconds plus either a 3% oxygen desaturation or an arousal from sleep to score a hypopnea.
BresoDX derived AHI correlated strongly with that from PSG according to both the TV50 criteria (R=0.94) and AASM criteria (R=0.93). Based on a cutoff of AHI-p≥10, overall accuracy of BresoDx AHI was 90% and negative predictive value was 100%.
Acoustic analysis of BSs via a portable monitoring system is a reliable method for quantifying the AHI and diagnosing sleep apnea compared to simultaneous PSG. Scoring hypopneas with or without oxygen desaturation and arousals had a negligible effect on the correlation and accuracy of the portable monitoring system in determining the AHI.