Evaluation of a 2-channel portable device and a predictive model to screen for obstructive sleep apnea in a laboratory environment

Abstract

BACKGROUND: Various portable monitors for identifying obstructive sleep apnea (OSA) have been investigated and reported to enable accurate recording of OSA severity. However, more information is needed from different populations. This study was conducted to evaluate the efficiency of a portable 2-channel sleep apnea device (SleepView) for screening OSA in the Chinese population. METHODS: Ninety-three consecutive subjects underwent simultaneous SleepView testing and laboratory polysomnography (PSG) monitoring. Data were collected and blindly analyzed. The efficiency of the SleepView device and a newly established predictive model for identifying OSA was evaluated in comparison with PSG. RESULTS: Good agreement was evident between the SleepView and PSG based on the apnea-hypopnea index (AHI; r2= 0.84, P 15, and > 30 was 0.923, 0.924, and 0.979, respectively. When the AHI and oxygen desaturation index calculated by the SleepView were combined with waist circumference, the new predictive model showed a higher sensitivity of 92.96% and a specificity of 95.45% for a PSG AHI of ≥ 5, and the corresponding area under the receiver operating curve was 0.983. CONCLUSIONS: The SleepView device exhibited acceptable diagnostic accuracy for OSA in the Chinese population, especially in the severe OSA group. A practical predictive model comprising waist circumference, AHI, and oxygen desaturation index obtained with the SleepView was highly effective for screening even mild OSA. This simple and practical device may serve as a useful tool to screen for OSA. Further studies are required to validate the diagnostic efficiency of the SleepView in the home environment and in different populations.