Evaluation of a portable respiratory recording device for detecting apnoeas and hypopnoeas in subjects from a general population.

Abstract

This study was designed to validate a new home portable respiratory recording device (PRRD) to identify sleep apnoea and hypopnoea in a group of subjects (n=116), from a sample of the general population. Full night polysomnography (PSG) was used as the gold standard and simultaneously performed with PRRD. PRRD measurements included oronasal airflow (thermistry), chest wall impedance, oxygen saturation, snoring and body position. The sensors were unique for each recording system. Data obtained was blindly reviewed and analysed. A high level of agreement between both methods apnoea/hypopnoea index by PSG and the respiratory disturbance index (RDI) by PRRD was observed. Accuracy of the PRRD was evaluated in terms of sensitivity and specificity for different RDI-PRRD cut-off points with respect to AHI-PSG >10 and AHI-PSG >30. A logistic regression model was performed to estimate the chance per unit of RDI of apnoeas. A received operating characteristic (ROC) curve was drawn to obtain the sensitivity/specificity profile for each observed RDI value obtained. From the ROC curve the authors identified the better cut-off points, which represent a balanced sensitivity/specificity. Through a classification table defined by the cut-off point, the post-odds to exhibit the disease was calculated. For a full PSG cut-off point of 10 a PRRD of six showed a balanced sensitivity of 95% and a specificity of 92%. For a full PSG cut-off point of 30 a PRRD of 16 shows a balanced sensitivity/specificity (100% and 97%, respectively). Post odds of apnoea were calculated for each cut-off point. In conclusion, these data suggest that the portable respiratory recording device is an effective device to identify apnoeas and hypopnoeas in a general population and is therefore a suitable device to be used in epidemiological studies.