Evaluation of three approaches used for respiratory measurement in healthy subjects

Abstract

Abstract Objective. Respiration is one of the critical vital signs of human health status, and accurate respiratory monitoring has important clinical significance. There is substantial evidence that alterations in key respiratory parameters can be used to determine a patient’s health status, aid in the selection of appropriate treatments, predict potentially serious clinical events and control respiratory activity. Although various approaches have been developed for respiration monitoring, no definitive conclusions have been drawn regarding the accuracy of these approaches because each has different advantages and limitations. In the present study, we evaluated the performance of three non-invasive respiratory measurement approaches, including transthoracic impedance (IMP), surface diaphragm electromyography-derived respiration (EMGDR) and electrocardiogram-derived respiration (ECGDR), and compared them with the direct measurement of airflow (FLW) in 33 male and 38 female healthy subjects in the resting state. Approach. The accuracy of six key respiratory parameters, including onset of inspiration (Ion ), onset of expiration (Eon ), inspiratory time (It ), expiratory time (Et ), respiratory rate (RR) and inspiratory–expiratory ratio (I:E), measured from the IMP, EMGDR and ECGDR, were compared with those annotated from the reference FLW. Main results. The correlation coefficients between the estimated inspiratory volume and reference value were 0.72 ± 0.20 for IMP, 0.62 ± 0.23 for EMGDR and 0.46 ± 0.21 for ECGDR (p < 0.01 among groups). The positive predictive value and sensitivity for respiration detection were 100% and 100%, respectively, for IMP, which were significantly higher than those of the EMGDR (97.2% and 95.5%, p < 0.001) and the ECGDR (96.9% and 90.0%, p < 0.001). Additionally, the mean error (ME) for Ion , Eon , It , Et and RR detection were markedly lower for IMP than for EMGDR and ECGDR (p < 0.001). Significance. Compared with EMGDR and ECGDR, the IMP signal had a higher positive predictive value, higher sensitivity and lower ME for respiratory parameter detection. This suggests that IMP is more suitable for dedicated respiratory monitoring and parameter evaluation.