Computerized respiratory sounds are a reliable marker in COPD

Abstract

Introduction: Computerized respiratory sounds (RS) have shown potential to monitor respiratory status in patients with COPD. However, variability and reliability of this promising marker in COPD are unknown. Therefore, this study assessed the variability and reliability of RS at distinct airflows and standardized anatomic locations in patients with COPD. Methods: A two-part study was conducted. Part one assessed the intra-subject reliability of RS at spontaneous and target (0.4-0.6L/s and 0.7-1L/s) airflows in 13 outpatients (69.3±8.6yrs; FEV1 70.9±21.4% predicted). Part two characterized the inter-subject variability and intrasubject reliability of RS at each standardized anatomic location, using the most reliable airflow, in a sample of 63 outpatients (67.3±10.4yrs; FEV1 75.4±22.9% predicted). RS were recorded simultaneously at seven anatomic locations (trachea, right and left: anterior, lateral and posterior chest). Airflow was recorded with a pneumotachograph. Normal RS intensity, mean number of crackl and wheezes were analyzed with developed algorithms. Inter-subject variability was assessed with the coefficient of variation (CV) and intra-subject reliability with Intraclass Correlation Coefficient (ICC) and Bland and Altman plots. Results: Relative reliability was moderate to excellent for normal RS intensity and mean number of crackles (ICCs .66-.89) and excellent for mean number of wheezes (ICCs .75-.99) at the three airflows. Absolute reliability was greater at target airflows; especially at 0.4-0.6L/s. Intersubject variability was high for all RS parameters and across locations (CV .12-2.22). RS parameters had acceptable relative and absolute intra-subject reliability at the different anatomic locations. The only exception was the mean number of crackles at trachea, which relative and absolute reliability was poor. Conclusions: RS parameters are more reliable at an airflow of 0.4-0.6L/s and overall reliable at all anatomic locations. This should be considered in future studies using computerized auscutation.