Computerized Respiratory Sounds Are a Reliable Marker in Subjects With COPD

Area of Science:

• Pulmonary Medicine
• Biomedical Engineering
• Medical Informatics

Background:

• Computerized respiratory sound analysis shows promise for monitoring chronic obstructive pulmonary disease (COPD) patients.
• However, the variability and reliability of these digital markers in COPD require thorough investigation.

Purpose of the Study:

• To assess the variability and reliability of computerized respiratory sounds.
• To evaluate these parameters at distinct airflow rates and standardized anatomical locations in COPD patients.

Main Methods:

• A two-part study involving 13-63 COPD outpatients.
• Respiratory sounds recorded at 7 anatomical locations during spontaneous and target airflows (0.4-0.6 L/s and 0.7-1 L/s).
• Analysis of sound intensity, crackles, and wheezes using validated algorithms; reliability assessed via ICC and Bland-Altman plots.

Main Results:

• Moderate to excellent relative reliability for sound intensity and crackles (ICC 0.66-0.89); excellent for wheezes (ICC 0.75-0.99).
• Absolute reliability was higher at target airflows, especially 0.4-0.6 L/s.
• High inter-subject variability observed across all parameters and locations; crackle reliability was poor at the trachea.

Conclusions:

• Respiratory sound parameters demonstrate greater reliability at an airflow of 0.4-0.6 L/s.
• Overall reliability is acceptable across all anatomical locations, with the exception of tracheal crackles.
• Findings should inform the design of future studies employing computerized auscultation for COPD management.