Characterization of crackles from patients with fibrosis, heart failure and pneumonia

Med Eng Phys. 2013 Apr;35(4):448-56. doi: 10.1016/j.medengphy.2012.06.009. Epub 2012 Jul 11.

Abstract

The spectral analysis of crackles sounds has been carried out based on the assumption that they are stationary signals, and the majority of the work on the crackles was accomplished before the publication of the Computerized Respiratory Sound Analysis (CORSA) guidelines. This works characterizes crackles acquired from patients with fibrosis, heart failure, and pneumonia, breathing at a constant rate, with a system developed according to the CORSA guidelines. Their maximum frequency was obtained by applying discrete pseudo Wigner-Ville distribution, suitable for non-stationary signals, and an objective method to estimate the maximum frequency, the modified geometric method. The effects of the breathing rate and the tidal volume on the spectra of the crackles were also investigated. The role of the high-pass filter cutoff frequency of the acquisition system on the characteristics of the acquired crackles was also assessed in this present study. Higher high-pass filter cutoff frequency allows for higher amplification which modifies the maximum frequency and the 2CD index. It is shown that the crackles acquired according to the CORSA guidelines have higher frequencies and shorter 2CD indexes than those previously reported, highlighting the need for the standardization and detailed report of the acquisition setup when quantifying lung sounds. The results pointed out that the maximum frequency and the 2CD indexes may allow crackles generated by fibrosis to be distinguished from the ones generated by the heart failure and pneumonia. It is not possible, however, by means of these two indexes, to differentiate between pneumonia and heart failure crackles.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Aged
  • Fibrosis
  • Heart Failure / physiopathology*
  • Humans
  • Lung / physiopathology
  • Middle Aged
  • Pneumonia / physiopathology*
  • Respiration
  • Respiratory Sounds / diagnosis*
  • Signal Processing, Computer-Assisted