Computer assisted sound analysis of arteriovenous fistula in hemodialysis patients.

The purpose of this study was to reveal the unique sound characteristics of the bruit produced by arteriovenous fistulae (AVF), using a computerized method. An electronic stethoscope (20 Hz to 20 000 Hz sensitivity) was used, connected to a portable laptop computer. Forty prevalent hemodialysis patients participated in the study. All measurements were made with patients resting in supine position, prior to the initiation of mid-week dialysis session. Standard color Doppler technique was used to estimate blood flow. Clinical examination revealed the surface where the perceived bruit was more intense, and the recording took place at a sample rate of 22 000 Hz in WAV lossless format. Fast Fourier Transform (FFT) mathematical algorithm, was used for the sound analysis. This algorithm is particularly useful in revealing the periodicity of sound data as well as in mapping its frequency behavior and its strength. Produced frequencies were divided into 40 frequency intervals, 250 Hz apart, so that the results would be easier to plot and comprehend. The mean age of the patients was 63.5 ± 14 years; the median time on dialysis was 39.6 months (mean 1 month, max. 200 months). The mean blood flow was 857.7 ± 448.3 ml/min. The mean sound frequency was approximately 5 500 Hz ± 4 000 Hz and the median, which is also expressing the major peak of sound data, was 750 Hz, varying from 250 Hz to 10 000 Hz. A possible limitation of the study is the relatively small number of participants.