Eine Pilotstudie zur akustischen Langzeitanalyse nächtlicher respiratorischer Symptome bei dekompensierter chronischer Herzinsuffizienz mittels digitaler Auskultation
Loading...
Files
Date
Authors
Publisher
Philipps-Universität Marburg
Supervisors
Abstract
INTRODUCTION: Digital auscultation provides an objective, ubiquitous, and cost-effective means of detecting pulmonary and cardiac diseases based on their characteristic breath sounds. Previous studies have focused on the acoustic analysis of pathognomonic breath sounds. Common to their results is that diseases have specific differential diagnostic characteristics that can aid in diagnosis. It is only in recent years that binding standards have been established regarding the technical requirements of an acoustic recording system, so that the comparison of individual research results was only possible to a limited extent. With the LEOSound long-term recorder, it has become possible to record respiratory sounds in the sense of a long-term stethoscope and to evaluate them. Chronic heart failure is one of the most common and cost-intensive diseases today. This work investigates by means of digital long-term auscultation the prevalences of crackles and sleep-related respiratory disorders in patients with decomp. chronic heart failure. To date, there has been no comparable study using this method over a nocturnal period of 8 hours. The collected data will be used to train an automatic crackles algorithm in the future.
METHODS: The respiratory and breath sounds of 13 patients with decomp. chronic. Heart failure (LVEF <40% or BNP >35pg/ml) were recorded by digital auscultation over eight hours at night. Subsequently, an analysis of the sound spectrograms was performed epoch by epoch by an expert. For this purpose, 37,440 recorded 30-second epochs were audiovisually videotaped, and pathological events were recorded accordingly. The data set examined included spectrograms of the trachea, as well as those of the basal portions of the left and right lungs, in each patient. The prevalences of crackles, apnea, and Cheyne- Stokes events were determined and statistically analyzed. Statistical analysis regarding correlations with LVEF, BNP value, and BMI was performed.
RESULTS: All patients exhibited rales in the long-term acoustic recordings. A median of 1614 crackles per patient per night could be detected (Q1 = 625; Q3 = 2087). The median crackle rate/h was 202 (Q1 = 70; Q3 = 261). In 22% of thenocturnal epochs (30s sections) we found crackles in the recording. Phasically, there were epochs without crackles events. The mean BMI of the study participants was 28.1 kg/m² ± 6.2. Pearson's correlation showed an association between increased BMI and more frequent occurrence of crackles events (rP = 0.62; p = .03), which was not statistically robust to extreme values (rS = 0.44; p = .15). The strongest significant correlations were found between BMI and Cheyne-Stokes respiration (rS = 0.88; p < .001), and BMI and overall prevalence of sleep-related breathing disorders (rS = 0.77; p = .003). The prevalence of sleep-related breathing disorders with an AHI cutoff of ≥ 15 was 54%. Apnea and Cheyne-Stokes episodes occurred at a median frequency of 121 times per patient per night (Q1 = 75; Q3 = 284). Hourly, a median of 15 respiratory episodes were recorded (Q1 = 9; Q3 = 36). Nonspecific apneas were noted in 21% of epochs. Lower LVEF correlated significantly with more sleep-related breathing disorders (rS = -0.55; p = .05). Subgroup analysis showed a significant correlation of LVEF with the frequency of single nonspecific apneas lasting longer than 10s (rS = - 0.68; p = .01). The NYHA group did not correlate significantly with either rales or apnea events in this series of measurements.
CONCLUSION: In this work, it was shown that long-term measurements of breath sounds and respiratory sounds could be practically performed with the LEOSound recorder and crackles could be reliably measured. There is an increased prevalence of rales and sleep-disordered breathing in patients with severe heart failure as measured by LVEF. In particular, nonspecific apneas lasting >10s correlate positively with impaired left ventricular dysfunction. There is evidence that the number of crackles is positively related to BNP levels and the severity of a patient's left ventricular failure. A high BMI appears to be associated with more crackles events and sleep-disordered breathing - especially Cheyne- Stokes breathing. Long-term digital auscultation offers great potential in the field of monitoring patients with cardiac and pneumological diseases. It represents a cost-effective, ubiquitous and non-invasive tool for established medicine.
Review
Metadata
Contributors
Supervisor:
Dates
Created: 2023Issued: 2023-06-05Updated: 2023-06-05
Faculty
Medizin
Publisher
Philipps-Universität Marburg
Language
ger
Data types
DoctoralThesis
Keywords
dekompensierte chronische HerzinsuffizienzAtemgeräuscheRasselgeräuschedigital auscultationChronic heart failureAtemnebengeräuscheLung soundsCHFLungenauskultationpilot study,akustische Langzeitanalysecracklesrales
DFG-subjects
Digitale Auskultation
DDC-Numbers
610
show more
Lumm, Julian: Eine Pilotstudie zur akustischen Langzeitanalyse nächtlicher respiratorischer Symptome bei dekompensierter chronischer Herzinsuffizienz mittels digitaler Auskultation. : Philipps-Universität Marburg 2023-06-05. DOI: https://doi.org/10.17192/z2023.0343.