Monitoring respiration rate in sleep-disordered breathing patients using chest belts or ECG

Abstract

Background: For patients with suspected sleep-disordered breathing respiration rate (RR) is typically measured using a nasal airflow sensor. Additionally, two belts measure the movement of the chest ribcage and upper abdominal walls. At least one channel of ECG is also monitored. Instead of airflow, which could be considered the gold standard, RR may be computed using belts or ECG-derived respiration (EDR) techniques. In this study we compare the performance of such techniques to using airflow. Methods: Evaluation was done on 24 adult polysomnograms from the publicly available PhysioNet/ucddb database. Subjects were randomly selected from patients referred to a sleep lab for possible diagnosis of sleep apnea. Subjects had no known cardiac disease, autonomic dysfunction, and were not on medication known to interfere with heart rate. ECG was automatically analyzed using the Philips ST/AR algorithm to detect normal and exclude abnormal QRS complexes. The EDR waveform was then constructed by measuring the peak to trough amplitude in each normal QRS complex, and interpolating the measurements using cubic splines. A frequency-domain method was developed to calculate RR in a 30-second sliding window with a 5-second step by finding the fundamental frequency of oscillations. This method was applied to all four waveforms: airflow, EDR, ribcage, and abdominal belts. The RR values computed from airflow, used as reference, were compared with RRs from the other waveforms. Results: A total of 119,698 respiration cycles were analyzed. The average difference in RR from airflow was 0.7 ± 1.8 breaths-per-minute (bpm) for abdominal belt, 1.0 ± 2.4 for ribcage belt, and 1.8 ± 2.7 for EDR. Percentage distributions of RR differences are listed below.