One Minute Heart Rate Variability Quantification in Airway Obstruction Model

Following approval by the Institutional Review Board, the study was performed in a prospective cross-over, longitudinal cohort manner. Healthy volunteers aged between 20 and 40 years were included. Each participant signed an informed consent. Exclusion criteria were respiratory or cardiac pathology, diabetes, a body mass index lower than 18 or higher than 26 kg/m2, chronic medical conditions with significant involvement of the sympathetic or parasympathetic systems (e.g. diabetes), obstructive sleep apnea, use of medications with either cardiovascular or respiratory effects, and ECG abnormalities including heart rate other than sinus rhythm, atrioventricular conduction blocks, intrafascicular conduction delays or prolonged QT interval (the time between consecutive Q and T waves on the electrocardiogram).

Initially, each volunteer underwent an ultrasound evaluation (SonoSite™ M-Turbo™ ultrasound machine, SonoSite, Bothell, Washington, USA) of the lungs using a high frequency linear probe at mid clavicular lines to exclude pneumothorax, an apical view of the heart using a curve-linear probe to exclude pericardial effusion. A twelve lead ECG was obtained and analyzed to exclude rate or conduction abnormalities. Noninvasive blood pressure and oxygen saturation were recorded in all subjects throughout the study to identify and prevent any complication.

A three lead ECG and spirometry were obtained in supine position with the upper body raised by 30 degrees. Data was collected using a Datex AS/3 monitor (Datex Ohmeda Medical Equipment, GE Healthcare, USA). The data was recorded using a digital to analog acquisition card (NI-6008, National Instruments™, Austin, Texas, USA) and a Biosignal Logger of National Instruments™ Biomedical Workbench™ at a sampling rate of 500 Hertz (Hz). All experiments were performed at the same time of day (early afternoon), and under the same conditions (same place and experimental setup). Volunteers were instructed to refrain from smoking for 4 hours prior to participating in the study.

The airway obstruction was simulated by an 18 cm long, 4 mm internal diameter endotracheal tube, connected to a spirometry adaptor and an antimicrobial filter. During the obstructed breathing phase, the volunteers were directed to seal their lips tightly around the filter to prevent air leak and encouraged to reach a peak pressure of 30-40 cm H2O, using the instantaneous display on the spirometry monitor. Three sets were recorded for each volunteer; each set was comprised of one minute of normal unobstructed breathing that served as control, immediately followed by one minute of obstructed breathing. Following each set, the volunteers were allowed at least one minute of rest period to recover and return to their baseline breathing before the next set.

HRV analysis A detailed description of the means to measure and evaluate the significance of HRV can be found in the European Society of Cardiology and the North American Society of Pacing Electrophysiology Task Force on heart rate variability, standards of measurement and physiological interpretation and clinical use guidelines14. Briefly, the raw ECG signal was preprocessed (including high pass filtering to remove Baseline wandering and ECG feature identification) and the R-R intervals were extracted from the raw ECG signal using ECG Features Extractor of National Instruments™ Biomedical Workbench™ with threshold adjust factor of 0.1, a rough highest heart rate of 60 beats per minute, QRS frequency of 10-25 Hz and middle QRS onset and offset.

Evaluation of HRV is predominantly performed using time and frequency domains. It may be also performed using nonlinear methods; however, here the investigators concentrated on the more commonly used time and frequency domain methods. Time domain measures are based on the statistical analysis of the time interval between two adjacent QRS waves on the electrocardiogram complexes, referred to as R-R intervals (time between two consecutive R waves on the electrocardiogram). R-R intervals standard deviation (SD), root mean square of successive differences (RMSSD) between adjacent R-R intervals, number of pairs of successive R-R intervals that differ by more than 50 millisecond (NN50) and proportion of NN50 divided by total number of R-R intervals (pNN50) are routinely used to quantify HRV. Frequency domain employ mathematical manipulation to the signal, such as the fast Fourier transform (FFT), which converts the time function into a sum of sine waves of different frequencies. These are used to calculate the power spectral density in very low (VLF), low (LF) and high frequency (HF) ranges and provide a quantification of the physiological HRV-related effects. In this report the investigators did not include the VLF, as this value cannot be reliably measured with a brief measurement window of one minute. As HRV is clearly related to the heart rate, the investigators normalized the HF power and LF power to the heart rate15. The normalized values are referred to as HFnorm and LFnorm.

HRV parameters were calculated over one minute sampling window using the Heart Rate Variability Analyzer of National Instruments™ Biomedical Workbench™. Fast Fourier analysis was employed using a Hanning Window of 1024 samples with a 50% overlap, and with a 2 Hz interpolation rate and 1024 frequency bins. As suggested in the literature for the power spectral density calculation, HF was defined as 0.15-0.4 Hz, LF as 0.04-0.15 Hz and VLF as below 0.04 Hz14.

Respiratory rate calculation To evaluate the changes in respiration during the obstructed breathing the investigators calculated the respiratory rate during the control and obstructed breathing periods. the investigators used the raw ECG traces to calculate the respiratory rate16. This was done primarily based on the R wave amplitude, and calculating the number of local peaks in the sampling window. To obtain meaningful results the investigators chose the control and the obstructed breathing segments with the most obvious changes in the R wave amplitude. Using this methodology the investigators could reliably evaluate the respiratory rate in 33 subjects.