Oscillations of Hemodynamic Parameters Induced by Negative Pressure Breathing in Healthy Humans
INTRODUCTION: Negative pressure breathing is breathing with decreased pressure in the respiratory tract without lowering pressure acting on the torso. We lowered air pressure only during inspiration (NPBin). NPBin, used to increase venous return to the heart, is considered a countermeasure against redistribution of body fluids toward the head during spaceflight. We studied NPBin effects on circulation in healthy humans with an emphasis on NPBin-induced oscillations of hemodynamic parameters synchronous with breathing. We propose an approach to analyze the oscillations based on coherent averaging. METHODS: Eight men ages 24–42 yr participated in the NPBin and control series. During the series, to reproduce fluids shift observed under microgravity, subjects were supine and head down (−8°). Duration of NPBin was 20 min, rarefaction −20 cm H2O. Hemodynamic parameters were measured by Finometer. Electrical impedance measurements were used to estimate changes in blood filling of cerebral vessels. RESULTS: Mean values of hemodynamic parameters virtually did not change under NPBin, but NPBin induced oscillations of the parameters synchronous with respiration. Peak-to-peak amplitude under NPBin were: mean arterial pressure, 4 ± 1 (mmHg); stroke volume, 7 ± 3 (mL); and heart rate, 4 ± 1 (bpm). Electrical impedance of the head increased during inspiration. The increase under NPBin was three times greater than under normal breathing. DISCUSSION: Analysis of oscillations gives more information than analysis of mean values. NPBin induces short-term decrease in left ventricle stroke volume and arterial blood pressure during each inspiration; the decrease is compensated by increase after inspiration. NPBin facilitates redistribution of body fluids away from the head. Semenov YS, Melnikov IS, Luzhnov PV, Dyachenko AI. Oscillations of hemodynamic parameters induced by negative pressure breathing in healthy humans. Aerosp Med Hum Perform. 2024; 95(6):297–304.
Negative pressure breathing (NPB) involves the use of various techniques and devices that decrease the mean pressure in the respiratory tract below the mean pressure acting on the torso during the breathing cycle. Since the 1990s, NPB has been considered a method for alleviating microgravity effects, which include redistribution of human body fluids in the cranial direction and an increase in intracranial pressure.1–3 One of the NPB types is controlled additional negative pressure applied only during inspiration (NPBin). As well as NPB, NPBin is suggested as a method for prevention of redistribution of human body fluids in the cranial direction under microgravity conditions.4 In addition, NPBin is proposed for correction of reduced venous return and treatment of cerebral hemodynamic disorders.5,6
Changes in mean values of hemodynamic parameters are considered the main criterion of NPBin effectiveness (the averaging is performed for time intervals of several minutes or more). However, the influence of the respiratory pump on the circulation significantly increases under NPBin, which is confirmed by some illustrative fragments of raw recordings.5,6 The recordings clearly show oscillations of central hemodynamic parameters synchronous with oscillations of intrathoracic pressure or airway pressure. Despite the fact that the oscillations can provide more information on the effect of NPBin on the circulatory system than the mean values of hemodynamic parameters, NPBin-induced oscillations of hemodynamic parameters were not analyzed in detail.
The aim of our study was to investigate the effect of NPBin on blood circulation in healthy humans. An emphasis was made on the analysis of NPBin-induced oscillations of central hemodynamic parameters synchronous with breathing. Moreover, we propose an approach for analysis of the oscillations based on the coherent averaging. In contrast to methods that use the Fourier transform, the approach allows analysis of oscillations of values that do not have a constant sampling frequency (e.g., heart rate) and, in addition, it provides more illustrative results. The approach allows analyzing changes in hemodynamic parameters with time resolution sufficient to investigate dynamics of the parameters within the respiratory cycle and provides new opportunities for studying the relationships between respiration and circulation.
METHODS
Subjects
Eight healthy nonsmoking men of average build ages 24–42 yr (mean ± SD = 32.4 ± 6.5) volunteered to participate in our study. Before inclusion, all subjects underwent a medical history and physical examination by a physician to ensure that they had no previous or current medical conditions that might preclude their participation. In accordance with the Declaration of Helsinki, all volunteers gave their written informed consent to participate in the study. The study was approved by the Bioethics Commission of the Institute of Biomedical Problems of the Russian Academy of Sciences (protocol No. 373 of 31.10.2014). When planning the study, we focused primarily on the potential use of NPBin during spaceflight. As most astronauts are men, there are no sex effects on NPBin-induced circulatory reactions,7 and to obtain a more uniform group of volunteers, we did not include women in the group.
Procedure
Each volunteer participated in the NPBin and control series. The NPBin series was conducted first. The rest between NPBin and the control series was 4 wk. Volunteers were acquainted with the measuring equipment and NPBin several days before the start of the study.
Each series consisted of three stages: breathing through the breathing circuit without NPBin (hereafter referred to as “free breathing”) for 10 min, NPBin (or continuing free breathing in the control series) for 20 min, and free breathing for 6 min. For convenience, although there was no NPBin in the control series, we refer to the second stage of the control series as the NPBin stage.
During the series, subjects were positioned supine on a tilted couch with a lowered headboard (the angle to the horizon was 8°). The tilt was used to reproduce a hemodynamic state close to that observed under microgravity conditions.8,9 Breathing with atmospheric air was performed through a face mask connected to a valve box that separated inspiratory and expiratory flows. An additional adjustable valve was installed in the inspiratory part of the breathing circuit to apply NPBin. The adjustable valve was structurally similar to a standard gas reducer and opened at a preset pressure difference between the inspiratory part of the circuit and an atmosphere. In the NPBin series, the valve was set to −20 cm of water (here and below all pressure values are given in relation to barometric pressure). In the control series, the valve was set to zero pressure difference.
It has previously been shown that the decrease in intrathoracic pressure is about half of the decrease in pressure in the respiratory tract under NPB.2 Since the hydrostatic increase in blood pressure at the level of the head under the tilt of −8° is about 10 cm of water relative to the horizontal position, rarefaction of −20 cm of water was chosen to compensate for the increase.
Equipment and Materials
Heart rate (HR) and blood pressure (pulse wave; systolic: SAP; diastolic: DAP; and mean arterial pressure: MAP) were continuously recorded using a modified Peňáz method (Finometer Pro, Finapress Medical Systems, Enschede, The Netherlands). The device evaluated stroke volume (SV), cardiac output (CO), total peripheral vascular resistance (TPR) by pulse wave of arterial blood pressure, and baroreflex sensitivity (BRS)10 by interbeat interval and arterial pressure data. Also, pressure inside the breathing mask (Pm) was recorded continuously. Respiratory rate (BF) and inspiratory time were calculated from Pm data.
An electrical bioimpedance technique was used to estimate the changes in blood filling of the cerebral vascular bed. The modulus of electrical impedance was measured at a frequency of 40 kHz using the frontal-mastoidal lead. This lead measures impedance between an electrode located on the forehead above the eye and an electrode located behind the ear on the mastoid process. One channel of the impedance meter (RPKA2-01, Medass, Moscow, Russia) recorded impedance from the left side and another channel from the right side of the head. A signal representing an absolute value of the impedance was analyzed in the frequency range from 0.07–0.5 Hz (slowR). The frequency range was cropped from the top to remove impedance oscillations produced by the changes in tissue blood volume associated with a pulse pressure wave passing through the vascular bed. The frequency range was cropped from the bottom to remove components of the signal associated with changes in skin-electrode contact characteristics that inevitably occur during prolonged registration (for example, due to gel drying or perspiration). The remaining part of the signal spectrum predominantly reflects respiratory (associated with breathing) changes in volume of blood and cerebrospinal fluid, since the volume of other fluids, including intracellular fluids, cannot change quickly enough to track breathing.
Statistical Analysis
The analysis of data consisted of two parts: analysis of the mean values of hemodynamic parameters at every stage of a series and analysis of the oscillations of hemodynamic parameters synchronous with breathing. To calculate mean values, 1 min at the start and end of each stage was cropped, so calculation was performed for the remaining part. Then we tested differences in the mean value of each parameter between stages using the Wilcoxon signed-rank test. The analysis was performed for each series separately. We also compared NPBin and the control series by analysis of individual (calculated for each volunteer) changes in the mean values between the “NPBin” and “before NPBin” stages (“NPBin” minus “before NPBin”) using the Wilcoxon signed-rank test.
To analyze the oscillations of hemodynamic parameters induced by breathing, we used an approach based on the coherent averaging method. In contrast to spectral methods based on the Fourier transform, coherent averaging can be applied to parameters that do not have a constant sampling frequency. For example, the values of R-R interval duration or SV are not determined until the end of the current cardiac cycle. Moreover, the duration of the cardiac cycle is not a constant value. It is noteworthy that such physiological parameters are often analyzed using spectral methods. This involves filling the gaps in continuous recordings with some value, for example, obtained in the preceding cardiac cycle. However, this approach inevitably leads to spectrum distortions associated with a surge in the signal at the end of each cardiac cycle. In addition, this procedure causes shifts in the mean value of the signal in favor of longer cardiac cycles. For these reasons, we did not fill the gaps in recordings to equalize the sampling rate. We analyzed the original data without introducing any additional values, as the coherent averaging method allows.
At the first step of analysis, the synchronous recording of a studied parameter and Pm was divided into frames corresponding to separate respiratory cycles. The edges of respiratory cycles were determined by the threshold value of the Pm signal. If the duration of an inspiration in a particular frame significantly differed from that in other frames (i.e., was out of the 10…90-percentile range), this frame was excluded from further analysis.
At the second step, all frames were superimposed on one graph so that the starting points of each inspiration were aligned. If the recording is of a sufficient length (i.e., contained 20 or more breathing cycles), the points (values of a parameter) form a cloud which evenly covers the entire time interval corresponding to the frame duration. If a parameter oscillates coherently with breathing (Pm oscillation), the cloud follows these oscillations with some scatter and shift along the time axis. If a parameter does not oscillate at all, or oscillations are not coherent with respiration, the cloud is horizontal and not oscillating. The coherence of oscillations (i.e., the constant difference in phases) is a stricter requirement than the coincidence of the oscillation frequencies of two signals. Therefore, oscillations of the cloud obtained with this method indicate a causal relationship between breathing and changes in a studied parameter more reliably than the coincidence of frequencies.
According to the classic coherent averaging method, a simple calculation of the mean value of a studied parameter at a given moment of time is performed. The averaging is performed over different frames, but for a fixed time value starting from the beginning of the frame. In case a parameter has no constant sampling frequency, each moment of time corresponds to only one value or no values from other frames. Therefore, instead of the mean, we calculated the curve approximating the obtained cloud using the robust locally weighted scatterplot smoothing (RLOWESS) method.11 The resulting curve represented an average response of a studied parameter to an average inspiration observed in an individual volunteer at a single stage of the study. Fig. 1 shows the first two steps of the analysis using HR recording as an example.
Citation: Aerospace Medicine and Human Performance 95, 6; 10.3357/AMHP.6419.2024
Next, we calculated the curve for oscillations of a parameter averaged over the group of volunteers. Since the duration of a breathing cycle and the mean value of a parameter can differ in individual volunteers, values were normalized before calculating the mean curve for the group. Since cardiovascular parameters change most notably during inspiration, we normalized the time axis to the mean value of the individual inspiration duration. After normalization, the starting point of an inspiration corresponded to 0, and the ending point of an inspiration to 1 on the x-axis. The y-axis was not scaled, but the mean value at a particular stage was subtracted from a parameter, so that only changes from the mean value were analyzed. After subtraction of the mean value and normalization of the time axis, the curves obtained for a particular parameter in different volunteers were superimposed on one graph. The calculation of the mean curve and ± SD interval was performed; in this case, the resulting curve represents an average response of a parameter to an average inspiration in the group of volunteers. Analysis was performed separately for each stage of each series.
An example of the final result of the analysis of the parameter oscillations (in this case, HR) in response to Pm changes during breathing is shown in Fig. 2. Hereinafter, we will refer to the curves obtained by such analysis as respiratory oscillations of a parameter.
Citation: Aerospace Medicine and Human Performance 95, 6; 10.3357/AMHP.6419.2024
The significance of the differences in the obtained curves can be roughly estimated using SD. For arbitrary points of a curve or points of different curves, the differences are considered to be significant if they exceed 2 SD.
RESULTS
The mean values of hemodynamic parameters and respiratory pattern are presented in Table I. BF decreased under NPBin in all subjects. In some of them, a decrease in BF was partially retained for several minutes after the end of NPBin. A decrease in BF was mainly due to an increase in the expiration time and the pause between expiration and a subsequent inspiration. Inspiratory time did not change significantly under NPBin (P = 0.141, W = 7, N = 8, NPBin vs. control series).
The mean values of hemodynamic parameters changed little or did not change at all under NPBin. More pronounced NPBin effects upon circulation were observed during inspirations. This can be clearly seen in the pattern of respiratory oscillations (the mean results for the group of volunteers are shown in Fig. 3). It is noteworthy that different volunteers demonstrated unidirectional changes in each parameter during inspiration at each stage of each series. This allowed averaging over the entire group of volunteers. Moreover, the respiratory oscillations at any stage of the control series were similar and virtually coincided with the oscillations before and after NPBin in the NPBin series (Fig. 2 provides HR oscillations as an example). A similar pattern was observed for other hemodynamic parameters. Therefore, a detailed comparison was performed for the curves related to the second stage of the control and the NPBin series (i.e., the NPBin stage).
Citation: Aerospace Medicine and Human Performance 95, 6; 10.3357/AMHP.6419.2024
Respiratory oscillations were observed in the NPBin and in the control series in all parameters excluding BRS (BRS values reflect baroreflex parameters averaged over approximately 20 s). The magnitude of oscillations in the NPBin series was higher than in the corresponding period of the control series. Nevertheless, waveforms of all parameters, except CO, differed little. CO demonstrated a rapid biphasic dynamic during inspiration under NPBin: a decrease in the first half of an inspiration and an equally large increase in the second half of an inspiration.
SAP, DAP, MAP, and SV decreased during inspiration under NPBin and free breathing. Compared with free breathing, the decrease in parameters under NPBin was approximately 1.5-fold greater. Mostly, the decrease occurred in the first half of inspiration under NPBin.
TPR also decreased during inspiration under free breathing and under NPBin. Under NPBin the decrease was more pronounced, and mainly occurred in the second half of inspiration. HR during inspiration increased under free breathing and NPBin, reaching its highest value in the middle or in the second half of an inspiration.
The variable component of the impedance modulus signal (slowR) reflects changes in fluids volume caused by processes with frequencies close to the respiration rate. Respiratory oscillations of slowR were detected under free breathing and under NPBin. Absolute values of impedance increased during inspiration, indicating a decrease in the fluids volume during inspiration. Peak-to-peak amplitude of slowR oscillations increased about three times under NPBin. It was approximately 150 mOhm under NPBin vs. approximately 50 mOhm under free breathing (for comparison, peak-to-peak amplitude of pulse oscillations of impedance modulus value were 70-150 mΩ). During measurements, the constant component of impedance modulus was in the range of 190–250 Ω. If we assume that the relative change in fluids volume (all fluids, not only blood) is equal to the relative change in the impedance taken with the opposite sign, then NPBin produces a decrease in fluids volume in the head by about 0.05–0.1% or 0.5–1 mL (for rough estimation we took the value of fluids volume as 1000 mL). A decrease in the intracranial fluids volume by this value leads to a decrease in intracranial pressure by several mmHg.12
DISCUSSION
Mean values of hemodynamic parameters under NPBin virtually do not change compared to the values observed under free breathing. At the same time, NPBin induces pronounced oscillations of hemodynamic parameters synchronous with respiratory movements. The magnitude of oscillations can dramatically exceed changes in the mean values of parameters. The waveform of oscillations under NPBin is similar to those observed under free breathing. Nevertheless, their amplitude is greater under NPBin than under free breathing. When NPBin stops, characteristics of oscillations return to those observed under free breathing almost immediately. Apparently, respiratory oscillations of hemodynamic parameters under free breathing and NPBin are caused by the same processes, although the effects are more pronounced under NPBin due to a greater decrease in intrathoracic pressure.
SV, HR, and blood pressure are the first to respond to an inspiration under NPBin; SAP and MAP respond somewhat faster and more pronouncedly than DAP. This relationship between the time of change in the parameters and their magnitude indicates that the heart is the first to react to NPBin: SV decreases, whereas HR increases simultaneously, although it takes slightly longer. A decrease in SV produces a decrease in SAP and, consequently, MAP. DAP changes less under NPBin, as it is more dependent on vascular resistance.
Oscillations of SV could be produced by alterations in blood filling of the ventricles during diastole. Taking into consideration rapid response of SV on an inspiration under NPBin and its restoration after the inspiration, alterations of blood filling are most likely produced by stretching of the pulmonary vessels and intrathoracic sections of systemic veins. Decrease in intrathoracic pressure inevitably causes a short-term decrease in blood pressure inside these vessels and their filling, which leads to a short-term decrease in venous return to the left ventricle, hence a decrease in SV.
Already by the end of an inspiration the blood return is restored, leading to the recovery of SV, arterial pressure, and HR. Immediately after the end of an inspiration, SV and blood pressure values slightly exceed the initial ones, which compensates for the decrease observed during inspiration. So, the long-term mean value of these parameters remains virtually unchanged.
Therefore, the direct mechanical effect of reduced intrathoracic pressure during inspiration under NPBin on intrathoracic vessels is crucial for producing the observed changes. Changes in SV seem to trigger changes in hemodynamic parameters and related reactions of cardiovascular regulatory mechanisms.
The decrease in SV during inspiration under NPBin does not contradict the proposal to use NPBin to improve central hemodynamics during hypovolemia or other circulatory disorders. NPBin enhances the respiratory pump, and the respiratory pump is responsible for the observed oscillations of SV, but it is also a physiological mechanism that contributes to increasing venous return to the heart and maintaining cardiac output in patients. The decrease in SV during inspiration is just another manifestation of the respiratory pump.
Rapid biphasic dynamics of CO during inspiration under NPBin is associated with a lag between HR and SV changes. A decrease in CO during the first half of an inspiration under NPBin is caused by a decrease in SV, and its increase during the second half of an inspiration is the consequence of increase in HR.
The decrease in TPR appears to reflect a compensatory response to a decrease in arterial pressure. It is not possible to conclude reliably from the available data whether this reaction is caused by baroreflexes or is mediated by the local mechanisms of blood flow regulation. The decrease in TPR with decrease in MAP is quite intriguing.
Respiratory oscillations of HR under NPBin can be caused by various reflex influences as well as by the direct mechanical effect of reduced intrathoracic pressure on the heart. Based on the available data, it is not possible to discern between the contributions of different physiological mechanisms to changes in HR.
The observed increase in absolute value of head electrical impedance indicates a decrease in blood filling of the cerebral vascular bed. It is consistent with the data on the association between a decrease in intracranial pressure and a decrease in intrathoracic pressure observed in an animal study13 and in patients with clinical conditions associated with cerebral edema.14 Therefore, NPBin may be considered a countermeasure against visual impairment intracranial pressure syndrome, which is believed to be a major risk for future long-duration spaceflight.15
We did not find changes in the mean values of the hemodynamic parameters under NPBin. NPBin does not always change the mean values of the parameters. The effect of NPBin on the mean values may be absent under conditions close to rest, while the effect is dramatic when the circulatory system is challenged by lower body negative pressure, blood loss, etc. (see Ryan et al.16 as an example). There is no answer to the question of why this happens. Speculatively, one can assume that the reason lies in the ability of the circulatory regulation systems to maintain normal values of parameters under external influences. At rest, a considerable part of the reserve of regulatory systems is not involved, as regulatory mechanisms successfully compensate for changes in hemodynamics caused by external actions, including NPBin. But when the influence is too strong, for example, during blood loss or lower body negative pressure, there is no unused reserve. In this case, additional actions such as NPBin will have a profound effect on hemodynamics, regardless of whether the effect is negative or positive.
Previously,17,18 as in this work, an increase in the amplitude of oscillations of hemodynamic parameters was observed under NPBin. The novelty of our work is introduction of the data processing method that gives an intuitive result and allows, using respiratory oscillations, studying in detail the processes in the circulatory system associated with breathing. In contrast to frequency-domain methods or analysis of average values, the approach based on the coherent averaging makes it possible to study the dynamics of the parameters within the respiratory cycle.
However, methods based on coherent averaging have limitations. To successfully apply the method, it is necessary to collect at least two dozen respiratory cycles with similar characteristics so that the average oscillation can be calculated. Accordingly, the duration of recordings is limited from below. Also, the method does not work well if the base value of the parameter being studied changes greatly during the recording.
In conclusion, mean values of hemodynamic parameters virtually did not change under NPBin, but NPBin induced pronounced oscillations of hemodynamic parameters synchronous with respiration. The approach based on coherent averaging made it possible to analyze the oscillations and investigate in detail processes associated with the respiratory pump mechanism. We believe it will be a useful tool for research in the field of basic and applied physiology. The results allow us to state that NPBin facilitates redistribution of body fluids away from the head. Another effect of NPBin on circulation is the short-term decrease in left ventricle stroke volume and arterial blood pressure observed during each inspiration. We should note the observed cardiovascular reactions to NPBin are characteristic of healthy humans under conditions close to resting. Cardiovascular reactions to NPBin in patients or in healthy individuals under extreme conditions may be profoundly different.
An example of heart rate (HR) and mouth pressure (Pm) recordings illustrating the method for data analysis. A) A fragment of the original recording; B) overlaid frames that correspond to individual breathing cycles obtained from the fragment shown in panel A (markers of points on panels A and B correspond to individual frames); and C) overlaid frames of complete recording, the vertical lines mark the edges of an average inspiration.
The mean respiratory oscillations of heart rate (HR) in the group of volunteers. The results are obtained A) before NPBin, B) under NPBin, and C) after NPBin in the control series (dashed) and NPBin series (solid). The vertical lines mark the edges of an inspiration. Thin lines indicate the ± SD interval and bold lines indicate the mean curve of the group. The prefix “d” indicates that calculation was performed for changes in the parameter from its mean value.
The mean respiratory oscillations of central hemodynamic and head electrical bioimpedance parameters in the group of volunteers at the NPBin stage. The control series is shown as a dashed line, the NPBin series is shown as a solid line. Vertical lines mark the edges of an inspiration. Thin lines indicate the ± SD interval and bold lines indicate the mean curve for the group of volunteers. The prefix “d” indicates that calculation was performed for changes in the parameter from its mean value. SAP: systolic arterial pressure; DAP: diastolic arterial pressure; MAP: mean arterial pressure; SV: stroke volume; CO: cardiac output; TPR: total peripheral resistance; slowR: variable component (0.07–0.5 Hz) of the impedance modulus signal, left and right side of the head, respectively.
Contributor Notes
