Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Aug 14.
Published in final edited form as: Auton Neurosci. 2022 Jun 11;241:103007. doi: 10.1016/j.autneu.2022.103007

The reciprocal relationship between cardiac baroreceptor sensitivity and cerebral autoregulation during simulated hemorrhage in humans

Alexander J Rosenberg a,b,c, Victoria L Kay a, Garen K Anderson a, My-Loan Luu a, Haley J Barnes a, Justin D Sprick a,d, Hannah B Alvarado a,e, Caroline A Rickards a,*
PMCID: PMC10424721  NIHMSID: NIHMS1919421  PMID: 35716525

Abstract

A reciprocal relationship between the baroreflex and cerebral autoregulation (CA) has been demonstrated at rest and in response to acute hypotension. We hypothesized that the reciprocal relationship between cardiac baroreflex sensitivity (BRS) and CA would be maintained during sustained central hypovolemia induced by lower body negative pressure (LBNP), and that the strength of this relationship would be greater in subjects with higher tolerance to this stress. Healthy young adults (n = 51; 23F/28M) completed a LBNP protocol to presyncope. Subjects were classified as high tolerant (HT; completion of −60 mmHg LBNP stage, ≥20-min) or low tolerant (LT; did not complete −60 mmHg LBNP stage, <20-min). R-R intervals (RRI), systolic arterial pressure (SAP), mean arterial pressure (MAP), and middle cerebral artery velocity (MCAv) were measured continuously. Cardiac BRS was calculated in the time domain (ΔHR/ΔSAP) and frequency domain (RRI-SAP low frequency (LF) transfer function gain), and CA was calculated in the time domain (ΔMCAv/ΔMAP) and frequency domain (MAP-mean MCAv LF transfer function gain). There was a moderate relationship between cardiac BRS and CA for the group of 51 subjects in both the time (R = −0.54, P < 0.0001) and frequency (R = 0.61, P < 0.001) domains; there was a stronger relationship in the HT group (R = 0.73) compared to the LT group (R = 0.31) in the frequency domain (P = 0.08), but no difference between groups in the time domain (HT: R = −0.73 vs. LT: R = −0.63; P = 0.27). These findings suggest that an interaction between BRS and CA may be an important compensatory mechanism that contributes to tolerance to simulated hemorrhage in young healthy adults.

Keywords: Cardiac baroreceptor sensitivity, Cerebral autoregulation, Cerebral blood velocity, Lower body negative pressure

1. Introduction

Tolerance to reduced central blood volume (e.g., via hemorrhage) depends upon cardiovascular and cerebrovascular mediated compensatory mechanisms to preserve adequate cerebral blood flow and perfusion (Rickards, 2015). Compensation to acute hypotension requires integrated control of arterial pressure and cerebral blood flow via the baroreflex and cerebral autoregulation (CA) (Ogoh et al., 2010; Tzeng et al., 2010). Baroreflex-mediated increases in heart rate and systemic vascular resistance provide a protective mechanism to buffer temporary reductions in arterial pressure, subsequently maintaining a stable perfusion pressure gradient across the vital organs, including the cerebral circulation (Rickards, 2015; Cooke et al., 2004; Levine et al., 1994; Rickards et al., 2015). To further support adequate cerebral blood flow with changes in perfusion pressure, adjustment of local arterial diameter through myogenic and neurogenic mechanisms are initiated as part of a cerebral autoregulatory response (Zhang et al., 2002; Willie et al., 2014). Accordingly, with reductions in central blood volume and subsequent hypotension, both the baroreflex and CA can protect against cerebral hypoperfusion, the onset of presyncopal symptoms, and syncope.

Prior studies have demonstrated a reciprocal relationship between the baroreflex and CA at rest, and in response to acute hypotension (Ogoh et al., 2010; Tzeng et al., 2010; Nasr et al., 2014; Witter et al., 2017; Gelpi et al., 2021). In young and older human subjects, individuals with relatively low cardiac baroreflex sensitivity (BRS) exhibit stronger metrics of CA and vice versa (Tzeng et al., 2010; Nasr et al., 2014). These findings indicate that there is an individualized capacity to regulate cerebral blood flow, with some relying on CA to buffer changes in perfusion pressure, while others rely on the baroreflex to maintain perfusion pressure. However, if both of these control systems fail, cerebral perfusion may be impaired, potentially leading to syncope. Indeed, it has been demonstrated that reductions in both cardiac BRS and CA occur prior to the onset of syncope during prolonged head-up tilt testing (Faes et al., 2013). Interestingly, in subjects with low tolerance to head-up tilt testing, presyncope was associated with impaired cardiac BRS and an inability maintain adequate CA compared to subjects not susceptible to syncope (Bari et al., 2017). It is unknown if the reciprocal relationship between BRS and CA persists during progressive and sustained hypotension, simulating hemorrhage, and how tolerance to this stress impacts this relationship.

Individuals with high tolerance to progressive reductions in central blood volume induced via lower body negative pressure (LBNP) exhibit greater compensatory responses in systemic vascular resistance (Levine et al., 1994; Convertino and Sather, 2000; Convertino et al., 2012), heart rate (Convertino et al., 2012; Rickards et al., 1985; Kay and Rickards, 2016), and release of vasoactive hormones (Convertino and Sather, 2000; Greenleaf et al., 2000), which may contribute to protection of cerebral blood flow (or velocity) (Levine et al., 1994), although this is not always the case (Rickards et al., 1985; Kay and Rickards, 2016). Interestingly, in a large LBNP study of 135 healthy human subjects, cardiac BRS decreased (assessed via low frequency (LF) transfer function gain between systolic arterial pressure (SAP) and R-R intervals, SAP-RRI LF), while cerebral autoregulation increased (assessed via LF transfer function gain between mean arterial pressure (MAP) and mean middle cerebral artery velocity (MCAv), MAP-mean MCAv LF) along similar trajectories in both high and low tolerant subjects (Rickards et al., 1985). At presyncope, both SAP-RRI LF gain and MAP-mean MCAv LF gain were lower in the high tolerant subjects vs. low tolerant subjects, suggesting that the reduced baroreflex sensitivity (i.e., an attenuated increase in heart rate for a decrease in arterial pressure) was compensated by enhanced CA capacity (i.e., an attenuated decrease in MCAv for a decrease in arterial pressure). The interaction between BRS and CA was not explored in this study, however, and may be an important compensatory mechanism that contributes to tolerance to simulated hemorrhage in young healthy adults. Additionally, elucidating potential differences in the compensatory interaction between BRS and CA responses to hemorrhage, may provide information critical to the development and application of monitoring devices and therapeutic interventions to improve survival from these injuries.

Accordingly, the objective of the current study was to determine the functional relationship between cardiac BRS and CA during a simulated hemorrhage via LBNP, and to assess whether tolerance to LBNP is associated with the strength of this relationship. We hypothesized that the reciprocal relationship between cardiac BRS and CA would be maintained during a simulated hemorrhage induced by maximal LBNP, and that the strength of this relationship would be greater in subjects with higher tolerance to this stress.

2. Methods

2.1. Subjects

Healthy young adults completed this study, conducted at the University of North Texas Health Science Center (UNTHSC) in Fort Worth, TX as part of three independent studies. Data from these subjects has been reported in prior publications that focused on independent research questions (Kay and Rickards, 2016; Kay and Rickards, 2015; Kay et al., 2017; Rosenberg et al., 2021a; Rosenberg et al., 2021b). The UNTHSC Institutional Review Board reviewed and approved the three experimental protocols (Protocol Numbers: 2012–163, 2014–127, 2018–120), with all subjects providing written informed consent prior to participation in the studies. All subjects were non-smokers and/or nicotine users, and free from diagnosed cardiovascular, cerebrovascular, respiratory, metabolic, or inflammatory diseases. All subjects attended a familiarization session, where they were given a verbal briefing and written description of all the measurements and risks associated with the protocol, and were familiarized to the laboratory, personnel, procedures, and monitoring equipment. A medical history, and standing and seated arterial pressure and ECG measurements were also obtained for physician approval prior to participation in the study. Female subjects underwent a urine pregnancy test during the familiarization session and were excluded if pregnant. When included in the study, female subjects performed an additional urine pregnancy test immediately prior to the experimental session. Additional exclusion criteria were body mass index (BMI) >30 kg/m2, resting systolic arterial pressure (SAP) >140 mmHg or diastolic arterial pressure (DAP) >90 mmHg, use of prescription medications (except for oral contraceptives), orthostatic hypotension, donation of blood within 60 days prior to testing, bleeding disorders, and a history of abdominal hernias.

Due to potential confounding effects on cardiovascular and autonomic regulation, all subjects were asked to arrive for testing at the laboratory having abstained for the 24-h before each session from alcohol consumption, stimulants (e.g., caffeine, and cold medications), prescription and non-prescription drugs, herbal medications, and strenuous exercise. Subjects were also instructed to remain hydrated (ad libitum water consumption) and maintain routine sleep patterns prior to the familiarization and experimental session. All female subjects were tested within days 1–4 of the start of menses, or if taking oral contraceptives, during the blank pill or no pill days. Experimental sessions were performed in the morning for all subjects in a temperature-controlled laboratory (22–24 °C).

2.2. Instrumentation

Subjects were positioned in a supine posture inside the LBNP chamber (VUV Analytics Inc., Austin, TX), straddling a bicycle seat with their waist (at the iliac crest) in line with the opening of the chamber. An airtight seal between the subject and the LBNP chamber was achieved by wrapping a plastic sleeve around the subject’s waist, which was then secured in place with a neoprene band. All subjects were instrumented for continuous measurement of heart rate (HR) via a standard three lead ECG (shielded leads, cable, and amplifier, AD Instruments, Bella Vista, NSW, Australia), and beat-to-beat arterial pressure and stroke volume (SV; via ModelFlow®) measurements were obtained via infrared finger photoplethysmography (Finometer, Finapres Medical Systems, Amsterdam, The Netherlands). Respiration rate and end tidal CO2 (etCO2) were measured on a breath-by-breath basis through either a facemask or an oral-nasal cannula via capnography (ML206 Gas Analyzer, AD Instruments, Bella Vista, NSW, Australia). A standardized approach, as outlined in the literature (Willie et al., 2011), was used to acquire cerebral blood velocity signals from a middle cerebral artery (MCA) with a transcranial Doppler ultrasound probe (2-MHz probes; ST3, Spencer Technologies, Seattle, WA), which was held in place over the temporal window with a cushioned and adjustable headframe (Marc 600, Spencer Technologies, Seattle, WA). Efforts were made to ensure that MCA blood velocity (MCAv) recordings were obtained from the right side of the head in all subjects.

2.3. Protocol

All subjects completed a LBNP test to the point of maximal tolerance (i.e., presyncope). The LBNP protocol consisted of a 5-min baseline followed by application of progressively decreasing chamber pressure every 5-min to −15, −30, −45, −60, −70, −80, −90 and −100 mmHg, or until the onset of presyncopal symptoms (Rickards et al., 1985). The protocol was terminated when subjects reached one or more of the following termination criteria: 1) instantaneous SAP below 80 mmHg; 2) sudden relative bradycardia, and/or; 3) voluntary subject termination due to subjective presyncopal symptoms such as grey-out, nausea, sweating, dizziness, blurred vision, or general discomfort (Kay and Rickards, 2016; Kay et al., 2017). The chamber pressure was released immediately at the onset of presyncope, or upon reaching the end of 5-min of −100 mmHg LBNP. Presyncopal symptoms normally resolve within 30–60 s following release of the chamber pressure, and subjects remained in the chamber for a 10-min recovery period.

2.4. Data analysis

All continuous waveform data (e.g., ECG, arterial pressure, SV, MCAv, respiratory rate, etCO2) were collected at 1000 Hz (LabChart, AD Instruments, Bella Vista, NSW, Australia) and analyzed offline via specialized software (WinCPRS, Absolute Aliens, Turku, Finland). Timing of each cardiac cycle was determined from the ECG signal by detection of R-waves. Beat-to-beat SAP, DAP, and systolic and diastolic cerebral blood velocities were then detected and marked from the continuous arterial pressure and MCAv tracings. Mean arterial pressure (MAP) and mean MCAv were automatically calculated as the area under the arterial pressure and cerebral blood velocity waveform via the WinCPRS software.

Oscillatory patterns of R-to-R intervals (RRI), SAP, MAP and mean MCAv were determined via power spectral analysis. Data was made equidistant by interpolating linearly and resampling at 5 Hz, then passed through a low-pass filter with a cutoff frequency of 0.5 Hz. Three-minute data sets were fast Fourier transformed to obtain power spectra (Mahdi et al., 2017), and are expressed as the integrated area within the low frequency (LF, 0.04–0.15 Hz) range. Coherence between MAP and mean MCAv and between RRI and SAP in the LF range were calculated by dividing the squared cross-spectral densities of the two signals by the product of the individual autospectra for all subjects at each time point. Coherence values are reported for all subjects. However, the absolute transfer function gains between SAP and RRI, and between MAP and mean MCAv were calculated only when coherence values were ≥ 0.5.

BRS and CA gains were calculated from the maximal absolute changes (Δ) in each variable of interest from baseline to presyncope (i.e., HR, SAP, MAP, mean MCAv, SAP-RRI LF gain, MAP-MCAv LF gain). In the time domain, BRS and CA gains were calculated as ΔHR/ΔSAP, and CA gain was calculated as ΔMCAv/ΔMAP. In the frequency domain, ΔSAP-RRI LF gain was used as the frequency domain index of BRS, while ΔMAP-MCAv LF gain was the frequency domain index of CA.

2.5. Statistical analysis

Subjects were classified as high tolerant (HT) if they completed at least the −60 mmHg stage of LBNP (≥ 20-min), and low tolerant (LT) if they became presyncopal prior to completing the −60 mmHg stage of LBNP (< 20-min). JMP (Pro 12, SAS Institute, Cary, NC) was used for all statistical analyses, unless otherwise stated. All time domain variables were calculated from the final 1-min of each LBNP stage, and all frequency domain variables were calculated from the final 3-min of each LBNP stage. The absolute change from baseline to pre-syncope (Δ) was calculated for each for the key variables of interest. Normality was assessed using the Shapiro-Wilk test. Normally distributed data were compared between HT and LT groups using separate unpaired t-tests, and non-normally distributed data were compared between HT and LT groups using the Kruskal-Wallis Test. The relationships between BRS and CA (in the time domain or frequency domain) were assessed with Pearson’s correlation coefficients for the entire group of 51 subjects, and separately within the HT and LT sub-groups, utilizing the absolute change from baseline to pre-syncope values. To compare the correlation coefficients between the HT and LT groups (independent groups), Fisher’s r-to-Z transformations were performed, and the resulting Z value assessed via a one-tailed test based on the a priori hypothesis that the strength of this relationship would be higher in HT subjects vs. LT subjects (Weiss, n.d.). To compare slopes between the HT and LT groups, we used the ‘cocor’ package in R (v. 4.0.1) for two independent groups that uses Zou’s (Zou, 2007) confidence interval estimation method. If the confidence interval included zero, the null hypothesis was retained. Additionally, our primary outcome variables were analyzed using two-factor (tolerance, time) linear mixed model analyses with repeated measures from baseline to pre-syncope, followed by Holm-corrected post-hoc tests for multiple comparisons (which were run on the least squared means generated by the linear mixed-model analysis). All subject data are reported as means ± standard deviation (SD). Exact P values are reported for all comparisons, allowing the reader to make their own judgments and interpretation of the results, rather than selecting an arbitrary threshold and the dichotomous use of the term “significant” (Curran-Everett, 2020; Curran-Everett and Benos, 2004).

3. Results

3.1. LBNP tolerance

Sixty-one (N = 61) subjects completed the LBNP protocol, however data were only included for the 51 subjects (28 males and 23 females) who reached true presyncope (defined as mean SAP ≤100 mmHg for the 1-min prior to presyncope, and/or minimum SAP ≤90 mmHg within the 1-min prior to presyncope), and/or bradycardia >10 beats per minute within the 1-min prior to presyncope), as we have previously reported (Kay and Rickards, 2016; Kay et al., 2017). Additionally, all but four subjects exhibited at least one subjective presyncopal symptom (i.e., blurred vision, sweating, nausea, dizziness, lightheaded) within the 1-min prior to presyncope. Of the final 51 subjects who completed the LBNP protocol, 30 were classified as HT and 21 were classified as LT. The LBNP protocol was terminated at −15 mmHg for 1 subject, at −30 mmHg LBNP for 1 subject, at −45 mmHg LBNP for 7 subjects, at −60 mmHg LBNP for 12 subjects, at −70 mmHg LBNP for 19 subjects, at −80 mmHg LBNP for 7 subjects, and at −90 mmHg LBNP for 4 subjects.

Baseline characteristics are compared between HT and LT subjects in Table 1. There were no differences between groups in terms of age, height, weight, or BMI (P ≥ 0.13). By design, the average maximal LBNP at presyncope was 75.0 ± 7.3 mmHg for HT and 51.4 ± 12.2 mmHg for LT (P < 0.0001), with a difference in time to presyncope between groups (HT: 1775 ± 225 s vs. LT: 1184 ± 253 s, P < 0.0001; Fig. 1A).

Table 1.

Demographic data for high tolerant (HT) and low tolerant (LT) subjects at baseline.

HT LT P-Value
N 30
(18M; 12F)		 21
(11M; 10F)		 –
Age, y 27 ± 5 25 ± 3 0.13
Height, cm 170 ± 8 167 ± 9 0.36
Weight, kg 73 ± 15 73 ± 13 0.96
BMI, kg/m2 25.3 ± 3.7 25.9 ± 2.4 0.49
Open in a new tab

Data are means ± SD. BMI, body mass index. Unpaired t-tests were used to compare HT and LT subjects.

Fig. 1.

Fig. 1.

Open in a new tab

By design, tolerance to LBNP was higher in the high tolerant (HT) vs. low tolerant (LT) group (Panel A). HT subjects demonstrated a greater ΔSV (Panel B) and a greater ΔHR (Panel C) in response to maximal LBNP compared with LT subjects. There were no differences in ΔSAP (Panel D), ΔMAP (Panel E), or ΔMCAv (Panel F) between groups. Comparsions between HT and LT groups were made using unpaired t-tests. Exact P values are reported for all comparisons.

3.2. Central and cerebral hemodynamic responses to LBNP

As expected, in response to maximal LBNP, HT subjects demonstrated a greater absolute reduction in SV (HT: −51 ± 17 ml vs. LT: −41 ± 13 ml; P = 0.03; Fig. 1B) and a greater absolute increase in HR (HT: +56 ± 17 bpm vs. LT: +34 ± 20 bpm; P < 0.0001; Fig. 1C). Interestingly, the absolute change from baseline in SAP (HT: −33 ± 11 mmHg vs. LT: −32 ± 12 mmHg; P = 0.65; Fig. 1D), MAP (HT: −18 ± 7 mmHg vs. LT: −19 ± 9 mmHg; P = 0.51; Fig. 1E), and MCAv (HT: −17 ± 8 cm/s vs. LT: −19 ± 10 cm/s; P = 0.52; Fig. 1F) were similar between HT and LT groups in response to maximal LBNP. For reference, all absolute data at baseline and pre-syncope are presented in Table 2. Note that SV, etCO2, and respiration rate data were analyzed with N = 20 LT subjects, due to loss of SV signal in one subject and loss of etCO2 signal in another subject during data collection.

Table 2.

Time domain responses to progressive lower body negative pressure (LBNP) to presyncope (PS) in high tolerant (HT) and low tolerant (LT) subjects.

LBNP Stage LMM P-Value
0 PS 1-min Time Tolerance Interaction
HR, bpm
HT 61.8 ± 8.2 117.6 ± 14.9*,† <0.0001 0.01 <0.0001
 LT 65.5 ± 7.0 99.3 ± 21.4*
SAP, mmHg
HT 128.1 ± 11.8 94.9 ± 5.8* <0.0001 0.81 0.63
 LT 127.7 ± 8.7 96.1 ± 7.2*
MAP, mmHg
HT 94.0 ± 8.8 76.4 ± 5.7* <0.0001 0.59 0.50
 LT 94.0 ± 8.0 74.8 ± 5.8*
SV, ml
HT 91.6 ± 19.6 40.3 ± 10.2* <0.0001 0.43 0.02
 LT (n = 20) 89.4 ± 18.6 48.2 ± 15.4*
Mean MCAv, cm/s
HT 62.0 ± 12.3 45.0 ± 9.5* <0.0001 0.35 0.54
 LT 66.4 ± 15.7 47.8 ± 15.3*
etCO2, mmHg
HT 39.6 ± 4.2 28.6 ± 5.5* <0.0001 0.29 0.71
 LT (n = 20) 40.5 ± 4.0 30.2 ± 7.4*
Respiratory Rate, breaths/min
HT 14.3 ± 4.6 15.3 ± 7.0 0.65 0.89 0.11
 LT (n = 20) 15.4 ± 4.3 13.7 ± 4.5
Open in a new tab

Data are presented as absolute means ± SD. LMM, linear mixed model; HR, heart rate; SAP, systolic arterial pressure; MAP, mean arterial pressure; SV, stroke volume; MCAv, middle cerebral artery velocity; etCO2, end tidal carbon dioxide. PS-1 time point refers to the 1-min prior to pre-syncope. A two-factor linear mixed model analysis with repeated measures from baseline to PS-1 was performed, followed by Holm-corrected post-hoc tests for multiple comparisons (run on the least squared means generated by the linear mixed model analysis).

*

P ≤ 0.004 compared to baseline within a group.

†

P ≤ 0.001 between HT and LT groups. Exact P values are reported for all comparisons.

3.3. Time domain cardiac BRS and CA responses to LBNP

Decreases in ΔHR/ΔSAP indicate reductions in cardiac BRS, while decreases in ΔMCAv/ΔMAP indicate enhanced CA capacity. HT subjects demonstrated a greater absolute decrease in time domain cardiac BRS (ΔHR/ΔSAP) from baseline to presyncope compared with LT subjects (HT: −1.9 ± 0.9 bpm/mmHg vs. LT: −1.2 ± 0.8 bpm/mmHg; P = 0.008), but there was no difference in the absolute increase in time domain CA (ΔMCAv/ΔMAP) between groups (HT: 1.2 ± 0.9 (cm/s)/mmHg vs. LT: 1.5 ± 1.8 (cm/s)/mmHg; P = 0.45 (Table 4).

Table 4.

Comparison of changes in baroreflex sensitivity (BRS) and cerebral autoregulation (CA) in the time and frequency domain between high tolerant (HT) and low tolerant (LT) subjects.

Time domain metrics t-test P-Value
tBRS (ΔHR/ΔSAP), bpm/mmHg
HT −1.9 ± 0.9 P < 0.01
LT −1.2 ± 0.8
tCA (ΔMCAv/ΔMAP), (cm/s).mmHg−1
HT 1.2 ± 0.9 P = 0.45
LT 1.5 ± 1.8
Frequency domain metrics t-test P-Value
fBRS (ΔSAP-RRI LF Gain), ms/mmHg
HT (n = 14) −13.0 ± 9.6 P = 0.47
LT (n = 13) −10.8 ± 5.6
fCA (ΔMAP-mean MCAv LF Gain), (cm/s).mmHg−1
HT (n = 14) −0.29 ± 0.41 P = 0.13
LT (n = 13) −0.06 ± 0.32
Open in a new tab

Data are presented as the mean ± SD change from baseline to pre-syncope. HR, heart rate; SAP, systolic arterial pressure; MAP, mean arterial pressure; MCAv, middle cerebral artery velocity; RRI, R-to-R interval; LF, low frequency; tBRS, time domain cardiac baroreceptor sensitivity; tCA, time domain cerebral autoregulation; fBRS, frequency domain cardiac baroreceptor sensitivity; fCA, frequency domain cerebral autoregulation. Group differences in the absolute change from baseline to presyncope were assessed using unpaired t-tests (tBRS, tCA, and fBRS) and Kruskal-Wallis test (fCA). Exact P values are reported for all comparisons.

There was a moderate reciprocal relationship (B = −0.83; R = −0.54; P < 0.0001; Fig. 2A) between cardiac BRS (ΔHR/ΔSAP) and CA (ΔMCAv/ΔMAP) in the time domain in the group of 51 subjects. In summary, these results suggest that reductions in BRS (i.e., an attenuated increase in HR for a decrease in SAP) was compensated by enhanced CA (i.e., an attenuated decrease in MCAv for a decrease in MAP). When independently assessing this relationship within each tolerance group, there was a stronger relationship in both the HT group (B = −0.80; R = −0.73; P < 0.0001; Fig. 2B) and the LT group (B = −1.44; R = −0.63; P < 0.01; Fig. 2C), with no differences in the R values (Fisher-Z = 0.62; P = 0.27) and slopes (Zou’s confidence interval = −0.4202 – 0.2436; null hypothesis retained) between groups.

Fig. 2.

Fig. 2.

Open in a new tab

There is a moderate reciprocal relationship between baroreflex sensitivity (BRS) and cerebral autoregulation (CA) in the group of 51 subjects in both the time and frequency domain (Panels A and D). This relationship was stronger in the high tolerant (HT) group (Panels B and C) compared to low tolerant (LT) group (Panels E and F) in both the time and frequency domain. The relationships between BRS and CA were assessed with Pearson’s correlation coefficients for the entire group of 51 subjects, and within the HT and LT sub-groups. Exact P values are reported.

3.4. Frequency domain cardiac BRS and CA responses to LBNP

All LF power spectral density data used to calculate our indices of BRS and CA in the frequency domain are presented in Table 3. Due to low coherence (<0.5) between signals throughout LBNP for some subjects, SAP-RRI LF gain data were analyzed with 23 HT and 17 LT subjects, and MAP-mean MCAv LF gain data were analyzed with 25 HT and 18 LT subjects. Accordingly, BRS and CA responses in the frequency domain were only compared in 14 HT and 13 LT subjects who had data at both baseline and pre-syncope for both metrics. As shown in Table 4, the absolute change in frequency domain cardiac BRS (HT: −13.0 ± 9.6 ms/mmHg vs. LT: −10.8 ± 5.6 ms/mmHg; P = 0.47) and CA (HT: −0.29 ± 0.41 cm/s.mmHg−1 vs. LT: −0.06 ± 0.32 cm/s.mmHg –1; P = 0.13) were similar between HT and LT subjects. Decreases in ΔSAP-RRI LF gain indicate reductions in cardiac BRS, however, decreases in ΔMAP-mean MCAv LF gain indicate enhanced CA capacity.

Table 3.

Frequency domain responses to progressive lower body negative pressure (LBNP) to presyncope (PS) in high tolerant (HT) and low tolerant (LT) subjects.

LBNP Stage LMM P-Value
0 PS 3-min Time Tolerance Interaction
RRI LF, ms2
HT 2448.1 ± 4911.5 322.6 ± 425.1* 0.01 0.49 0.21
LT 1369.5 ± 1088.9 633.5 ± 717.4
SAP LF, mmHg2
HT 7.6 ± 6.6 20.8 ± 15.7* <0.0001 0.28 0.25
LT 7.0 ± 7.1 15.4 ± 15.6*
SAP-RRI LF coherence
HT 0.61 ± 0.13 0.61 ± 0.15 0.62 0.44 0.70
LT 0.64 ± 0.15 0.62 ± 0.14
SAP-RRI LF Gain, ms/mmHg
HT (n = 23) 15.4 ± 8.8 3.2 ± 2.0* <0.0001 0.36 0.17
LT (n = 17) 14.7 ± 5.8 6.4 ± 5.4*
MAP LF, mmHg2
HT 5.6 ± 3.6 15.7 ± 11.9* <0.0001 0.30 0.06
LT 6.4 ± 5.4 11.0 ± 9.2
MCAv LF, (cm/s)2
HT 3.5 ± 2.8 4.8 ± 3.5 0.14 0.11 0.78
LT 5.0 ± 3.7 5.8 ± 4.7
MAP-mean MCAv LF coherence
HT 0.63 ± 0.12 0.71 ± 0.11 0.23 0.71 0.17
LT 0.66 ± 0.18 0.66 ± 0.19
MAP-mean MCAv LF gain, (cm/s).mmHg−1
HT (n = 25) 0.75 ± 0.38 0.55 ± 0.14*,† 0.02 0.13 0.13
LT (n = 18) 0.80 ± 0.21 0.75 ± 0.39
Open in a new tab

Data are presented as absolute means ± SD. LF, low frequency; RRI, R-to-R interval; SAP, systolic arterial pressure; MAP, mean arterial pressure; MCAv, middle cerebral artery velocity. PS 3-min refers to the 3-min prior to presyncope. A two-factor linear mixed model analysis with repeated measures from baseline to PS-3 was performed, followed by Holm-corrected post-hoc tests for multiple comparisons (run on the least squared means generated by the linear mixed model analysis).

*

P ≤ 0.03 compared to baseline within a group.

†

P ≤ 0.09 between HT and LT groups. Exact P values are reported for all comparisons.

There was a moderate reciprocal relationship (B = 0.03; R = 0.61; P < 0.001; Fig. 2D) between the absolute changes in frequency domain cardiac BRS (ΔSAP-RRI LF gain) and CA (ΔMAP-MCAv LF gain) in the group of 51 subjects. Similar to the time domain data, these results demonstrated that reductions in the response of BRS from baseline to presyncope are indicative of suppressed BRS that was compensated for by improved CA. Due to the use of absolute change values from baseline to presyncope for indexes of BRS (ΔSAP-RRI LF gain) and CA (ΔMAP-MCAv LF gain) creates a direct relationship, statistically, in the Frequency domain (Fig. 2D–F), however, it is interpreted as an inverse/or reciprocal relationship physiologically. However, when assessing this relationship within each sub-group independently, there was only a strong reciprocal relationship in HT subjects (B = 0.03; R = 0.73; P = 0.003; Fig. 2E), and a weak relationship in the LT subjects (B = 0.02; R = 0.31; P = 0.31; Fig. 2F), with a difference in this this relationship between groups (Fisher-Z = 1.39; P = 0.08), but no difference in the slopes between groups (Zou’s confidence interval = −0.1670–1.0466; null hypothesis retained).

4. Discussion

This is the first study to examine the reciprocal relationship between cardiac BRS and CA in healthy young humans during pre-syncopal limited LBNP. Consistent with our hypothesis, we found that in both the time and frequency domains, CA was inversely related to cardiac BRS during progressive hypotension, and this relationship was stronger in subjects with higher tolerance to this stress. These findings suggest that subjects with an attenuated cardiac baroreflex response to hypotension tend to have a greater cerebral autoregulatory response (and vice versa), and the strength of this relationship can contribute to greater tolerance to simulated hemorrhage. Interestingly, the maximal reductions in both arterial pressure and cerebral blood flow were similar between tolerance groups, but the time (and subsequent magnitude of LBNP) required to reach these maximal responses was related to the strength of the BRS-CA relationship.

This study extends upon prior work from Rickards et al. (1985) who demonstrated an increase in dynamic cerebral autoregulatory function (indexed by reduced transfer function gain between MAP and mean MCAv) in both HT and LT subjects during progressive LBNP, while cardiac baroreflex function decreased (indexed by reduced transfer function gain between SAP and RRI). In the current study, we included metrics of cardiac BRS and CA in both the frequency and time domains and examined the interaction between these two compensatory mechanisms. The overall reciprocal relationship between BRS and CA was maintained in both time and frequency domains but was generally stronger in individuals with higher tolerance to the central hypovolemia induced by LBNP.

Others have also explored the relationship between BRS and CA using hypotensive stimuli. Tzeng et al. examined cardiac BRS during hypotension induced by intravenous bolus injection of sodium nitroprusside, and CA during hypotension induced by rapid bilateral thigh cuff inflation and release (via the rate of regulation and autoregulatory index) but did not simultaneously assess BRS and CA during any one stimulus. Despite this difference in approach compared with the current study, these investigators did demonstrate reciprocal relationships between BRS and CA (R values of −0.69 and −0.72) that are similar to those we show in HT subjects. Witter et al. (2017) expanded upon these findings by assessing the relationship between sympathetic BRS and CA, again during two independent hypotensive stimuli. Sympathetic BRS was examined by quantifying the relationship between DAP and muscle sympathetic nerve activity (MSNA) during pharmacological decreases (via sodium nitroprusside) and increases (via phenylephrine) in arterial pressure, while the autoregulatory index was again used for assessment of dynamic CA during hypotension induced by the thigh cuff technique. Again, the strength of the reciprocal relationship between sympathetic BRS and CA (R = 0.64) was similar to the cardiac BRS to CA relationship shown in prior work, and in the current study.

In these aforementioned studies, compensatory responses of the baroreflex and CA were intact in healthy human subjects. To further examine the contribution of the cardiac baroreflex to cerebral autoregulatory control, Ogoh et al. (2010) pharmacologically suppressed the baroreflex-mediated response to acute hypotension (induced by the thigh cuff technique), then examined the subsequent cerebral blood flow responses. Cerebral autoregulation was impaired (via the rate of regulation), and the reduction in cerebral blood flow was magnified when the baroreflex-mediated increase in heart rate was suppressed by cardiac autonomic blockade. These data indicate that cerebral autoregulation cannot completely compensate for transient reductions in arterial pressure without contributions from the cardiac baroreflex.

4.1. Methodological considerations

We elected to measure cardiac BRS and CA in the time domain using simple delta values from baseline, and in the frequency domain via transfer function analysis, but these are just two approaches of many for assessing these reflex responses. While we showed relatively similar relationships between cardiac BRS and CA with both approaches within the group of 51 subjects (R values of 0.54 and 0.61) and in the HT group (both R values of 0.73), they were not consistent in the LT group (R values of 0.63 and 0.31). It has been shown that many independent metrics of CA do not demonstrate convergent validity, so cannot be used interchangeably (Tzeng et al., 2012), and the same is likely true for measures of cardiac BRS. This may be related to the different physiological mechanisms underpinning these responses. For example, our time domain approaches incorporate the maximal absolute changes in each variable of interest from baseline to presyncope (i.e., HR, SAP, MAP, and mean MCAv), while our frequency domain approaches utilize the beat-to-beat variations in the input and output parameters at baseline and within the final 3-min prior to presyncope. These analytical differences may account for some of the observed disparities in responses we observed, such as the greater decrease in time domain BRS in HT subjects when compared with the LT subjects, but no difference in the transfer function gain responses. Ultimately, the utility of any of these measurements alone, or in combination, will only become apparent once applied to patient populations who have impairments in cardiovascular reflex function.

In a related topic, we only examined the relationship between cardiac BRS and CA in young healthy adults, so it is not clear if these findings are translatable to clinical populations or older individuals with impairments in one or both reflex mechanisms. In patients with carotid stenosis, resting cardiac BRS and CA were lower than healthy control subjects, but the reciprocal relationship between these measurements persisted, although it was weaker (R = 0.58 for patients vs. R = 0.94 for healthy subjects) (Nasr et al., 2014). Interestingly, baseline resting cardiac BRS and CA were similar between HT and LT subjects in the current study, with differences in regulatory capacity only becoming apparent under cardiovascular stress. This observation supports the practice of examining physiological responses to perturbations to unveil impairments, rather than relying on resting conditions only.

We only examined the cardiac baroreflex, without considering the contribution of vascular compensation to the hypotension induced by LBNP. It has been well established that sympathetic activity increases with progressive onset of LBNP (Cooke et al., 2009; Ichinose et al., 2006), to a greater magnitude in individuals with high tolerance to this stress (Convertino et al., 2012), and the DAP-MSNA relationship is impaired prior to the onset of syncope (Ichinose et al., 2006). Based upon the findings of Witter et al. (2017), it is certainly conceivable that the reciprocal relationship between vascular BRS and CA during LBNP would be similar to that observed between cardiac BRS and CA, but this should be tested empirically.

Additionally, it is important to note that the same LF frequency bands were utilized for the analysis of both the cardiac baroreflex and CA (i.e., 0.04–0.15 Hz), compared with the current recommendation of using 0.07–0.20 Hz for CA (Claassen et al., 2016). The rationale for this approach was based on previous studies using the same analytical methods when assessing cardiac BRS and CA with maximal LBNP (Rickards et al., 1985). As the central frequency is around 0.1 Hz for both frequency bands, it is highly unlikely that the outcomes of our study would differ if we used the 0.07–0.20 Hz range for the CA analysis.

5. Conclusion

In summary, we demonstrated a reciprocal relationship between cardiac BRS and CA during progressive central hypovolemia to the onset of presyncope in healthy human subjects, and this relationship was stronger in individuals with higher tolerance to this stress. These findings highlight the inter-related compensatory mechanisms required to respond to perturbations that threaten vital organ perfusion, although the relevance of this relationship to preservation of tissue integrity should be further explored in clinical and animal models. Additionally, this study provides further support that LBNP experiments offer important insight into mechanisms underlying hemodynamic decom-pensation during hemorrhage in humans, and subsequently can provide information critical to the development and application of potential monitoring approaches and therapeutic interventions to improve survival from these injuries.

Acknowledgments

The authors would like to thank our subjects for their time and cheerful cooperation, Dr. Andrew Yockey for assistance with the statistical analysis of the slope data, and Drs. Albert Yurvati, Levi Rice, and Sibi Thomas for their assistance with subject medical examinations.

Funding

Funding for this study was provided, in part, by the U.S. Army MRMC Combat Casualty Care Research Program (Grant # W81XWH-11-2-0137; CAR), the William & Ella Owens Medical Research Foundation (CAR), a contract with Pendar Medical LLC (CAR), and training fellowships awarded to GKA through a National Institutes of Health-supported Neurobiology of Aging Training Grant (T32 AG020494, Principal Investigator: N. Sumien), and an American Heart Association Predoctoral Fellowship (20PRE35210249), to AJR through a Ruth L. Kirchstein NRSA F32 Postdoctoral Fellowship (1F32 HL144082-01A1), and to JDS through a National Institutes of Health-supported Neurobiology of Aging Training Grant (T32 AG020494, Principal Investigator: S. Singh) and a Ruth L. Kirchstein NRSA F31 Predoctoral Fellowship (1 F31 HL134242-01A1). The content is solely the responsibility of the authors and does not necessarily represent the official views the US Department of Defense.

Footnotes

Declaration of competing interest

None to declare for any authors.

References

  1. Bari V, De Maria B, Mazzucco CE, Rossato G, Tonon D, Nollo G, Faes L, Porta A, 2017. Cerebrovascular and cardiovascular variability interactions investigated through conditional joint transfer entropy in subjects prone to postural syncope. Physiol. Meas 38, 976–991. [DOI] [PubMed] [Google Scholar]
  2. Claassen JA, Meel-van den Abeelen AS, Simpson DM, Panerai RB, International Cerebral Autoregulation Research N, 2016. Transfer function analysis of dynamic cerebral autoregulation: a white paper from the International Cerebral Autoregulation Research Network. J. Cereb. Blood Flow Metab 36, 665–680. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Convertino VA, Sather TM, 2000. Vasoactive neuroendocrine responses associated with tolerance to lower body negative pressure in humans. Clin. Physiol 20, 177–184. [DOI] [PubMed] [Google Scholar]
  4. Convertino VA, Rickards CA, Ryan KL, 2012. Autonomic mechanisms associated with heart rate and vasoconstrictor reserves. Clin. Auton. Res 22, 123–130. [DOI] [PubMed] [Google Scholar]
  5. Cooke WH, Ryan KL, Convertino VA, 2004. Lower body negative pressure as a model to study progression to acute hemorrhagic shock in humans. J. Appl. Physiol 96, 1249–1261. [DOI] [PubMed] [Google Scholar]
  6. Cooke WH, Rickards CA, Ryan KL, Kuusela TA, Convertino VA, 2009. Muscle sympathetic nerve activity during intense lower body negative pressure to presyncope in humans. J. Physiol 587, 4987–4999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Curran-Everett D, 2020. Evolution in statistics: P values, statistical significance, kayaks, and walking trees. Adv. Physiol. Educ 44, 221–224. [DOI] [PubMed] [Google Scholar]
  8. Curran-Everett D, Benos DJ, 2004. Guidelines for reporting statistics in journals published by the American Physiological Society. Am. J. Physiol. Regul. Integr. Comp. Physiol 287, R247–R249. [DOI] [PubMed] [Google Scholar]
  9. Faes L, Porta A, Rossato G, Adami A, Tonon D, Corica A, Nollo G, 2013. Investigating the mechanisms of cardiovascular and cerebrovascular regulation in orthostatic syncope through an information decomposition strategy. Auton. Neurosci 178, 76–82. [DOI] [PubMed] [Google Scholar]
  10. Gelpi FB, Vlasta, Cairo Beatrice, De Maria Beatrice, Tonon Davide, Rossato Gianluca, Faes Luca, Porta Alberto, 2021. Correlation between baroreflex sensitivity and cerebral autoregulation index in healthy subjects. Comput. Cardiol 48. [Google Scholar]
  11. Greenleaf JE, Petersen TW, Gabrielsen A, Pump B, Bie P, Christensen NJ, Warberg J, Videbaek R, Simonson SR, Norsk P, 2000. Low LBNP tolerance in men is associated with attenuated activation of the renin-angiotensin system. Am. J. Physiol. Regul. Integr. Comp. Physiol 279, R822–R829. [DOI] [PubMed] [Google Scholar]
  12. Ichinose M, Saito M, Fujii N, Kondo N, Nishiyasu T, 2006. Modulation of the control of muscle sympathetic nerve activity during severe orthostatic stress. J. Physiol 576, 947–958. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kay VL, Rickards CA, 2015. Reproducibility of a continuous ramp lower body negative pressure protocol for simulating hemorrhage. Physiol. Rep 3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Kay VL, Rickards CA, 2016. The role of cerebral oxygenation and regional cerebral blood flow on tolerance to central hypovolemia. Am. J. Physiol. Regul. Integr. Comp. Physiol 310, R375–R383. [DOI] [PubMed] [Google Scholar]
  15. Kay VL, Sprick JD, Rickards CA, 2017. Cerebral oxygenation and regional cerebral perfusion responses with resistance breathing during central hypovolemia. Am. J. Physiol. Regul. Integr. Comp. Physiol 313, R132–R139. [DOI] [PubMed] [Google Scholar]
  16. Levine BD, Giller CA, Lane LD, Buckey JC, Blomqvist CG, 1994. Cerebral versus systemic hemodynamics during graded orthostatic stress in humans. Circulation 90, 298–306. [DOI] [PubMed] [Google Scholar]
  17. Mahdi A, Nikolic D, Birch AA, Payne SJ, 2017. At what data length do cerebral autoregulation measures stabilise? Physiol. Meas 38, 1396–1404. [DOI] [PubMed] [Google Scholar]
  18. Nasr N, Czosnyka M, Pavy-Le Traon A, Custaud MA, Liu X, Varsos GV, Larrue V, 2014. Baroreflex and cerebral autoregulation are inversely correlated. Circ. J 78, 2460–2467. [DOI] [PubMed] [Google Scholar]
  19. Ogoh S, Tzeng YC, Lucas SJ, Galvin SD, Ainslie PN, 2010. Influence of baroreflex-mediated tachycardia on the regulation of dynamic cerebral perfusion during acute hypotension in humans. J. Physiol 588, 365–371. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Rickards CA, 2015. Cerebral blood-flow regulation during hemorrhage. Compr. Physiol 5, 1585–1621. [DOI] [PubMed] [Google Scholar]
  21. Rickards CA, Ryan KL, Cooke WH, Convertino VA, 1985. Tolerance to central hypovolemia: the influence of oscillations in arterial pressure and cerebral blood velocity. J. Appl. Physiol 111 (1048–1058), 2011. [DOI] [PubMed] [Google Scholar]
  22. Rickards CA, Johnson BD, Harvey RE, Convertino VA, Joyner MJ, Barnes JN, 2015. Cerebral blood velocity regulation during progressive blood loss compared with lower body negative pressure in humans. J. Appl. Physiol 119, 677–685. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Rosenberg AJ, Kay VL, Anderson GK, Luu ML, Barnes HJ, Sprick JD, Rickards CA, 2021. The impact of acute central hypovolemia on cerebral hemodynamics: does sex matter? J. Appl. Physiol 130, 1786–1797. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Rosenberg AJ, Kay VL, Anderson GK, Sprick JD, Rickards CA, 2021. A comparison of protocols for simulating hemorrhage in humans: step versus ramp lower body negative pressure. J. Appl. Physiol 130, 380–389. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Tzeng YC, Lucas SJ, Atkinson G, Willie CK, Ainslie PN, 2010. Fundamental relationships between arterial baroreflex sensitivity and dynamic cerebral autoregulation in humans. J. Appl. Physiol 108, 1162–1168. [DOI] [PubMed] [Google Scholar]
  26. Tzeng YC, Ainslie PN, Cooke WH, Peebles KC, Willie CK, MacRae BA, Smirl JD, Horsman HM, Rickards CA, 2012. Assessment of cerebral autoregulation: the quandary of quantification. Am. J. Physiol. Heart Circ. Physiol 303, H658–H671. [DOI] [PubMed] [Google Scholar]
  27. Weiss BA. Fisher’s r-to-Z transformation calculator to compare two independent samples [Computer software]. https://blogs.gwu.edu/weissba/teaching/calculators/fishers-z-transformation/.
  28. Willie CK, Colino FL, Bailey DM, Tzeng YC, Binsted G, Jones LW, Haykowsky MJ, Bellapart J, Ogoh S, Smith KJ, Smirl JD, Day TA, Lucas SJ, Eller LK, Ainslie PN, 2011. Utility of transcranial Doppler ultrasound for the integrative assessment of cerebrovascular function. J. Neurosci. Methods 196, 221–237. [DOI] [PubMed] [Google Scholar]
  29. Willie CK, Tzeng YC, Fisher JA, Ainslie PN, 2014. Integrative regulation of human brain blood flow. J. Physiol 592, 841–859. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Witter T, Tzeng YC, O’Donnell T, Kusel J, Walker B, Berry M, Taylor CE, 2017. Inter-individual relationships between sympathetic arterial baroreflex function and cerebral perfusion control in healthy males. Front. Neurosci 11, 457. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Zhang R, Zuckerman JH, Iwasaki K, Wilson TE, Crandall CG, Levine BD, 2002. Autonomic neural control of dynamic cerebral autoregulation in humans. Circulation 106, 1814–1820. [DOI] [PubMed] [Google Scholar]
  32. Zou GY, 2007. Toward using confidence intervals to compare correlations. Psychol. Methods 12, 399–413. [DOI] [PubMed] [Google Scholar]

RESOURCES