Attenuated cardiac autonomic function in humans with high-affinity hemoglobin and compensatory polycythemia

Abstract

During hypoxic exposure, humans with high-affinity hemoglobin (and compensatory polycythemia) have blunted increases in heart rate compared with healthy humans with typical oxyhemoglobin dissociation curves. This response may be associated with altered autonomic control of heart rate. Our hypothesis-generating study aimed to investigate cardiac baroreflex sensitivity and heart rate variability among nine humans with high-affinity hemoglobin [6 females, O₂ partial pressure at 50% $\text{S}{\text{a}}_{{\text{O}}_2}$ (P₅₀) = 16 ± 1 mmHg] compared with 12 humans with typical affinity hemoglobin (6 F, P₅₀ = 26 ± 1 mmHg). Participants breathed normal room air for a 10-min baseline, followed by 20 min of isocapnic hypoxic exposure, designed to lower the arterial partial pressure O₂ ($\text{P}{\text{a}}_{{\text{O}}_2}$) to ∼50 mmHg. Beat-by-beat heart rate and arterial blood pressure were recorded. Data were averaged in 5-min periods throughout the hypoxia exposure, beginning with the last 5 min of baseline in normoxia. Spontaneous cardiac baroreflex sensitivity and heart rate variability were determined using the sequence method and the time and frequency domain analyses, respectively. Cardiac baroreflex sensitivity was lower in humans with high-affinity hemoglobin than controls at baseline and during isocapnic hypoxic exposure (normoxia: 7 ± 4 vs. 16 ± 10 ms/mmHg, hypoxia minutes 15–20: 4 ± 3 vs. 14 ± 11 ms/mmHg; group effect: P = 0.02, high-affinity hemoglobin vs. control, respectively). Heart rate variability calculated in both the time (standard deviation of the N-N interval) and frequency (low frequency) domains was lower in humans with high-affinity hemoglobin than in controls (all P < 0.05). Our data suggest that humans with high-affinity hemoglobin may have attenuated cardiac autonomic function.

INTRODUCTION

Humans with rare genetic hemoglobin mutations may demonstrate chronically elevated oxyhemoglobin-binding affinity, which manifests in a lower partial pressure of oxygen required to achieve to 50% oxyhemoglobin saturation (P₅₀), and compensatory polycythemia. The compensatory polycythemia likely stems from increased erythropoietin levels as a result of tissue hypoxia detected by the kidney from hindered release of oxygen due to the hemoglobin’s high affinity for oxygen (1). Recently, our laboratory investigated the impact of reduced P₅₀ on integrative physiological responses to conditions of hypoxic stress at rest and during exercise among a cohort of humans with high-affinity hemoglobin (P₅₀: 15.6 ± 1.0 mmHg, [Hb]: 18.8 ± 2.6 g/dL) (2–4). These studies demonstrated that humans with high-affinity hemoglobin have greater tolerance to hypoxic exercise due to an attenuated reduction in arterial oxygen saturation ($\text{S}{\text{a}}_{{\text{O}}_2}$) and arterial oxygen content ($\text{C}{\text{a}}_{{\text{O}}_2}$), which is commonly observed in other animals (5–7). However, little is known regarding the impact of high-affinity hemoglobin on cardiac autonomic control.

Previous observations from our laboratory (2) and others (8, 9) suggest that humans with high-affinity hemoglobin exhibit altered cardiovascular regulation during hypoxic exposure. Specifically, during 20 min of isocapnic hypoxia [partial pressure of arterial O₂ ($\text{P}{\text{a}}_{{\text{O}}_2}$): ≈50 mmHg], we found that the peak heart rate response was 30–40% lower in humans with high-affinity hemoglobin compared with healthy age-matched controls, despite similar systemic vascular resistance and arterial blood pressure responses (3). The blunted heart rate response is consistent with earlier experiments of participants with high-affinity hemoglobin exposed to short-term (5 min) (9) and longer-term (10 days) hypoxia (8). Mechanisms contributing to the blunted heart rate responses to hypoxia in humans with high-affinity hemoglobin could include reduced activation of aortic chemoreceptors that may be sensitive to $\text{C}{\text{a}}_{{\text{O}}_2}$ in addition to $\text{P}{\text{a}}_{{\text{O}}_2}$ (3, 10–13). Blunted heart rate responses to hypoxia may also be attributed to lower resting parasympathetic control, which results in less parasympathetic withdrawal and attenuated increases in heart rate during hypoxia. Furthermore, the blunted heart rate responses to hypoxia among humans with high-affinity hemoglobin may arise from altered baroreflex regulation of heart rate. This idea is consistent with the observation that humans with polycythemia, but typical affinity hemoglobin, demonstrate lower resting cardiac baroreflex function and elevated sympathetic outflow (14, 15). One mechanism that may cause a reduction in cardiac baroreflex sensitivity in humans with high-affinity hemoglobin is increased arterial stiffness stemming from a lifetime of polycythemia (16). Increased arterial stiffness of baroreceptor-harboring vessels reduces the transduction of arterial pressure changes into baroreceptor sensory input to the brainstem, rendering the arterial baroreflex less sensitive. However, cardiac autonomic function remains to be studied in humans with genetic mutations producing high-affinity hemoglobin.

Therefore, the current study aimed to investigate the impact of high-affinity hemoglobin on cardiac autonomic function in humans. We compared cardiac autonomic function [as assessed by cardiac baroreflex sensitivity (BRS) and indices of heart rate variability (HRV) between humans with rare high-affinity hemoglobin mutations and age-matched humans with typical affinity hemoglobin. Because of our earlier observation that humans with high-affinity hemoglobin demonstrated greater resting heart rate and blunted heart rate responses despite similar changes in mean arterial pressure to isocapnic hypoxia (3) relative to controls, we hypothesized that cardiac baroreflex sensitivity and heart rate variability are reduced in humans with high-affinity hemoglobin compared with in humans with typical affinity hemoglobin.

METHODS

This study was approved by the Mayo Clinic Institutional Review Board (16–007719), and each participant provided their written informed consent after study protocol and safety risks were presented. The study adhered to the Declaration of Helsinki except in regard to registration in a database.

Participants

Nine participants (6 female, 3 male; age range: 20–62 yr; P₅₀: 15.6 ± 1.0 mmHg) with high-affinity hemoglobin were recruited from a Mayo Clinic database. Eight of the high-affinity hemoglobin participants were consanguineous and had the Hb-Malmӧ variant: β97His → Gln (17). One high-affinity hemoglobin participant had the Hb-San Diego variant: β109Val → Met (18). The hemoglobin associated with both variants demonstrates normal Bohr effects but decreased heme-heme interactions (19, 20). Twelve age-matched participants with typical affinity hemoglobin (6 female, 6 male; age range: 25–56 yr; P₅₀: 26.2 ± 1.3 mmHg) were recruited.

Experimental Design

This study represents a novel, hypothesis-generating analysis of previously published data (3). After participants arrived, venous blood samples were taken to measure P₅₀. Measurements of P₅₀ for these participants have been described previously (3) and were conducted using a dual-wavelength spectrophotometry and a Clark electrode (Hemox Analyzer; TCS Medical Products, Huntington Valley, PA). Once completed, an arterial catheter was inserted to allow for arterial blood gas sampling and beat-by-beat blood pressure monitoring. The participants then laid supine and were fit with a face mask interfaced with a mass flow sensor proximal to a two-way nonrebreathing valve (2700 Series, Hans Rudolph, Inc., Shawnee, KS). After resting with the mask on while breathing room air, participants were exposed to 20 min of isocapnic hypoxia during which the target end-tidal oxygen pressure ($\text{P}{\text{ET}}_{{\text{O}}_2}$) was 50 mmHg while maintaining their end-tidal carbon dioxide pressure ($\text{P}{\text{ET}}_{{\text{CO}}_2}$) level at baseline ($\text{P}{\text{ET}}_{{\text{CO}}_2}$: ≈40 mmHg). A custom-built dynamic end-tidal forcing system integrated with a rapid sampling mass spectrometer (MGA-1100; Perkin-Elmer, Kansas City, MO) was used to adjust the inspirate on a breath-to-breath basis (mixing from 3 individual tanks of 100% O₂, 100% CO₂, and 100% N₂) to adjust the end-tidal gases to the desired levels (fraction of inspired O₂ averaged 11.9 ± 0.0% in both groups) (21, 22).

Exercise Testing

As previously reported (2), participants completed a step-wise cycling exercise test to determine maximal oxygen uptake (V̇o_2max). V̇o_2max is reported here as a demographic variable to address the role of fitness in the differences of autonomic function.

Data Collection

Three-lead electrocardiography (ECG) allowed for continuous monitoring of heart rate. Respiratory flow and volume were measured using a pneumotachograph (MCG Diagnostics, St. Paul, MN). Breath-by-breath end-tidal oxygen and carbon dioxide were measured from a sampling port in the mass flow sensor and connected to the mass spectrometer. A pulse oximeter placed on a distal phalanx of the hand was used to monitor hemoglobin saturation (Cardiocap/5; Datex-Ohmeda, Louisville, CO). Heart rate, blood pressure, peripheral oxygen saturation, respiratory flow, and respiratory volume were analog-to-digital converted (PowerLab; AD Instruments, Colorado Springs, CO) and recorded (1,000 Hz) using data acquisition software (LabChart v8.1, AD Instruments, Colorado Springs, CO).

Arterial Blood Pressure and Samples

An arterial catheter (model RA-04020; Arrow International, Reading, PA) was placed by a trained physician either in the brachial or the radial artery using a local anesthetic (2% lidocaine) and sterile technique. The catheter was connected to a commercially available pressure transducer (Single TruWave model, Edwards LifeSciences Corporation, Irvine, CA) that allowed for continuous recording of blood pressure (Cardiocap/5; Datex-Ohmeda, Louisville, CO) and arterial blood sampling. The pressure transducer was placed at heart level and pressurized with saline to ensure patency. Arterial blood gas samples were collected at baseline and after 5, 12, and 20 min of hypoxic exposure. Blood samples were analyzed with a commercial blood gas analyzer (ABL90 COOX; Radiometer, Copenhagen, Denmark) to measure hemoglobin content, $\text{P}{\text{a}}_{{\text{O}}_2}$, and $\text{S}{\text{a}}_{{\text{O}}_2}$, as well as to calculate $\text{C}{\text{a}}_{{\text{O}}_2}$.

Cardiac Baroreflex Sensitivity

Cardiac baroreflex sensitivity was analyzed using a custom program (23) based on the widely used cardiac baroreflex sensitivity sequence method (24–26) using commercially available software (MATLAB R2021a, MathWorks, Natick, MA). This analysis includes sequences of three or more consecutive beats in which systolic blood pressure and R-R interval changed in the same direction (i.e., increased or decreased). All data were analyzed using a zero-beat lag (i.e., systolic blood pressure and R-R interval from the same heartbeat). For each sequence, systolic blood pressure and R-R interval pairs were plotted, and a linear regression was performed to determine the slope of the line. The mean regression slope from all sequences detected in a condition was taken as a measure of spontaneous cardiac baroreflex sensitivity. This analysis was performed on 5-min analysis segments for all study conditions.

Heart Rate Variability

Heart rate variability was measured using 5-min segments to calculate time and frequency domains. Heart rate variability domains were analyzed with the Lomb–Scargle periodogram using commercially available software (MLS310/8 HRV 2.0 m, LabChart v8.1.18, ADInstruments, Colorado Springs, CO). Time domain calculations included the standard deviation of N-N interval (SDNN) and root mean square of successive difference of the R-R interval (RMSSD). Frequency domain measures included low-frequency (LF; 0.04–0.15 Hz) and high-frequency (HF; 0.15–0.40 Hz) bands. Both low frequency and high frequency were expressed as absolute (ms²) and normalized units (nu). Normalized units were calculated as the absolute power for the selected frequency band divided by the total absolute values of low and high frequency.

Data Analysis and Statistics

Data were averaged in 5-min segments, beginning with the last 5 min of baseline in normoxia and throughout the 20-min isocapnic hypoxia period. All variables were evaluated for statistical outliers and considered an outlier if they fell beyond ±2 standard deviations from the mean. Two-tailed, independent sample t tests were used to compare demographics, arterial blood gases, and resting hemodynamics between groups under baseline conditions. Linear mixed modeling analysis with random intercept for each participant was performed for the main outcome variables: Hb Type (Control, High Affinity) × Time (Baseline, 0–5, 5–10, 10–15, 15–20 min). Q-Q plots were created to depict the normality of the residuals and can be found in the data supplement (Supplemental Figs. S1, S2, and S3). Fisher’s least significant difference post hoc test was performed to determine group differences at designated time points. Spearman’s correlations (Spearman’s ρ) were performed to test bivariate relationship between main outcome variables (i.e., BRS and HRV) and 1) aerobic fitness and 2) body composition. Pearson’s correlation test (r) was performed to test the bivariate correlation between R-R interval and heart rate. All statistical analyses were performed using SPSS Version 28 (IBM Corp, Armonk, NY). Significance was set at P < 0.05. All data are presented as means ± SD.

RESULTS

Participant Characteristics

Baseline demographics, arterial blood gases, and resting hemodynamics are presented in Table 1. There were no between-group differences in demographics, $\text{P}{\text{a}}_{{\text{O}}_2}$, arterial partial pressure of CO₂ ($\text{P}{\text{a}}_{{\text{CO}}_2}$), or aerobic fitness (V̇o_2max and % predicted V̇o_2max). Humans with high-affinity hemoglobin had a higher resting heart rate than the control group. The controls were also taller than humans with high-affinity hemoglobin, but there were no differences in body mass index (BMI; range: control: 20.7–31.2 kg/m² vs. high-affinity hemoglobin: 18.8–53.2 kg/m²). As anticipated, at baseline, humans with high-affinity hemoglobin had lower P₅₀, higher hemoglobin concentration, higher $\text{S}{\text{a}}_{{\text{O}}_2}$, and higher $\text{C}{\text{a}}_{{\text{O}}_2}$ (both while breathing room air).

Table 1.

Demographics, baseline arterial blood gases, and resting hemodynamics

	Control (n = 12)	High-Affinity Hemoglobin (n = 9)
Sex	6M/6F	3M/6F	P Value
Age, yr	41 ± 10	43 ± 13	0.676
Mass, kg	82 ± 15	91 ± 30	0.437
Height, cm	174 ± 11	164 ± 8	0.031
BMI, kg/m2	27 ± 3	34 ± 11	0.110
P50, mmHg	26 ± 1	16 ± 1	<0.001
$\text{S}{\text{a}}_{{\text{O}}_2}$, %	96 ± 1	97 ± 1	0.005
$\text{P}{\text{a}}_{{\text{O}}_2}$, mmHg	89 ± 6	87 ± 7	0.486
$\text{P}{\text{a}}_{{\text{CO}}_2}$, mmHg	40 ± 3	39 ± 2	0.442
[Hb], g/dL	13.7 ± 1	18.8 ± 2	0.005
$\text{C}{\text{a}}_{{\text{O}}_2}$, mL/dL	18 ± 1	25 ± 3	<0.001
SBP, mmHg	125 ± 14	129 ± 17	0.509
DBP, mmHg	73 ± 10	76 ± 13	0.603
MAP, mmHg	91 ± 10	94 ± 14	0.565
Heart rate, beats/min	60 ± 15	75 ± 12	0.029
V̇o2max, mL/kg/min	33 ± 7	28 ± 8	0.185
% Predicted V̇o2max, %	88 ± 15	72 ± 18	0.088

Data for V̇o_2max are from n = 11 (5 M/6F) controls and n = 7 (3 M/4F) individuals with high-affinity hemoglobin. P values are from a two-tailed independent sample t test. Bold signifies a P < 0.05. Values are presented as mean ± SD. Data collected from Dominelli et al. (3). BMI, body mass index; $\text{C}{\text{a}}_{{\text{O}}_2}$, arterial oxygen content; DBP, diastolic blood pressure; [Hb], hemoglobin concentration; MAP, mean arterial pressure; P₅₀, arterial oxygen pressure at which 50% of hemoglobin is saturated; $\text{P}{\text{a}}_{{\text{CO}}_2}$, arterial carbon dioxide pressure; $\text{P}{\text{a}}_{{\text{O}}_2}$, arterial oxygen pressure; $\text{S}{\text{a}}_{{\text{O}}_2}$, arterial oxygen saturation; SBP, systolic blood pressure; V̇o_2max, maximal oxygen consumption.

Cardiorespiratory Response to Hypoxia

Cardiorespiratory variables in response to isocapnic hypoxia are presented in Table 2. Diastolic, mean arterial blood pressures, tidal volume, breathing frequency, and minute ventilation all increased during the isocapnic hypoxia exposure. There were no between-group differences in the cardiorespiratory response to hypoxia except for breathing frequency, which was greater in humans with high-affinity hemoglobin compared with in controls. There was a significant interaction between group and time in diastolic blood pressure, such that humans with high-affinity hemoglobin had an increase in diastolic blood pressure that was not seen in controls in response to isocapnic hypoxia. For Fisher’s least significant difference test results, see Supplemental Table S1.

Table 2.

Cardiorespiratory variable changes during 20-min isocapnic hypoxia

Variables	Baseline	Isocapnic Hypoxia				Time Effect	Hb Type Effect	Interaction
		5 Min	10 Min	15 Min	20 Min
DBP, mmHg
Control	82 ± 10	84 ± 11	82 ± 10	82 ± 10	83 ± 11	P < 0.001	P = 0.371	P = 0.008
High-affinity hemoglobin	84 ± 10	90 ± 9	87 ± 8	87 ± 8	85 ± 8
MBP, mmHg
Control	108 ± 12	112 ± 15	111 ± 15	111 ± 15	111 ± 13	P < 0.001	P = 0.703	P = 0.461
High-affinity hemoglobin	109 ± 10	116 ± 12	114 ± 11	114 ± 10	112 ± 12
VT, L/breath
Control	0.72 ± 0.18	1.18 ± 0.35	1.22 ± 0.48	1.12 ± 0.42	1.10 ± 0.44	P < 0.001	P = 0.653	P = 0.719
High-affinity hemoglobin	0.69 ± 0.12	1.07 ± 0.21	1.15 ± 0.36	1.08 ± 0.32	1.04 ± 0.32
Fb, breath/min
Control	15 ± 3	15 ± 4	17 ± 4	17 ± 4	18 ± 4	P < 0.001	P = 0.050	P = 0.611
High-affinity hemoglobin	18 ± 3	19 ± 3	20 ± 4	21 ± 4	21 ± 4
V̇e, L/min
Control	10 ± 2	17 ± 6	20 ± 9	19 ± 9	20 ± 11	P < 0.001	P = 0.376	P = 0.758
High-affinity hemoglobin	12 ± 3	20 ± 5	22 ± 6	22 ± 6	21 ± 6

Linear mixed modeling analysis was performed for the variables. Bold signifies a P < 0.05. Values are presented as means ± SD. DBP, diastolic blood pressure; F_b, breathing frequency; Hb, hemoglobin; MBP, mean arterial pressure; V̇e, minute ventilation; V_T, tidal volume.

Changes in arterial blood gases in response to isocapnic hypoxia are presented in Table 3. In response to hypoxia, $\text{S}{\text{a}}_{{\text{O}}_2}$, $\text{P}{\text{a}}_{{\text{O}}_2}$, and $\text{C}{\text{a}}_{{\text{O}}_2}$ all decreased. The $\text{P}{\text{a}}_{{\text{O}}_2}$ response to hypoxia was similar between groups, whereas $\text{S}{\text{a}}_{{\text{O}}_2}$ and $\text{C}{\text{a}}_{{\text{O}}_2}$ were different by group, such that humans with high-affinity hemoglobin had higher $\text{S}{\text{a}}_{{\text{O}}_2}$ and $\text{C}{\text{a}}_{{\text{O}}_2}$ than controls. The decreases in both $\text{S}{\text{a}}_{{\text{O}}_2}$ and $\text{C}{\text{a}}_{{\text{O}}_2}$ were greater in controls compared with in humans with high-affinity hemoglobin. $\text{P}{\text{a}}_{{\text{CO}}_2}$ was unchanged from baseline in both groups. For Fisher’s least significant difference test results, see Supplemental Table S2.

Table 3.

Arterial blood gases variable change during 20-min isocapnic hypoxia

Variable	Baseline	Isocapnic Hypoxia			Time Effect	Hb Type Effect	Interaction
		5 Min	12 Min	20 Min
$\text{S}{\text{a}}_{{\text{O}}_2}$, %
Control	95 ± 1	83 ± 2	84 ± 5	82 ± 2	P < 0.001	P < 0.001	P < 0.001
High-affinity hemoglobin	96 ± 2	91 ± 2	91 ± 2	91 ± 3
$\text{P}{\text{a}}_{{\text{O}}_2}$, mmHg
Control	90 ± 7	49 ± 3	48 ± 3	48 ± 2	P < 0.001	P = 0.284	P = 0.451
High-affinity hemoglobin	86 ± 11	48 ± 1	48 ± 1	49 ± 3
$\text{P}{\text{a}}_{{\text{CO}}_2}$, mmHg
Control	40 ± 4	40 ± 3	40 ± 3	40 ± 4	P = 0.681	P = 0.274	P = 0.787
High-affinity hemoglobin	39 ± 2	39 ± 1	39 ± 1	39 ± 1
$\text{C}{\text{a}}_{{\text{O}}_2}$, mL/dL
Control	18 ± 2	15 ± 1	15 ± 1	15 ± 1	P < 0.001	P < 0.001	P < 0.001
High-affinity hemoglobin	26 ± 3	24 ± 3	24 ± 3	24 ± 3

Linear mixed modeling analysis was performed for the variables. Bold signifies a P < 0.05. Values are presented as means ± SD. $\text{C}{\text{a}}_{{\text{O}}_2}$, arterial oxygen content; Hb, hemoglobin. $\text{P}{\text{a}}_{{\text{CO}}_2}$, arterial carbon dioxide pressure; $\text{P}{\text{a}}_{{\text{O}}_2}$, arterial oxygen pressure; $\text{S}{\text{a}}_{{\text{O}}_2}$, arterial oxygen saturation.

Mean and individual data heart rate, N-N interval, and systolic blood pressure are presented in Fig. 1. With all data from participants included, heart rate, N-N interval, and systolic blood pressure changed in response to isocapnic hypoxia. All variables were similar between groups. The change in heart rate and N-N interval during isocapnic hypoxia was different between groups.

Figure 1.

Heart rate, N-N interval, and systolic blood pressure during normoxia and 20 min of isocapnic hypoxia in humans with high-affinity hemoglobin and controls. A: heart rate. B: N-N interval. C: systolic blood pressure. Twelve controls (6 F/6M) and nine humans with high-affinity hemoglobin (6 F/3M) were included in the analyses. Linear mixed modeling analysis was used for each panel. For Fisher’s least significant difference test results, see Supplemental Table S3. Data are presented as individual data points, the bars are group means, and the error bars represent standard deviation. Humans with high-affinity hemoglobin are represented by gray bars/triangles and controls are represented by black bars/circles. Females are represented by open symbols and males are represented by solid symbols.

Cardiac Baroreflex Sensitivity

Mean and individual data for cardiac baroreflex sensitivity are presented in Fig. 2. With all data included, cardiac baroreflex sensitivity was not different between humans with high-affinity hemoglobin and those with typical affinity hemoglobin. Cardiac baroreflex sensitivity was not reduced with isocapnic hypoxia. As highlighted in Fig. 2, outlier analysis identified an outlier with high cardiac baroreflex sensitivity among the high-affinity hemoglobin group (circled value). This outlier was excluded from baroreflex sensitivity analysis because 1) this human displayed a baroreflex gain greater than two standard deviations above the mean, and 2) this human had a different hemoglobin mutation (Hb-San Diego) and higher P₅₀ (16.5 mmHg) compared with the other eight participants with high-affinity hemoglobin (Hb-Malmo, P₅₀ ∼15 mmHg). With the outlier excluded, cardiac baroreflex sensitivity was lower in humans with high-affinity hemoglobin compared with in controls. Also, with the outlier removed, there was no reduction in cardiac baroreflex sensitivity with isocapnic hypoxia. The changes in cardiac baroreflex sensitivity with hypoxia did not differ between groups. Among humans with high-affinity hemoglobin, there was no relationship between BRS and BMI (outlier included: Spearman’s ρ = −0.301, P = 0.426; outlier excluded: Spearman’s ρ = −0.119, P = 0.779) or BRS and V̇o_2max (outlier included: Spearman’s ρ = 0.378, P = 0.404; outlier excluded: Spearman’s ρ = 0.289, P = 0.577). In humans with typical affinity hemoglobin, there was an association between BRS and BMI (Spearman’s ρ = −0.711, P = 0.019) and BRS and V̇o_2max (Spearman’s ρ = 0.808, P = 0.004).

Figure 2.

Cardiac baroreflex sensitivity during normoxia and 20 min of isocapnic hypoxia in humans with high-affinity hemoglobin and controls. Twelve controls (6 F/6M) and nine humans with high-affinity hemoglobin (6 F/3M) were included in the analyses. Linear mixed modeling analysis was used in the panel. For Fisher’s least significant difference test results, see Supplemental Table S3. For Q-Q plots depicting the residuals of the linear mixed modeling analysis, see Supplemental Fig. S1. Data are presented as individual data points, the bars are group means, and the error bars represent standard deviation. Humans with high-affinity hemoglobin are represented by gray bars/triangles and controls are represented by black bars/circles. Females are represented by open symbols and males are represented by solid symbols. There was no significant main effect of time (P = 0.080) or main effect of Hb type (P = 0.109). There was no interaction of Hb type-by-time (P = 0.503). We also ran the analysis with the human circled in the figure excluded from the analysis due to this data point being two standard deviations greater than the mean within the high-affinity hemoglobin group at baseline. This human also had a different Hb mutation than the rest of the high-affinity hemoglobin group (Hb-San Diego vs. Hb-Malmo). With the outlier excluded, 12 controls (6 F/6M) and 8 humans with high-affinity hemoglobin (5 F/3M) were included in the analysis. There was a significant main effect of Hb type (P = 0.021) but no significant main effect of time (P = 0.136). No interaction between the two groups was found (P = 0.502). F, female; Hb, hemoglobin; M, male.

R-R interval was chosen to analyze cardiac baroreflex sensitivity, as there is a significant linear relationship between R-R interval and heart rate found within this study. For the range of heart rates in this study (range 35–105 beats/min), there is a strong linear relationship between R-R interval and heart rate (Pearson’s r = −0.96), suggesting that the nonlinear relationship that exists at very low and very high heart rates did not strongly affect our findings.

Heart Rate Variability

Time domain.

Data for time domain heart rate variability indices are illustrated in Fig. 3. With data from all participants included, humans with high-affinity hemoglobin demonstrated lower standard deviation of N-N interval (SDNN) compared with controls, and SDNN decreased with isocapnic hypoxia. Among humans with high-affinity hemoglobin, there was no relationship between SDNN and BMI (Spearman’s ρ = 0.117, P = 0.775) or SDNN and V̇o_2max (Spearman’s ρ = 0.072, P = 0.887). Humans with typical affinity hemoglobin also exhibited no relationship between SDNN and BMI (Spearman’s ρ = −0.182, P = 0.573) or SDNN and V̇o_2max (Spearman’s ρ = 0.304, P = 0.361). RMSSD was not different between groups and was unaffected by isocapnic hypoxia.

Figure 3.

Time domain indices of heart rate variability during normoxia and 20 min of isocapnic hypoxia in humans with high-affinity hemoglobin and controls. A: standard deviation of N-N interval (SDNN). B: root mean square of successive difference (RMSSD). Twelve controls (6 F/6M) and nine humans with high-affinity hemoglobin (6 F/3M) were included in the analyses. Linear mixed modeling analysis was used for each panel. For Fisher’s least significant difference test results, see Supplemental Table S3. For Q-Q plots depicting the residuals of the linear mixed modeling analysis, see Supplemental Fig. S2. Data are presented as individual data points, the bars are group means, and the error bars represent standard deviation. Humans with high-affinity hemoglobin are represented by gray bars/triangles and controls are represented by black bars/circles. Females are represented by open symbols and males are represented by solid symbols.

Frequency domain.

Frequency domain indices of heart rate variability are presented in Fig. 4. With data from all participants included, frequency domain variables did not change in response to the isocapnic hypoxia. Low-frequency heart rate variability, in absolute units, was lower in humans with high-affinity hemoglobin compared with in controls. Among humans with high-affinity hemoglobin, there was no relationship between low-frequency heart rate variability, in absolute units, and BMI (Spearman’s ρ = 0.000, P > 0.999) or low-frequency heart rate variability, in absolute units, and V̇o_2max (Spearman’s ρ = 0.396, P = 0.377). Humans with typical affinity hemoglobin also exhibited no relationship between low-frequency heart rate variability, in absolute units, and BMI (Spearman’s ρ = 0.161, P = 0.619) or low-frequency heart rate variability, in absolute units, and V̇o_2max (Spearman’s ρ = 0.548, P = 0.084). All other variables were similar between groups.

Figure 4.

Frequency domain indices of heart rate variability during normoxia and 20 min of isocapnic hypoxia in humans with high-affinity hemoglobin and controls. Low-frequency power is depicted in absolute (A) and relative (C) units, and high-frequency power is depicted in absolute (B) and relative (D) units. Twelve controls (6 F/6M) and nine humans with high-affinity hemoglobin (6 F/3M) were included in the analyses. Linear mixed modeling analysis was used for each panel. For Fisher’s least significant difference test results, see Supplemental Table S3. For Q-Q plots depicting the residuals of the linear mixed modeling analysis, see Supplemental Fig. S3. Data are presented as individual data points, the bars are group means, and the error bars represent the standard deviation. Humans with high-affinity hemoglobin are represented by gray bars/triangles and controls are represented by black bars/circles. Females are represented by open symbols and males are represented by solid symbols.

Potential outlier analysis.

Exclusion of the human with Hb-San Diego from the Hb variant group did not change the findings for any of the heart rate variability metrics. There were several outliers in the heart rate variability data. Outlier exclusion did not affect any other statistical outcomes.

DISCUSSION

This novel, hypothesis-generating analysis provides evidence of lower cardiac autonomic function in humans with high-affinity hemoglobin compared with age-matched controls. Our primary findings are twofold. First, when compared with controls, humans with high-affinity hemoglobin demonstrated attenuated cardiac baroreflex sensitivity on removal of outlier data from one participant with high-affinity hemoglobin. Second, when compared with controls, some, but not all, indices of heart rate variability were lower in humans with high-affinity hemoglobin. We interpret these data to suggest that high-affinity hemoglobin mutations may be associated with altered cardiac autonomic function but that reflex autonomic responses that exist to maintain cardiac output and blood pressure homeostasis during hypoxia appear intact.

Cardiac Baroreflex Sensitivity

Our data support the hypothesis that humans with high-affinity hemoglobin demonstrate lower cardiac baroreflex sensitivity compared with controls. These data are in agreement with earlier evidence indicating that humans with high-affinity hemoglobin have blunted heart rate responses to both acute (3, 9) and chronic (8) hypoxic exposure, which may have been attributed to impaired cardiac baroreflex mechanisms. We offer the caveat that we found lower cardiac baroreflex sensitivity in humans with high-affinity hemoglobin only when we excluded an outlier within the high-affinity hemoglobin group. We chose to exclude this outlier for two reasons. First, they were deemed a statistical outlier due to their cardiac baroreflex sensitivity in the resting condition existing greater than two standard deviations above the mean for this group and condition (27). Second, they were biologically different from the rest of the high-affinity hemoglobin cohort, as they had a different hemoglobin mutation (Hb-San Diego vs. Hb Malmö). As such, we conclude that baroreflex sensitivity may be blunted in this population. Lesser cardiac baroreflex sensitivity in humans with high-affinity hemoglobin may be related to some feature associated with high-affinity hemoglobin or compensatory polycythemia.

The mechanisms contributing to blunted cardiac baroreflex sensitivity among the high-affinity hemoglobin group are unclear, but one explanation may include elevated arterial stiffness secondary to compensatory polycythemia. Studies have found that humans with high-affinity hemoglobin exhibit compensatory polycythemia (28, 29), which is likely due to the increased hemoglobin-oxygen affinity that causes reduced oxygen offloading to the tissues (30). This is especially important at the kidney, as the low tissue oxygen partial pressure activates the erythropoietin pathway to increase erythrocyte production (4, 31). Previously, Keyl et al. (15) observed a lower spontaneous cardiac baroreflex sensitivity in a cohort of Andean natives (∼4,300 m elevation) with polycythemia and chronic mountain sickness compared with Andean natives solely with polycythemia. However, when using the modified Oxford method to assess baroreflex control of heart rate during large changes in blood pressure, Simpson et al. (32) showed that Andean natives with polycythemia and chronic mountain sickness had greater cardiac baroreflex sensitivity than Andean natives solely with polycythemia. We speculate that divergent results between these two studies may stem from the different method used to assess cardiac baroreflex sensitivity. Similar to the study by Keyl et al., our methods of analyzing spontaneous cardiac baroreflex sensitivity show a decreased sensitivity in humans with high-affinity hemoglobin and compensatory polycythemia.

The effect of blood viscosity on arterial stiffness remains poorly understood, with some studies showing no significant effect in normal humans (33) and others showing some correlation in disease state (34, 35). Earlier work showed that patients with polycythemia demonstrate augmented pulse-wave velocity, indicating greater arterial stiffness (16). Elevated stiffness of baroreceptor-harboring vessels (i.e., carotid sinuses and aortic arch) would reduce distension and afferent baroceptor discharge in responses to blood pressure fluctuations, reducing the sensitivity of the arterial baroreflex. Increases in pulse pressure can provide indirect evidence of arterial stiffness. However, the present study did not evaluate arterial stiffness, and to our knowledge, the vascular health of humans with high-affinity hemoglobin mutations and compensatory polycythemia remains to be studied. Thus, given the scarcity of data, elevated arterial stiffness related to compensatory polycythemia represents the most direct explanation for the present finding of lower baroreflex sensitivity among humans with high-affinity hemoglobin mutations. Further studies must be performed to isolate the contribution of high-affinity hemoglobin and polycythemia on cardiac baroreflex function either via hemodilution of rare patients with hemoglobin mutations or pharmacologically shifting the oxygen-hemoglobin dissociation curve acutely.

Heart Rate Variability

The present study evaluated several indices of heart rate variability, including time and frequency domain measures. Although not all indices revealed between-group differences, SDNN and low-frequency heart rate variability, expressed in absolute units, were lower in the high-affinity hemoglobin group compared with in controls shown by the main effect of Hb type. These findings were not affected by the inclusion or exclusion of the outlier present in the cardiac baroreflex sensitivity analysis. Evidence of lower heart rate variability among the high-affinity hemoglobin group is directionally consistent with our finding regarding cardiac baroreflex sensitivity, discussed earlier. Thus, we conclude that modest altered cardiac autonomic function exists in humans with high-affinity hemoglobin. Because heart rate variability represents an indirect measure of cardiac autonomic function, we are unable to determine whether altered autonomic function stems from elevated cardiac sympathetic discharge, reduced parasympathetic activity toward the heart, or some feature at the postganglionic neuron-cardiac junctions (e.g., muscarinic receptor density). Heart rate variability also did not seem to be impacted by the small differences in respiratory volume or frequency between groups. Future pharmacological and microneurographic studies should investigate this question.

Consistent with the abovementioned discussion related to cardiac baroreflex function, it is not possible to disentangle the effects of high-affinity hemoglobin versus compensatory polycythemia on heart rate variability in this study. Lower heart rate variability among individuals with high-affinity hemoglobin may stem from compensatory polycythemia. This is consistent with the observation that Andean highlanders with elevated hemoglobin concentrations demonstrated elevated sympathetic outflow compared with normal healthy humans (14). However, these data are difficult to interpret, as the Andean highlanders were only studied at a high altitude, which also elevates sympathetic outflow. The finding of reduced heart rate variability among the high-affinity hemoglobin cohort aligns with the chronic reduction in heart rate variability and increase in sympathetic outflow that occur when healthy lowlanders ascend to high-altitude hypoxic environments (14, 36). The only insight to the impact of high-affinity hemoglobin and compensatory polycythemia on heart rate variability arises from the comparative biology literature. During high-altitude exposure, cinnamon teal ducks demonstrated a leftward shift in the oxyhemoglobin dissociation curve, an 8-mmHg reduction in P_50, and a 10% increase in hematocrit (37). These changes were associated with reductions in the heart rate variability at a moderate or high altitude compared with a low altitude (37). Again, it is difficult to ascertain whether these cardiac autonomic changes are due to the reduction in P₅₀ or increased hematocrit or the arterial chemoreflex-mediated cardiac autonomic adjustments to low arterial oxygen pressure at a high altitude. Given that the abovementioned studies feature an association between elevated hemoglobin concentration and reductions in heart rate variability, we suggest that the polycythemia represents the simplest explanation for the present observations.

Potential Role of Aerobic Fitness and Body Composition

Aerobic fitness and body composition may have contributed to lower autonomic function among individuals with high-affinity hemoglobin in the present study. This study recruited control participants to match aerobic fitness and body composition to the high-affinity hemoglobin group (Table 1). However, the groups were no longer matched for these variables on removing the outlier in the cardiac baroreflex sensitivity analysis. When the outlier was removed, the high-affinity hemoglobin group demonstrated lower percent predicted V̇o_2max [high-affinity hemoglobin: 67 ± 12%; controls: 88 ± 15% (P = 0.011)] and a trend toward greater body mass index [high-affinity hemoglobin: 35 ± 11 kg/m²; controls: 27 ± 3 kg/m² (P = 0.069)]. Earlier studies suggested that cardiac baroreflex sensitivity is positively associated with aerobic fitness and negatively associated with body composition (38–41). In the present study, these associations were observed among the control group but not the high-affinity group. These findings suggest that although we cannot eliminate the contribution of fitness or BMI to the present findings, it is likely that some other factor, such as high-affinity hemoglobinopathy and/or compensatory polycythemia, explained a larger proportion of variance in cardiac baroreflex sensitivity in this study. Thus, we cannot eliminate a contribution of aerobic fitness and body composition to our findings related to cardiac baroreflex sensitivity. Regarding heart rate variability, earlier studies have suggested that there are weak associations for both aerobic fitness and body composition (42, 43). In the present study, no associations were observed among either group. Nonetheless, some, but not all, indices of heart rate variability were lower in the high-affinity hemoglobin group regardless of outlier inclusion/exclusion, suggesting that a yet unknown mechanism related to high-affinity hemoglobin and/or polycythemia likely contributes to lower cardiac autonomic function that does not relate to aerobic fitness or body composition.

Limitations

This study has several limitations that should be discussed. First, we cannot differentiate between the effects of compensatory polycythemia and the high-affinity hemoglobin mutation per se on cardiac autonomic function. As discussed, polycythemia alone seems to produce the same effects on cardiac autonomic function that were observed in humans with high-affinity hemoglobin in this study. Thus, the genetic mutation causing high-affinity hemoglobin per se may have no impact on autonomic function beyond polycythemia. We recognize that a second “control” group of patient with typical binding affinity and polycythemia may be most appropriate to isolate the impacts of each of these factors. Second, the number of humans with high-affinity hemoglobin who participated in the study is low. The population of humans with this genetic mutation is very rare, comprising less than 3% of all cases of secondary polycythemia (44), making it difficult to achieve a large number of participants. Thus, the rarity of the mutation precluded our ability to rigorously case-match humans with high-affinity hemoglobin to controls. Third, although statistically case-matched for body composition, further investigations should improve this case-matching to help better elucidate the potential impact of body composition on cardiac baroreflex sensitivity and heart rate variability. However, the lack of relationship between BMI or V̇o_2max and BRS within the high-affinity group supports the concept that some other factor, such as high-affinity hemoglobinopathy and/or compensatory polycythemia, explained a larger proportion of variance in BRS in this study. Fourth, the present study indirectly assessed cardiac autonomic regulation using ECG. To gain greater insight to the impact of high-affinity hemoglobin mutations on the physiological underpinnings of blood pressure homeostasis, future studies should consider microneurographic measurements of muscle sympathetic nerve activity or pharmacological techniques that expose the discrete contributions of the sympathetic and parasympathetic nervous systems. Fifth, this study focused solely on isocapnic hypoxia to impose physiological stress. The reflex-mediated responses to other forms of physiological stress (e.g., baroreflex unloading, physical exercise, or mental stress) among humans with high-affinity hemoglobin mutations remain unknown. Sixth, although some indices of heart rate variability were significant, this finding was not attributed to all indices analyzed. Finally, despite widespread sex differences in neurocardiovascular regulation in humans, this study did not have adequate statistical power to examine the impact of biological sex on the relationship between hemoglobin-oxygen affinity, polycythemia, and cardiac autonomic function (45, 46).

Conclusions

We evaluated cardiac autonomic function in humans with genetic mutations resulting in high-affinity hemoglobin. We found that humans with high-affinity hemoglobin have lower cardiac baroreflex sensitivity and lower heart rate variability compared with humans with typical affinity hemoglobin. Overall, this hypothesis-generating work suggests that humans with high-affinity hemoglobin may demonstrate lower cardiac autonomic function due to elevated sympathetic outflow and/or reduced parasympathetic activity. Future work should examine whether these neurocardiovascular maladaptations increase cardiovascular risk in this cohort and should attempt to understand the discrete contributions of compensatory polycythemia and the high-affinity hemoglobin mutation.

Perspectives and Significance

Cardiac baroreflex sensitivity and heart rate variability are noninvasive measures of cardiac autonomic function. In this study, we observed that humans with high-affinity hemoglobin demonstrate reduced cardiac baroreflex sensitivity and heart rate variability during normoxia and isocapnic hypoxia compared with humans with typical affinity hemoglobin. High-affinity hemoglobin mutations have been proposed to be advantageous for individuals living at high altitude. Our data suggest that there may be a negative consequence of having a high-affinity hemoglobin mutation, as reduced cardiac baroreflex sensitivity and heart rate variability have weaker autonomic modulation of heart rate. Future studies should aim to determine whether this patient population exhibits augmented sympathetic outflow using microneurography. Additional potential investigations include hemodiluting this patient group to determine whether the findings described herein are a function of the hemoglobin mutation or the compensatory polycythemia.

DATA AVAILABILITY

Data will be made available on reasonable request.

SUPPLEMENTAL DATA

Supplemental Figs. S1–S3: https://doi.org/10.6084/m9.figshare.22177151.

Supplemental Tables S1–S3: https://doi.org/10.6084/m9.figshare.22159169.

GRANTS

S.A. Klassen was supported by a Natural Sciences and Engineering Research Council of Canada (NSERC) postdoctoral fellowship award PDF-532926-2019 and an NSERC Discovery Grant award RGPIN-2022–05293. C.C. Wiggins was supported by National Institute of Diabetes and Digestive and Kidney (NIDDK) Grant T32-DK-007352. J.W. Senefeld was supported by National Heart, Lung, and Blood Institute (NHLBI) Grant F32-HL-154320. The work was funded by NHLBI Grant R-35-HL-139854 to M.J. Joyner. S.E. Baker was supported by NHLBI Grant K01-HL-148144.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

W.W.P., S.A.K., and S.E.B. conceived and designed research; P.B.D., C.C.W., and S.E.B. performed experiments; W.W.P., S.A.K., and S.E.B. analyzed data; W.W.P., S.A.K., P.B.D., C.C.W., J.W.S., T.K.R., M.J.J., and S.E.B. interpreted results of experiments; W.W.P. and S.A.K. prepared figures; W.W.P., S.A.K., and S.E.B. drafted manuscript; W.W.P., S.A.K., P.B.D., C.C.W., J.W.S., T.K.R., M.J.J., and S.E.B. edited and revised manuscript; W.W.P., S.A.K., P.B.D., C.C.W., J.W.S., T.K.R., M.J.J., and S.E.B. approved final version of manuscript.