Effects of acute hyperoxia on autonomic function and coronary tone in patients with peripheral artery disease

Abstract

Numerous studies have shown that oxidative stress plays an important role in peripheral artery disease (PAD). Prior reports suggested autonomic dysfunction in PAD. We hypothesized that responses of the autonomic nervous system and coronary tone would be impaired in patients with PAD during exposure to acute hyperoxia, an oxidative stressor. In 20 patients with PAD and 16 healthy, sex- and age-matched controls, beat-by-beat heart rate (HR, from ECG) and blood pressure (BP, with Finometer) were recorded for 10 min during room air breathing and 5 min of hyperoxia. Cardiovagal baroreflex sensitivity and HR variability (HRV) were evaluated as measures of autonomic function. Transthoracic coronary echocardiography was used to assess peak coronary blood flow velocity (CBV) in the left anterior descending coronary artery. Cardiovagal baroreflex sensitivity at rest was lower in PAD than in healthy controls. Hyperoxia raised BP solely in the patients with PAD, with no change observed in healthy controls. Hyperoxia induced an increase in cardiac parasympathetic activity assessed by the high-frequency component of HRV in healthy controls but not in PAD. Indices of parasympathetic activity were lower in PAD than in healthy controls throughout the trial as well as during hyperoxia. Hyperoxia induced coronary vasoconstriction in both groups, while the coronary perfusion time fraction was lower in PAD than in healthy controls. These results suggest that the response in parasympathetic activity to hyperoxia (i.e., oxidative stress) is blunted and the coronary perfusion time is shorter in patients with PAD.

NEW & NOTEWORTHY Patients with peripheral artery disease (PAD) showed consistently lower parasympathetic activity and blunted cardiovagal baroreflex sensitivity compared with healthy individuals. Notably, hyperoxia, which normally boosts parasympathetic activity in healthy individuals, failed to induce this response in patients with PAD. These data suggest altered autonomic responses during hyperoxia in PAD.

INTRODUCTION

Peripheral artery disease (PAD) refers to obstructive atherosclerosis that mostly affects arteries of the lower extremities and is clinically characterized by claudication and critical limb ischemia. In PAD the risk of myocardial infarction and stroke is increased twofold (1). Oxidative stress has been suggested as a contributor to the conditions of PAD (2). Hyperoxia promotes the generation of reactive oxygen species (ROS) (3). Short-term administration of 100% oxygen is an effective method to examine oxidative stress physiologically (4). In healthy subjects, hyperoxia lowers heart rate (HR) and cardiac output while systemic vascular resistance increases (5). In patients with hypertension or sleep apnea, breathing oxygen results in lower HR and blood pressure (BP) due to the underlying enhanced chemoreceptor activity (6). These hemodynamic changes are related to alterations in autonomic nervous system activity initiated by chemoreceptor/baroreceptor stimulation. Hyperoxia has long been recognized to increase vagal tone and attenuate peripheral chemoreceptor activity in healthy subjects (7). No previous study had addressed the impact of oxidative stress on autonomic function in PAD. Given the chronic/repeated exposure of PAD to oxidative stress (8, 9), alongside indications from prior research suggesting alternation in resting autonomic function (10, 11), we hypothesized that autonomic responses to acute oxidative stress (e.g., with hyperoxia) would be altered in PAD.

Although patients with PAD have a high risk of myocardial infarction (1), the effects of acute oxidative stress on the coronary vasoactivity are unknown. Inhalation of 100% oxygen leads to decreases in coronary blood flow (CBF) and coronary blood velocity (CBV) and increases coronary vascular resistance in healthy subjects (12). However, it is unclear whether the response of coronary vascular tone to acute hyperoxia is altered in PAD. Therefore, the purpose of the present study was to examine the physiological impact of oxidative stress induced by breathing oxygen on autonomic function and coronary vascular tone in patients with PAD.

METHODS

Subjects

The data from 20 patients with PAD and 16 healthy control subjects were included in this study (Table 1). All subjects underwent thorough medical screening to assess eligibility. All patients with PAD were classified as Fontaine stages I or II and had an ankle-brachial pressure index (ABI) < 0.9. Healthy control subjects were selected so that each subject with PAD had a corresponding match based on sex, age, and body mass index (BMI). The study protocol was approved by the Institutional Review Board of the Milton S. Hershey Medical Center (Study 5330) and conformed with the Declaration of Helsinki. Written informed consent was obtained from all subjects.

Table 1.

Group characteristics and resting hemodynamics

	Control	PAD	P Value
n	16	20
Age, yr	65 ± 6	66 ± 7	0.46
Male/female	10/6	14/6	0.63
Height, cm	173 ± 9	172 ± 8	0.71
Weight, kg	78 ± 16	85 ± 18	0.30
Body mass index, kg/m2	26 ± 3	28 ± 5	0.09
ABI exercising leg	1.12 ± 0.20	0.57 ± 0.10	<0.001
ABI nonexercising leg	1.09 ± 0.12	0.76 ± 0.19	<0.001
Heart rate, beats/min	63 ± 5	68 ± 8	0.27
Systolic BP, mmHg	118 ± 10	133 ± 12	<0.001
Diastolic BP, mmHg	75 ± 6	74 ± 7	0.85
MAP, mmHg	88 ± 7	94 ± 8	0.058
Medications
Anticoagulants		2
Clopidogrel		8
Aspirin		13
Other antiplatelet agents		4
Statins		13
Antihypertensives		10
β-Blockers		4
Other drugs for PAD		3

Values are means ± SD; n, number of subjects. ABI, ankle-brachial index; BP, blood pressure; MAP, mean arterial pressure; PAD, peripheral artery disease.

Measurements and Protocols

Patients with PAD were allowed to adhere to their normal medication (Table 1). The subjects were tested in a supine position. ECG and beat-by-beat blood pressure (BP) (Finometer, Finapress Medical system) were recorded simultaneously throughout the trial. Resting beat-by-beat BP was verified by brachial cuff pressure. After an acclimation period, resting baseline data were collected for 10 min when the subjects breathed room air. Thereafter, the subjects breathed 100% oxygen through a facemask for 5 min.

Transthoracic echocardiography was performed using a GE Vivid 7 echocardiography system. As described in prior reports from our laboratory (13), peak diastolic CBV was obtained in the left anterior descending coronary artery. These measurements were performed sporadically during the 10-min baseline and 5-min hyperoxia periods.

Since the recovery time for hyperoxia-induced changes in autonomic function remains uncertain, and the site of the CBV measurement might vary across different visits, the order of the conditions was not randomized in this study.

Data Analysis

Beat-to-beat HR, cardiac RR interval (RRI), and systolic BP (SBP) were obtained from ECG and the Finometer recordings. The slope of the relationship between SBP and RRI was used as an index of cardiovagal baroreflex sensitivity (CBRS) (14) with the sequence technique using CardioSeries v2.7 (15). Briefly, the slopes of the sequences of three or more consecutive beats where SBP and RRI changed in the same direction (up or down) with a R² > 0.80 of a linear regression were calculated (15). Because the data from the up sequences, the down sequences, and all BRS sequences yielded comparable statistical results, only the CBRS from all BRS sequences are reported.

The HRV indices were analyzed over each of 5-min data segments during the 15-min recordings using LabChart (v.8, AD Instruments, Castle Hill, Australia). The averaged values from the two 5-min data segments of baseline were used as baseline values. The reported HRV indices include root-mean-square of differences of adjacent normal-to-normal R-R intervals (RMSSD) and high frequency (HF, 0.15–0.4 Hz) power, which reflect parasympathetic activity (16).

When the recording quality in a subject was not sufficient for HRV and/or BRS analysis, that subject was excluded from this report. Three patients with PAD and one healthy subject were excluded due to arrhythmias. Thus, HRV and CBRS data were derived from 17 patients with PAD and 15 healthy controls.

The rate pressure product (RPP, the product of SBP and HR) was calculated as an index of myocardial oxygen demand. Peak CBV was used as an index of myocardial oxygen supply. Coronary vascular conductance (CVC, i.e., peak CBV/DBP) and coronary perfusion time fraction (CPTF, i.e., coronary perfusion time/R-R interval) were calculated.

Statistical Analysis

All statistical analyses were performed using SPSS (v.27, IBM). Independent t tests were used to compare differences between the groups. For each dependent variable, two-way mixed measures ANOVA (RMANOVA) was conducted by two time points for the stages as within-subject factor (baseline, hyperoxia) and as between-subject factor between groups (healthy control subjects and patients with PAD). The Bonferroni post hoc test was applied when appropriate. Paired t tests were used to compare CBV and CVC between baseline and hyperoxia within each group. All values were presented as means ± SD and values of P < 0.05 (2-tail) were considered statistically significant.

RESULTS

Baseline SBP in PAD was higher than that in controls (Table 1). The HR and mean arterial pressure (MAP) during baseline and hyperoxia in the two groups with RMANOVA analysis are shown in Fig. 1. The main effect of the stage on HR was significant (P < 0.001). MAP in healthy control subjects did not significantly change along the stages, whereas MAP in the patients with PAD significantly raised during hyperoxia (P < 0.001, post hoc). The change in MAP by hyperoxia in PAD was significantly greater than that in control subjects (Δ4.4 ± 4.1 vs. −Δ0.8 ± 2.3 mmHg, unpaired t test, P < 0.001).

Figure 1.

Blood pressure and heart rate (HR) responses to hyperoxia in healthy control subjects and patients with peripheral artery disease (PAD). MAP, mean arterial pressure. n, number of patients. Symbols represent individual data.

CBRS, RMSSD, and HF power during baseline and hyperoxia in the two groups with RMANOVA analysis are shown in Fig. 2. The effect of group on CBRS was significant (P = 0.002). Resting CBRS was significantly lower than in healthy controls (P = 0.001, post hoc). CBRS did not change along the stages in either group. Resting RMSSD in PAD was significantly lower than in healthy controls (P = 0.032, post hoc). RMSSD and HF power in healthy subjects significantly increased along the stages (both P < 0.005, post hoc). Either RMSSD or HF power did not significantly change along the stages.

Figure 2.

Baroreflex function and parasympathetic activity indices during hyperoxia in healthy control subjects and patients with peripheral artery disease (PAD). CBRS, cardiovagal baroreflex sensitivity; HF, high-frequency power; RMSDD, root-mean-square of differences of adjacent normal-to-normal R-R intervals. n, number of patients. Symbols represent individual data.

The RPP in PAD (Table 2) was significantly higher than in healthy controls during baseline (P = 0.005, post hoc) and hyperoxia (P = 0.001, post hoc). The CPTF in PAD was significantly lower than in the healthy controls during baseline (P = 0.035, post hoc) and hyperoxia (P = 0.001, post hoc). CBV and CVC in each group significantly decreased during hyperoxia (all P < 0.005, paired t test; Table 2).

Table 2.

Rate pressure product and coronary circulatory parameters responses to hyperoxia in healthy control subjects and patients with PAD

				P Values (F Values)
	n	Baseline	Hyperoxia	Group	Stage	Int
RPP, mmHg·(beats/min)
Healthy	16	6,765 ± 867	6,560 ± 711	(12.4)	(0.04)	(2.4)
PAD	20	8,219 ± 1,546	8,378 ± 1,590	0.002	0.844	0.130
P (group)		0.005	0.001
CPTF
Healthy	16	0.55 ± 0.03	0.57 ± 0.03	(9.3)	(2.2)	(1.2)
PAD	20	0.52 ± 0.06	0.52 ± 0.05	0.004	0.146	0.288
P (group)		0.035	0.001
CBV, cm/s
Healthy	16	20.19 ± 3.50	17.09 ± 4.45		<0.001
PAD	20	25.11 ± 10.14	22.16 ± 8.92		<0.001
CVC, cm/s/mmHg
Healthy	16	0.22 ± 0.03	0.19 ± 0.05		<0.001
PAD	20	0.30 ± 0.11	0.26 ± 0.09		=0.001

Values are means ± SD; n, number of subjects. CBV, coronary blood flow velocity; CPTF, coronary perfusion time fraction; CVC, coronary conductance; PAD, peripheral artery disease; RPP, rate pressure product; Stage, the effect of “stage” (baseline and hyperoxia); Group, the effect of “group” (PAD or healthy); Int, interaction between “group” and “stage.”

P (group), P values between groups in the same “stage.”

DISCUSSION

The novel finding of this study is that the autonomic neural responses to hyperoxia were altered in patients with PAD. Under resting conditions, the parasympathetic nervous activity indices and CBRS were significantly lower in patients with PAD than in healthy controls. Hyperoxia induced significant increases in the indices for parasympathetic activity in healthy individuals, but this was not seen in patients with PAD. Hyperoxia induced a significant increase in BP only in PAD but not in healthy individuals. Collectively, these observations suggest that resting autonomic function, as well as autonomic responses to acute oxidative stress, are impaired in PAD. PAD was associated with impaired coronary perfusion, especially during hyperoxia based on higher RPP and shorter CPTF in PAD compared with controls. To the best of our knowledge, this is the first investigation that examines autonomic neural responses to hyperoxia in patients with PAD.

Resting Autonomic Activity and Baroreflex Sensitivity in PAD

Our data showed that resting RMSSD was lower in PAD than those in healthy individuals. Moreover, the main effects of group in RMANOVA analysis show that RMSSD and HF power were lower in PAD than those in healthy controls. These results suggest that cardiac parasympathetic function is impaired in PAD, consistent with some prior reports (10, 11). In contrast, other studies reported that resting HRV in PAD was higher than the patients only with cardiovascular disease (17) or not different from healthy controls (18). Differences in these observations could be due to differences in patient population [e.g., older males only (17)] and/or conditions during the HRV assessment [e.g., unknown posture (17), etc.]. In our study, the data were obtained from both males and females in a quiet environment after an acclimation period.

Our data showed that CBRS in patients with PAD was significantly lower than those in healthy controls under resting conditions. This finding suggests that CBRS is impaired in PAD and agrees with prior studies (19, 20). Cardiac parasympathetic activity is influenced by CBRS (21). The lower resting cardiac parasympathetic activity reflected by the indices of HRV, and the lower resting CBRS in PAD in our study support this concept.

Hemodynamic and Autonomic Responses to Hyperoxia in PAD

In the present study, HR in healthy subjects significantly decreased during hyperoxia whereas MAP did not change. This is consistent with a prior report (22). Breathing oxygen causes vasoconstriction by altered endothelial function (3). When baroreflex function is preserved, baroreflex could reduce HR via increased parasympathetic activity and normalize BP in healthy individuals (22, 23). Therefore, in healthy individuals, increased parasympathetic activity, decreased HR, and peripheral vasoconstriction contribute to the unchanged BP during hyperoxia.

Notably, the presented data showed an increase in BP during hyperoxia in patients with PAD, in contrast to the lack of effect in healthy individuals. This differed from observations in patients with hypertension or sleep apnea, in which breathing oxygen lowers HR and BP, presumably due to underlying enhanced chemoreceptor activity (6). In the present study, hyperoxia did not induce a significant increase in parasympathetic nerve activity (indicated with RMSSD and HF power) in PAD, unlike healthy controls. These data suggest that the autonomic responses to hyperoxia are blunted in PAD. We speculate this to be one of the factors that contribute to the BP increase seen during hyperoxia. In addition, the CBRS in patients with PAD was lower than that in healthy controls. These could indicate the impaired capacity for BP control in PAD, which could contribute to the BP increase during hyperoxia in PAD. Importantly, these effects may contribute to the higher incidence of cardiac and cerebral vascular events in PAD (1).

Effects of Hyperoxia on Coronary Circulation

Because the precise anatomic location of the coronary vessels used for the measurement could vary among subjects and the resolution of the transthoracic echocardiography measurements was not high enough, we did not calculate absolute CBF. For these technical reasons, there was less meaning for the comparison of the CBV and CVC between the groups in this study. However, the changes in CVC within subjects can serve as an index of the vasomotion, since the location of the measurement was kept within each visit. Hyperoxia induced significant decreases in CBV and CVC within each of the groups, which indicates a vasoconstriction. On the other hand, CPTF, a measure of the fraction of time during the cardiac cycle, is not affected by the anatomic location of the measurement, and thus this index can be used for the comparison between the groups. Diastolic time fraction and coronary perfusion are inversely related (24). In the current study, the significantly shorter CPTF in PAD both at baseline and during hyperoxia suggests a shorter duration of myocardial blood flow during each cardiac cycle in PAD. Furthermore, we found an elevated myocardial oxygen demand (indicated with RPP) in PAD throughout the study. Collectively, these findings indicate that PAD is associated with impaired coronary perfusion, which may contribute to the high risk of adverse cardiovascular events in this population (25).

Limitations

The patients in this study continued their daily medications including β-blockers and angiotensin-converting enzyme inhibitors. These drugs may potentially affect HRV and CBRS. However, considering only a portion of the patients were on these medications, the observed results are not likely caused by the medications. Thus, we speculate that PAD was the main factor responsible for the observations reported here. In addition, we acknowledge that the lack of randomization of the conditions was a limitation. We did not consider randomization because the recovery period for autonomic activity after hyperoxia exposure was unknown and might therefore have complicated the interpretation. We also acknowledge that CBV was just an index for the coronary flow since it was not the mean value of the whole cardiac cycle.

In summary, our study revealed that resting autonomic activity is altered and cardiovagal baroreflex function is impaired in patients with PAD. Moreover, unlike in healthy individuals, hyperoxia does not induce an increase in parasympathetic nerve activity in PAD. Importantly, hyperoxia induces an increase in BP in PAD but not in healthy individuals. We speculate that these changes in autonomic responses to hyperoxia presumably mediated by oxidative stress, in hemodynamic variables, and in coronary perfusion, may contribute to the high risk of myocardial ischemia observed in PAD.

DATA AVAILABILITY

The data sets generated and analyzed during the current study are available from the corresponding author upon reasonable request.

GRANTS

This project was funded by National Institutes of Health Grants R01 HL141198 (to J. Cui), R01 HL164571 (to J. Cui), P01 HL134609 (to L. I. Sinoway), and UL1 TR002014 (to L. I. Sinoway).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

U.A.L., L.I.S., and J.C. conceived and designed research; U.A.L., Z.G., F.A., D.J.-K.K., J.C.L., C.B., A.E.C., and J.C. performed experiments; M.H., Z.G., D.J.-K.K., J.C.L., C.B., A.E.C., and J.C. analyzed data; M.H., U.A.L., Z.G., F.A., L.I.S., and J.C. interpreted results of experiments; M.H. prepared figures; M.H., U.A.L., and J.C. drafted manuscript; M.H., U.A.L., Z.G., F.A., D.J.-K.K., J.C.L., C.B., A.E.C., L.I.S., and J.C. edited and revised manuscript; M.H., U.A.L., Z.G., F.A., D.J.-K.K., J.C.L., C.B., A.E.C., L.I.S., and J.C. approved final version of manuscript.