Arterial Stiffening with Exercise in Patients with Heart Failure and Preserved Ejection Fraction

Abstract

BACKGROUND

Aortic stiffening and reduced nitric oxide (NO) availability may contribute to the pathophysiology of heart failure with preserved ejection fraction (HFpEF).

OBJECTIVES

We assessed indices of arterial stiffness at rest and during exercise in subjects with HFpEF and hypertensive controls, to examine their relationship to cardiac hemodynamics and determine whether exertional arterial stiffening can be mitigated by inorganic nitrite.

METHODS

Twenty-two hypertensive controls and 98 HFpEF subjects underwent hemodynamic exercise testing with simultaneous expired gas analysis. Invasively measured radial artery pressure waveforms were converted to central aortic waveforms by transfer function to assess integrated measures of pulsatile aortic load, including arterial compliance, resistance, elastance, and wave reflection.

RESULTS

Arterial load and wave reflections were similar in HFpEF and controls at rest. During submaximal exercise, HFpEF subjects displayed reduced total arterial compliance and higher effective arterial elastance despite similar mean arterial pressures to controls. This was directly correlated with higher ventricular filling pressures and depressed cardiac output reserve (both p <0.0001). With peak exercise, increased wave reflections impaired compliance, resistance and elastance were observed. A subset of HFpEF subjects (n = 52) received sodium nitrite or placebo in a 1:1 double blind, randomized fashion. Compared to placebo, nitrite decreased aortic wave reflections at rest and improved arterial compliance, elastance, and improved central hemodynamics during exercise.

CONCLUSION

Abnormal pulsatile aortic loading during exercise occurs in HFpEF independent of hypertension, and is correlated with classical hemodynamic derangements that develop with stress. Inorganic nitrite mitigates arterial stiffening with exercise and improves hemodynamics, indicating that arterial stiffening with exercise is at least partially reversible. Further study is required to test effects of agents that target the NO pathway in reducing arterial stiffness in HFpEF.

INTRODUCTION

Human senescence is characterized by an increase in aortic stiffness (1). This causes systolic hypertension via reductions in arterial compliance and increases in systolic wave reflections (1,2). Aging, along with hypertension and obesity, are the strongest risk factors for the development of heart failure with preserved ejection fraction (HFpEF) (3). These comorbidities are believed to promote the development of HFpEF by causing deficiency in nitric oxide-guanosine monophosphate (NO-cGMP) signaling, which may alter ventricular mechanical properties that result in classical hemodynamic findings of heart failure (HF) (3–5).

In addition to ventricular abnormalities, prior studies have demonstrated that arterial stiffness is increased in people with HFpEF above levels expected from aging alone (6–11). This has historically been ascribed to chronic systemic hypertension, but recent data suggest it may be related in large part to comorbid conditions observed with HFpEF such as metabolic syndrome and obesity (3,4). Few studies have compared HFpEF subjects to a hypertensive control group (7,8), and no invasive data is available relating indices of arterial stiffness at rest and during exercise to central hemodynamics in people with HFpEF.

We undertook the current study to compare invasive measures of arterial load in patients with invasively proven HFpEF with a matched hypertensive control group at rest and during exercise. We hypothesized that arterial stiffness would be increased with exercise in HFpEF compared to hypertensive controls without HF, that this stiffening would be associated with abnormal cardiac hemodynamics, and that acute treatment with inorganic sodium nitrite, a novel NO-cGMP providing agent (12–14), would partially reverse arterial stiffening with exercise.

METHODS

STUDY SUBJECTS

The study population includes subjects participating in prospective trials conducted in our laboratory all using the same uniform invasive supine exercise testing protocol (13–18). Data from these cohorts has been published previously, but not in total or as they relate to vascular indices in the current analysis. In the first cohort, HFpEF and control subjects underwent exercise testing in the evaluation of exertional dyspnea with simultaneous echocardiographic imaging (15–18). In the second and third cohorts, HFpEF subjects underwent exercise testing prior to and following administration of either intravenous or inhaled sodium nitrite administered acutely in a 1:1 randomized, double-blind, placebo-controlled fashion (13,14).

For the initial analysis comparing arterial load in HFpEF and controls, hemodynamic data acquired prior to receiving study drug was included for cohorts 2 and 3, and was analyzed with cohort 1 at matched workloads during cycle ergometry. For the next part of our analysis we assessed the effects of sodium nitrite, a novel nitric oxide (NO) donor, on arterial load in HFpEF. All studies were approved by the Mayo Clinic Institutional Review Board and were registered (NCT01418248, NCT01932606 and NCT02262078).

CASE DEFINITIONS

HFpEF and hypertensive control subjects were defined as previously specified (13–18). HFpEF was defined as symptoms of HF (fatigue and/or dyspnea) and preserved left ventricular ejection fraction ≥50% with elevated pulmonary capillary wedge pressure (PCWP) at rest (>15mmHg) and/or with exercise (≥25mmHg). Controls had systemic hypertension but no evidence of HF, with normal rest and exercise pulmonary artery (PA) pressures (rest <25mmHg, exercise <40mmHg) and normal PCWP (rest <15mmHg, exercise <25mmHg). Exclusion criteria included significant valvular heart disease (> mild stenosis, > moderate regurgitation), significant pulmonary disease, congenital heart disease, left-to-right shunt, unstable coronary artery disease, myocardial infarction within 60 days, hypertrophic or restrictive cardiomyopathy, high-output heart failure, or constrictive pericarditis.

CATHETERIZATION PROTOCOL

Subjects were studied in the postabsorptive state on chronic medication in the supine position at rest and during supine cycle ergometry as previously described (13–18). The radial artery was cannulated using a 5- or 6-French sheath for continuous recording of arterial pressure waveforms and blood sampling. Right heart catheterization was performed via a 9-French sheath inserted in the internal jugular vein. Right atrial (RA) pressure, PA and PCWP were measured at end-expiration using 2-French high fidelity micromanometer-tipped catheters (Millar Instruments, Houston, TX) advanced through the lumen of a 7-French fluid-filled balloon catheter (Arrow International Inc., Reading PA, USA).

Oxygen consumption (VO₂) was measured using breath-by-breath expired gas analysis (MedGraphics, St. Paul, MN, USA). Because exercise was performed via supine ergometry, peak VO₂ values achieved are 40 to 50% lower than what is observed with upright treadmill exercise. Arterial-venous O₂ difference (A-VO₂diff) was measured directly as the difference between systemic and PA O₂ contents (O₂ saturation•hemoglobin•1.34). Cardiac index (CI) was calculated using the direct Fick method (CI=VO₂/A-VO₂diff•BSA). Stroke volume index (SVI) was calculated as CI/heart rate (HR).

ARTERIAL WAVEFORM ANALYSIS

Radial artery pressure tracings were digitized (240 Hz) and stored for offline analysis. Central aortic pressure waveforms were determined from radial artery pressure tracings at rest and during exercise by mathematical transfer function using custom software (SphygmoCor, AtCor, NSW, AUS) as previously described and validated at rest, during exercise, and with drug therapy (19–21). Arterial waveform analysis was performed in a blinded fashion without knowledge of the subject group, hemodynamics or clinical data.

Steady, non-pulsatile arterial load was quantified by systemic vascular resistance index (SVRI = 80•[mean central blood pressure (BP) – RA]•BSA/CO). Pulsatile arterial load was assessed by pulse pressure (PP), effective arterial elastance index (EaI) and total arterial compliance index (TACI). EaI, a lumped measure of the total “stiffness” of the arterial system, was assessed by end-systolic central BP/SVI (22). TACI, which is a linear approximation of the pressure-volume relationship for the lumped arterial system, was calculated as SVI/central PP (23–25). Peripheral Pulse pressure amplification (PPA) was calculated as the ratio of peripheral to central PP. An increase in pulsatile load is evidenced by increases in Ea and PP and decreases in TACI and PPA.

The contribution of wave reflections to arterial load was also assessed. Forward (Pf) and backward (Pb) pressure waves were isolated from the composite central aortic waveform as previously described (26). Wave reflections were quantified by the reflection magnitude (RM = Pb/Pf) and aortic augmentation index (AIx) (21,26).

STATISTICAL ANALYSIS

Results are reported as mean ± standard deviation (SD), median (interquartile range [IQR]) or percentage. Differences between HFpEF and controls at rest and exercise were tested using Chi square, Student’s t-test or Wilcoxon rank sum test as appropriate. All comparisons were also adjusted for baseline differences in age and body mass index (BMI) using multivariable linear regression. Pearson correlation and linear regression were performed to detect the correlation between variables of interest. All tests were two-sided, with a p-value <0.05 considered significant. Analyses were performed using JMP 10.0.0 SAS Institute, Cary, NC, USA.

RESULTS

Compared to controls (n = 22), subjects with HFpEF (n = 98) were older and heavier (Table 1). Comorbidities including hypertension, diabetes and coronary disease were common and similarly prevalent, with no group difference. As expected, HFpEF subjects had higher NT-proBNP levels, increased E/e′ ratio, and higher right ventricular systolic pressure on echocardiography, consistent with increased filling pressures.

Table 1.

Baseline characteristics

	Control (n = 22)	HFpEF (n = 98)	p value
Age, years	62 ± 12	68 ± 10	0.01
Female, %	45	56	0.5
Body mass index, kg/m2	26.9 ± 4.1	34.1 ± 7.9	<0.0001
Comorbidities
Coronary disease, %	27	35	0.5
Diabetes Mellitus, %	23	31	0.5
Hypertension, %	100	98	1.0
Medications
ACE or ARB, %	41	59	0.1
Beta blocker, %	50	58	0.5
Calcium channel blocker, %	32	23	0.4
Loop diuretic, %	13	43	0.01
Laboratories
Estimated GFR, ml/	86 ± 29	86 ± 48	0.9
NT-proBNP, (pg/ml)	83 [42,295]	422 [122,1022]	0.005
Echocardiography
Ejection Fraction, %	62 ± 9	63 ± 8	0.6
E/e′ velocity, m/s	9 ± 3	15 ± 8	0.0003
RVSP, mmHg	30 ± 7	40 ± 14	0.004

Values represent mean ± standard deviation, or median [interquartile range]

ACE, angiotensin converting enzyme inhibitors; ARB = angiotensin receptor blockers; E/e′ = ratio of early diastolic transmitral filling velocity (E) and early diastolic mitral annular tissue velocity (e′); NT-proBNP = N-terminal pro Brain Natriuretic Peptide; RVSP = right ventricular systolic pressure; GFR = Glomerular Filtration Rate.

Because age and adiposity are well known to affect arterial properties and were different in HFpEF and controls, all subsequent comparisons between HFpEF and control subjects were adjusted for these covariates.

BASELINE VENTRICULAR AND VASCULAR FUNCTION

At rest both HFpEF and control subjects displayed peripheral and central arterial hypertension, with no group differences (Table 2). There were also no differences in resting measures of arterial afterload including systemic arterial resistance, elastance, compliance, or wave reflections. As expected, biventricular filling pressures and PA pressure were higher at rest in the HFpEF group.

Table 2.

Resting Arterial Properties and Central Hemodynamics

	Control (n = 22)	HFpEF (n = 98)	p value	Adjusted* p value
Radial Pressures
Radial Systolic BP, mmHg	163 ± 22	160 ± 25	0.6	0.1
Radial Diastolic BP, mmHg	73 ± 6	71 ± 10	0.3	0.5
Radial Mean BP, mmHg	104 ± 11	103 ± 14	0.6	0.4
Radial PP, mmHg	90 ± 18	89 ± 21	0.9	0.2
Aortic Pressures
Aortic Systolic BP, mmHg	145 ± 23	145 ± 24	0.9	0.4
Aortic Diastolic BP, mmHg	75 ± 6	72 ± 10	0.3	0.5
Aortic Mean BP, mmHg	104 ± 11	103 ± 14	0.6	0.4
Aortic PP, mmHg	71 ± 20	73 ± 20	0.6	0.5
Arterial Afterload
Ea Indexed, mmHg.m2/ml	3.26 ± 0.92	3.43 ± 1.02	0.5	0.8**
SVRI, dyne-sec.m2/cm5	3318 ± 1039	3059 ± 809	0.2	0.02**
TAC index, ml/mmHg.m2	0.61 ± 0.24	0.55 ± 0.19	0.2	0.8**
Pf, mmHg	47 ± 9	51 ± 14	0.2	0.7
Pb, mmHg	35 ± 12	36 ± 12	0.7	0.6
RM, % [IQR]	74[61,86]	68[58,91]	0.8	0.1
Aortic AIx, %	28 ± 16	31 ± 15	0.3	0.7
Peripheral PPA	1.31 ± 0.21	1.26 ± 0.17	0.2	0.0001
Central Hemodynamics
Heart Rate, bpm	69 ± 15	69 ± 11	0.9	0.5
RAP, mmHg	4 ± 2	11 ± 5	<0.0001	<0.0001
PCWP, mmHg	8 ± 3	18 ± 6	<0.0001	<0.0001
PA systolic pressure, mmHg	27 ± 6	44 ± 15	<0.0001	0.008
PA mean pressure, mmHg	16 ± 4	28 ± 9	<0.0001	0.0003
Stroke volume index, ml/m2	40 ± 10	37 ± 10	0.3	0.7**
Cardiac index, L/min.m2	2.7 ± 0.8	2.5 ± 0.6	0.2	0.8**

BP = blood pressure; PP = Pulse Pressure; HR = heart rate; Ea = Effective arterial elastance; SVRI = Systemic Vascular Resistance Index; TAC = Total Arterial Compliance; RAP = Right Atrial Pressure; PCWP = Pulmonary Capillary Wedge Pressure; Pf = Forward Wave; IQR = Interquartile range; Pb = Backward Wave; RM = Reflection Magnitude; Aix = Augmentation Index; PPA = Pulse Pressure Amplification; PA = Pulmonary Artery.

^*p value adjusted for age and body mass index.

^**Data that was already indexed to body weight was only adjusted for age. p values are not adjusted for multiple hypothesis testing.

ARTERIAL RESERVE WITH SUBMAXIMAL EXERCISE

During submaximal exercise (20 W), radial and aortic pressures were similar in HFpEF and controls, though aortic pulse pressure tended to increase more in HFpEF (Table 3, Figure 1). In contrast, direct measures of arterial afterload failed to decrease as much in HFpEF subjects as compared to controls during 20W exercise, manifest by higher arterial elastance, lower total arterial compliance, and decreased PPA (Table 3, Figure 1). Each of these differences persisted after adjusting for age and BMI.

Table 3.

Submaximal (20W) Exercise Arterial and Ventricular Function

	Control (n = 22)	HFpEF (n = 98)	p value	Adjusted* p value
Radial Pressures
Radial Systolic BP, mmHg	178 ± 18	182 ± 30	0.5	0.7
Radial Diastolic BP, mmHg	76 ± 5	76 ± 12	0.9	0.8
Radial Mean BP, mmHg	112 ± 11	114 ± 19	0.6	0.8
Radial PP, mmHg	102 ± 16	107 ± 24	0.4	0.5
Aortic Pressures
Aortic Systolic BP, mmHg	151 ± 19	160 ± 27	0.1	0.5
Aortic Diastolic BP, mmHg	80 ± 6	78 ± 12	0.4	0.7
Aortic Mean BP, mmHg	112 ± 11	114 ± 19	0.6	0.8
Aortic PP, mmHg	71 ± 16	83 ± 22	0.02	0.3
Arterial Afterload
Ea Indexed, mmHg.m2/ml	2.64 ± 0.71	3.26 ± 0.92	0.004	0.04**
SVRI, dyne-sec.m2/cm5	2052 ± 389	2239 ± 569	0.2	0.5**
TAC index, ml/mmHg.m2	0.70 ± 0.22	0.50 ± 0.18	<0.0001	0.0009**
Pf, mmHg [IQR]	55 [51,60]	62 [51–70]	0.01	0.6
Pb, mmHg	32 ± 9	36 ± 12	0.1	0.3
RM, %	58 ± 14	59 ± 19	0.8	0.4
Aortic AIx, %	20 ± 11	21 ± 18	0.8	0.4
Peripheral PPA	1.46 ± 0.17	1.35 ± 0.19	0.02	0.02
Central Hemodynamics
Heart Rate, bpm	91 ± 15	88 ± 15	0.3	0.6
RAP, mmHg	9 ± 3	21 ± 8	<0.0001	<0.0001
PCWP, mmHg	14 ± 5	32 ± 7	<0.0001	<0.0001
PA systolic pressure, mmHg	38 ± 11	67 ± 17	<0.0001	<0.0001
PA mean pressure, mmHg	25 ± 6	43 ± 11	<0.0001	<0.0001
Stroke volume index, ml/m2	47 ± 10	39 ± 10	0.0006	0.005**
Cardiac index, L/min.m2	4.2 ± 0.8	3.4 ± 0.9	0.0005	0.004**

BP = blood pressure; PP = Pulse Pressure; HR = heart rate; Ea = Effective arterial elastance; SVRI = Systemic Vascular Resistance Index; TAC = Total Arterial Compliance; RAP = Right Atrial Pressure; PCWP = Pulmonary Capillary Wedge Pressure; Pf = Forward Wave; Pb = Backward Wave; RM = Reflection Magnitude; Aix = Augmentation Index; PPA = Pulse Pressure Amplification; PA = Pulmonary Artery.

^*p value adjusted for age and body mass index.

^**Data that was already indexed to body weight was only adjusted for age. p values are not adjusted for multiple hypothesis testing.

Figure 1. Arterial load during exercise.

With exercise, HFpEF subjects, compared to controls, demonstrated [A] increased Aortic Pulse Pressure (PP), mmHg [B] – higher Effective arterial elastance indexed (EaI), mmHg.m2/ml [C] – lower Total Arterial Compliance Index (TACI), ml/mmHg.m2 and [D] – lower Peripheral Pulse Pressure Amplification (PPA)

As expected, HFpEF subjects developed typical cardiac abnormalities with exercise including pulmonary venous and arterial hypertension and reduced cardiac output reserve (Table 3). Lower total arterial compliance and higher arterial elastance with exercise were correlated with higher PCWP and lower cardiac output (CO) during exertion (Figure 2). These relationships remained significant after adjusting for age and BMI (both p <0.001).

Figure 2. Correlation between arterial load and central hemodynamics.

A higher exercise Pulmonary Capillary Wedge Pressure (PCWP), mmHg was associated with [A] lower Total Arterial Compliance Index (TACI), ml/mmHg.m2 and [B] higher Effective Arterial Elastance Index (EaI), mmHg.m2/ml. In addition, a lower Cardiac Output (CO) response was associated with lower exercise [C] arterial compliance and [D] elastance.

ARTERIAL RESERVE WITH PEAK EXERCISE

All of the control subjects (n = 22) and 42 of the HFpEF subjects (43%) exercised past the submaximal 20W workload to volitional exhaustion. Peak exercise capacity was roughly 40% lower in HFpEF subjects compared to controls (peak VO₂ 8.6 ± 2.3 vs. 14.8 ± 3.8 ml/min/kg, p <0.0001). As with 20W exercise, systemic pressures were similar but measures of arterial afterload decreased less at peak exercise in HFpEF as compared to controls, with higher arterial elastance and systemic vascular resistance and lower total arterial compliance (Table 4). Unlike submaximal exercise, there was also evidence for increased wave reflection-associated pressure load at peak exercise in HFpEF subjects as compared to controls. This was evidenced by higher AIx, RM and lower PPA. Among the measures of reflected pressure waves, AIx at peak exercise was correlated with reduced total arterial compliance, higher PCWP, EaI, and lower peak CO (Figure 3).

Table 4.

Peak Exercise Arterial and Ventricular Function

	Control (n = 22)	HFpEF (n = 42)	p value	Adjusted* p value
Radial Pressures
Radial Systolic BP, mmHg	188 ± 19	192 ± 25	0.5	0.4
Radial Diastolic BP, mmHg	76 ± 6	75 ± 10	0.9	0.4
Radial Mean BP, mmHg	113 ± 10	114 ± 17	0.8	0.6
Radial PP, mmHg	112 ± 17	116 ± 21	0.3	0.5
Aortic Pressures
Aortic Systolic BP, mmHg	149 ± 15	158 ± 24	0.1	0.6
Aortic Diastolic BP, mmHg	81 ± 6	79 ± 11	0.4	0.2
Aortic Mean BP, mmHg	113 ± 10	114 ± 17	0.8	0.6
Aortic PP, mmHg	68 ± 13	79 ± 20	0.02	0.2
Arterial Afterload
Ea Indexed, mmHg.m2/ml	2.45 ± 0.71	3.16 ± 0.92	0.003	0.02**
SVRI, dyne-sec.m2/cm5	1473 ± 440	1923 ± 494	0.0007	0.007**
TAC index, ml/mmHg.m2	0.77 ± 0.24	0.55 ± 0.21	0.0006	0.006**
Pf, mmHg	62 ± 19	66 ± 14	0.4	0.6
Pb, mmHg	27 ± 11	35 ± 13	0.04	0.05
RM, % [IQR]	41 [34,54]	50 [43,63]	0.08	0.05
Aortic AIx, %	8 ± 11	17 ± 13	0.007	0.009
Peripheral PPA	1.66 ± 0.17	1.51 ± 0.22	0.007	0.006
Central Hemodynamics
Heart Rate, bpm	121 ± 16	99 ± 16	<0.0001	<0.0001
RAP, mmHg	8 ± 4	22 ± 6	<0.0001	<0.0001
PCWP, mmHg	14 ± 5	34 ± 7	<0.0001	<0.0001
PA systolic pressure, mmHg	42 ± 9	70 ± 13	<0.0001	<0.0001
PA mean pressure, mmHg	28 ± 7	49 ± 8	<0.0001	<0.0001
Stroke volume index, ml/m2	99 ± 29	85 ± 28	0.06	0.02**
Cardiac index, L/min.m2	12.0 ± 4.0	8.4 ± 2.9	<0.0001	<0.0001**
Peak VO2, ml/kg/min	14.8 ± 3.8	8.6 ± 2.3	<0.0001	<0.0001**

BP = blood pressure; PP = Pulse Pressure; HR = heart rate; Ea = Effective arterial elastance; SVRI = Systemic Vascular Resistance Index; TAC = Total Arterial Compliance; RAP = Right Atrial Pressure; PCWP = Pulmonary Capillary Wedge Pressure; Pf = Forward Wave; Pb = Backward Wave; RM = Reflection Magnitude; IQR = Interquartile range; Aix = Augmentation Index; PPA = Pulse Pressure Amplification; PA = Pulmonary Artery.

^*p value adjusted for age and body mass index.

^**Data that was already indexed to body weight was only adjusted for age. p values are not adjusted for multiple hypothesis testing.

Figure 3. Correlation between wave reflection and central hemodynamics at peak exercise.

Increased systolic pressure augmentation due to wave reflection (AIx) during peak exercise was correlated with [A] higher Effective Arterial Elastance Index (EaI), [B] lower total arterial compliance index (TACI), [C] higher Pulmonary Capillary Wedge Pressures (PCWP), and [D] depressed cardiac output (CO) reserve.

EFFECT OF NITRITE ON ARTERIAL PROPERTIES

We next evaluated the effects of sodium nitrite, a novel NO providing agent, on the observed abnormalities in central and peripheral arterial loading at rest and with exercise in HFpEF subjects. Nitrite or matching placebo was administered in a randomized, blinded fashion intravenously in cohort 2 subjects (n = 13) and via inhaled nebulization in cohort 3 subjects (n = 14). Because plasma NO₂ levels and hemodynamic effects were similar with intravenous and inhaled nitrite administration (Online Table), we combined both nitrite and placebo groups together to analyze the effects on arterial load.

At rest, nitrite modestly reduced arterial pressures, with greater effects noted on central pressure waveforms (Table 5). As compared to placebo, nitrite decreased aortic wave reflections at rest (lower Pb, RM, and AIx), but had no effect on arterial elastance or compliance as compared to placebo. In contrast, with exercise (20W), nitrite reduced central aortic pressures, decreased arterial elastance, reduced systemic vascular resistance and AIx, and increased total arterial compliance as compared to placebo (Figure 4). These favorable arterial effects of nitrite were coupled with salutary reductions in biventricular filling pressures and PA pressures at rest and to greater extent with exercise, along with an improvement in cardiac output (Table 5).

Table 5.

Effect of Nitrite on Arterial Properties at Rest and Exercise

	Rest		20W exercise
	Placebo-corrected Nitrite Effect (n = 52)	p value	Placebo-corrected Nitrite Effect (n = 52)	p value
Radial Pressures
Radial Systolic BP, mmHg	−4 ± 5	0.5	−2 ± 3	0.6
Radial Diastolic BP, mmHg	−3 ± 2	0.09	−4 ± 1	0.01
Radial Mean BP, mmHg	−6 ± 3	0.04	−6 ± 2	0.02
Radial PP, mmHg	−1 ± 4	0.8	+2 ± 3	0.6
Aortic Pressures and Flow
Aortic Systolic BP, mmHg	−10 ± 5	0.05	−9 ± 3	0.01
Aortic Diastolic BP, mmHg	−3 ± 2	0.07	−4 ± 1	0.02
Aortic Mean BP, mmHg	−6 ± 3	0.04	−6 ± 2	0.02
Aortic PP, mmHg	−8 ± 4	0.04	−5 ± 2	0.03
Arterial Afterload
Ea Indexed, mmHg.m2/ml	−0.13 ± 0.20	0.5	−0.38 ± 0.14	0.008
SVRI, dyne-sec.m2/cm5	−90 ± 173	0.6	−204 ± 75	0.009
TAC index, ml/mmHg.m2	+0.06 ± 0.04	0.1	+0.10 ± 0.02	0.0005
Pf, mmHg	+2 ± 3	0.5	−0 ± 2	0.8
Pb, mmHg	−7 ± 3	0.009	−3 ± 2	0.06
RM, %	−19 ± 5	0.0003	−5 ± 4	0.1
Aortic AIx, %	−8 ± 5	0.002*	−8 ± 2	<0.0001*
Peripheral PPA	+0.12 ± 0.03	0.0001	+0.12 ± 0.03	<0.0001
Central Hemodynamics
Heart rate	+3 ± 2	0.1	+3 ± 2	0.1
RAP, mmHg	−1 ± 1	0.01	−4 ± 1	<0.0001
PCWP, mmHg	−3 ± 1	0.001	−8 ± 1	<0.0001
Cardiac output, L/min	−0.01 ± 0.3	0.9	+0.8 ± 0.3	0.01

Table shows placebo-corrected values (change with study drug minus change with placebo).

Values represent mean ± Standard Error of Difference.

^*by Wilcoxon rank sum test

BP = blood pressure; PP = Pulse Pressure; HR = heart rate; Ea = Effective arterial elastance; SVRI = Systemic Vascular Resistance Index; TAC = Total Arterial Compliance; RAP = Right Atrial Pressure; PCWP = Pulmonary Capillary Wedge Pressure; Pf = Forward Wave; Pb = Backward Wave; RM = Reflection Magnitude; Aix = Augmentation Index; PPA = Pulse Pressure Amplification. p values are not adjusted for multiple hypothesis testing.

Figure 4. Effects of inorganic nitrite on resting and exercise arterial load.

Percentage improvement with nitrite therapy in measures of [A] Resting reflective load: Backward Wave (Pb), Reflection Magnitude (RM), Peripheral Pulse Pressure Amplitude (PPPA) [B] Exercise arterial load: Effective Arterial Elastance Indexed (EaI), Total Arterial Compliance Index (TACI), Systemic Vascular Resistance Index (SVRI)

DISCUSSION

We comprehensively examined arterial properties in patients with invasively-proven HFpEF as compared to hypertensive controls at rest and during exercise using gold standard invasive hemodynamic assessments while assessing the effects of a novel NO-cGMP providing agent that we show partially reverses arterial stiffening. We show that as compared to hypertensive controls, subjects with HFpEF displayed similar indices of arterial afterload when measured at rest. In contrast, exercise unmasked significant limitations in arterial compliance and vasodilatory reserve that were correlated with classical hemodynamic abnormalities observed in HFpEF, including elevated ventricular filling pressures and inadequate CO. Arterial stiffening was present despite similar arterial pressures measured centrally and peripherally. Arterial stiffening at rest and with exercise was partially reversed with inorganic sodium nitrite, a novel NO-providing therapy, and this was coupled with favorable improvements in central hemodynamics. These data emphasize that arterial stiffening and impaired arterial vasodilator reserve with exercise plays an important role in the pathophysiology of HFpEF that is independent of hypertension and mean blood pressure alone. These data emphasize the importance of central aortic stiffening in HFpEF as a viable therapeutic target that merits further prospective study in this cohort of patients for whom few treatment options exist.

CENTRAL ARTERIAL LOAD AND BLOOD PRESSURE

Aging causes changes in the mechanical properties of the aorta and conduit arteries (1,2). One of the key sequelae of this change is a decrease in total arterial compliance, which can be conceptualized as the ability of the arteries to store blood during systole without excessive increases in pressure. The aorta contributes the majority of compliance in the systemic arterial bed, unlike the lungs where compliance is distributed more evenly throughout the vasculature. Decreases in aortic distensibility heighten left ventricular (LV) afterload directly by augmenting the amplitude of outgoing (incident) pressure waves, and indirectly by increasing wave velocity, which enhances early return of reflected pressure waves that sum with incident waves to increase systolic pressure load (1,2). Acutely, these arterial changes adversely affect LV ejection performance and diastolic relaxation (27,28). Chronically, these arterial changes may contribute to concentric chamber remodeling, fibrosis, and changes in the mechanical properties of the LV (28).

As shown in the current study, arterial stiffening is often not apparent in the peripheral arteries, and requires careful assessment of central aortic pressure and flow characteristics (2). In the young and healthy arterial system, there is augmentation of systolic pressure in the peripheral arteries due to the effects of wave reflection. With aortic stiffening, these reflected pressure waves travel more rapidly, arriving in the central aorta during systole rather than diastole, to increase aortic pressure load. This results in a decrease in PPA as vascular stiffness increases along with other characteristic changes in the central aortic waveform.

VASCULAR STIFFENING IN HFpEF

Systemic hypertension is highly prevalent in HFpEF and has been extensively implicated in its pathogenesis (3). A number of studies employing noninvasive techniques have reported that arterial stiffness is increased in HFpEF as compared to healthy controls, and some have correlated this stiffening with decreased exercise capacity (6,9–11). However, elevation in blood pressure is associated with reduced arterial distensibility in and of itself, because the arterial pressure-volume relationship is not linear and varies with ambient pressure. Thus, arterial stiffening may not be specific to HFpEF but common to all or most hypertensive patients. In this regard, 2 studies comparing HFpEF subjects to carefully matched hypertensive controls have reported little to no difference in most measures of arterial stiffness between groups (7,8).

The current data are consistent with these studies, showing that compared to a hypertensive control group, there was no discernable difference in resting arterial afterload in subjects with HFpEF. However, with the increases in arterial pressure and blood flow associated with exercise, abnormalities became apparent, revealing vascular stiffening that was not clearly identifiable at rest. Older age and obesity are 2 of the strongest risk factors for HFpEF and are believed to play major roles in the pathogenesis (3). Importantly, the differences in arterial stiffness observed in this study during exercise persisted even after adjusting for age and BMI, which are known to directly affect arterial properties (29,30), and these differences were not related to differences in arterial blood pressure. This important observation demonstrates that arterial stiffening and reduced arterial reserve is specific to the HFpEF phenotype, and is thus a potentially important therapeutic target.

This study is also unique due to the invasive determination of arterial and reflected load with exercise, which allowed for direct correlation with simultaneous filling pressures and CO during exercise. One prior non-invasive study measured and demonstrated increased proximal aortic impedance and decreased arterial compliance with exercise, but they did not measure simultaneous reflected load or filling pressures (9). While causality cannot be proven from this study design, the fact that abnormalities in arterial compliance, elastance and wave reflection were present in HFpEF, associated with abnormal hemodynamics, and partly reversed with nitrite in tandem with improved hemodynamics, strongly supports the notion that arterial stiffening plays an important role in the pathophysiology of HFpEF, especially during exercise. This is thus an important therapeutic target for future studies.

THERAPEUTIC IMPLICATIONS

The finding of greater arterial afterload with exercise in HFpEF suggests a role for drugs that enhance arterial compliance and reduce wave reflection. Different antihypertensives have varying effects on wave reflection and aortic compliance properties. For example, beta blockers increase wave reflection and central blood pressure compared to vasodilators. This has been suggested as a mechanism for the inferior outcomes seen with beta-blocker therapy for hypertension in the ASCOT trial compared to an amlodipine vasodilator regimen (31). While this effect has been ascribed to negative chronotropic effects, this may not apply to all heart rate-lowering drugs. Ivabradine, which selectively lowers the sinus rate, improves aortic compliance in patients with HFrEF (32). Alternative vasodilators may have similar effects on lowering arterial stiffness and improving central hemodynamics, but comparative data on their hemodynamic effects in HFpEF is lacking, and we cannot assume that the current observations will apply to all vasodilators. Future studies evaluating both cardiac and arterial hemodynamic effects of vasoactive medicines in HFpEF would be valuable to address this important question.

Deficiency of NO and its downstream second messenger cGMP have been repeatedly implicated in the pathogenesis of HFpEF and are being targeted in a variety of ongoing trials (3,4). Neprilysin inhibitors increase intracellular cGMP by decreasing the breakdown of endogenous natriuretic peptides, and treatment with the neprilysin inhibitor omapatrilat has been shown to improve central aortic distensibility and reduce systolic wave reflections and characteristic impedance in hypertensives (33). It remains unknown if similar vascular effects will be seen with the newer neprilysin antagonist sacubitril, which is currently being tested in the PARAGON-HF (Prospective comparison of ARni with Arb Global Outcomes in heart failure with preserved ejectioN fraction) trial in patients with chronic HFpEF (NCT01920711).

The inorganic nitrate/nitrite/NO pathway represents an alternative method to improve NO-cGMP availability in HFpEF. Acute administration of inorganic nitrite (or its precursor nitrate) decreases conduit vessel stiffness in healthy volunteers (12,34), enhances exercise capacity and vasodilation in HFpEF (13,14,35,36), and improves rest and exercise hemodynamics in HFpEF (13,14,37). Zamani and colleagues recently found that inorganic nitrate decreased AIx in HFpEF patients when measured at rest, while improving peak exercise capacity (35). The current data importantly extend upon these previous findings, confirming salutary effects of nitrite on wave reflection, while demonstrating for the first time direct improvements in arterial compliance, elastance and wave reflection at rest and during exercise, in which hemodynamic perturbations contribute to symptoms of dyspnea. Effects of nitrite in hypertensive controls were not examined in this study, but might also produce favorable effects on exercise tolerance. Larger scale clinical trials sponsored by the NHLBI are currently underway to evaluate the effects of longer-term nitrite therapy in HFpEF (NCT02742129) (NCT02713126), and further study is warranted using other novel therapies targeting arterial stiffness.

LIMITATIONS

Central aortic pressures were not directly measured but were derived mathematically from the directly measured radial artery tracings. However, this method has been previously validated against directly-measured central aortic pressures (19), and the use of directly-measured pressures from an arterial cannula is a unique strength as compared to prior studies that relied on noninvasive applanation tonometry to measure radial waveforms. While imputation of central pressures from radial waveforms has been validated following nitroglycerin therapy (19), there is less validation data available using other drug therapies. Correction for multiple hypothesis testing was not performed. Arterial stiffening may be related to structural remodeling and changes in the material properties of the vasculature, or to endothelial dysfunction and vasoconstriction, or both. We cannot distinguish which components explained the greater stiffening during exercise in HFpEF. Future studies evaluating effects of nitrite on individual components, such as aortic pulse wave velocity and endothelium-dependent vasodilation, would be important to help sort out the mechanisms by which nitrites improves arterial stiffening. Subjects with HFpEF were older and had higher BMI than controls, which may influence arterial properties independent of HFpEF status, but all key differences remained highly significant after adjusting for these baseline differences.

CONCLUSIONS

People with HFpEF display impaired arterial compliance, resistance and elastance reserve that is provoked by the physiologic stress of exercise. These abnormalities are directly correlated with greater hemodynamic severity of HF and worse functional capacity, even after controlling for the presence of hypertension. Sodium nitrite mitigates these vascular perturbations in tandem with salutary effects on the hemodynamic abnormalities that contribute to effort intolerance. Further study is warranted to investigate whether therapies targeting central aortic stiffness can improve clinical outcomes in HFpEF.

Central Illustration. Arterial load and wave reflections in HFpEF at rest and during exercise.

At rest (upper panel), the aortic pressure waveform is shown as a composite of the forward wave (color) and reflected wave (color). Wave reflections, which develop at the points of impedance mismatch along the arterial tree, are reflected back to the aorta causing systolic pressure augmentation. Total arterial compliance, which reflects the ability of the arteries to store blood during systole without untoward elevation in pressure, is not significantly compromised, and pulmonary capillary wedge pressure (PCWP) is near-normal. During exercise (bottom panel), venous return and cardiac output increase. Stiffening of the aorta, along with a lack of small vessel vasodilation in the periphery (inadequate reduction in systemic vascular resistance), augments pressure wave reflections (color) and pressure augmentation of in the central aorta during mid to late systole. Total arterial compliance reserve becomes saturated, such that increases in stroke volume cause greater aortic pulse pressure increases, further augmenting left ventricular load. These changes are then correlated with pathologic increases in PCWP that dyspnea and impairment in forward cardiac output reserve limiting oxygen transfer to the body.

PERSPECTIVES.

Competency in Medical Knowledge

Arterial stiffness and wave reflections increase during exercise in patients with heart failure and preserved left ventricular ejection fraction (HFpEF). These abnormalities are associated with higher cardiac filling pressures and lower cardiac output, but improve after administration of inorganic nitrite medications.

Translational Outlook

Prospective studies are needed to quantify the potential benefit of therapies that target nitric oxide deficiency and ameliorate central arterial stiffening and wave reflections in patients with HFpEF..

Abbreviations

AIx

Augmentation Index

A-VO₂diff

Arterial-venous O₂ difference

BMI

Body mass index

BP

Blood pressure

CI

Cardiac index

CO

Cardiac output

EaI

Effective Arterial Elastance Index

HF

Heart failure

HFpEF

Heart Failure with Preserved Ejection Fraction

HR

Heart rate

IQR

Interquartile range

LV

Left ventricle

NO

Nitric oxide

NO-cGMP

Nitric oxide-guanosine monophosphate

PA

Pulmonary artery

Pb

Backward Wave

PCWP

Pulmonary Capillary Wedge Pressure

Pf

Forward Wave

PP

Pulse pressure

PPA

Pulse pressure amplification

TACI

Total Arterial Compliance Index

RM

Reflection Magnitude

SD

Standard deviation

SVI

Stroke volume index

SVRI

Systemic Vascular Resistance Index

VO₂

Oxygen consumption