Author + information
- Received April 14, 2008
- Revision received October 13, 2011
- Accepted October 17, 2011
- Published online January 31, 2012.
- Lamia Tartière-Kesri, MD⁎,⁎ (, )
- Jean-Michel Tartière, MD†,‡,
- Damien Logeart, MD, PhD‡,§,
- Florence Beauvais, MD‡,§ and
- Alain Cohen Solal, MD, PhD‡,§∥
- ↵⁎Reprint requests and correspondence:
Dr. Lamia Tartière-Kesri, Hôpital Léon Bérard, Réadaptation Cardiaque, Avenue Docteur Marcel Amanet, 121 BP 10, 83418 Hyeres, France
Objectives This study sought to demonstrate that arterial stiffness is probably underestimated in patients with heart failure with preserved ejection fraction (HFpEF) at rest and may be revealed with moderate exercise.
Background HFpEF is associated with ventriculoarterial stiffening.
Methods We compared 23 patients with stable chronic HFpEF, left ventricular ejection fraction >45%, and impaired relaxation with 15 controls without cardiac disease. Patients were compared at rest and during a 30-W exercise. The following variables were measured or calculated by Doppler echocardiography and tonometry: left ventricular volumes and end-systolic elastance (Ees), peripheral resistance, arterial elastance (Ea), arterial compliance, aortic pulse wave velocity, and carotid Peterson modulus (Ep).
Results Patients with HFpEF were comparable to controls in age, sex ratio, blood pressure, and heart rate. Ventriculoarterial coupling, assessed by Ees/Ea and Ees/Ep ratios, was moderately impaired at rest in patients compared with controls (both p < 0.01). HFpEF was associated during exercise with a major increase in Ep (+155 ± 193% vs. −5 ± 28%), pulse wave velocity (+20 ± 30% vs. −7 ± 24%), and Ea (+12 ± 15% vs. −5 ± 10%), and a lower decrease in peripheral resistance (−17 ± 12% vs. −26 ± 12%) (p < 0.05 for all). In addition, HFpEF patients showed a lower increase in stroke volume (+10 ± 16% vs. +21 ± 12%) despite a greater increase in Ees (+20 ± 18% vs. +3 ± 12%) (p < 0.05 for all). Also during exercise, adaptation of proximal ventriculoarterial coupling was impaired in HFpEF patients (Ees/Ep: −26 ± 47% vs. +20 ± 47% for controls) (p < 0.01), with no difference in Ees/Ea.
Conclusions In HFpEF patients, moderate exercise leads to a steep increase in proximal afterload that is underestimated at rest and is associated with unfavorable ventriculoarterial coupling and exercise intolerance.
Heart failure (HF) is a major cause of morbidity and mortality. It is predominantly a disease of elderly people (1). Nearly half of affected patients have an apparently preserved systolic function (2). Rehospitalization rates and prognosis are similar to those for patients with systolic failure (2). In contrast to the improvement in survival observed for patients with HF and reduced left ventricular ejection fraction (LVEF), mortality for patients with HF and preserved LVEF (HFpEF) has remained constant, which emphasizes the lack of therapy with proven benefit for these patients (3).
Diastolic dysfunction is thought to play a dominant role in HFpEF and exercise intolerance (4,5). However, the pathophysiology of the exercise intolerance in this syndrome is incompletely understood and likely multifactorial (6–8). Recently, some authors drew attention to the fact that in older people with HFpEF, impaired exercise tolerance correlates strongly with aortic stiffening (9). Another factor that may contribute to HF pathophysiology is abnormal ventricular-arterial interaction due to stiffening of both systems (6,10), thus contributing to blood pressure lability and higher energy cost to increase cardiac output (6).
In this study, we hypothesized that the arterial stiffness in patients with HFpEF at rest is probably underestimated and could be revealed with moderate exercise.
Twenty-three consecutive patients referred to our laboratory with chronic HFpEF, LVEF >45%, sinus rhythm, and stable hemodynamic conditions were included. The diagnosis of HFpEF was based on criteria defined by the European Society of Cardiology (3). Patients with severe valvular disease or severe wall-motion abnormalities were excluded. In the same period, 15 consecutive patients referred to our laboratory for chest pain with normal electrocardiography stress-test results were included. Medications were not withdrawn for this study. Investigators were not blinded to subject assignment. All patients gave their informed consent. The study was approved by the local hospital institutional review board.
Central pressure measurement
Carotid pulse wave was recorded at rest and during exercise by aplanation tonometry (SphygmoCor Px PWA System, AtCor Medical, West Ryde, Australia) on the right carotid artery, and calibration involved the use of brachial mean blood pressure and diastolic blood pressure (11). Augmentation pressure was defined as carotid systolic blood pressure minus the pressure at the first peak shoulder of the carotid pulse wave.
Determination of vascular geometry
Images were taken by using a commercially available sonographer (Vivid 7, GE Vingmed, Horten, Norway) with a 14-MHz transducer.
Right common carotid measurements involved transversal and longitudinal scanning for measuring the maximum lumen diameter, end-diastolic diameter, and intima-media thickness. At least 10 cycles were averaged. Cross-sectional area of the carotid artery was calculated as follows: [(maximum lumen diameter/2)2 × π] − (maximum lumen diameter/2 − intima-media thickness)2 × π].
Determination of vascular function
Effective arterial elastance (Ea) was estimated as end-systolic pressure/stroke volume (SV) (12). End-systolic pressure was estimated as carotid systolic pressure × 0.9 (12,13). Arterial compliance was estimated by the ratio of SV to pulse pressure (PP) (14) and peripheral vascular resistance (PVR) by mean arterial pressure/cardiac output × 79.9.
Peterson's modulus (Ep) (in kPa · 102) was defined as: (carotid PP × carotid end-diastolic diameter)/ΔD. Young's elastic modulus (YEM) (in kPa · 103) was (carotid PP × carotid end-diastolic diameter2)/(ΔD × 2 × intima-media thickness), with ΔD indicating the mean carotid distention.
Aortic pulse wave velocity (PWV) was assessed by use of pulsed Doppler and defined as the transit time (in seconds) between the aortic arch and the abdominal aorta divided by the distance (in meters) between these 2 points.
Determination of LV geometry and function
Images were taken using a 3.5-MHz transducer (GE Vingmed). The left ventricular (LV) parameters were recorded or calculated according to the American Society of Echocardiography recommendations (15).
The modified single-beat method was used to estimate end-systolic elastance (Ees) from carotid pressure, SV, and pre-ejection and total systolic periods determined on Doppler analysis of the aortic flow, Simpson LVEF, and an estimated normalized ventricular elastance at arterial end-diastole, as previously validated (13).
Peak velocities of the mitral inflow were derived from pulsed Doppler recordings with the sample volume placed at the tip of the mitral leaflets. Peak E-wave and peak A-wave velocities were measured. The mitral annular (mean of lateral and septal annulus) early diastolic velocity (E′) was determined by color-decoded tissue Doppler imaging. The constant tau was calculated as previously published (16). E/E′ was used as a surrogate of LV end-diastolic pressure (EDP) and E/E′ divided by the end-diastolic volume (E/E′/EDV) as an index of absolute chamber compliance. A single-beat approach was used to characterize the EDP volume relationship (where EDP = αEDVβ; α is a curve-fitting constant and β is a diastolic stiffness constant) (17). To account for covariance in α and β (18), both of which are indicative of the shape and position of the EDP volume relationship, derived α and β in each subject were used to predict the EDV at a common EDV of 20 mm Hg. Comparison of EDP of 20 mm Hg indexed to body surface area (V20i) was then used as a comparison of overall diastolic stiffness between groups (8).
All recordings were taken by 2 operators (J.-M.T., L.T.-K.) at rest and during exercise, and analysis was performed offline using the EchoPAC system (GE Vingmed). All Doppler echocardiographic measurements were taken at least 4 times and averaged. Measurements were indexed to body surface area as appropriate. Carotid parameters and PWV were assessed on the basis of the mean of at least 10 measurements. Intra- and interobserver reproducibility at rest and during exercise was analyzed for cardiac and vascular Doppler echocardiographic measurements and tonometry, with correlations >0.89 except for E′ interobserver reproducibility (r = 0.81) and the percentage of difference <5%.
Exercise test and symptoms assessment
Measurements were performed at rest and during 30-W exercise (Ergoline 900, Schiller Medical, Wissembourg, France) with the patient in a semirecumbent position after heart rate stabilization, before fusion between E and A waves. Data were acquired during exercise and stored after heart rate stabilization.
Twenty HFpEF patients and 14 controls underwent the 6-min walking test (6-MWT) without encouragement, and 23 HFpEF patients and 14 controls performed a maximal exercise test in an upright position on a cycloergometer following the Bruce protocol.
Categorical data are presented as percentages, and continuous data as mean ± 1 SD. E/E′ was log-transformed to achieve normal distribution, and values were given by using the geometric mean and 95% lower and upper confidence limits. Ees values were compared after normalization for the LV mass and adjustment for V0 (Ees-LVM-V0). The Student t test and chi-square test were used as appropriate. Analysis of variance with 2 factors was used to determine the effect of exercise (rest vs. exercise) or the effect of group (HFpEF vs. control). This analysis was adjusted for sex. Spearman rank test was used for calculating partial correlation coefficients adjusted for sex, and another multivariate analysis was performed with sex and height as partial variables. The values in Figure 1 are shown without adjustment, but an asterisk appears next to the unadjusted p value when statistical significance was changed by adjusting for sex. A 2-sided p value of <0.05 was considered significant. Analyses involved the use of NCSS 6.0.21 software (J.L. Hintze, Kaysville, Utah).
Baseline characteristics (Tables 1 and 2) were different between groups with regard to cardiovascular risk factors, which were more frequent in HFpEF patients except for age, sex, and body mass index. Of HFpEF patients, 43% had a history of coronary heart disease.
Patients with HFpEF showed significant functional impairment compared with controls: lower exercise duration ability, lower maximum workload, and shorter 6-MWT (p < 0.001 for all). They were more frequently taking antihypertensive agents, aspirin, and statins.
After adjustment for sex, brachial blood pressure was not significantly different between the groups (Table 2). HFpEF patients had a thicker right carotid artery, with increased cross-sectional area, but no difference in YEM. The left ventricle was thicker, with an increased LV mass index, despite no difference in LV end-diastolic diameter index or relative wall thickness. Patients also had an enlarged left atrium and a lower LVEF.
Cardiac adaptation during exercise
After adjustment for sex, the increase of heart rate was similar between groups during exercise (Tables 3 and 4). SV increased significantly during exercise in both groups with a greater increase of EDV in controls compared with HFpEF patients (+8.6 ± 2.7% vs. −0.9 ± 2.2%, respectively; sex-adjusted p value = 0.010). LV relaxation (E′ and tau) was impaired in HFpEF patients, but exercise was associated with improved LV relaxation in both groups (Table 3). HFpEF patients showed increased chamber stiffness (β) without significant modification by exercise. The log-transformed E/E′ ratio was statistically higher in HFpEF patients but was not increased during exercise. However, after adjusting for sex and controlling for covariance in α and β, overall diastolic LV stiffness was not different among groups (same V20i) even if another parameter such as the E/E′/EDV was indicative of an increased diastolic LV stiffness in the HFpEF group (Table 3, Fig. 2A). Moreover, the increase in V20i during exercise, expressed as a percentage, was significantly higher in controls compared with HFpEF patients (+10.7 ± 3.4% vs. −1.1 ± 2.7%, respectively; sex-adjusted p value = 0.016), but not variations of β, tau, or E/E′/EDV (p > 0.05 for all).
Regarding contractile response and after adjustment for sex, the magnitude of increase in Ees from rest to exercise was greater in the HFpEF group than in controls (Fig. 1A). Ees-LVM-V0 was statistically different between groups (p = 0.011), and the magnitude of increase during exercise was greater in HFpEF group (+20 ± 18% vs. +3 ± 12% in controls; p = 0.004).
Vascular adaptation during exercise
Exercise was characterized by increased carotid systolic blood pressure, PP and PP amplification in both groups, with higher values of carotid systolic blood pressure and PP, and lower PP amplification in HFpEF group (Table 4, Fig. 2B). A significant interaction was shown in terms of carotid PP, demonstrating a steeper slope of increased PP in the HFpEF group. Modifications of vascular afterload substantially differed between the groups, especially regarding the proximal arterial compartment. Exercise was associated with a significant decrease in PVR and a nonsignificant decrease in Ea (p = 0.066). However, the magnitude of the PVR decrease during exercise seems to be different among groups with a blunted phenomenon in HFpEF patients (−26 ± 3% vs. −17 ± 3% in the HFpEF group; p < 0.05). When vascular load was assessed only from its proximal compartment (PWV, arterial compliance, and Ep), vascular load seemed to be impaired in HFpEF patients, especially regarding Ep (−5 ± 39% vs. 155 ± 32% for HFpEF patients; p < 0.01; and interaction, p = 0.011), arterial compliance (6 ± 4% vs. −19 ± 4% for HFpEF patients; p < 0.001; interaction, p = 0.059) or PWV (Fig. 1B). HFpEF patients showed a steeper slope of the carotid pressure-to-diameter ratio during exercise (Fig. 2B).
Ventriculoarterial coupling during exercise
Ventriculoarterial coupling was impaired in the HFpEF group when globally assessed by the Ees/Ea or Ees/Ep ratios (Table 4). Moreover, the magnitude and direction of change of Ees/Ep were extremely unfavorable in the HFpEF group during exercise (+20 ± 47% vs. −26 ± 47% for HFpEF patients; p = 0.005; interaction, p = 0.025) (Fig. 1C).
Relationship between cardiovascular function and exercise capacity
Sex-adjusted baseline values of impaired LV relaxation, high E/E′ or β and low E′ were correlated with a lower maximal workload and shorter 6-MWT duration (Table 5). Low Ees was correlated only with lower maximal workload. No significant correlation was found with other parameters at rest. There was no modification after adjustment for sex and height.
After controlling for sex, modifications of vascular parameters during exercise, but not modifications of LV diastolic parameters (E′, E/E′, and β), were correlated with maximum workload (Ea, PVR) or 6-MWT (Ep). An increased Ees was also related to greater maximal workload (r = −0.455, p < 0.01). Ep was no longer related to 6-MWT after adjustment for sex and height, but other correlations were not modified.
This noninvasive study examines for the first time ventriculoarterial coupling and its adaptation during exercise in patients with HFpHF. In these patients, a moderate level of exercise revealed a major increase in proximal arterial stiffness, which was absent at rest, compared with age- and sex-matched controls.
In this study, at rest, when considering vascular load, groups are comparable in proximal and peripheral load. However, moderate exercise is associated with a major increase in proximal arterial stiffness in HFpEF patients, corresponding to a dramatic increase in proximal arterial load. This result contrasts with a less impaired peripheral vascular system. This proximal arterial stiffening is probably associated with a steeper slope in the arterial pressure-volume relationship during moderate exercise. Moreover, impairment of arterial load during exercise is related to lower exercise performance in our study. This relationship between aortic stiffening and impaired exercise tolerance was previously demonstrated by Hundley et al. (9) in older individuals with HFpEF. These modifications are revealed despite the frequent use of angiotensin-converting enzyme/angiotensin receptor blocker agents in HFpEF patients, suggesting a more important role of these drugs in peripheral than proximal arteries. These results could suggest a different composition of the arterial wall in HFpEF patients and controls, but no significant difference was found between the groups in our study in terms of YEM. However, if YEM is theoretically more representative of the intrinsic properties of the arterial wall, its interpretation depends on accepting strict homogeneity of the studied material, an assumption that cannot be made with human arteries. Therefore, a different composition of arterial wall between the groups can probably not be excluded in this study.
Several factors are implicated in arterial stiffening and include structural changes with, for example, diabetes, renin-angiotensin-aldosterone system activation, metalloprotease imbalance, and advanced glycation end-product accumulation (19). These factors affect LV afterload, ventriculoarterial coupling, and LV relaxation, as well as kidney function as a fluid-retention regulator (7,20).
In this study, HFpEF patients had a rigid ventricular chamber, with an impaired pre-load reserve. These baseline characteristics are related to exercise intolerance, as previously described (5,20). Theoretically, SV increases during low-level exercise via the Frank-Starling mechanism, by the increased EDV with increased venous return, without significant change in the EDP, as in controls when LV chamber compliance is normal (4). In HFpEF patients, the diastolic pressure-volume relationship, described by an exponential equation, implies that in a rigid ventricular chamber, little increase in EDV occurs, with a major increase in EDP. In this study, several noninvasive parameters demonstrate the failure of the ventricle to dilate (less increase in EDV and V20i in HFpEF) and to relax (increase in β, tau, E/E′/EDV), which, when coupled with a failure to reduce afterload, permits the delineation of the complex mechanisms underlying the failure to increase SV, despite the use of contractile reserve. In this setting, exercise limitation in HFpEF appears to be the consequence of a preload reserve failure and of an afterload mismatch, resulting in the use of the left ventricular work more to generate pressure than eject volume. Moreover, adaptation of the LV diastolic function is probably blunted by this overcoming of afterload, as previously described (21).
There are some limitations to this analysis. The small sample size could have led to a low-powered analysis, and small differences in terms of hypertension or diabetes between groups could have been statistically significant in a more powered analysis, especially for multivariate analysis. Confounding effects of medications may not be excluded because they were not withdrawn in this study, even if these medications are usually used in HFpEF patients.
This study provides another contributing variable to the phenotype of HFpEF patients by revealing a major increase in proximal afterload and arterial stiffening with moderate exercise. This contributes to abnormal ventriculoarterial coupling and exercise intolerance.
This study was supported by a grant from the French Society of Cardiology. Dr. Beauvais is a consultant for MSD France, AstraZeneca France, and Menarini France. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- peak velocity of the early diastolic mitral inflow divided by E′
- mitral annular early diastolic velocity
- arterial elastance
- end-diastolic pressure
- end-diastolic volume
- end-systolic elastance
- Ees after normalization for the left ventricular mass and V0
- Peterson's modulus
- heart failure
- heart failure with preserved ejection fraction
- left ventricular
- left ventricular ejection fraction
- 6-min walking test
- pulse pressure
- peripheral vascular resistance
- pulse wave velocity
- stroke volume
- volume at end-diastolic pressure of 20 mm Hg indexed to body surface area
- Young's elastic modulus
- Received April 14, 2008.
- Revision received October 13, 2011.
- Accepted October 17, 2011.
- American College of Cardiology Foundation
- Ho K.K.,
- Pinsky J.L.,
- Kannel W.B.,
- Levy D.
- Swedberg K.,
- Cleland J.,
- Dargie H.,
- et al.
- Kitzman D.W.,
- Higginbotham M.B.,
- Cobb F.R.,
- Sheikh K.H.,
- Sullivan M.J.
- Kawaguchi M.,
- Hay I.,
- Fetics B.,
- Kass D.A.
- Maurer M.S.,
- Burkhoff D.,
- Fried L.P.,
- Gottdiener J.,
- King D.L.,
- Kitzman D.W.
- Lam C.S.,
- Roger V.L.,
- Rodeheffer R.J.,
- et al.
- Hundley W.G.,
- Kitzman D.W.,
- Morgan T.M.,
- et al.
- Kelly R.P.,
- Ting C.T.,
- Yang T.M.,
- et al.
- Chen C.H.,
- Fetics B.,
- Nevo E.,
- et al.
- Chemla D.,
- Hebert J.L.,
- Coirault C.,
- et al.
- Schiller N.B.,
- Shah P.M.,
- Crawford M.,
- et al.
- Ommen S.R.,
- Nishimura R.A.,
- Appleton C.P.,
- et al.
- Burkhoff D.,
- Mirsky I.,
- Suga H.
- Yano M.,
- Kohno M.,
- Kobayashi S.,
- et al.