Background Exercise capacity in patients with hypertrophic cardiomyopathy (HCM) varies despite similar diastolic dysfunction, left ventricular outflow tract (LVOT) obstruction and mitral regurgitation (MR). Pulse wave velocity (PWV), determined by cardiac magnetic resonance (CMR), measures aortic stiffness and is abnormal in patients with HCM in comparison with controls.
Objective To determine potential clinical and imaging predictors of peak oxygen consumption (pVO2) in patients with HCM.
Methods Fifty newly referred patients with HCM (62% men, 44±13 years, 90% receiving optimal drugs, 18% hypertensive) underwent Doppler echocardiography (transthoracic echocardiography (TTE)), cardiopulmonary exercise testing and CMR for symptom evaluation. TTE variables (diastology, post exercise MR and LVOT gradient (mmHg)), pVO2 (ml/kg/min) and CMR variables (PWV (aortic path length between mid- and descending aorta/time delay between arrival of the foot of the pulse wave between two points, m/s), and LV volumetric indices) were measured.
Results After exercise LVOT gradient, MR, deceleration time and pVO2 were 104±52, 1±1, 240±79 ms, and 25±6, respectively. Mean basal septal thickness (cm), PWV, EF, ESV index (ml/m2), EDV index (ml/m2) and LV mass index (g/m2) were 1.9±0.5, 9.3±7, 64%±7, 32±9, 87±17 and 112+36, respectively. Multiple regression analyses showed that only age (β=−0.38, p=0.004) and PWV (β=−0.33, p=0.01) predicted pVO2.
Conclusion In patients with HCM, age and PWV are predictors of pVO2, independent of LV thickness, LVOT gradient and diastolic indices. Aortic stiffness potentially has a role in evaluation of symptoms of patients with HCM.
- Aortic stiffness
- pulse wave velocity
- exercise capacity
- hypertrophic cardiomyopathy
- exercise testing
- cardiomyopathy hypertrophic
Statistics from Altmetric.com
- Aortic stiffness
- pulse wave velocity
- exercise capacity
- hypertrophic cardiomyopathy
- exercise testing
- cardiomyopathy hypertrophic
Hypertrophic cardiomyopathy (HCM) is characterised by progressive left ventricular (LV) hypertrophy and myocardial disarray/fibrosis, with a substantial proportion of patients demonstrating resting or provokable LV outflow tract (LVOT) obstruction.1 2 With progression of disease, there is often a reduction in exercise capacity, thought in large part to be due to diastolic dysfunction, mitral regurgitation (MR) and dynamic LVOT obstruction.3 However, commonly measured parameters of disease progression such as LV hypertrophy and LV ejection fraction (LVEF) have been shown to be poor predictors of symptoms and exercise capacity in such patients.4–6 Also, it is not an uncommon clinical situation to encounter patients with varying exercise capacity despite a similar degree of provokable LVOT obstruction, diastolic dysfunction or MR.
In recent years, increased aortic stiffness in the setting of hypertension and other cardiovascular diseases has been shown to have incremental value in predicting long-term outcomes.7–10 In a recent study, our group demonstrated that patients with HCM had abnormal aortic stiffness (measured as pulse wave velocity (PWV)) as compared with controls using velocity-encoded cardiac magnetic resonance (VENC-CMR).11 However, the association of aortic stiffness with exercise capacity has not been studied in patients with HCM. Peak oxygen consumption (pVO2) is an accurate marker of exercise capacity, which can be used to stratify patient groups at higher risk for adverse cardiac outcomes.12 13 These properties make it a commonly used method to measure exercise capacity in patients with HCM as compared with traditional clinical assessment of functional class.5 We sought to test the potential clinical and imaging (CMR and echocardiographic) predictors of pVO2 in patients with HCM.
This was an observational study of 50 consecutive, newly referred patients with HCM who underwent comprehensive evaluation, including outpatient clinic visit, Doppler echocardiography, CMR and metabolic stress testing within 24 h in 2007–8. HCM was defined by a disproportionately hypertrophied and non-dilated left ventricle in the absence of another cardiac or systemic disease that could produce a similar magnitude of hypertrophy.14 15 Patients unable to undergo entire multimodality evaluation owing to comorbid conditions such as significant pulmonary disease or other disorders (precluding exercise) or presence of pacemakers, defibrillators or aneurysm clips (precluding CMR) were excluded from the study. Patients with an LVEF <50% were not included in the study. We also excluded patients >65 years of age with LVOT obstruction who were deemed to have hypertensive heart disease of the elderly with dynamic LVOT obstruction. This diagnosis was based upon clinical presentation, history of longstanding hypertension and presence of a characteristic sigmoid septum and ovoid LV cavity on echocardiography and CMR.16 17 Clinical, echocardiographic, CMR and metabolic stress test measurements were recorded. All patients belong to a registry which is approved by local institutional review board with waiver of individual informed consent.
Resting and stress echocardiography
Transthoracic echocardiography was performed using commercially available HDI 5000 (Philips Medical Systems, NA, Bothell, Washington, USA) and Acuson Sequoia (Siemens Medical Solution USA, Malvern, Pennsylvania, USA) machines. The mitral inflow velocity pattern was recorded from the apical four-chamber view with the pulsed-wave Doppler sample volume positioned at the tips of mitral leaflets during diastole. Deceleration time, measured as the distance from peak of the E wave in the mitral inflow view to the baseline, and peak velocities of E and A waves were measured (in patients in normal sinus rhythm). Tissue Doppler imaging, including mitral annulus septal E′ velocity, was also measured. LV end-diastolic stiffness was non-invasively estimated as follows: pulmonary capillary wedge pressure (3.2 + (1.1 × E/E′))LV end-diastolic volume (measured on CMR, as described below).18 19 Grades of diastology were assigned to each patient, based upon multiple standard criteria, including mitral inflow Doppler pattern, pulmonary venous inflow Doppler pattern, tissue Doppler data, left atrial size and LVEF.20 All diastolic data were acquired over 10 consecutive beats using sweep speeds of 50 and 100 cm/s in a standard fashion. The myocardial performance index was calculated using a previously described and validated marker of systolic and diastolic function measured in each patient as follows: ((isovolumic contraction time+isovolumic relaxation time)/aortic ejection time).21 All Doppler measurements were made over three cardiac cycles and averaged. Resting LVOT peak velocity was measured by continuous-wave Doppler echocardiography, and resting unprovoked LVOT pressure gradient was estimated by using a simplified Bernoulli equation.22 Care was taken to avoid contamination of the LVOT waveform by the MR jet. Peak provoked LVOT gradients were measured in those patients with unprovoked LVOT gradients <30 mmHg after provocative manoeuvres, including Valsalva and amyl nitrite, had been implemented. Peak post-exercise LVOT gradient was measured in the study group in a similar fashion. The degree of resting MR was assessed by colour Doppler and quantified by a combination of visual assessment and flow convergence method,23 on a scale of 0–4+ (0=none, 1+=mild, 2+=moderate, 3+=moderately severe and 4+=severe). Resting and post-exercise wall motion analysis were performed in all patients in a standard manner to assess for myocardial ischaemia.24 The Doppler echocardiographic measurements were made by an independent observer (BAA) without knowledge of patient outcomes and confirmed by the corresponding author (MYD).
Cardiopulmonary exercise testing
All patients underwent maximal, symptom-limited exercise on a treadmill using a standard ramping protocol. Patients were encouraged to exercise until symptoms of fatigue, chest discomfort or dyspnoea prevented further testing. Data were collected during each stage of exercise on symptoms, rhythm and blood pressure. Standard measurements, including pVO2, CO2 production and heart rate, were made at rest and after every minute during exercise and recovery. We also recorded the ratio of pVO2 and predicted VO2 (based on age and gender). The respiratory exchange ratio or ventilatory response to exercise was defined as the value of VCO2/VO2 at peak exercise.25
Cardiac magnetic resonance acquisition and analysis
The CMR examinations were performed on 1.5 T MR Scanner (Avanto, 45 mT/m maximum gradient strength, 200 T/m/s maximum slew rate, Siemens Medical Solutions, Erlangen, Germany). The various CMR sequences used to image patients with HCM has been described previously by our group.11 26 27 Briefly, following scout imaging to locate the cardiac axes, non-breath-hold black blood prepared HASTE (Half Fourier Acquisition in Steady State) images were obtained, followed by a comprehensive assessment of cardiac volumes in multiple short- and long-axis orientations using a balanced steady-state free precession sequence. For patients who could suspend respiration, breath-hold duration was 10–15 s, depending on the heart rate; otherwise, images were acquired using three signal averages. This was followed by acquisition of VENC-CMR images at the level of the pulmonary trunk to measure through-plane flow in the mid-ascending and mid-descending aorta. Finally, the presence and amount of scar was assessed, using phase-sensitive delayed hyperenhancement CMR images.
Maximal end-diastolic basal septal thickness was measured in all patients (MYD). Offline analysis was performed using Argus analytical software (Siemens Medical Solutions) to assess LV volumes and LVEF, in a standard fashion. LV mass index was also calculated. Left atrial area was averaged between two- and four-chamber views. All volumetric parameters were indexed to body surface area. End-systolic and end-diastolic ascending aortic areas were measured on balanced steady-state free precession images and aortic distensibility (10−3 mmHg−1) was calculated using the following formula28: (end-systolic aortic area−end-diastolic aortic area/pulse pressure × end-diastolic aortic area). Subsequently, the contours of the mid-ascending and mid-descending aorta were drawn on VENC-CMR images by the corresponding author blinded to the clinical and echocardiographic data (MYD). The flow (m/s) at these two levels was obtained from the velocity data of each voxel in all phases of the cardiac cycle. From the corresponding flow–time curves, the arrival of the foot of the pulse wave was measured as the point of interception of the linear extrapolation of the steep early systolic slope and the baseline. Multiplanar reconstructions of the axial HASTE images were performed to measure the aortic path length (MYD). The centre line was drawn on a reconstructed sagittal view from the level of the mid-ascending aorta to the mid-descending aorta, corresponding to the same level at which the VENC-CMR image was acquired. The PWV assessed between the mid-ascending and mid-descending aorta was calculated according to the following formula: PWV=Δx/ Δt (m/s), where Δx is the aortic path length between the mid-ascending and mid-descending aorta, and Δt is the time delay between the arrival of the foot of the pulse wave at these levels.29 30 PWV assessment using CMR has been recently demonstrated to be highly reproducible.11 Finally, scar was measured and graded as none, mild, moderate and severe, as previously described.27 CMR assessment was blinded from all other analyses.
Baseline demographics, risk factors and clinical variables are descriptively summarised. Continuous variables are expressed as mean±standard deviation. Interquartile ranges are presented, where appropriate. Categorical data are presented as percentage frequency. Differences between groups were compared with the use of the Student t test and analysis of variance for continuous variables and the χ2 test for categorical variables. The association between pVO2 and pVO2/predicted VO2 ratio and potential predictors was initially performed using univariable linear regression analysis. Subsequently, predictors with a p value of <0.10 were selected and forward-stepwise multivariable regression analysis was performed. Because PWV and aortic distensibility are two different methods of measuring aortic stiffness, only one parameter was entered in the final multivariable model. Receiver operator characteristic curve analysis was performed to assess the ability of PWV to predict median pVO2. Data assembly and basic statistical comparisons were performed with JMP Software version 6.0.2 (Cary, North Carolina, USA). Advanced statistical analysis was performed using SPSS, version 11 (Chicago, Illinois, USA). A p value <0.05 was considered significant.
The study population consisted of 50 consecutive patients with HCM (62% men), of which the majority (90%) were receiving maximally tolerated doses of optimal drugs (including β blockers and calcium channel blockers). Mean respiratory exchange ratio was 1.1±0.09, reflective of adequate exercise effort to calculate pVO2. Four patients had a blood pressure drop >10 mmHg at peak exercise. Median pVO2 was 25 ml/kg/min (lowest to highest range 15–42 ml/kg/min). A significantly higher proportion of women and older patients had pVO2 ≤median. At the time of image acquisition, no patients were in atrial fibrillation. Clinical and stress testing characteristics are shown in table 1.
In the study population, only 40% had an elevated (>29 mmHg) resting unprovoked LVOT gradient. A vast majority (80%) of patients had an elevated (>49 mmHg) LVOT gradient following amyl nitrite. Similarly, 72% patients had an elevated (>49 mmHg) LVOT gradient at peak stress. Also, no patients had evidence of ischaemia on stress echocardiography. As shown in table 2, there was borderline difference in E/E′ and LV stiffness between the two subgroups. The only parameters that were significantly different between groups were PWV and ascending aortic distensibility. Figures 1 and 2 represent examples of two patients with similar degree of basal septal hypertrophy, LVOT gradients and diastology, but with significant differences in exercise tolerance and PWV.
Subsequently, we performed univariable and multivariable regression analysis, testing the association between pVO2 and various predictors. The results are shown in table 3.
Only age and aortic stiffness were significantly associated with pVO2 (figure 3). Of note, in our study population, there was no correlation between age and PWV (r=0.11, p=0.4). For the multivariable regression analysis, only PWV was entered into the model as a determinant of aortic stiffness as it is independent of geometric assumptions, unlike ascending aortic distensibility. However, ascending aortic distensibility was also significantly associated with pVO2 on univariable analysis (β=0.37, p=0.008). Finally, using the median cut-off point for pVO2, we tested its association with PWV using receiver operating characteristic curve analysis. PWV was significantly associated with pVO2 (area under the curve=0.70, p=0.001, figure 4).
We also tested the association between various clinical and imaging parameters and the pVO2 /predicted VO2 ratio. The results are shown in table 4. In this group of patients, only markers of aortic stiffness remained significant predictors of this ratio.
Finally, we also tested the various associations versus pVO2 in a subgroup of patients without documented hypertension (n=41). There remained a significant univariable association between VO2 and PWV and ascending aortic distensibility (both β −0.46, p=0.003). However, the association between age and pVO2, although significant, was weaker (β=−0.35, p=0.03). No other variables were significant. On multivariable regression analysis, PWV (β −0.41, p=0.005) was highly significant and age (β −0.29, p=0.04) remained weakly significant.
To the best of our knowledge, this study is one of the first to demonstrate that aortic stiffness (measured non-invasively using CMR) is an independent predictor of exercise capacity in patients with documented HCM. The results are similar to a recent study of non-ischaemic cardiomyopathy, which demonstrated that there is a significant association between exercise capacity and aortic stiffness, measured by ultrasound.31 However, unlike that study, there was no significant association between pVO2 and LV diastolic variables, probably owing to substantially different patient populations. We chose pVO2 as the exercise end point, as it is a highly accurate marker of exercise capacity and can be used to stratify patient groups at higher risk for adverse cardiac outcomes.12 13 Measurement of exercise capacity is significantly more accurate than the traditional assessment of the functional class or metabolic equivalents.5 With progression of disease, a significant proportion of patients with HCM demonstrate reduced exercise capacity. This is traditionally thought to be at least partially due to diastolic abnormalities, LV stiffening and dynamic LVOT obstruction.3 However, patients with HCM with similar degrees of LV hypertrophy, diastolic dysfunction or LVOT obstruction often have significant variation in their exercise capacity.4–6 Indeed, in our study, a wide range of pVO2 was achieved despite similar provokable/stress LVOT gradient. Hence, it is apparent that there are additional, underappreciated factors influencing exercise capacity in such patients. In recent years, aortic stiffness has been shown to provide incremental prognostic value in some populations, including hypertensive patients and those with coronary artery disease.7–10 In a recent study, our group showed that patients with HCM had increased aortic stiffness compared with normal controls.11 Hence, abnormal vascular function may be a novel parameter for risk stratification in patients with HCM. However, its association with exercise capacity, especially in HCM, has not been previously described.
In this study, we demonstrate that aortic stiffness, measured using CMR, and age were strongest predictors of exercise capacity in this group of consecutively referred patients with HCM who underwent multimodality imaging. This association is independent of standard ventricular parameters, including diastolic function, LV stiffness, myocardial performance index, LV volumes and degree of myocardial fibrosis. The predictive variables did not change even after excluding patients with documented hypertension. In order to evaluate the differences in the population for their clinical and imaging characteristics versus their exercise capacity, we also divided the population into two subgroups, based upon median pVO2. As expected, women and older patients had a lower pVO2 than men and younger patients. However, there was no difference between the groups in other clinical and traditional imaging characteristics. In the subgroup with pVO2 ≤median, aortic stiffness parameters were significantly worse (i.e. higher PWV and lower ascending aortic distensibility) than in those with pVO2 >median. To further reduce the influence of age and gender, we also tested the associations against the pVO2/predicted O2 ratio. Indeed, the only significant parameters with a strong association with this ratio were measures of aortic stiffness.
In this study, there was no significant association between age and aortic stiffness. This is unlike previous reports demonstrating a strong association between the two.7–10 However, in our study population, in order to include only patients with true HCM, we excluded patients with dynamic LVOT obstruction who were >65 years of age and were deemed to have hypertensive heart disease of the elderly.16 17 This artificially limited the upper age range, as compared with other publications where this was not a concern. Also, in our study, there was no association between myocardial fibrosis and PWV (as demonstrated in our previous study11) or exercise capacity. As discussed above, in this study, as compared with the previous study,11 we excluded elderly patients >65 years of age who had hypertensive heart disease of the elderly. The characteristic fibrosis pattern seen in true HCM is not as often seen in hypertensive patients.32 Hence, limiting the analysis to patients with true HCM only (as also reflected by a much younger average age of the current study population compared with the previous study11), significantly reduces the likelihood of finding significant association between fibrosis and other parameters. Finally, this study, unlike a previous report,3 demonstrated that there was no association between LV stiffness and exercise capacity. Again, this was probably due to differences in patient population between the two studies. Our population was much younger, had a higher proportion of patients with New York Heart Association class ≥II, and a higher proportion of patients with an elevated resting or peak LVOT gradient.
We calculated aortic stiffness using two different measurement techniques—aortic distensibility and PWV. Both had a significant association with exercise capacity on univariable analysis. However, we did not include aortic distensibility for the multivariable analysis because of its reliance on area and blood pressure measurement. In a clinical setting, it is usually not feasible to have invasively measured central pressure and we have to rely on non-invasive methods like arm-cuff pressure acting as a surrogate of central aortic pressure. This technique is a fairly imprecise approximation, as pressure amplification may be a significant hindrance to equating peripheral to central pressure.33 Further, this technique assumes a lack of haemodynamically significant stenoses between the central and peripheral vasculature. In addition, cardiovascular risk correlates better with central rather than peripheral pressure.34 In contrast, PWV is a highly reproducible marker of aortic stiffness without the central pressure assumption.35 PWV assessment by CMR has inherent advantages over ultrasound or tonometric techniques. It allows visualisation of the anatomy of vessels in any plane (enabling precise measurement of propagation distance), as opposed to other imaging techniques that are subject to errors in determining the aortic length between measurement sites, particularly in the case of tortuous aortas. VENC-CMR sequences can reliably and non-invasively measure blood flow in any direction or orientation.30 36 However, it needs to be recognised that the temporal resolution of VENC-CMR (which is in the 30 ms range) is relatively low in comparison with other invasive (and some non-invasive) techniques. Also, because it is part of a routine comprehensive CMR evaluation, potentially useful information about aortic stiffness can be obtained in a truly non-invasive manner, with no additional cost, time or equipment.
A definitive pathophysiological link between LV function and aortic stiffness, especially in the context of HCM, has not yet been established. However, there are many potential explanations. Indeed, increased aortic stiffness has been linked to structural alterations (including increased collagen content37) of the arterial wall and rearrangement of its three-dimensional structure.38 HCM is characterised by myocardial fibre disorganisation, which leads to the speculation of a common pathway reflecting the phenotypic characteristics of the same disease process. Recent reports have also implicated neurohumoral pathways such as the renin–angiotensin system or plasma norepinephrine (active in the setting of heart failure), which may cause vasoconstriction and sodium retention in the vessel wall, resulting in increased stiffness.39 40 In another recent report, aortic stiffness has also been associated with diastolic function in end-stage renal failure (also resulting in significant LV hypertrophy), probably through the rise of the LV afterload.41 Indeed, aortic stiffness might contribute to exercise capacity in several ways. Increased stiffness affects the proximal aorta-cushioning effect, thus increasing LV afterload and further compromising the ability to generate adequate cardiac output during exercise.42 This probably has important implications for ventricular–arterial coupling, which is at the core of LV performance and cardiac energetics. Alteration of cardiac performance by the arterial system (in the form of increased stiffness) probably affects stroke work and energy efficiency, hence adversely affecting exercise capacity. Indeed, our group has recently demonstrated an association between ventricular–arterial coupling and exercise capacity in a group of patients with severe ischaemic cardiomyopathy.43 Finally, in HCM, aortic stiffness also probably contributes to a reduction in exercise capacity via the inability to generate adequate cardiac output and reduction in skeletal muscle perfusion during exercise.44 45
From a clinical utility standpoint, if aortic stiffness is associated with exercise capacity in patients with HCM, it might potentially become a target for therapeutic intervention. PWV assessment during CMR is easily obtained and highly reproducible, with the potential to play an important part in further symptom evaluation of patients with HCM. However, prospective studies evaluating the role of PWV on long-term outcomes need to be conducted.
This study is limited by a relatively small sample size; however, the elaborate multimodality imaging evaluation which was a prerequisite for entry precludes a significantly larger sample without substantial extension of the enrolment period. The sample size was further limited both owing to the exclusion of patients >65 years of age, as well as standard CMR contraindications. Our institution is a large tertiary care centre which results in the possibility of a selection bias, as there is a chance that only ‘sicker' patients were referred to us. Hence, the data cannot be extrapolated to the entire spectrum of patients with HCM, including a sizeable proportion of asymptomatic patients in the very early phase of their disease process. However, only about one-third of these patients required definitive treatment (ie, surgery).
In the final study population, we did include a small proportion of patients with a known diagnosis of hypertension, in whom the degree of hypertension was not deemed sufficient to cause the extent of evident LV hypertrophy. We did not analyse torsion and strain as part of the echocardiographic assessment, as it was not uniformly performed in all patients. Data on A-wave duration and pulmonary vein retrograde flow (a marker of LV end-diastolic pressure) were not uniformly available in the study population, and hence not included.18 The potential methodological limitations related to assessment of PWV by VENC-CMR were discussed earlier. This study is hypothesis generating and the findings need to be validated in a larger, possibly multicentre, cohort. The association between PWV and exercise capacity is relatively modest, probably reflecting the fact that exercise capacity in HCM has multiple determinants. Also, this study tests associations, not causality. Longer-term follow-up studies ascertaining the ability of aortic stiffness to predict outcomes in patients with HCM are necessary.
In patients with HCM referred for symptom evaluation, aortic stiffness, measured non-invasively on CMR, predicts exercise capacity independent of LVOT gradient, septal hypertrophy, volumetric indices and diastolic parameters. PWV assessment might play an important part in further symptom evaluation of patients with HCM. However, prospective studies evaluating the role of PWV on long-term outcomes are needed.
Funding No individual disclosures. The institution receives modest research support from Siemens Medical Solutions.
Competing interests None.
Ethics approval This study was conducted with the approval of the Cleveland Clinic.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.