Impaired Right Ventricular Mechanics at Rest and During Exercise Are Associated With Exercise Capacity in Patients With Hypertrophic Cardiomyopathy

Background Impaired right ventricular (RV) function indicates RV involvement in patients with hypertrophic cardiomyopathy (HCM). We aimed to assess RV function at rest and during exercise in HCM patients and to examine the association between impaired RV mechanics and exercise capacity. Methods and Results A total of 76 HCM patients (48 without and 28 with RV hypertrophy) and 30 age‐ and sex‐matched controls were prospectively recruited. RV function was evaluated at rest and during semisupine bicycle exercise by conventional echocardiography and 2‐dimensional speckle‐tracking imaging. Exercise capacity was measured by metabolic equivalents. RV functional reserve was calculated as the difference of functional parameters between peak exercise and rest. Compared with controls, HCM patients had significantly higher RV free wall thickness, lower RV global longitudinal strain and RV free wall longitudinal strain at rest and during exercise, and reduced RV systolic functional reserve. Compared with those with HCM without RV hypertrophy, patients with HCM with RV hypertrophy had lower metabolic equivalents. Among HCM patients, an effective correlation was seen between exercise capacity and peak exercise RV global longitudinal strain and peak exercise RV free wall longitudinal strain. A binary logistic regression model revealed several independent predictors of exercise intolerance in HCM patients, but receiver operating characteristic curve analysis indicated exercise RV global longitudinal strain had the highest area under the curve for the prediction of exercise intolerance in HCM patients. Conclusions HCM patients have RV dysfunction and reduced contractile reserve. Exercise RV global longitudinal strain correlates with exercise capacity and can independently predict exercise intolerance. In addition, patients with HCM with RV hypertrophy exhibit more reduced exercise capacity, suggesting more severe disease and poorer prognosis.

H ypertrophic cardiomyopathy (HCM) is a genetically heterogeneous cardiomyopathy characterized by left ventricular (LV) hypertrophy with a prevalence of %1/500 in the general population and associated with sudden cardiac death, heart failure, atrial fibrillation, and stroke. 1 Given the importance of LV function in the evaluation of HCM patients' clinical status, the vast majority of previous studies performed on HCM patients have focused on the abnormality of LV structure and function but overlooked the involvement of the right ventricle, which may also be dysfunctional in HCM patients. 2,3 Indeed, as previously reported, a large number of HCM patients display right ventricular (RV) dysfunction at rest, which may result from an extension of myopathic process, sarcomere-related protein gene mutations, augment of RV afterload, and/or shared anatomically hypertrophic interventricular septum (IVS). [4][5][6] Although the precise mechanism of RV dysfunction in HCM patients remains to be explored, RV dysfunction in HCM patients may be associated with a higher risk of cardiovascular mortality. 7 Consequently, determination of RV involvement in HCM is important for evaluation of disease status and the prognosis of HCM patients. 8,9 Impaired exercise capacity, namely, exercise intolerance, is the primary clinical feature in HCM patients and may be used to determine risk and prognosis of these patients. 10,11 Although some studies demonstrate impaired RV systolic function at rest in HCM patients, 12,13 few studies have examined RV function during exercise and explored the association of RV function and exercise capacity in HCM patients. 4,14 Exercise stress echocardiography has an important role in predicting the development of symptoms, revealing liable obstruction of the LV outflow tract, and evaluating functional capacity, heart rate (HR), and blood pressure. 15 In addition, exercise stress echocardiography is safe and has become an essential component of the standard management of HCM, as recommended by the current guidelines defined by the European Society of Cardiology. However, given the complex anatomy of the right ventricle, accurate assessment of RV function remains a challenge. A more recently developed imaging technique, 2-dimensional speckle-tracking imaging (2D-STI), allows for earlier detection of subclinical RV dysfunction and is less affected by angle dependency. 16 Therefore, 2D-STI is an effective tool for assessing RV function.
This study was designed to evaluate RV function in HCM patients with and without RV hypertrophy (RVH) at rest and during exercise and to determine the association between RV function and exercise capacity in these patients.

Data Availability
The data that support the findings of this study are available from the corresponding author on reasonable request.

Study Population
A total of 93 patients with HCM who were referred to the Department of Echocardiography, Heart Center, Beijing Chao Yang Hospital, for their risk stratifications from June 2015 to January 2018 were consecutively enrolled in this study, and their RV function was assessed at rest and during exercise. These patients fulfilled the previously published HCM diagnostic criteria, 17 which were mainly based on the echocardiographic manifestation of a maximal LV wall thickness ≥15 mm in the absence of other cardiac or systemic disease that may produce a similar degree of LV hypertrophy. In addition, HCM was also diagnosed for a patient who had a positive family history of HCM and a maximal wall thickness between 13 and 14 mm without other definite causes. Patients with 1 of the following were excluded from our study: a history of coronary artery disease with percutaneous coronary intervention and/or coronary artery bypass surgery, diabetes mellitus, New York Heart Association functional class III or IV, pulmonary hypertension, atrial fibrillation, poor imaging quality, a history of septal reduction therapy with surgical myectomy or alcohol septal ablation, significant valvular disease, or valvular prostheses. After application of exclusion criteria, patients with septal reduction therapy (n=4), coronary artery disease (n=3), valvular prostheses (n=1), and poor image quality (n=9) were excluded. Ultimately, 76 patients were included in this study. In addition, 30 ageand sex-matched healthy participants were recruited as controls. The participants were divided into 3 groups: the control group, the HCM-without-RVH group, and the HCMwith-RVH group. This study was approved by the ethics committee of Beijing Chao Yang Hospital, and written informed consent was obtained from all participants.

Conventional measurements
Conventional echocardiography was performed on all participants to assess their cardiac function at rest and during exercise. At the standard transducer position, images were obtained with the patient at the left lateral decubitus position, using a commercially available ultrasound machine (EPIQ 7C; Philips Healthcare) equipped with an X5-1 multiphase-array probe. Conventional measurements were performed at rest in accordance with the current recommendations. 18 The RV wall was observed from different echocardiographic views to identify the maximal wall thickness without including RV trabeculae. RV free wall thickness (RVWT) was measured at the end-diastole below the tricuspid annulus at a distance approximating the length of the anterior tricuspid leaflet when it was fully open in a zoomed subcostal image with the focus on the RV midwall. 18  clearly displayed, RVWT was evaluated in the left parasternal long-axis view. A value of RVWT >5 mm was defined as the existence of RVH, as recommended. 18 The RV basal maximal transversal dimension was measured at the end-diastole in the RV-focused view. Peak early and late transtricuspid filling velocities were measured at the level of tricuspid tips. Tricuspid annular plane systolic excursion was acquired in the M-mode, and RV fractional area change was calculated as the RV area difference (diastolic-systolic) divided by the RV enddiastolic area. RV tricuspid annular peak velocities, including systolic peak velocity of tricuspid annulus and early and latediastolic velocities, were acquired with pulsed tissue Doppler imaging from the lateral. LV internal diameters at the enddiastole and end-systole, IVS (interventricular septum), and LV posterior wall thickness were measured in the parasternal long-axis 2D view. LV ejection fraction was calculated using the modified biplane Simpson method. LV mass was determined with the transthoracic 3D full-volume data sets obtained from the standard apical 4-chamber view with a frame rate >25 frames/s, and the LV mass index (LVMI) was acquired by indexing for body surface area. Peak early (shown as E) wave velocity of mitral filling and average of septal and lateral mitral annular early-diastolic peak velocity (shown as e 0 ) were measured, and then E/ e 0 was calculated. Left atrial volume index was calculated by the maximum volume of the left atrium indexing to body surface area.

Deformation measurements
The RV-focused apical 4-chamber view was obtained at rest and during exercise by recoding 3 consecutive heart cycles, with the frame rate optimized to 50 to 100 frames/s. 2D-STI offline analyses were performed using the dedicated software QLAB 10.3 software (Philips Healthcare). Three RV anchor points were identified at the lateral and septal tricuspid annulus level and the RV apex by a point-and-click approach aiming the software to automatically track the endocardial contour. If the entire RV myocardial wall was not included in the region of interest, tracking was further adjusted manually for optimization. In addition, the pericardium was excluded in case of underestimation of measured strain. RV global and segmental strain values were obtained after software analysis ( Figure 1). RV global longitudinal strain (RVGLS) was measured from RV free wall and 3 septal (basal, middle, and apical) segments, and RV free wall longitudinal strain (RVFWLS) was obtained from 3 segments of the free wall. RV contractile reserve was defined as the difference in RV strain values between peak exercise and rest.

Exercise stress echocardiography
All study participants underwent multistage, symptom-limited, semisupine exercise testing using a bicycle ergometer (Ergoselect 1200, Stress Echo Couch Ergometer; Ergoline) after the resting echocardiography was performed. Workload began at 25 W and was increased by 25 W every 2 minutes.
Participants were asked to maintain a rate of 55 to 60 rounds/min. The protocol required that b-blockers or calcium channel blockers be withheld ≥24 hours before the exercise stress test. Heart rhythm, HR, and blood pressure were continuously monitored during exercise. The exercise test was promptly interrupted when a patient had 1 of the following conditions: achievement of the age-related maximum HR, presence of significant arrhythmias, severe hypertension (blood pressure ≥240/120 mm Hg), leg muscle fatigue, or symptom intolerance such as severe chest pain, dizziness, and breathlessness. Exercise capacity was defined as metabolic equivalents (METs) and METs <7 was defined as the presence of reduced exercise capacity, as reported previously. 19,20 The following parameters were recorded at peak exercise: maximal HR, maximal systolic blood pressure, METs, and rate-pressure product (maximal HR times maximal systolic blood pressure).

Evaluation of Reproducibility of Echocardiographic Measurements
To assess the data reproducibility, 5 patients were randomly selected from each group at least 1 month after the initial analysis, and the data sets of 15 patients were analyzed by the original investigator and a second experienced echocardiographer who were blinded to each other's measurements.
Inter-and intraobserver variability of exercise RVGLS (RVGLSexe) was evaluated by intraclass correlation coefficients and the coefficient of variation.

Statistical Analysis
Continuous variables are expressed as meanAESD and categorical variables as absolute values and percentages. Normal distribution was evaluated with the Kolmogorov-Smirnov test. Differences between groups were compared with one-way ANOVA for normally distributed data, with the Mann-Whitney U test or Kruskal-Wallis test for skewed data, and with the v 2 test for categorical data. Correlations between different parameters were performed with the Pearson product moment correlation or Spearman rank correlation, as appropriate. Multivariable regression analysis was used to identify independent predictors of exercise intolerance in HCM patients. Receiver operating characteristic curves were used to determine the optimal cutoff values of chosen variables, and the area under the curve was used for predicting exercise intolerance. All tests were 2-sided, and P<0.05 was considered statistically significant. All analyses were performed with SPSS 23.0 (IBM SPSS Statistics, v23) and MedCalc 15.6 (MedCalc Software).

Comparison of Baseline Characteristics and Resting RV Function of Controls and HCM Patients With and Without RVH
The baseline clinical and echocardiographic characteristics of study participants at rest are summarized in Table 1

HCM Patients Exhibited Abnormal RV Function During Exercise
We next examined RV function in participants during exercise. As shown in  HCM without RVH). These findings together suggest that HCM patients, regardless of whether they have RVH, exhibited dysfunctional RV function during exercise and that HCM patients with RVH had more severe RV dysfunction than HCM patients without RVH.

HCM Patients Exhibited More Impaired Hemodynamic Characteristics During Exercise
We next evaluated hemodynamic characteristics of these study participants during exercise. Both groups of HCM patients, with and without RVH, underwent exercise stress and did not show any severe complications (

Determination of the Independent Predictor of Exercise Intolerance
Pearson product moment correlation or Spearman rank correlation was used to assess correlations between exercise capacity, as measured by METs and RV functional parameters. These analyses revealed moderate correlations of METs with resting RVGLS (r=À0.56), RVGLS-exe (r=À0.62), resting RVFWLS (r=À0.53), and RVFWLS-exe (r=À0.59) in HCM patients (P<0.001 for all; Figure 2). In addition, exercise capacity in HCM patients showed a moderate inverse correlation with LVMI (r=À0.46, P<0.05). Other correlations between exercise capacity and RV echocardiographic parameters at rest or during exercise are presented in Table 4. We next used univariate regression analyses to assess the association between exercise intolerance (METs <7) and RV functional parameters, LVMI, and clinical characteristics, and parameters with significant correlations identified by this approach were incorporated into a multivariable logistic regression model to identify independent predictors of exercise intolerance. This analysis revealed that RVGLS-exe was the independent predictor of exercise intolerance (Figure 3).

Identification of Parameters Associated With Exercise Intolerance
Receiver operating characteristic analysis (Figure 4) revealed that RVGLS-exe had better capability to identify patients with exercise intolerance compared with other parameters in HCM patients, with an area under the curve of 0.832 (P<0.05), sensitivity of 63.6%, and specificity of 90.5% ( Table 5).

Determination of Reproducibility
Inter-and intraobserver variability of RVGLS-exe was presented by intraclass correlation coefficients and the coefficients of variation in the subjects randomly selected from each group (n=15). Intraclass correlation coefficients of inter-and intraobserver variability were 0.81 and 0.86, respectively, for RVGLSexe. The coefficients of variation for inter-and intraobserver variability were À5.1AE1.2% and À4.8AE1.1%, respectively, for RVGLS-exe. These observations document high reproducibility of the measurements obtained in our study.

Discussion
To the best of our knowledge, this study is the first to assess the relationship of RV function and its contractile reserve with exercise capacity in HCM patients with or without RVH, both at rest and during exercise, in the same experimental setting. We had 4 main findings. First, RV myocardial function was impaired in HCM patients at rest and during exercise, and RV contractile reserve was reduced during peak exercise, regardless of whether HCM patients had concomitant RVH. Second, HCM patients with RVH had lower RVGLS, RVFWLS, and exercise capacity compared with HCM patients without RVH and controls. Third, RV wall thickness, RVGLS, and RVFWLS at rest and during exercise were associated with exercise capacity. Fourth, RVGLS-exe had the strongest correlation with exercise capacity, even after adjustment for other variables, and was the most powerful predictor for identifying HCM patients with reduced exercise capacity compared with other parameters. Our findings provide a significant reference for the clinical management of HCM patients.

Development of RVH in HCM Patients and Potential Mechanisms
Although HCM is a genetically heterogeneous cardiomyopathy characterized by LV hypertrophy, a number of studies have shown abnormal RV morphology and function in HCM patients Although the exact mechanisms underlying the development of RVH in HCM patients are not clear, multiple factors have been proposed to play a role. A mutation on the MYBPC3 (myosin binding protein C, cardiac) gene, for example, was found to be linked to the reduced synthesis and abnormal assembly of sarcomere-associated proteins, and this affected the phenotypic expression of HCM. 23 Increased calcium sensitivity, which correlates to myocardial hypercontractility, was also shown to contribute to the pathogenesis of RVH. 24 In addition, depletion of myocardial energy, which is fundamental for effective cardiac muscle contraction and efficient cardiac output, was reported to regulate RV remodeling. 25 Further understanding the mechanisms that promote the pathogenesis of RVH in HCM patients will facilitate the design of strategies aimed at diminishing adverse outcomes and improving the quality of life of HCM patients.

RV Diastolic and Systolic Dysfunction in HCM Patients
As mentioned, >50% of HCM patients had RV diastolic and systolic dysfunction. The E/e 0 ratio is a reliable parameter and has been recommended by the current guideline for evaluating RV diastolic function. 18 In the present study, we also found that the RV filling pattern was significantly more restrictive in HCM patients, both with and without RVH, than in controls, as demonstrated by a higher E/e 0 ratio in patient groups. Consistent with those previous studies, our observations suggest impairment in RV diastolic function in HCM patients, regardless of the presence of RVH. Because RV diastolic dysfunction has been considered an independent predictor of adverse outcome in patients with HCM, the risk of death from heart failure for HCM patients with RV diastolic dysfunction and an increased E/e 0 value is increased by 1.6 times. 26 It will be interesting to examine the prognostic value of RV diastolic function in these patients for a longer period of time.
Recently, 2D-STI, which can provide early detection of subclinical myocardial dysfunction, 27 has offered a sensitive method for the study of RV myocardial mechanics at rest and during exercise. In this study, we used 2D-STI to appraise RV systolic function based on RV strain parameters (RVGLS, RVFWLS) and showed impaired RV function in patients with HCM. However, we did not observe any significant differences in conventional echocardiographic parameters such as tricuspid annular plane systolic excursion and RV fractional area change between HCM patients and controls, which was in agreement with findings from a previous study. 28 Several mechanisms have been proposed to be linked to the pathogenesis of RV dysfunction in HCM patients. First, myocardial disarray and interstitial fibrosis may contribute to the progression of RV diastolic or systolic dysfunction. A study based on cardiac magnetic resonance has shown that late gadolinium enhancement, a manifestation of replacement fibrosis, was associated with abnormal RV mechanics. 3 Microcirculatory ischemia, 29 which is mainly attributed to abnormal intramural arteries characterized by intimal proliferation and luminal narrowing, was also correlated with RV dysfunction. It is well established that the RV wall is composed of superficial layer fibers that are aligned circumferentially and deep layer fibers arranged longitudinally, the latter of which are most vulnerable to ischemia. In addition, energy deficits also contribute to the pathogenesis of RV systolic dysfunction. 30

Correlation Between RV Function and Exercise Capacity
RV dysfunction is an important cause of exercise intolerance, which is an independent predictor of adverse outcomes in patients with HCM. 31,32 Exercise stress echocardiography has been proposed to assess exercise capacity in symptomatic HCM patients who failed to induce LV outflow tract obstruction ≥50 mm Hg by bedside maneuvers, according to the current guideline. 17 In our study, HCM patients either with or without RVH had higher RV wall thickness and tricuspid E/e 0 both at rest and at peak exercise compared with controls, indicating the presence of RV remodeling and increased RV filling pressure. These abnormal RV functions may contribute to reduced exercise capacity. Conversely, we showed that increased IVS thickness and LVMI were also correlated with reduced exercise capacity, as measured by METs. Because of the close correlation of exercise capacity with age and sex, we further analyzed these correlations in the multivariable regression model. We found that it was not LVMI; rather, maximal systolic blood pressure during exercise and RVGLS-exe were independent factors affecting exercise intolerance, suggesting that HCM patients with an abnormal BP response to exercise, which was demonstrated as a risk factor of sudden cardiac death in HCM patients, 15 exhibit reduced exercise capacity.
RV strain values were sensitive markers of exercise capacity. 31 Indeed, in the present investigation, we found that RVFWLS and RVGLS, either at rest or during exercise, were strongly correlated with exercise capacity assessed by METs in HCM patients. Moreover, RVGLS-exe was an   independent predictor of exercise intolerance. We also showed that RV strain absolute values were significantly lower in patients with HCM than in controls, further supporting the existence of RV subclinical myocardial dysfunction in HCM patients, as reported by others. 14,28 Although global RV systolic function comprises 3 functional parts-RV longitudinal systolic function, RV radial function, and anteroposterior systolic function 33,34 -80% of the total stroke volume was shown to be dependent on shortening of the longitudinal axis. 33,35 Consequently, we hypothesize that impaired exercise capacity in HCM patients may result from decreased RV longitudinal strain due to subclinical myocardial dysfunction, leading to improper increase in RV contractile reserve during exercise according to the Frank-Starling law. In addition, we found that RVGLS-exe had the highest area under the curve for predicting exercise intolerance among all echocardiographic variables and that an RVGLS-exe cutoff value of À18.4% identified exercise intolerance in HCM patients with 63.6% sensitivity and 90.5% specificity. Moreover, RVGLS was more reliable in predicting reduced exercise capacity in HCM patients compared with other predictors such as RVFWLS.
Mechanistically, because of hypertrophic IVS, the amplitude of ventricular septum bulging into the right ventricle was reduced, and the force of stretching the RV free wall over the septum was weakened during systole, both of which impaired RV pump function. Because a reduction in exercise capacity was an independent predictor of poor prognosis in HCM patients, 31 our identification of RVGLS-exe measured by 2D-STI as an evaluator of exercise capacity points to the potential application of RVGLS-exe in the risk stratification and prognostic assessment of HCM patients.

Clinical Implications
Our findings present a novel understanding of HCM characteristics. RV systolic or diastolic dysfunction may increase risk of cardiovascular adverse events in HCM patients and indicates poor prognosis, as demonstrated by previous studies. 2,26 In clinical practice, we emphasize the importance of the evaluation of RV function in patients with HCM and the rapid identification of patients with RV dysfunction, with the goal of eventually reducing the incidence of clinical adverse events, especially in HCM patients with RVH.

Limitations
Several limitations of this study should be acknowledged. First, it was a single-center observational study with a relatively small sample size; therefore, the inherent sampling bias of this kind study could not be precluded. Second, LVspecific software was used to analyze RV function because no 2D-STI dedicated software was available for RV deformation measurements during the course of this study. Third, as mentioned, global RV systolic function has 3 functional parts, but we evaluated only RV longitudinal systolic function because it contributed to 80% of the RV total stroke volume. Nevertheless, we cannot rule out the possibility that the other 2 RV functions may be compromised in HCM patients with and/or without RVH, and this may need to be explored further. Fourth, the gold standard for imaging evaluation of RV structure and function is cardiac magnetic resonance, which was not performed in our study given its time-consuming nature and high cost. Fifth, in this study, we mainly evaluated the correlation between RV function and exercise capacity and did not examine the effect of LV function on exercise capacity in patients with HCM. Finally, the predictive value of RV strain parameters at peak exercise for predicting the prognosis of HCM patients needs to be further confirmed by future multicenter prospective studies.

Conclusions
In this study, we demonstrated that patients with HCM have reduced RV systolic and diastolic function at rest and at peak exercise and have significantly impaired RV systolic function reserve during exercise. In addition, RV function is associated with exercise capacity in HCM patients, and HCM patients with RVH exhibit more severe reductions in exercise capacity than HCM patients without RVH. RVGLS-exe is an independent predictor of exercise intolerance in HCM patients, indicating that it may be used for risk stratification and prognostic assessment of HCM patients.