Changes in left atrial deformation in hypertrophic cardiomyopathy: Evaluation by vector velocity imaging

Abstract: Objectives: Hypertrophic cardiomyopathy (HCM) represents a generalized myopathic process affecting both ventricular and atrial myocardium. We assessed the global and regional left atrial (LA) function and its relation to left ventricular (LV) mechanics and clinical status in patients with HCM using Vector Velocity Imaging (VVI). Methods: VVI of the LA and LV was acquired from apical four- and two-chamber views of 108 HCM patients (age 40 ± 19years, 56.5% men) and 33 healthy subjects, all had normal LV systolic function. The LA subendocardium was traced to obtain atrial volumes, ejection fraction, velocities, and strain (ϵ)/strain rate (SR) measurements. Results: Left atrial reservoir (ϵsys,SRsys) and conduit (early diastolic SRe) function were significantly reduced in HCM compared to controls (P < .0001). Left atrial deformation directly correlated to LVϵsys, SRsys and negatively correlated to age, NYHA class, left ventricular outflow tract (LVOT) gradient, left ventricular mass index (LVMI), LA volume index and severity of mitral regurge (P < 0.001). Receiver operating characterist was constructed to explore the cutoff value of LA deformation in differentiation of LA dysfunction; ϵsys < 40% was 75% sensitive, 50% specific, SRsys < 1.7s− 1 was 70% sensitive, 61% specific, SRe> − 1.8s− 1 was 81% sensitive and 30% specific, SRa> − 1.5s− 1 was 73% sensitive and 40% specific. By multivariate analysis global LVϵsys and LV septal thickness are independent predictors for LAϵsys, while end systolic diameter is the only independent predictor for SRsys, P < .001. Conclusion: Left atrial reservoir and conduit function as measured by VVI were significantly impaired while contractile function was preserved among HCM patients. Left atrial deformation was greatly influenced by LV mechanics and correlated to severity of phenotype.


INTRODUCTION
Hypertrophic cardiomyopathy (HCM) is a familial disease with an autosomal dominant pattern of inheritance caused by mutations in genes encoding for sarcomeric proteins. The condition usually results in a hetrogenous myopathic disease affecting both the ventricular and atrial myocardium [1,2]. It is thought to be a progressive disease that most often begins with left ventricular (LV) diastolic dysfunction and/or structural remodeling of the atria, including chamber enlargement and interstitial fibrosis [3]. Better understanding of atrial structure and function could lead to improvements in identifying early signs of diastolic impairment, heart failure and to enhanced prediction of risk related to the development of atrial fibrillation (AF). Such enhanced knowledge would also allow practitioners to better estimate the response to treatments in patients with this arrhythmia.
Noninvasive assessment of structure and function of the atria has been limited by a lack of suitable methods for making these measurements. Left atrial (LA) dysfunction is mostly diagnosed by mitral and pulmonary vein Doppler echocardiography. A complementary method is tissue Doppler imaging (TDI), which more directly estimates myocardial tissue velocities and thus provides a relatively load-independent measure of function [4]. The addition of maximal LA volume may improve diagnostic accuracy [5].
The atrium has multiple functions, acting as a reservoir and a conduit in addition to its contractile function. Thus, there is a need for a more detailed analysis of its pathophysiological importance, and accordingly, for techniques that may supplement available technology in identifying early signs of mechanical impairment.
Recently, 2-dimensional (2D) strain and strain rate (SR) measurements from grayscale images have been introduced based on speckle tracking (SPT), a method in which ultrasound speckles within the image are tracked and strain is determined from the displacement of speckles in relation to each other. The method provides an angle-independent parameter of myocardial function [6,7]. A new feature-tracking echocardiographic method using vector velocity imaging (VVI) is achieved through the combination of SPT, mitral annulus motion, tissue-blood border detection, and the periodicity of the cardiac cycle using R-R intervals. It can measure myocardial strain, SR, and velocity of the regional endocardium [8]. Because VVI tracks moving tissue the area or volume changes in the investigated heart chamber can be calculated automatically frame by frame. 2D-derived LA strain rate facilitates comprehensive evaluation of LA contractile, reservoir, and conduit function.
The aim of the present study was to explore left atrial myocardial properties using VVI as a method to quantify and detect global and regional LA dysfunction and its relation to LV mechanics and clinical status in patients with HCM.

PATIENTS AND METHODS Study population HCM group
Between January 2011 and March 2012, we prospectively included 108 HCM patients between the ages of 8 and 70 years, who were referred to our echocardiographic laboratories for risk stratification. They were examined in a single center (Yacoub Research Unite, Menoufiya University, Egypt and as a part of the BA-HCM National Program). Patients were enrolled in the study after their informed consent, and approval of Ethics Committee of Menoufiya University Hospitals was obtained.
The diagnosis of HCM was based on conventional echocardiographic demonstration of a nondilated, hypertrophic LV ($ 15 mm) in the absence of other cardiac or systemic diseases capable of producing the magnitude of hypertrophy evident, [2] all patients have sinus rhythm and normal ejection fraction (EF% . 55%). Exclusion criteria were diabetes mellitus, arterial hypertension, and evidence of coronary artery disease, atrial fibrillation, lung disease, and inadequate echocardiograms.

Control group
We studied 33 age-and sex-matched healthy subjects without detectable cardiovascular risk factor or receiving any medication. Volunteer controls were all selected from departments of pediatric and adult cardiology among subjects investigated for work eligibility.

Conventional echocardiography
Echocardiographic exams were performed in the left lateral decubitus, in the parasternal long, short-axis, apical two-and four-chamber views using standard transducer positions. Esaote Mylab Gold 30 ultrasound system (Esaote S.p.A, Florence, Italy) equipped with a multi-frequency 2.5-3.5 MHz phased-array transducer was utilized [8,9]. Left ventricular end diastolic (EDD), end systolic diameter (ESD), septum(ST), posterior wall thickness (PWT), ejection fraction (EF%) and LA diameter and volume were measured in accordance with the recommendations of the American Society of Echocardiography [2]. Color flow mapping and continuous-wave Doppler were used to define resting left ventricular outflow tract obstruction (LVOTO) and to estimate pulmonary artery pressure (PAP) from tricuspid regurgitation velocity (Bernolli equation). Peak early (E) and late (A) transmitral (E and A) filling velocities were measured from mitral inflow velocities. Peak systolic (S m ), early diastolic (E m ) and atrial diastolic (A m ) velocity as well as isovolumetric relaxation time (IRT) were obtained by placing a pulsed tissue Doppler (TDI) sample volume at the lateral mitral annulus in the apical four-chamber view. From this, the E/E m ratio was calculated.

Analysis of LA and LV deformation
Border tracking of the LA and LV was manually traced from the digitized 2D video clips recorded during breath holding and with good quality ECG signals, which were acquired and stored for off-line analysis using XStraine software with a frame rate between 40-80 fps. The ''Zoom/RES'' feature on the echocardiographic machine was used to improve the accuracy of atrial measurements. A circular region of interest was traced on the endocardial cavity interface of the apical four-chamber view at end diastole (LA minimum cavity area) using a point-and-click approach. Time-volume curves were extracted from LA wall tracking that provided automatically indexed maximum and minimum LA volume and left atrium ejection force (LAEF) [5,6]. We measured longitudinal (LNG) peak velocities achieved by LA walls 1 cm above the mitral annulus in systole (S am ), early (E am ) and late diastole (A am ).
Definition of the LA endocardial border enabled the system to calculate regional longitudinal deformation of the LA walls. Peak systolic strain (esys) and LA systolic SR (SR sys ) were measured as a positive curve at LV systole (representing reservoir function), early diastole (SR e ) (representing conduit function), and atrial diastole (SR a ) (representing contractile function). Image processing algorithm automatically subdivides the atrial wall into 12 segments distributed in septum and lateral and posterior LA wall -"roof". The graphs for each segment were displayed and averaged to calculate global LA function [10] (Figure 1, 2).
For LV deformation, the same parameters were measured for the basal, mid and apical segments of the septal, lateral, anterior and inferior wall, from apical four and two-chamber views. To reduce random noise, each sample was obtained by averaging three consecutive heart cycles. To estimate mechanical dyssynchrony, the index of myocardial systolic activation was calculated from regional strain curves for each segment, as time from the beginning of a Q wave of ECG to the peak LNG esys (TTP). Left atrial and LV electromechanical delay were measured as the difference between TTP (d-TTP) in 12 LA and 12 LV segments, respectively (difference between the longest and shortest cycle) [10,11]. Intra-atrial and intra-ventricular dyssynchrony were defined as the standard deviation of the averaged time-to-peak strain (TTP-SD) [12,13].
Inter and Intra-observer variability Two independent observers performed two separate quantitative esys and SR analyses of LA and LV images blindly in 35 participants. Inter-observer and intra-observer agreement for esys data was: for LV esys, R ¼ 0.87 and 0.92, LA esys, R ¼ 0.89 and 95, TTP, R ¼ 0.89 and 0.95, respectively; and SR sys, inter-observer R ¼ 0.88, intra-observer R ¼ 92. Both inter and intra-observer agreements were lower for diastolic SR. For SR e : R ¼ 0.84 and 0.87, respectively and for SR a : R ¼ 0.82 and 0.85, respectively.

Statistical analyses
Data were presented as numbers (%) or as mean and standard deviation values. The distribution of qualitative variables was analyzed by Chi-Square test or Fisher's exact test. Quantitative variables were correlated by the use of Pearson's correlation coefficient "r." All tests were two-tailed, and a P value , 0.05 was considered statistically significant. To identify significant independent predictors of global LA strain and SR, variables that were statistically significant in univariate analysis were introduced in a multivariate regression model; the overall fitness of the model was evaluated with the calculation of the coefficient R þ SE. Receiver operating characteristic (ROC) curve analysis was performed to select optimal cut-off values of LA deformation measurements. The analysis was performed by the IBM SPSS statistics software package [19].

Conventional echocardiographic analysis
There was no significant difference between the two groups in LV EF%, E m inflow velocity. Left atrial dimension, volume index, ST and PWT, LVMI and LVOT gradiant, A m inflow velocity, were significantly greater, whereas left ventricular end-systolic diameters (LVESD), left ventricular end diastolic diameter (LVEDD), E m /A m were significantly reduced in HCM group (P, .001). E/E m was significantly elevated when compared to control (P, .001) ( Table 1). (Table 1) In HCM, 2D strain analysis detected lower global and regional peak LA e sys , SR sys and SR e (P,.001) at the level of all analyzed segments in comparison to control. Despite the significant difference of SR a at some segmental levels, LV global atrial diastolic SR did not differ from control. Similarly, electromechanical delay (d-TTP) and intra-V dyssynchrony was considerably prolonged between LV segments compared with its corresponding segments in healthy individuals (P, .001).

LV deformation analysis
Regional and Global LA function (Table 2) As compared to healthy subjects, HCM patients had lower LNG peak LA wall velocities S am , E am and A am P, .001. Similar findings were observed for regional and averaged LA wall esys (P, 0.001) and for reservoir and conduit function derived by SR measured during the systolic and early diastolic period (P,.001). Contrarily, LA contractile function, as expressed by SR measured during late diastole, remains conserved and did not differ between the studied groups (P, .0001) ( Table 4). Left atrial volumes were significantly increased (P, .01, .003) and left atrial ejection force (LAEF) estimated by time volume curve was significantly reduced in HCM compared with control (P, .001).
For electromechanical delay between LA segments, controls showed homogeneous systolic activation of the atrial walls. Conversely, the HCM group showed significant delay between segments (d-TTP; P, .001) and dispersion of electromachanical activation between LA segments (TTP-SD: 46.5^38.2 vs. 29.9^21, P , .001).
To explore the cutoff points that discriminate LA dysfunction, we constructed ROC curves for LA esys, SR sys ,SR e and SR a in HCM (Figure 7). For atrial reservoir function; LA esys , 40% shows 75% sensitivity and 50% specificity respectively. AUC

DISCUSSION
The central finding of our study is a quantitative assessment of LA function in patients with HCM. The LA reservoir and conduit function derived by VVI is significantly compromised while LA contractile function is preserved in HCM with normal ejection fraction. Global LA 1 sys and SR is strongly related to functional class and severity of phenotypic expression. Moreover, LV SR sys and septal thickness are independent predictors to LA 1 sys in patients with HCM. Atrial function is an integral part of cardiac function that is often neglected. The recognition of the upper limits of atrial size and function in HCM may be of clinical relevance by assisting in distinguishing cardiac remodeling and help with patients' risk stratification [1,2]. In the present study we applied VVI, [13,14] that allows for a multidirectional analysis of myocardial motion in an angle-independent fashion, to explore the possibility of improving the diagnosis of early LA dysfunction beyond that accomplished by traditional Doppler echocardiography and TDI. The results of the present study demonstrate the usefulness of VVI in analyzing LA myocardial function in patients with HCM. In this study, besides LA volume measurements, which showed significant increase in HCM, and LA strain analysis, phasic LA SR was used to describe the three components of atrial function that provided incremental information pertaining to LA function [15].

LA function in HCM
The LA operates as a reservoir and conduit compartment for blood flow from the pulmonary veins to LV and as a contracting chamber enhancing LV filling [13]. Left atrial reservoir function is critical for LV    filling by conserving energy during ventricular systole, emitted after MV opening [16,17]. This phase is influenced by atrial compliance, atrial contraction, and LV basal systolic descent [17,18]. During the LA reservoir period, maximal positive LA strain occurs at the end of LV systole, representing a measure of the maximal stretching of LA. Left atrial peak lengthening during ventricular systole-represented by positive strain and SR-are important indicators of LA compliance [16 -19]. Several authors have recently shown that LA esys, could be used as an index of LA reservoir function [20][21][22]. In view of that, our study verified that LA esys was significantly correlated to LA SR sys and gave similar relationship to LV function and LV phenotype.
In our cohort, quantification of longitudinal myocardial LA deformation during this phase showed significant reduction in HCM, which was strongly related to LA volume index, LV mass index in addition to LV systolic deformation as measured by global LV esys and SR. Meanwhile, LA reservoir was inversely correlated to severity of mitral regurgitation. As previously documented, chronic MR provokes volume overload, LA remodeling and impairment of LA elastic properties and compliance with subsequent elevation of LA pressure and could intensify the deterioration of reservoir function in HCM patients, [22][23][24][25] as was the case in the current study in which 98% displayed mild to severe mitral regurgitation.
Left atrial esys decreased linearly with the increasing severity of LV diastolic dysfunction, as expressed by E m , A m using TDI and LV filling pressure estimated by E/E m . Consequently, LA esys emerged as a promising variable for the expression of the intrinsic LA function when investigating patients with different stages of diastolic dysfunction. In the case of HCM, myocardial hypertrophy, disarray and fibrosis are associated with increased stiffness or noncompliance of the LV, and LA pressure rises to maintain adequate LV filling. The resultant increased atrial wall tension leads to chamber dilation and stretch of the atrial myocardium [26,27]. This explains why, early in the course of disease, LA stiffness increases whereas the LA booster pump is augmented causing work mismatch. This is could explain the preserved contractile function (SR a ) in our studied HCM population and might imply incipient LV dysfunction in patients with apparently normal LV ejection fraction.
Left atrial conduit function, which reflects passive LV filling in early diastole, is predominantly governed by the rate of LV relaxation [15,16]. In this study, evidence of impaired conduit function in our HCM patients was manifested by decreased transmitral E wave and mitral annular E m velocity. Using VVI indices, passive emptying LA velocity (E am ) and SR during early diastole (SR e ) showed significant reduction compared with the controls. This reduction in conduit function is worse with increased severity of hypertrophy and presence of LVOTO and is associated with pulmonary hypertension. It is firmly related to LV diastolic function as measured by deformation indexes.
Atrial booster pump function, which reflects atrial contraction at end-diastole, and is seen as negative SR, is controlled by preload, afterload, and contractility [28,29]. In our study of HCM patients, indexes of LA booster pump function were preserved and did not differ from control. In addition to LA SR a , LA systolic function has been evaluated utilizing a variety of indicators, such as transmitral Doppler flow (A), and TDI of the mitral annulus during atrial systole (A m ) which showed augmented values in relation to control. Nonetheless, the main inconvenience in the later parameters was their high sensitivity to autonomic status, loading conditions, and LV systolic function [9,10].
The present study clarified that the deformation in the atria is reciprocally related to the deformation of the ventricles, both reflecting the motion of the atrio-ventricular plane. In systole, the ventricle shortens while the atria expand. This is a function of ventricular contraction. In early diastole, there is elongation of the ventricles and shortening of the atria; the active constituent of this is the ventricular relaxation. While in late diastole, the active component is the atrial contraction. In our HCM cohort, LA  reservoir and conduit function deteriorated due to reduction of LV deformation as measured by VVI. At the same time, LA contractile function remained intact in patients without evidence of heart failure.

LA functions in previous studies
Previous reports have already noted that newer applications of echocardiography, such as strain imaging, may be a useful technique to evaluate LA myocardial function in different pathological states (13,22,30,31).
In agreement with our results, Roşca et al [32] investigated 34 HCM with speckle tracking and reported that peak LAe and SR parameters were significantly reduced and related to LV dysfunction. Left atrial booster pump function emerged as an independent correlate of heart failure symptoms.
Paraskevaidis et al [27] quantified LA longitudinal function by TDI and 2D strain in 43 HCM patients-21 patients with non-HCM LVH, and 27 healthy volunteers. Left atrial longitudinal function is reduced in HCM compared to non-HCM LVH and healthy controls. Adding 2D contractile atrial strain to atrial diameter and volume index, inter-ventricular septal thickness, and E/A ratio and E/E 0 ratios increased its prognostic value in differentiating HCM from non-HCM LVH (p, .001). The authors concluded that 2D atrial is more reproducible and less time consuming than TDI strain.
Additionaly, Shin et al [33] reported that through using real-time 3D echocardiography the maximal LA volume index was larger, and LA active emptying fraction was lower in 26 HCM patients when compared with control subjects, and the increased LA volume was related to decreased LA contraction in HCM. LA passive emptying was related to LV relaxation whereas LA active contraction was related to LV stiffness.
Myocardial function using VVI Unlike previous methodology, we explored in this study indices of LV systolic and diastolic function using VVI acquired from the LV myocardium. We believe that this approach represents a new paradigm in evaluating LV function because it is based on measurements obtained directly from the ventricular myocardium and not mitral recordings during early LV filling. Therefore, problems related to annular and valvular pathology can be circumvented. Second, we obtained an index that was derived from all LV segments and is therefore more representative of global LV performance than classic EF%.
In this study we evaluated LA function quantification and assessed the characteristics of LA regional segments especially in presence of remodeling and dilation. Left atrial roof necessitates an angleindependent method; in addition, the very thin LA wall posed challenges to pure SPT techniques while VVI simplified the automated endocardial border tracking and resulted in a LA time-volume curve generated from volume measurements at every frame.
In the present study, analysis of LA deformation raises the question about its relationship with other functional parameters; surprisingly we found obvious relationships with TDI parameters but not with mitral inflow velocities, which may indicate on more load-dependent and "autonomic" LA function.

Clinical implications
This study provides further insight into the influence of HCM on mechanical function. Our figures and cutoff values put forward that LA deformation indices could be considered both diagnostic and prognostic adjuncts that facilitate unmasking of incipient myocardial dysfunction in HCM. We recommend serial measurement of LA strain/SR to detect the onset of LA contractile dysfunction and impaired LA compliance, known to take place in more advanced disease. These indices may prove to be useful in treatment decision making; however, the significance of these findings and their possible application will require further study.

STUDY LIMITATIONS
Good quality images are needed as VVI is based on 2D grayscale imagery, leading to exclusion of many patients from the study. The myocardial deformation is 3D but, at present, VVI permits only 2D images. The use of DICOM data with relatively low frame rates might be challenging and less reproducible, which may question the feasibility of this method in standard clinical practice. Only longitudinal function parameters were explored in this study. Vector velocity imaging can potentially quantify circumferential and radial deformation, but we found it difficult to track the acoustic kernels associated with circumferential and radial movement of the thin LA wall from the apical position. Additionally, the influence of myectomy in our cohort, with LVOT obstruction, on LA function has not been studied. The reported LV reverse remodeling following this operation, [34] could, in theory, have an effect on atrial function. Finally, the relatively small sample size may preclude strong statistical inference. However, we excluded patients who had recordings that could not be adequately interpreted; accordingly, the results are representative of a population among whom it is possible to use the VVI technique.

CONCLUSION
Left atrial reservoir and conduit function as measured by VVI-derived strain/strain rate is significantly impaired while contractile function is preserved in HCM patients with normal systolic function. Left atrial deformation is greatly influenced by LV mechanics and correlated to severity of phenotype. Vector velocity imaging has a discriminative power as a single measure to detect LA dysfunction in HCM.