|Home | About | Journals | Submit | Contact Us | Français|
Genetic testing identifies sarcomere mutation carriers (G+) before clinical diagnosis of hypertrophic cardiomyopathy (HCM), allowing characterization of initial disease manifestations. Prior studies demonstrated that impaired relaxation develops before left ventricular hypertrophy (LVH). The precise impact of sarcomere mutations on systolic function in early and late disease is unclear.
Comprehensive echocardiography with strain imaging was performed on 146 genotyped individuals with mutations in 5 sarcomere genes. Contractile parameters were compared in 68 preclinical (G+/LVH−), 40 overt (G+/LVH+) HCM subjects, and 38 mutation (−) normal control relatives. All subjects had normal LV ejection fraction (EF). In preclinical HCM, global and regional peak systolic strain (εsys) and longitudinal systolic strain rate (SSR) were not significantly different from controls, but early diastolic mitral annular velocity (Ea) was reduced by 13%. In overt HCM, there was a significant 27% and 14% decrease in global longitudinal εsys and SSR respectively, compared with both preclinical HCM and controls (p<0.013 for all comparisons), and a 33% reduction in Ea.
Sarcomere mutations have disparate initial effects on diastolic and systolic function. Preclinical HCM is characterized by impaired relaxation but preserved systolic strain. In contrast, both diastolic and longitudinal systolic abnormalities are present in overt disease, despite normal EF. We propose that diastolic dysfunction is an early consequence of sarcomere mutations, whereas systolic dysfunction results from mutations combined with subsequent pathologic remodeling. Identifying mechanistic pathways triggered by these mutations may begin to reshape the clinical paradigm for treatment, based on early diagnosis and disease prevention.
Hypertrophic cardiomyopathy (HCM) is caused by dominant mutations in sarcomere genes.1, 2 Although the clinical manifestations are highly variable, HCM can be associated with serious consequences including sudden cardiac death and refractory heart failure. The diagnosis of HCM currently depends on identifying unexplained left ventricular hypertrophy (LVH), however the penetrance of LVH is age-dependent and LV wall thickness is typically normal early in life.3, 4 Genetic testing allows identification of a novel and important population comprised of individuals in families with HCM who have inherited the disease-causing sarcomere mutation and are therefore at high risk for disease development, but who have not yet manifest diagnostic features of HCM, specifically LVH. Better characterization of this intriguing cohort, denoted preclinical HCM, will identify early effects of sarcomere mutations prior to the pathologic cardiac remodeling that defines clinically overt disease.
The precise effects of mutations in sarcomere proteins on contractile function in HCM have not been fully characterized.5 Prior studies on animal models and preclinical human HCM demonstrated that diastolic function is impaired before LVH develops.6–9 While these findings suggest that sarcomere mutations have direct and early consequences on diastolic function, the impact on systolic function has not been fully elucidated.
Unlike traditional measures of systolic function such as fractional shortening or ejection fraction (EF), the new modality of echocardiographic strain analysis offers the advantage of measuring indices directly from the ventricular myocardium, allowing improved quantification of global and regional function.10 Systolic strain (εsys) is a measure of deformation, expressed as the percent change in length of a myocardial segment, relative to end-diastolic length.11 Systolic strain rate (SSR) is an early systolic index which represents the rate of myocardial deformation and correlates with end-systolic elastance and peak dP/dt12. Non angle-dependent speckle tracking techniques (two-dimensional strain, 2DS) allow evaluation of the 3 principal components of cardiac strain— radial, circumferential, and longitudinal.13 Furthermore, echocardiographic strain imaging has revealed systolic dysfunction in patients with cardiac amyloidosis14 or clinically-defined HCM15 despite preserved LV EF.
In this study we use echocardiographic strain imaging to better characterize the early and late effects of sarcomere mutations on systolic function by evaluating genotyped subjects with preclinical and overt HCM.
The study population consisted of genotyped HCM patients and relatives identified via research protocols or clinical evaluation at three centers in Boston, MA (n=63), Copenhagen, Denmark (n=56), and Minneapolis, MN (n=28). Genetic status was determined in all subjects by direct DNA sequencing of sarcomere genes. Families with at least 2 affected generations were recruited for participation and echocardiographic studies were performed on available subjects. Participants were assigned to 3 status groups designated overt HCM, preclinical HCM, and normal control based on genotypic status and LV wall thickness. Since borderline LVH may indicate mild or early disease in individuals who carry a pathogenic sarcomere mutation, we employed more rigorous criteria16 than used for a routine clinical diagnosis of HCM to avoid the inclusion of individuals with mild or emerging cardiomyopathy in the preclinical group. Overt HCM was defined by the presence of a sarcomere gene mutation and maximal LV wall thickness of ≥12 mm, or Z score ≥ 2 in children. The preclinical HCM group consisted of family members who carried a mutation but who lacked LVH (maximal LV wall thickness <12 mm or Z score <2 in children). Control subjects were healthy relatives who did not carry a sarcomere mutation. Individuals were excluded if they had systemic hypertension (SBP ≥140 or DBP≥90 mmHg or taking antihypertensive medication), LV EF<55%, coronary artery disease, valvular heart disease, prior septal myectomy or alcohol septal ablation, electronic ventricular pacing, or atrial fibrillation. Informed consent was obtained from all participants in protocols approved by the Institutional Review Boards of Brigham and Women’s Hospital and Minneapolis Heart Institute Foundation, and the Local Science Ethics Committee in Copenhagen, Denmark.
Vivid-7 ultrasound systems (GE Medical Systems, Milwaukee, WI) were used to obtain standard 2-dimensional (2D) and Doppler echocardiographic images. Recordings were stored digitally and analyzed offline by two observers (C.Y.H. and J.J.T), blinded to status group designation. Measurements were made on the average of 3 cardiac cycles according to criteria established by the American Society of Echocardiography,17 including cardiac dimensions, mitral inflow parameters, LV ejection fraction (modified Simpson’s method), and tissue Doppler myocardial velocities in systole (Sa) and early diastole (Ea). Global values of Ea and Sa represent the average of the lateral, septal, anterior, and inferior aspects of the mitral annulus, measured in the apical 4- and 2-chamber views.
Longitudinal, radial, and circumferential strain were determined from apical and mid-ventricular parasternal short axis images. During image acquisition for longitudinal strain, frame rates were maximized by narrowing the sector to isolate individual walls (range 40–85 frames per second). Offline image analysis was performed using commercial software (EchoPAC 6.0, GE Medical Systems) with speckle tracking methodology (two-dimensional strain (2DS), GE Medical Systems) which tracks the movement of natural acoustic speckles in the myocardium from two-dimensional gray-scale images.13,15,18,19 The endocardium was manually traced and myocardial motion was tracked with automated software. Tracking quality was verified manually and with the software’s automated quality grading scale. Segments were rejected if adequate quality could not be obtained despite manual correction. Representative 2DS tracings are shown in Figure 1.
Peak longitudinal systolic strain (εsys) and systolic strain rate (SSR) were measured in 12 segments from the basal, middle, and apical regions of the septal, lateral, anterior, and inferior walls from the apical 4- and 2-chamber views. Global values represent the average measurements from these 12 ventricular segments. Peak systolic radial and circumferential strain were measured in 6 segments from a mid-ventricular parasternal short axis image using 2DS. Global values represent the average measurements from 6 mid-ventricular segments.
To test for differences between the 3 status groups or between subjects with MYH7 and MYBPC3 mutations, analysis of variance (ANOVA) was performed adjusting for age and with clustering to adjust for the influence of relationships between family members. The GenMod procedure in SAS was used, assuming an exchangeable correlation structure. Basic clinical characteristics and conventional echocardiographic parameters are expressed as mean values ± standard deviations. Values for age-dependent parameters of tissue Doppler and strain are expressed as adjusted mean ± age and family-adjusted standard error. A p-value <0.017 was considered statistically significant to apply post-hoc Bonferroni correction for multiple comparisons across the 3 status groups following age-adjusted and family-correlated analysis. To evaluate multiple predictors simultaneously and adjust for confounders, the preclinical HCM group was compared to the control group with a logistic regression model using the GenMod procedure, assuming an exchangeable correlation structure to account for family relations. Statistical analysis was performed with SAS version 9.1 (SAS Institute Inc., Cary, NC, USA).
A total of 146 individuals from 47 families were analyzed, including 40 subjects with overt HCM, 68 subjects with preclinical HCM, and 38 mutation (−) healthy relatives as normal controls. Forty-one different mutations in 5 sarcomere genes were represented:β-myosin heavy chain (MYH7, n=51), cardiac myosin binding protein C (MYBPC3, n=40), cardiac troponin T (TNNT2, n=11), cardiac troponin I (TNNI3, n=5), and α-tropomyosin (TPM1, n=1) (see the Supplemental Table for a full listing of mutations). All mutations were known or presumed to be pathogenic by standard criteria of prior reports in the literature, appropriate co-segregation of the mutation, the absence of the mutation in large numbers of normal controls, alteration of evolutionarily conserved residues, and/or predicted impact on protein structure.
Basic clinical and echocardiographic parameters are summarized in Table 1. All preclinical and control subjects were asymptomatic (NYHA class I), not taking cardiac medications, and had normal conventional echocardiographic studies. Preclinical subjects were younger and had slightly lower blood pressure than controls. Standard cardiac dimensions were within normal limits for all preclinical and control subjects, although the preclinical HCM cohort had a slightly lower mean posterior wall thickness.
Consistent with its definition, the overt HCM cohort (n=40) had significantly greater LV wall thickness (mean maximal wall thickness 20.0 ± 5.0 mm) than the control and preclinical groups. Additionally, LV cavity size was smaller, and left atrial diameter larger. Seven percent of the overt HCM population had resting obstructive physiology (peak LV intracavitary gradient ≥30mmHg), and the majority (83%) had asymmetric septal hypertrophy. Subjects with overt disease had no or mild symptoms (85% NYHA class I; 15% NYHA class II). Thirty-two percent were taking cardiac medications (beta blockers, n=8; L-type calcium channel blockers, n=4; and disopyramide, n=1). Excluding individuals taking medications did not alter the results (data not shown). Thus, we feel that neither our results nor their interpretation were substantially influenced by cardioactive medication administration in a subset of the overt HCM cohort and the data presented reflect analyses of the full cohort.
Echocardiographic strain and tissue Doppler peak velocity data are summarized in Table 2. Interpretable 2D strain data were obtained in 89% of segments and 93% of walls; strain rate data were interpretable in 81% of segments and 82% of walls.
In comparing preclinical HCM to normal controls, there was no significant difference in systolic function as reflected by global longitudinal, radial, or circumferential systolic strain, or longitudinal SSR (Table 2). Sa velocities were minimally decreased in the preclinical cohort, but since Sa may be influenced by apex to base LV length20 and there was a wide variation in age and body size in our population, Sa was also normalized to LV length. Normalized values of Sa were not significantly different between the preclinical HCM and control cohorts. Results for global longitudinal strain and SSR are summarized in Figure 2. Evaluation of individual walls yielded the same results with no significant differences in regional longitudinal εsys or SSR between the preclinical and control cohorts.
Consistent with previous reports which focused on diastolic function,8, 9 standard parameters of mitral inflow did not discriminate between preclinical HCM and control groups (data not shown), but global Ea velocity was significantly reduced by 13% in preclinical HCM (12.3 ± 0.3 vs 14.2 ± 0.3 cm/sec, p<0.0001; Table 2), remaining significantly reduced if Ea was normalized for LV length. Multivariate logistic regression was performed, comparing preclinical HCM and controls, to identify predictors of carrying a gene mutation in these two cohorts with normal LV wall thickness. Age, gender, Ea, Sa, and global longitudinal εsys were included in the final model. No metrics of LV systolic function (Sa velocity, global longitudinal εsys or SSR) were significant multivariate predictors of carrying a sarcomere mutation. After adjustment, only age and Ea velocity (but not the E/Ea ratio) remained independently predictive of genotype status (data not shown). Preclinical HCM subjects were significantly younger and had a lower Ea velocity than genotype (−) normal control relatives.
The performance of reduced Ea velocity for identifying mutation carriers is summarized in Table 3. Receiver operator curve (ROC) analysis identified a global Ea velocity of 12 cm/sec as the optimal threshold value for individuals ≤ 25 years old. Ea ≤ 12 cm/sec had an 86% positive predictive value (PPV) for identifying young preclinical mutation carriers, however the negative predictive value (NPV) of Ea > 12 cm/sec was only 22%. As such, low Ea has high specificity but low sensitivity for differentiating apparently healthy family members with sarcomere mutations.
In subjects with overt HCM, longitudinal systolic strain and strain rate were substantially impaired (Figure 2), despite the presence of normal LV EF. Compared to normal controls, global longitudinal εsys was reduced by 27% (−15.8 ± 1.0% vs −21.5 ± 0.5% respectively, p<0.0001), and global longitudinal SSR was reduced by 14% (−1.21 ± 0.05 1/sec vs −1.41 ± 0.05 1/sec, respectively, p=0.0008; Table 2). Global circumferential and radial εsys did not significantly differ between any of the 3 cohorts. A further decline in diastolic function was noted with a 33% reduction global Ea velocity relative to controls, and a 23% reduction relative to preclinical HCM (Table 2).
Subjects with overt HCM were analyzed to determine if the degree of LVH correlated to strain abnormalities or other features of disease. There was a significant inverse correlation between maximal LV wall thickness and global longitudinal SSR (Pearson’s r =−0.52, p=0.002) and εsys (Pearson’s r = −0.53, p=0.0006). However there was no significant association between the degree of LVH and global Sa or Ea velocities, or LV EF.
Exploratory analyses were performed to compare subjects with mutations in the two most prevalent genes, myosin heavy chain and myosin binding protein C. Mutations in these genes accounted for 83% of all genotype positive subjects. In overt HCM, no significant differences were identified between individuals with mutations in MYH7 (n=17) and MYBPC3 (n=15), regarding age-adjusted metrics of diastolic (global Ea) or systolic function (global Sa, longitudinal εsys, SSR, or LV EF; Table 4). However, a greater degree of LVH was associated with mutations in MYBPC3 compared to MYH7 (maximal LV wall thickness 21.9 ± 0.2 vs 17.7 ± 0.1 cm, respectively, p=0.03).
In age-adjusted analysis of the preclinical HCM cohort (MYH7 n=35; MYBPC3 n= 24), subjects with MYH7 mutations had evidence of more impaired diastolic function, but enhanced systolic function relative to subjects with MYBPC3 mutations (Table 4). Global Ea was 9% lower in subjects with MYH7 mutations (12.8 ± 0.4 vs 14.1 ± 0.4 cm/sec, respectively, p=0.025) and global longitudinal εsys was 10% higher (22.92 ± 0.02 vs 20.82 ± 0.58%, respectively, p=0.004). Notably, global longitudinal εsys in preclinical MYH7 subjects was also 9% higher than control subjects (23.14 ± 0.43 vs 21.14 ± 0.50, p=0.001) whereas strain in preclinical MYBPC3 subjects did not differ from controls. However, there no significant differences in global Sa or global longitudinal SSR in preclinical MYBPC3 or MYH7 HCM were identified.
To better understand the early and late consequences of sarcomere mutations on cardiac function, echocardiographic strain parameters in preclinical mutation carriers without LVH were compared to individuals with overt HCM, and normal controls. Our results provide insights into the pathogenesis of HCM, indicating that different patterns of contractile abnormalities are present in early and late disease. Preclinical HCM is characterized by diastolic dysfunction but preserved LV systolic function. In contrast, both diastolic and longitudinal systolic function are substantially reduced in phenotypically overt HCM, despite normal LV ejection fraction. We propose that while diastolic dysfunction occurs as an early consequence of the underlying sarcomere mutation, the development of systolic dysfunction is associated with not only the mutation, but also the distinctive changes in myocardial architecture that accompany later development of clinical disease.
Building on prior reports,8, 9 our results provide further evidence that subtle myocardial dysfunction is present prior to the development of cardiac hypertrophy in preclinical HCM. Impaired relaxation is the predominant functional abnormality in preclinical HCM. In the context of family evaluation, reduced Ea velocity is the only identified parameter of cardiac function that predicts the presence of a sarcomere mutation in relatives without LVH. However despite the biological significance of this parameter, its diagnostic sensitivity is low. Identifying normal Ea velocity does not exclude the possibility that an apparently healthy member of a family with HCM carries a sarcomere mutation.
The disparate early impact of sarcomere mutations on systolic and diastolic function in preclinical human HCM is strikingly similar to findings in a mouse model of HCM produced by targeted insertion of the pathogenic Arg403Gln missense mutation in the myosin heavy chain gene.7 Detailed invasive hemodynamic studies directly assessed LV pressure-volume relations in intact animals. Prior to the development of LVH, young heterozygous mice showed decreased diastolic function (manifest as increased tau, time to peak filling, and end diastolic elastance; decreased –dP/dtmin), but no significant differences in systolic function (dP/dtmax, end systolic elastance) as compared to age-matched wild type controls. Although the molecular mechanisms by which sarcomere mutations impair diastolic function have not been fully elucidated, biophysical analyses of mutant proteins suggest that they result in slowed actin-myosin dissociation kinetics with decreased rates of crossbridge detachment.21 In animal models, these abnormalities have also been associated with altered intracellular calcium handling and decreased rates of calcium reuptake into the sarcoplasmic reticulum;22, 23 biochemical findings that could impair relaxation. Notably in our human overt HCM cohort, no correlation was found between LVH and Ea velocity, further supporting the hypothesis that the fundamental mechanisms underlying diastolic dysfunction in HCM are not driven by cardiac hypertrophy.
LV ejection fraction does not fully reflect contractile function in overt HCM. Despite the well recognized normal or increased LV ejection fraction typically seen in overt HCM and present in our population, we demonstrate that longitudinal systolic dysfunction is present, with significant reductions in longitudinal peak systolic strain and SSR. As opposed to echocardiographic strain analysis, LV EF may not reflect impaired long-axis function; a metric which may be more sensitive to myocardial disease processes and reveal earlier changes in cardiac performance.24 Our results for this population with genetically-confirmed HCM corroborate previous studies using tagged cardiac MRI and echocardiographic strain in more heterogeneous (non-genotyped) cohorts that identified reduced longitudinal systolic strain in individuals with a clinical diagnosis of HCM.15, 25–27 Similar findings have also been described in cardiac hypertrophy in a different disease, cardiac amyloidosis. Reduced longitudinal systolic strain and SSR were present in patients with normal ejection fraction and either no overt cardiac involvement or with clinical heart failure.14, 28
The cause of impaired systolic function in overt HCM and the links connecting sarcomere mutation, the development of LVH, and the development of systolic dysfunction are unclear. While it is possible that sarcomere mutations primarily result in slowly progressive or delayed decline in systolic function, triggering later development of compensatory LVH in overt HCM, we propose that the mutation in isolation is not sufficient to cause systolic abnormalities. Instead, both the mutation and the pathologic remodeling that typifies overt disease, namely myocyte disarray, hypertrophy, and fibrosis, lead to the downstream development of perturbed active force generation. We base this proposal on three lines of evidence. First, biophysical studies on isolated mutant myosin molecules and myofilaments suggest that sarcomere mutations which cause HCM are activating, resulting in higher actin sliding velocity, higher actomyosin ATPase activity, and increased maximal force generation.29, 30 This effect may be reflected in the enhanced systolic strain detected in our preclinical MYH7 subjects. Second, deficits in systolic strain are not an early consequence of sarcomere mutations as these parameters are normal in preclinical HCM. Third, animal models indicate that myocyte disarray and fibrosis develop later in life, coincidentally with the emergence of gross cardiac hypertrophy6 and by extrapolation we presume that preclinical HCM in humans lacks substantial histopathology. In aggregate, these data implicate changes in myocardial architecture in the development of systolic dysfunction.
An important limitation of this study is that our results reflect contractile function in two distinct stages of disease and therefore do not allow characterization of the time course between the onset of systolic dysfunction and the development of left ventricular hypertrophy. Longitudinal studies to follow preclinical sarcomere mutation carriers over time are needed to fully describe the temporal sequence of phenotypic evolution. Additionally, the small number of subjects MYH7 and MYBPC3 mutations limits the precision and power of exploratory subgroup comparisons to characterize gene-specific findings. As such, the failure to detect significant differences between these mutation groups does not prove that no differences are present. Future studies with larger numbers of subjects will be performed to better study gene-specific differences in phenotypic expression.
HCM is caused by mutations in contractile proteins, thus providing a remarkable opportunity to characterize how changes in the heart’s molecular motor influence myocardial structure and function. Studying a unique cohort with preclinical HCM, identified via genetic testing, allowed interrogation of disease pathogenesis and the intrinsic functional consequences of sarcomere mutations that occur in advance of cardiac remodeling. The predominant early effect of these mutations, prior to the development of cardiac hypertrophy, is diastolic dysfunction. Additionally, these studies demonstrate that the extent of the cardiomyopathic process in overt HCM is not fully appreciated by assessing left ventricular ejection fraction. Impaired longitudinal systolic function is present in overt HCM despite normal LV EF and may arise not only from the mutation, but also from the prototypic changes in myocardial architecture which develop later with expression LVH. These results have potentially important clinical implications, suggesting that future strategies aimed at diminishing the development of myocyte disarray, fibrosis, and hypertrophy, starting in the preclinical stage, may attenuate loss of systolic function and retard progression to symptomatic heart failure in HCM.
Funding Sources: This work was supported by grants from the GlaxoSmithKline Research & Education Foundation for Cardiovascular Disease and the National Institutes of Health (K23HL078901) (CYH).
There are no conflicts of interest to disclose