Clinical Phenotype and Outcome of Hypertrophic Cardiomyopathy Associated With Thin-Filament Gene Mutations

1Hypertrophic cardiomyopathy (HCM) is the most common genetic heart disease, generally caused by mutations in cardiac sarcomere genes. Most genotyped HCM patients harbor defects in the thick-filament genes, myosin heavy chain (MYH7) and myosin binding protein C (MYBPC3). However, in a distinct patient subgroup, the disease is caused by mutations in thin-filament genes, including cardiac troponin T (TNNT2) and I (TNNI3), α-tropomyosin (TPM1), and cardiac actin (ACTC),. Each accounts for a small proportion of HCM cohorts, with TNNT2, the most common, accounting for only 2% to 5%.

2The thin filament is a multisubunit, allosterically regulated molecular machine (Figure 1); thus, mutations in any of its components should exert similar biophysical effects and pathophysiological consequences,. However, a comprehensive assessment of the clinical presentation and outcome of patients carrying thin-filament mutations, compared with thick-filament disease, has not yet been performed. Initial clinical phenotype descriptions of TNNT2 and TNNI3 mutations were from families with severe HCM, characterized by high incidence of sudden cardiac death (SCD) despite relatively mild hypertrophy, often in children and adolescents,,,. Identification of mutations in these genes is therefore potentially relevant to clinical decision-making, including risk stratification for arrhythmic prophylaxis. However, subsequent reports of larger, less-selected cohorts show wide phenotypic and clinical variability for individual thin-filament genes, similar to thick-filament HCM,,. Consequently, whether thin-filament HCM has a truly distinct clinical profile from thick-filament HCM is unresolved. This study specifically addressed this issue by evaluating the clinical spectrum, echocardiographic features, and outcomes of a large, multicenter, genotyped cohort with HCM.

3All participants were unrelated index patients. HCM diagnosis was by 2-dimensional echocardiographic identification of a hypertrophied (≥13 mm), nondilated LV, in the absence of another cardiac or systemic disease capable of producing that magnitude of ventricular hypertrophy. The study included 80 HCM patients (8% of HCM patients genotyped during this time) with a pathogenic or likely pathogenic cardiac thin-filament gene mutation identified between January 2001 and December 2009 at 4 referral centers: Careggi University Hospital, Florence, Italy; Brigham and Women’s Hospital, Boston, Massachusetts; Stanford Medical Center, Palo Alto, California, and the University of Michigan Medical Center, Ann Arbor, Michigan (Table 1).

4For comparison, we evaluated 150 HCM patients with pathogenic or likely pathogenic mutations in the cardiac thick-filament genes MYH7 and MYBPC3 and the regulatory light chain (MYL2) consecutively identified in Florence during the same period. Clinical features of this reference group (Table 1) closely recapitulate published HCM cohorts from Europe and the United States,,, largely comprising thick-filament patients. Previous collaborative studies excluded significant discrepancy between cohorts from Florence and other centers,.

5After informed consent, patients were screened for mutations in protein-coding exons and splice sites of 8 myofilament genes, including the thin-filament genes TNNT2, TNNI3, TPM1, and ACTC; the thick-filament genes MYBPC3, MYH7, MYL2; and the essential light chain (MYL3). Genetic testing using established methods available at screening was performed by Clinical Laboratory Improvement Amendments (CLIA)–certified laboratories in the United States and at the Genetics Unit of Careggi University Hospital in Florence. Direct Sanger sequencing confirmed every variant. Variants were considered pathogenic if published as causative HCM mutations in at least 2 independent peer-reviewed studies. Novel mutations fulfilling the following internationally recommended criteria were considered likely to be pathogenic: 1) nonsynonymous variant causing an amino acid change in a residue highly conserved among species and predicted to significantly damage protein structure or function (Grantham, SIFT, and Polyphen scores), or truncating mutation; 2) the variant was absent in healthy control populations, including filtering for 1000 Genomes Project, National Heart, Lung and Blood Institute Exome Sequencing Project, and the Single Nucleotide Polymorphism database with minimal allelic frequency of <0.05; and 3) cosegregation with affected family members could be demonstrated for at least 1 patient. Before patient enrollment, the attending cardiologist and clinical geneticists evaluated this information on a case-by-case basis to confirm variant interpretation. Details of mutation distribution and classification in the thin- and thick-filament cohorts are found in Online Tables 1 and 2, respectively. At the beginning of the study, available published information was used to classify variant pathogenicity. Recent next-generation sequencing data led to subsequent downgrading of 3 thin-filament variants from likely pathogenic to variants of uncertain significance (TNNT2-Arg278Cys, TNNT2-Asn262Ser, and TNNI3-Arg162Pro). All remain potential disease-causing candidates, and there is evidence of Arg278Cys cosegregation in our cohort; thus these mutations were included in the analysis.

6To avoid bias related to founder effects, only the first identified patient carrying each of 2 highly recurrent mutations (i.e., E258K [Glu258Lys] in MYBPC3 and R869H [Arg869His] in MYH7 present in 52 and 19 Florence index patients, respectively) were included. Patients with complex genotypes, including thin-filament mutations associated with pathogenic or likely pathogenic MYBPC3 or MHY7 variants, were excluded.

7Echocardiographic studies were performed as described using commercially available instruments. LV filling patterns were assessed by pulsed-wave Doppler at the mitral tip level, and combined with tissue-Doppler evaluation of lateral mitral annulus velocity. We identified 4 LV filling patterns: (1 = normal; 2 = abnormal relaxation; 3 = pseudonormal; 4 = restrictive), defined according to existing guidelines,. Triphasic LV filling was considered present when a velocity peak of at least 0.2 m/s (an L-wave) was seen during diastasis, independent of the overall LV filling pattern.

8Cardiac magnetic resonance (CMR) imaging, including evaluation of late gadolinium enhancement (LGE), was performed as described in a subset of patients using commercially available 1.5-T scanners.

9Patients were followed up at yearly intervals or more often if clinically indicated, with review of history and symptoms, physical examination, echocardiographic examination, and 12-lead electrocardiography (ECG). If clinically indicated, ambulatory ECG monitoring for 24 to 48 h and CMR were performed. Established risk factors for SCD were defined as prior cardiac arrest or sustained ventricular tachycardia; family history of SCD at ≤40 years of age; nonvasovagal syncope; multiple episodes of nonsustained ventricular tachycardia (NSVT) during repeated ambulatory ECGs; maximal LV wall thickness ≥30 mm; and abnormal blood pressure response to exercise,.

10We documented major clinical outcomes including cardiovascular death, resuscitated cardiac arrest, nonfatal stroke, and progression to severe congestive symptoms (New York Heart Association [NYHA] functional class III or IV). Advanced LV dysfunction was defined by echocardiographic detection of systolic impairment with an ejection fraction <50% (generally used to identify end-stage HCM ) and/or restrictive LV filling pattern.

11Unpaired Student t tests were used to compare normally distributed data from patients with thick- and thin-filament HCM. Chi-square or Fisher exact tests were used to compare noncontinuous variables expressed as proportions. Survival curves were constructed according to the Kaplan-Meier method, and comparisons were performed using the log-rank test. Cox proportional hazard models were used to assess the effects of multiple clinical features on the risk of outcome events and to estimate survival curves of thin- versus thick-filament patients at net of covariates. Time of first clinical evaluation at each institution was considered as time 0. The probability values are 2-sided and considered significant when <0.05. Calculations were performed using SPSS version 20.0 software (SPSS, Chicago, Illinois).

12A total of 39 different pathogenic or likely pathogenic sequence variants were identified in 80 unrelated probands with thin-filament HCM (Figure 2, Online Table 1), including 35 missense, 2 truncation, and 2 insertion/deletion mutations. TNNT2 defects were the most common, with 15 distinct mutations identified in 43 of the 80 patients (53%). Additionally, we identified 15 different TNNI3 mutations in 24 patients (30%), 5 TPM1 mutations in 7 patients (9%), and 4 ACTC mutations in 6 patients (8%). None carried double thin-filament mutations. Patients carrying mutations in the most represented genes, TNNT2 and TNNI3, showed remarkably similar clinical features and outcome profiles (Online Tables 3 and 4). Among the 150 thick-filament HCM patients enrolled for comparison (age 42 ± 17 years, 44% female), 94 different mutations were identified (Online Table 2), including 49 in MYBPC3 (n = 83, 55%), 40 in MHY7 (n = 57, 38%), and 5 in MYL2 (n = 10, 7%).

13At initial evaluation, the mean age of the 80 thin-filament HCM patients was 44 ± 16 years; 45% were women. Most (66%) reported normal exercise tolerance (NYHA functional class I); however, 54% were symptomatic as a result of atrial fibrillation (AF) (30%), angina (20%), or syncope (18%) (Table 1). Overall, these features were comparable to the thick-filament cohort (Table 1). On 12-lead ECG, 37% of thin-filament patients showed inferolateral Q waves (vs. 9% in the 150 thick-filament patients, p < 0.001) and 67% showed inverted T waves in the precordial leads (vs. 44% of thick-filament patients, p = 0.002) (Table 1).

14Several differences between thick- and thin-filament HCM in LV morphology and function were noted. Patients with thin-filament mutations had lesser maximal LV wall thickness values than the thick-filament group (18 ± 5 mm vs. 24 ± 6 mm; p < 0.001) and more often exhibited atypically distributed hypertrophy (31%), including concentric and apical patterns, whereas 94% of thick-filament HCM presented as classic asymmetric LVH involving the basal septum and anterior wall (p < 0.01) (Table 1, Figure 3), consistent with the lower prevalence of resting LV outflow tract obstruction in thin-filament patients (19% vs. 34% in thick-filament HCM, p = 0.015). An apical or concentric distribution of hypertrophy was most likely in TNNI3 patients (41%) (Online Table 3). Aspects of LV noncompaction were uncommon; for example, none of the 6 patients with actin mutations showed regional noncompaction.

15Transmitral pulsed-wave interrogation showed a triphasic LV filling pattern characterized by an L-wave with prominent mid-diastolic flow velocity in 26% of thin-filament patients, but in only 11% with thick-filament disease (p = 0.002) (Figure 4). Furthermore, early diastolic lateral mitral annulus velocity (E′) was 24% lower in patients with thin compared with thick-filament mutations (p < 0.001) (Table 1).

16CMR studies in 47 thin-filament patients (59%) and 76 thick-filament patients (51%) highlighted significant differences between these cohorts. Thin-filament patients had a smaller LV mass index and a lower LV ejection fraction (Table 1). Although LGE was present in the majority of patients from both cohorts, the proportion of LV mass occupied was larger in thin-filament patients (20 ± 11% vs. 16 ± 8% in thick-filament patients). LGE exceeded 30% of LV mass in 12 thin-filament patients (27%), 10 of whom showed an EF <50% or a restrictive diastolic pattern at final evaluation. Only 8 thick-filament patients (11%) showed LGE exceeding 30% of the LV.

17Compared with the thick-filament cohort, thin-filament patients had a higher prevalence of NSVT, abnormal blood pressure response to exercise, and family history of SCD and were more likely to have at least 1 established SCD risk factor (74% vs. 59%; p = 0.031) (Table 1). However, the proportion of patients with 2 or more risk factors was similar in both groups (30% vs. 34%, respectively, p = 0.39).

18Mean follow-up for the thin-filament cohort was 4.7 ± 2.7 years, for a total of 361 patient-years, comparable to the thick-filament group (4.7 ± 3.0 years, p = 0.49). During this time, 2 patients from the thin-filament cohort (2.5%) died of cardiac causes (1 suddenly, 1 because of heart failure), 3 (4%) experienced nonfatal strokes, 3 (4%) had resuscitated cardiac arrests, and 5 (6%) had appropriate implantable cardioverter-defibrillator (ICD) shocks owing to rapid ventricular tachycardia or fibrillation (Table 2). All-cause mortality, cardiac mortality, and SCD rates did not differ between the cohorts (p > 0.05 for all comparisons) (Table 2), and their rates of malignant arrhythmias (including sudden death, resuscitated cardiac arrest, and appropriate ICD shocks) were also similar (Table 2).

19At final evaluation, moderate or severe congestive symptoms (NYHA functional class III/IV) were more prevalent in the thin compared with the thick-filament subgroup (19% vs. 10%; p = 0.034) (Table 2). Notably, 12 (15%) thin-filament patients with mild or no symptoms at initial evaluation progressed to NYHA functional class III/IV during follow-up at a mean age of 50 ± 9 years, 3 times the prevalence in the thick-filament cohort (5%; p = 0.013). Survival analysis showed a higher likelihood of developing moderate or severe congestive symptoms among thin-filament patients (Figure 5A). At multivariate analysis, performed on the 2 HCM cohorts combined, the presence of thin-filament disease more than doubled the likelihood of a final NYHA functional class III/IV (hazard ratio [HR]: 2.16, p = 0.040), independent of LV outflow obstruction (HR: 4.06; p <0.001) and AF (HR 2.74, p = 0.008) (Online Figure 1).

20Advanced LV dysfunction (defined as LVEF <50% or restrictive diastolic pattern) was present at final evaluation in 23 of the 80 thin-filament patients (29%) at ages ranging from 20 to 76 years, compared with 17 of the 150 thick-filament patients (p = 0.002). Of note, 10 of these 23 patients (43%) were younger than 50 years of age (Figure 6). The incidence of new LV dysfunction during follow-up proved higher in the thin-filament cohort (4.3% per annum) than in the thick-filament cohort (1.9% per annum; p = 0.013) (Figures 5B and 5C). In multivariate analysis, thin-filament mutations predicted LV dysfunction (HR: 2.28, p = 0.016) independent of female sex (HR: 2.08, p = 0.031), AF (HR: 2.33, p = 0.023), and NYHA functional class III symptoms at baseline (HR: 2.64, p = 0.044) (Online Figures 2 and 3). Consistent with the pronounced diastolic impairment, thin-filament patients more often had moderate or severe atrial dilation at final evaluation than thick-filament patients. Conversely, the incidence of new AF during follow-up was comparable in both groups (Table 2).

21During follow-up, most thin-filament patients (n = 75; 94%) received pharmacological treatment for HCM, including beta-blockers, verapamil, amiodarone, and disopyramide (Table 2). Predictably, they were more frequently treated with diuretics and angiotensin-converting enzyme inhibitors or angiotensin receptor blockers than the thick-filament cohort but less often received disopyramide (Table 2).

22Nineteen thin-filament patients (24%) received an ICD (Table 2), including 16 for primary and 3 for secondary prevention of SCD. Furthermore, 11 patients (14%) were referred for surgical septal myectomy (n = 7) or alcohol septal ablation (n = 4) for drug-refractory symptoms associated with LV outflow tract obstruction. Finally, 11 patients (14%) underwent radiofrequency catheter ablation for symptomatic, drug-refractory AF. Compared with thick-filament patients (Table 2), thin-filament patients more often underwent catheter ablation procedures for AF but were less frequently referred for invasive septal reduction therapies; ICD implantation rates were comparable (Table 2).

23This study supports the hypothesis that thin-filament HCM is phenotypically distinct from the more common thick-filament HCM,,,. Specific LV morphological, functional, and remodeling differences were identified, suggesting unique underlying pathophysiological mechanisms,. At initial evaluation, thin-filament patients showed lesser LV hypertrophy, often developing in apical or concentric patterns, whereas patients with thick-filament disease almost universally displayed classic asymmetric LV hypertrophy involving the basal septum and anterior wall. As a result, dynamic LV outflow tract obstruction was less common in the thin-filament cohort (19%, compared with 34% among thick-filament patients, p = 0.015), likely as a result of relative preservation of LV outflow morphology and function and reduced propensity to systolic anterior motion. Conversely, diastolic abnormalities were more common and pronounced in thin-filament HCM, including 16% of patients with restrictive LV pathophysiology. More than one-quarter of our thin-filament patients exhibited a triphasic LV filling pattern, compared with only 11% in the thick-filament subgroup, a disproportionate prevalence, consistent with severe diastolic impairment,. The presence of an L wave (a velocity peak ≥0.2 m/s during diastasis), believed to indicate elevated filling pressures and previously observed in HCM, has been associated with extensive septal fibrosis. Indeed, substantial LV remodeling occurred in thin-filament patients, mediated by progressive myocardial fibrosis (Figure 6). LGE was present in 85% of patients undergoing CMR and averaged 20% of the whole LV, reflecting greater prevalence and extension of fibrous tissue compared with our thick-filament patients and previously published, unselected HCM cohorts, largely reflecting thick-filament disease,.

24At the end of 4.7 years of follow-up, 29% of the 80 thin-filament patients had advanced LV dysfunction (defined as LVEF <50% or restrictive diastolic pattern), more than double the prevalence among thick-filament patients. Indeed, the incidence of newly occurring systolic dysfunction in our thin-filament patients was approximately 2.5% per year, compared with approximately 1% per year in our thick-filament subset; the latter value closely agrees with values in unselected HCM populations,. Furthermore, restrictive LV pathophysiology with preserved systolic function was observed in 11% of thin-filament patients during follow-up, consistent with prior reports emphasizing isolated, severe diastolic dysfunction in patients with troponin mutations, particularly troponin I, which occasionally presents as primary restrictive disease.

25At final evaluation, the adverse remodeling process observed in thin-filament HCM patients was paralleled by a considerable prevalence of moderate or severe congestive symptoms. Overall, 15% of thin-filament patients initially presenting with mild or no symptoms progressed to NYHA functional class III/IV during follow-up, 3 times the prevalence in the thick-filament cohort (5%; p = 0.013). In a multivariate model assessing established predictors of HCM outcome, thin-filament disease more than doubled the likelihood of a final NYHA functional class III/IV, independent of LV outflow obstruction and AF. Conversely, thick- and thin-filament HCM patients had comparably low rates of malignant ventricular arrhythmias, including SCD, resuscitated cardiac arrest, and appropriate ICD interventions. This limited arrhythmic propensity contrasts with prior reports suggesting increased risk of SCD in patients with troponin T or I mutations, likely because of the malignant profile of the highly selected families in early studies. Although individual arrhythmic risk may vary considerably, particularly in children and adolescents, thin-filament HCM in our cohort emerged as a progressive condition characterized by adverse LV remodeling and dysfunction, rather than by enhanced arrhythmogenicity. Clinical implications include the need for heightened attention to early signs of LV dysfunction and symptom progression in patients with thin-filament HCM. Conversely, aggressive strategies for primary prevention of SCD, including ICDs, are not warranted solely because a thin-filament mutation is present,.

26Preclinical studies using animal models support key features identified in our cohort of patients with thin-filament HCM. Transgenic mouse lines with thin-filament gene defects develop restrictive diastolic patterns and systolic dysfunction over time. Skinned myocardial tissue from patient samples and animal models with thin-filament mutations consistently show markedly increased myofilament Ca2+ sensitivity, closely related to abnormalities of cardiac relaxation and diastolic dysfunction. Furthermore, thin-filament defects can alter cardiac function by increasing the energy cost of contraction. Although several of these abnormalities are shared with defects in other HCM-related genes, their extent is generally greater in thin-filament HCM samples. The constellation of early impairment in excitation-contraction coupling, energetic derangement, abnormal cardiomyocyte signaling, and intrinsic abnormalities of sarcomeric relaxation caused by thin-filament mutations may collectively drive aggressive remodeling at the cellular and extracellular levels, resulting in impaired contractile and relaxation properties of the myocardium,; these ultimately account for the common occurrence of progressive LV dysfunction observed in our HCM cohort.

27Distinctive clinical and biophysical features characterize HCM associated with thin-filament mutations, at variance with the more common thick-filament disease. Thin-filament HCM is associated with less prominent and atypically distributed LV hypertrophy, increased LV fibrosis, higher likelihood of adverse LV remodeling leading to functional deterioration, and more frequent occurrence of triphasic LV filling, reflecting profound diastolic dysfunction (Central Illustration). Management strategies should consider adequate surveillance for early detection of LV dysfunction and symptomatic progression. Conversely, arrhythmic risk does not appear to increase solely as a result of a thin-filament genotype.

28COMPETENCY IN MEDICAL KNOWLEDGE: HCM associated with thin-filament mutations is characterized by less prominent and atypically distributed hypertrophy; increased fibrosis; and more adverse remodeling (hypokinetic or restrictive evolution), leading to congestive symptoms, triphasic left ventricular filling, and more severe diastolic dysfunction, compared with thick-filament HCM.COMPETENCY IN PATIENT CARE: Aggressive measures for primary prevention of sudden death with implanted cardiac defibrillators in patients with HCM should not be based solely on genotype.TRANSLATIONAL OUTLOOK 1: Better understanding of early phenotype development in thin-filament cardiomyopathy will require studies of index patients’ families, including monitoring of young relatives carrying thin-filament mutations over a relatively long period and comparison with carriers of thick-filament mutations.TRANSLATIONAL OUTLOOK 2: Preclinical studies involving transgenic animal models are warranted to assess mechanisms of disease progression and to test pharmacological strategies for controlling symptoms and reducing adverse myocardial remodeling in thin-filament HCM.

29COMPETENCY IN MEDICAL KNOWLEDGE: HCM associated with thin-filament mutations is characterized by less prominent and atypically distributed hypertrophy; increased fibrosis; and more adverse remodeling (hypokinetic or restrictive evolution), leading to congestive symptoms, triphasic left ventricular filling, and more severe diastolic dysfunction, compared with thick-filament HCM.

30COMPETENCY IN PATIENT CARE: Aggressive measures for primary prevention of sudden death with implanted cardiac defibrillators in patients with HCM should not be based solely on genotype.

31TRANSLATIONAL OUTLOOK 1: Better understanding of early phenotype development in thin-filament cardiomyopathy will require studies of index patients’ families, including monitoring of young relatives carrying thin-filament mutations over a relatively long period and comparison with carriers of thick-filament mutations.

32TRANSLATIONAL OUTLOOK 2: Preclinical studies involving transgenic animal models are warranted to assess mechanisms of disease progression and to test pharmacological strategies for controlling symptoms and reducing adverse myocardial remodeling in thin-filament HCM.