3-Dimensional Strain Analysis of Hypertrophic Cardiomyopathy

Abstract

BACKGROUND

Abnormal global longitudinal strain (GLS) has been independently associated with adverse cardiac outcomes in both obstructive and nonobstructive hypertrophic cardiomyopathy.

OBJECTIVES

The goal of this study was to understand predictors of abnormal GLS from baseline data from the National Heart, Lung, and Blood Institute (NHLBI) Hypertrophic Cardiomyopathy Registry (HCMR).

METHODS

The study evaluated comprehensive 3-dimensional left ventricular myocardial strain from cine cardiac magnetic resonance in 2,311 patients from HCMR using in-house validated feature-tracking software. These data were correlated with other imaging markers, serum biomarkers, and demographic variables.

RESULTS

Abnormal median GLS (> −11.0%) was associated with higher left ventricular (LV) mass index (93.8 ± 29.2 g/m² vs 75.1 ± 19.7 g/m²; P < 0.0001) and maximal wall thickness (21.7 ± 5.2 mm vs 19.3 ± 4.1 mm; P < 0.0001), lower left (62% ± 9% vs 66% ± 7%; P < 0.0001) and right (68% ± 11% vs 69% ± 10%; P < 0.01) ventricular ejection fractions, lower left atrial emptying functions (P < 0.0001 for all), and higher presence and myocardial extent of late gadolinium enhancement (6 SD and visual quantification; P < 0.0001 for both). Elastic net regression showed that adjusted predictors of GLS included female sex, Black race, history of syncope, presence of systolic anterior motion of the mitral valve, reverse curvature and apical morphologies, LV ejection fraction, LV mass index, and both presence/extent of late gadolinium enhancement and baseline N-terminal pro-B-type natriuretic peptide and troponin levels.

CONCLUSIONS

Abnormal strain in hypertrophic cardiomyopathy is associated with other imaging and serum biomarkers of increased risk. Further follow-up of the HCMR cohort is needed to understand the independent relationship between LV strain and adverse cardiac outcomes in hypertrophic cardiomyopathy.

Hypertrophic cardiomyopathy (HCM), characterized by myocardial hypertrophy, myofibrillar disarray, and interstitial fibrosis, manifests in a diverse set of phenotypic patterns (eg, reverse curvature, apical, concentric), normal or supra-normal ejection fraction, and impaired left ventricular (LV) relaxation.¹ LV strain analysis allows noninvasive evaluation of tissue deformation and both global and regional myocardial contractile performance. This is particularly important in HCM, which may be complicated by dynamic left ventricular outflow tract (LVOT) obstruction and mitral regurgitation, particularly in the isolated basal septal hypertrophy phenotype.² Abnormal global longitudinal strain (GLS) according to speckle tracking echocardiography and abnormal circumferential strain according to myocardial tagging with cardiac magnetic resonance (CMR) have been well described in HCM and associated with an adverse prognosis; however, prospective large-scale clinical validation data remain limited.^1,3–5 Moreover, uncertainty remains regarding the adverse pathophysiological processes, architectural tissue changes, and demographic associations that lead to abnormal GLS in HCM. Elucidation of these relationships may help incorporation of noninvasive strain measures for improved risk stratification and therapeutic decision-making in both obstructive and nonobstructive HCM.

The NHLBI (National Heart, Lung, and Blood Institute) HCMR (Hypertrophic Cardiomyopathy Registry) is the first large prospective international HCM registry to include rigorous CMR, genetic testing, and prospective collection of blood for biomarker analysis.⁶ We sought to evaluate comprehensive 3-dimensional (3D) myocardial LV strain measures by cine CMR using validated feature-tracking software in 2,755 patients from the NHLBI HCMR.

METHODS

The full study design for the prospective observational HCMR has been published previously.⁷ The relevant methods for the current study are summarized in the following sections. After providing written informed consent, all patients underwent standard clinical evaluation, echocardiography, CMR, and blood collection for genetic and biomarker analyses.

STUDY POPULATION.

HCMR is a prospective, international registry study of 2,755 HCM patients with comprehensive clinical, electrocardiographic (exercise treadmill testing [ETT] and Holter monitoring), imaging (clinical echocardiography and contrast-enhanced CMR), genetic, and biomarker analyses. The NHLBI HCMR study population, clinical characteristics, and inclusion and exclusion criteria have been described previously.^6,7 All patients underwent the written informed consent process, and the study protocol was approved by the institutional review board of each enrolling site.

CMR ACQUISITION.

CMR was conducted by using 1.5-T or 3.0-T scanners (General Electric, Philips Medical Systems, and Siemens Healthineers) using standardized protocol and multichannel phased-array chest coils with electrocardiographic gating. Short- and long-axis cine steady-state free precession imaging was conducted with typical parameters: TR/TE 3.1/1.2 milliseconds, in-plane resolution of 2 to 2.5 mm, and temporal resolution of 40 to 50 milliseconds. T1 mapping was typically acquired in 3 short-axis slices centered in the mid-left ventricle both before and after gadolinium contrast administration using a Shortened Modified Look-Locker Inversion recovery technique, as previously published.⁸ Gadolinium contrast was administered intravenously as a single 0.15 mmol/kg bolus dose. Late gadolinium enhancement (LGE) imaging was acquired in matching long- and short-axis steady-state free precession prescriptions postcontrast by using a segmented inversion-recovery sequence.

CMR AND STRAIN ANALYSES.

CMR analysis was conducted by the core laboratory (Brigham and Women’s Hospital, Boston, Massachusetts) using commercially available software (Medis Suite 3.0 and QMassMR, Medis). Measurements were conducted according to published guidelines from the Society for Cardiovascular Magnetic Resonance.⁹ In summary, LV and right ventricular volumes, mass, and function were calculated from manual tracing of endocardial and epicardial borders of end-diastolic and end-systolic short-axis cine images. Long-axis cine imaging was used for calculation of left atrial (LA) volumes and functional parameters.¹⁰ Degree of LGE was quantified by a core laboratory (Beth Israel Deaconess Medical Center, Boston, Massachusetts) using a 6 SD⁹ semi-quantitative threshold in addition to manual visual evaluation.¹¹ Native T1 and extracellular volume fraction (ECV) were calculated both globally and segmentally.¹²

Myocardial strain analysis was conducted by CMR myocardial strain core laboratories (University of Calgary, Calgary, Canada, and University of Oxford, Oxford, United Kingdom) using software developed and validated in-house.¹³ All core laboratory analyses were performed blinded to patients’ demographic, biomarker, and genetic data. In summary, an end-diastolic 3D mesh of the left ventricle was created from the subendocardial and subepicardial contours of long- and short-axis LV views after alignment. Feature tracking was performed for all myocardial pixels throughout the cardiac cycle. The 2-dimensional velocity field obtained for each view was used to reconstruct the 3D velocity field, taking into account the orientation of each acquired view, in-plane resolution, and the distance from each node of the 3D LV mesh. The motion of each node of the mesh throughout the cardiac cycle was derived from the reconstructed 3D velocity field. Using a finite-element approach, the deformation of each finite-element voxel was projected in radial, circumferential, longitudinal, and principal local directions, to compute radial, circumferential, longitudinal, and principal strain, respectively. Principal strains were defined as strains in the directions (principal directions) describing the totality of the deformation of each element of the mesh without neglecting the shear deformation.^14,15 Maximum and minimum principal strains are defined as the most positive and the most negative of the 3 principal strain amplitudes, respectively. Strain values were obtained for sub-endocardial, subepicardial, and transmural measures both globally and for the 16-segment American Heart Association segmental model.¹⁶ Lower values of radial and higher values (less negative) of longitudinal and circumferential strain are referred to as abnormal.

SERUM BIOMARKERS AND GENETICS.

All serum biomarkers were batch tested at the end of the study period by the Biomarker Research and Clinical Trials Laboratory (Brigham and Women’s Hospital). All genetic analysis was completed by the University of Oxford using amplicon-based sequencing for 36 cardiomyopathy-associated genes via the Illumina MiSeq platform. Bioinformatics analysis was performed by using the Genome Analysis Toolkit version 4 best practice guidelines. Further specific methods for both biomarker and genetic analyses have been described previously.⁶

STATISTICAL ANALYSIS.

Categorical variables are presented as counts with percentages, and continuous variables are expressed as mean ± SD. The nonparametric Kruskal-Wallis rank test was used to compare continuous variables due to nonnormal distribution of most data. Biomarker variables were log-transformed to improve normality. Categorical variables were compared by using the Fisher exact test or chi-square test. Based on the reported relationship between GLS and adverse outcomes within the HCM published data, the overall cohort was stratified according to median GLS to evaluate for any differences in demographic or imaging characteristics. Relationships between strain parameters and clinical/CMR characteristics were evaluated with multivariable linear regression, linear mixed effects modeling, and elastic net regression to evaluate for the effect size of independent predictors. All predictors were centered on their mean and scaled according to their SD for standardization. Univariable and multivariable logistic regression was used to evaluate for associations between clinical, CMR, and strain parameters with abnormal Holter and ETT results.

All statistical analysis was performed by using commercially available software (SAS version 9.4; SAS Institute, Inc). A 2-sided value of P < 0.05 was considered statistically significant.

RESULTS

Of the total 2,755 HCMR registry participants, 2,311 patient studies (84%) had diagnostic image quality suitable for strain analysis. The most common reasons for inadequate diagnostic quality included poor ECG-gating of cine images, inconsistent number of cardiac phases between short and long axis-imaging, and significant arrhythmia during cine imaging. Baseline characteristics stratified according to median GLS (−11.0%) are presented in Table 1. Abnormal median GLS was associated with higher LV mass index and maximal wall thickness, lower LV and right ventricular ejection fractions, lower LA total emptying function, and higher presence and myocardial extent of LGE. Patients with abnormal median GLS had higher baseline N-terminal pro-B-type natriuretic peptide (NT-proBNP) and troponin T levels. In the overall cohort, GLS exhibited positive correlations with LV mass index (r = 0.46), maximal wall thickness (r = 0.34), LGE by 6 SDs (r = 0.23), native T1 at both 1.5-T and 3.0-T (r = 0.09 for both), ECV (r = 0.17), baseline troponin T (r = 0.43), baseline NT-proBNP (r = 0.30), and LA volume (r = 0.11). GLS showed a negative correlation with LVEF (r = −0.35) and LA total contractile function (r = −0.19). Abnormal median global circumferential strain (GCS) (Supplemental Table 1) was associated with a lower presence of systolic anterior motion of the mitral valve (SAM), LVOT obstruction, and LVEF but higher LV mass index, and both the presence/extent of LGE. Abnormal median global radial strain (GRS) (Supplemental Table 2) was associated with the presence of sarcomere mutation status, lower LVEF, higher LV mass index, presence/extent of LGE, and LA volume.

TABLE 1.

Demographic Characteristics, Biomarkers, and CMR Characteristics Stratified According to Median Global Longitudinal Transmural Strain

	GLS < −11.0% (n = 1,155)	GLS > −11.0% (n = 1,156)	P Value
Demographic characteristic
Age, y	49.7 ± 11.5	49.2 ± 11.1	0.10
Female	36	22	<0.0001
Body mass index, kg/m2	28.5 ± 5.2	29.9 ± 5.8	<0.0001
Race			<0.0001
White	90	79
Black	4	11
East Asian	2	2
South Asian	4	6
Pacific Islander	0	1
American Indian/Alaskan	0	1
History of syncope	12	14	0.10
History of heart failure	4	6	0.09
History of hypertension	33	40	<0.01
First-degree relative with sudden death	13	11	0.26
Chest pain	32	33	0.63
Dyspnea	42	44	0.31
NYHA functional class			<0.05
I	68	66
II	26	26
III	6	7
IV	0	0.6
Systolic anterior motion of mitral valve	46	35	<0.0001
LV outflow tract obstruction
Resting	25	23	0.22
Resting, mm Hg	23.7 ± 31.9	21.7 ± 29.7	0.02
Provocable, mm Hg	49.3 ± 48.2	42.5 ± 45.7	<0.01
Abnormal Holter monitor	11	21	<0.0001
Exercise treadmill, METs	9.9 ± 3.8	9.8 ± 3.5	0.40
Abnormal exercise treadmill test	1	4	<0.001
Sarcomere mutation, positive	38	33	0.02
Morphology			<0.0001
Isolated basal septal	62	31
Reverse curvature	33	46
Apical	3	14
Concentric	0.4	2
Apical aneurysm	0.3	5
Other	1	1
AHA 2020 major risk factors for sudden death	24.9	39.3	<0.0001
Biomarkersa
NT-proBNP, pg/mL	5.2 ± 1.2	5.8 ± 1.3	<0.0001
Troponin T, ng/mL	2.3 ± 0.4	2.7 ± 0.7	<0.0001
Galectin-3, pg/mL	8.8 ± 0.4	8.8 ± 0.4	0.64
ST2, pg/mL	9.7 ± 0.5	9.8 ± 0.4	0.13
MMP-1, pg/mL	5.8 ± 0.8	5.8 ± 0.8	0.31
TIMP-1, pg/mL	11.4 ± 0.3	11.5 ± 0.3	0.06
CICP, ng/mL	2.2 ± 0.4	2.2 ± 0.4	0.84
Bone alkaline phosphatase, U/mL	2.9 ± 0.3	2.9 ± 0.3	0.30
CMR characteristics
LV mass index, g/m2	75.1 ± 19.7	93.8 ± 29.2	<0.0001
Maximal wall thickness, mm	19.3 ± 4.1	21.7 ± 5.2	<0.0001
Maximal wall thickness >30 mm	1.7	6.6	<0.0001
LV end-diastolic volume index, mL/m2	84.0 ± 16.6	86.0 ± 17.4	<0.01
LV end-systolic volume index, mL/m2	28.7 ± 10.1	33.4 ± 12.6	<0.0001
LVEF, %	66 ± 7	62 ± 9	<0.0001
LVEF <50%	1.7	9.5	<0.0001
RVEF, %	69 ± 10	68 ± 11	<0.01
RV mass index, g/m2	17.1 ± 4.9	18.0 ± 6.1	<0.001
RV end-diastolic volume index, mL/m2	77.1 ± 16.4	74.1 ± 15.6	<0.0001
RV end-systolic volume index, mL/m2	24.6 ± 10.8	24.7 ± 11.3	0.85
LGE presence	42	58	<0.0001
LGE 6 SDs, g	1.7 ± 4.4	4.4 ± 9.2	<0.0001
LGE 6 SDs, % of LV mass	1.2 ± 3.2	2.3 ± 5.0	<0.0001
LGE visual, g	1.8 ± 4.5	4.6 ± 9.1	<0.0001
LGE visual, % of LV mass	1.3 ± 3.3	2.4 ± 4.9	<0.0001
Native T1, ms
1.5-T	971 ± 71	980 ± 72	<0.001
3.0-T	1,166 ± 87	1,183 ± 88	<0.01
ECV, %	29 ± 4	31 ± 6	<0.0001
Left atrial volume, mL	113 ± 40	120 ± 44	<0.001
Reservoir function, %	36.3 ± 10.1	33.7 ± 10.2	<0.0001
Contractile function, %	48.9 ± 12.5	44.6 ± 13.9	<0.0001
Total function, %	70.0 ± 11.4	62.6 ± 13.2	<0.0001

Values are mean ± SD, n, or %.

^aBiomarker values are log-transformed.

AHA = American Heart Association; CICP = C-terminal propeptide type 1 procollagen; CMR = cardiac magnetic resonance; ECV = extracellular volume fraction; GLS = global longitudinal strain; LGE = late gadolinium enhancement; LV = left ventricular; LVEF = left ventricular ejection fraction; MMP = matrix metalloproteinase; NT-proBNP = N-terminal pro-B-type natriuretic peptide; NYHA = New York Heart Association; RV = right ventricle; RVEF = right ventricular ejection fraction; ST2 = suppression of tumorigenicity 2; TIMP = tissue inhibitor metalloproteinase.

Peak global and regional (basal, mid, apical) systolic strain values stratified according to HCM morphology are shown in Table 2; values for the entire cohort are presented in Supplemental Table 3. GLS and GCS were most preserved in patients with isolated basal septal hypertrophy, whereas GRS was highest among those with apical morphology. All measures of global peak systolic strain were lowest in those with apical aneurysm. Figure 1 illustrates the distribution of median GLS across the various HCM morphologies. (Supplemental Figure 1 displays the median GCS and GRS across the various HCM morphologic subtypes.) Of all strain measures, abnormal median GRS exhibited the highest proportion of reverse curvature morphology relative to isolated basal septal hypertrophy. Strain measures stratified according to the presence of LGE, sex, and sarcomere mutation status are shown in Tables 3 to 5. (Supplemental Tables 4 and 5 present strain measures stratified according to maximal wall thickness and resting LVOT obstruction.) All measures of global peak systolic strain were lower in male subjects and those with LGE but variable across sarcomere mutation status. All global peak systolic strain measures were more abnormal for HCM patients with maximal wall thickness ≥20 mm, whereas only circumferential strain was higher in those with resting LVOT obstruction. GLS stratified according to a maximal wall thickness <15 mm (Supplemental Figure 2) showed significant differences for both women and men. Women and men had a similar inverse relationship between GLS and maximal wall thickness (Supplemental Figure 3).

TABLE 2.

Global and Regional Peak LV Systolic Strain Stratified According to HCM Morphology

	Isolated Basal Septal (n = 1,066)	Reverse Curvature (n = 913)	Apical (n = 191)	Concentric (n = 36)	Apical Aneurysm (n = 65)	Other (n = 29)	P Value Trend
Peak systolic strain: global, %
Longitudinal
Transmural	−11.83 ± 2.32	−10.37 ± 2.51	−9.14 ± 2.21	−8.27 ± 3.01	−7.23 ± 2.52	−10.65 ± 2.57	<0.0001
Endocardial	−16.62 ± 3.19	−14.87 ± 3.47	−14.04 ± 3.88	−13.38 ± 4.87	−10.23 ± 3.65	−15.15 ± 3.63	<0.0001
Epicardial	−8.1 ± 1.95	−6.91 ± 2.12	−5.01 ± 1.51	−4.67 ± 2.09	−4.54 ± 2.18	−6.99 ± 2.07	<0.0001
Circumferential
Transmural	−12.24 ± 2.3	−11.28 ± 2.41	−10.2 ± 1.87	−9.65 ± 2.61	−9.39 ± 2.19	−11.45 ± 2.93	NS
Endocardial	−18.53 ± 3.48	−17.5 ± 3.5	−17.28 ± 3.51	−16.9 ± 4.36	−15.27 ± 3.63	−17.3 ± 4.19	<0.0001
Epicardial	−7.58 ± 1.69	−6.73 ± 1.78	−5.44 ± 1.16	−5.15 ± 1.7	−5.5 ± 1.73	−6.93 ± 1.85	<0.0001
Radial	37.27 ± 17.32	32.35 ± 16.28	39.58 ± 17.17	34.27 ± 20.39	27.55 ± 14.49	37.68 ± 20.67	NS
Minimum principal
Transmural	−23.44 ± 3.54	−22.82 ± 3.81	−21.85 ± 3.55	−20.81 ± 4.63	−20.56 ± 3.58	−22.38 ± 4.79	NS
Endocardial	−27.79 ± 3.94	−27.41 ± 4.17	−27.14 ± 4.18	−26.45 ± 5.64	−25.35 ± 4.11	−26.47 ± 5.34	NS
Epicardial	−18.96 ± 3.16	−18.07 ± 3.45	−16.54 ± 2.91	−15.35 ± 3.87	−16.08 ± 3.08	−18.13 ± 4.05	NS
Maximum principal	61.78 ± 21.65	56.3 ± 21.2	59.07 ± 21.22	51.6 ± 27.4	48.06 ± 16.32	59.2 ± 24.08	NS
Peak systolic strain: basal, %
Longitudinal
Transmural	−12.53 ± 3	−11.93 ± 2.93	−14.71 ± 2.59	−11.64 ± 3.32	−12.44 ± 3.41	−13.34 ± 2.83	<0.0001
Endocardial	−16.2 ± 4.31	−15.94 ± 4.15	−19.75 ± 3.27	−16.65 ± 4.91	−17.05 ± 4.73	−17.31 ± 3.94	<0.0001
Epicardial	−9.51 ± 2.18	−8.75 ± 2.22	−10.53 ± 2.17	−8.06 ± 2.48	−8.8 ± 2.64	−9.91 ± 1.88	<0.0001
Circumferential
Transmural	−11.31 ± 2.53	−10.9 ± 2.36	−12.62 ± 2.15	−10.36 ± 2.72	−11.7 ± 2.55	−11.96 ± 2.41	NS
Endocardial	−16.9 ± 3.77	−16.26 ± 3.4	−18.29 ± 3.11	−16.25 ± 4.51	−17.5 ± 3.57	−17.52 ± 3.78	NS
Epicardial	−7.34 ± 1.82	−7.02 ± 1.75	−8.34 ± 1.56	−6.51 ± 1.95	−7.42 ± 1.9	−7.89 ± 1.56	<0.0001
Radial	24.65 ± 11.66	23.85 ± 12.68	30.38 ± 13.07	19.49 ± 10.32	30.2 ± 15.73	25.73 ± 15.29	NS
Minimum principal
Transmural	−25.18 ± 4.41	−24.99 ± 4.43	−27.06 ± 3.95	−24.29 ± 5.36	−24.57 ± 4.34	−25.73 ± 4.94	NS
Endocardial	−27.36 ± 4.1	−27.02 ± 4.09	−29.17 ± 3.71	−26.49 ± 5.52	−27.09 ± 4.13	−27.96 ± 4.82	NS
Epicardial	−22.89 ± 4.74	−22.66 ± 4.89	−24.58 ± 4.35	−21.64 ± 5.36	−21.82 ± 4.65	−23.34 ± 4.85	NS
Maximum principal	43.06 ± 16.37	43.9 ± 18.39	51.14 ± 19.2	32.26 ± 14.61	50.12 ± 19.09	44.3 ± 20.34	<0.0001
Peak systolic strain: mid, %
Longitudinal
Transmural	−10.07 ± 2.77	−7.72 ± 3.13	−9.2 ± 3.49	−7.41 ± 3.8	−4.23 ± 4.05	−9.5 ± 3.18	<0.0001
Endocardial	−15.54 ± 4.58	−12.33 ± 5.71	−16.14 ± 5.71	−13.21 ± 6.94	−6.98 ± 8.12	−14.57 ± 4.7	<0.0001
Epicardial	−6.34 ± 2.2	−4.65 ± 2.26	−3.71 ± 2.48	−3.37 ± 2.41	−1.77 ± 2.77	−5.51 ± 2.39	<0.0001
Circumferential
Transmural	−10.96 ± 2.35	−9.71 ± 2.49	−10.72 ± 2.3	−8.96 ± 2.54	−8.37 ± 2.48	−10.53 ± 2.99	<0.0001
Endocardial	−18.1 ± 3.84	−16.77 ± 4.05	−18.54 ± 3.82	−16.86 ± 4.63	−15.17 ± 4.91	−16.88 ± 4.51	NS
Epicardial	−6.19 ± 1.78	−5.09 ± 1.73	−5.08 ± 1.56	−4.32 ± 1.57	−4.25 ± 1.71	−5.74 ± 1.84	<0.0001
Radial	39.44 ± 16.56	33.79 ± 15.47	43.35 ± 19.39	32.85 ± 18.68	32.35 ± 18.74	38.96 ± 18.87	NS
Minimum principal
Transmural	−20.94 ± 3.74	−19.95 ± 3.98	−20.83 ± 3.87	−19.02 ± 4.92	−18.37 ± 3.93	−20.35 ± 5.21	NS
Endocardial	−26.48 ± 4.37	−25.93 ± 4.76	−26.43 ± 4.5	−25.6 ± 5.9	−24.77 ± 5.02	−25.03 ± 6.24	NS
Epicardial	−15.5 ± 3.36	−13.98 ± 3.53	−14.26 ± 3.37	−12.58 ± 4.46	−12.19 ± 2.94	−15.25 ± 4.36	<0.0001
Maximum principal	61.7 ± 21.67	54.42 ± 20.27	60.19 ± 22.45	45.86 ± 22.86	51.63 ± 20.78	56.84 ± 23.47	NS
Peak systolic strain: apical, %
Longitudinal
Transmural	−13.41 ± 5.79	−12.03 ± 6.48	−0.69 ± 4.49	−4.49 ± 5.87	−3.9 ± 9.27	−8.34 ± 5.76	<0.0001
Endocardial	−18.88 ± 7.03	−17.09 ± 8.56	−2.34 ± 11.09	−8.73 ± 11.2	−4.87 ± 13.3	−12.76 ± 7.57	<0.0001
Epicardial	−8.61 ± 4.81	−7.52 ± 5.09	1.29 ± 2.95	−1.54 ± 3.52	−2.32 ± 7.18	−4.84 ± 5.84	<0.0001
Circumferential
Transmural	−15.57 ± 4.91	−14.21 ± 5.76	−5.8 ± 3.62	−9.61 ± 5.22	−7.45 ± 7.26	−12.07 ± 6.11	<0.0001
Endocardial	−21.62 ± 5.58	−20.45 ± 6.54	−13.89 ± 6.9	−17.96 ± 6.92	−12.09 ± 9.41	−17.61 ± 6.86	<0.0001
Epicardial	−10.03 ± 4.17	−8.76 ± 4.62	−1.64 ± 2.24	−4.36 ± 3.61	−4.48 ± 5.87	−7.3 ± 5.17	NS
Radial	52.93 ± 37.65	42.93 ± 34.24	47.71 ± 33.72	58.56 ± 44.29	16.37 ± 17.81	53.69 ± 46.4	<0.0001
Minimum principal
Transmural	−24.58 ± 6.05	−23.86 ± 6.89	−15.57 ± 5.17	−18.26 ± 6.5	−17.85 ± 9.4	−20.41 ± 7.01	<0.0001
Endocardial	−30.41 ± 6.08	−30.23 ± 6.65	−25.16 ± 7.18	−27.68 ± 8.75	−23.62 ± 9.32	−26.4 ± 7.01	NS
Epicardial	−18.24 ± 6.06	−17.31 ± 6.92	−7.9 ± 2.97	−10.06 ± 4.51	−13.31 ± 8.89	−14.62 ± 6.64	<0.0001
Maximum principal	89.97 ± 45.1	77.74 ± 44.68	69.27 ± 42.18	89.21 ± 60.66	39.59 ± 23.4	85.09 ± 51.27	<0.0001

Values are mean ± SD.

HCM = hypertrophic cardiomyopathy; NS = not significant; other abbreviation as in Table 1.

FIGURE 1. Proportion of Abnormal GLS According to HCM Morphology.

Abnormal median global longitudinal strain (GLS) is shown in green for hypertrophic cardiomyopathy (HCM) morphologies. P values shown per comparison.

TABLE 3.

Global Peak LV Systolic Strain Stratified According to Presence of LGE

	No LGE (n = 1,147)	LGE Present (n = 1,135)	P Value
Longitudinal, %
Transmural	−11.44 ± 2.58	−10.25 ± 2.58	<0.0001
Endocardial	−16.32 ± 3.55	−14.65 ± 3.56	<0.0001
Epicardial	−7.59 ± 2.19	−6.84 ± 2.25	<0.0001
Circumferential, %
Transmural	−12 ± 2.35	−11.15 ± 2.45	<0.0001
Endocardial	−18.48 ± 3.47	−17.33 ± 3.59	<0.0001
Epicardial	−7.24 ± 1.78	−6.7 ± 1.84	<0.0001
Radial	38.74 ± 17.63	31.8 ± 15.9	<0.0001
Minimum principal, %
Transmural	−23.33 ± 3.68	−22.56 ± 3.73	<0.0001
Endocardial	−27.82 ± 4.05	−27.18 ± 4.14	<0.01
Epicardial	−18.67 ± 3.33	−17.88 ± 3.37	<0.0001
Maximum principal, %	62.36 ± 22.24	55.43 ± 20.28	<0.0001

Values are mean ± SD.

Abbreviations as in Table 1.

TABLE 5.

Global Peak Systolic Strain Stratified According to Sarcomere Mutation Status

	Sarcomere Mutation Negative (n = 1,430)	Sarcomere Mutation Positive (n = 791)	P Value
Longitudinal, %
Transmural	−10.73 ± 2.68	−11.04 ± 2.53	<0.01
Endocardial	−15.52 ± 3.77	−15.45 ± 3.39	0.65
Epicardial	−7 ± 2.23	−7.6 ± 2.21	<0.0001
Circumferential, %
Transmural	−11.47 ± 2.42	−11.75 ± 2.47	<0.01
Endocardial	−17.94 ± 3.62	−17.82 ± 3.51	0.45
Epicardial	−6.8 ± 1.79	−7.27 ± 1.85	<0.0001
Radial	36.33 ± 17.25	33.17 ± 16.59	<0.0001
Minimum principal, %
Transmural	−22.66 ± 3.76	−23.46 ± 3.66	<0.0001
Endocardial	−27.3 ± 4.18	−27.86 ± 4.00	<0.01
Epicardial	−17.9 ± 3.35	−18.93 ± 3.35	<0.0001
Maximum principal, %	58.9 ± 21.7	59.9 ± 21.6	0.08

Values are mean ± SD.

Regional analysis evaluating segmental longitudinal strain vs LV wall thickness by both LV region (basal, mid, and septal) and wall (septal vs lateral) are shown in Figure 2 for the 3 most common HCM morphologic subtypes (isolated basal septal, reverse curvature, and apical). Similar analysis for segmental circumferential¹⁷ and radial strains are shown in Supplemental Figure 4. To test for possible confounding effects of hemodynamic loading, we built a linear mixed effects model that included as fixed effects the predictors in Figure 2 (morphology, level, free wall vs septum, and interactions), as well as mean arterial pressure and the resting outflow gradient. The positive associations of longitudinal strain with wall thickness remained highly significant. The strongest associations between segmental longitudinal strain and wall thickness were observed for apical portions of the left ventricle (both septum and free wall) in reverse curvature and mid septal portions of the left ventricle in apical morphology. Isolated basal septal morphology exhibited consistent correlations throughout the left ventricle.

FIGURE 2. Mean Transmural Longitudinal Strain vs Maximal Wall Thickness Stratified According to HCM Morphology and LV Location.

R and P values shown per comparison. LV = left ventricular; other abbreviation as in Figure 1.

Independent predictors of GLS were evaluated by using elastic net regression to determine the effect size of standardized predictors as illustrated in Figure 3. Coefficient estimates with 95% CIs are presented in Supplemental Table 6. A number of adjusted predictors showed an inverse relationship with GLS (female sex, LVEF, right ventricular end-diastolic volumes, and presence of SAM), whereas others showed a strong positive relationship (apical or other morphologic subtype, LV mass index, reverse curvature morphology, Black or Pacific Islander race, history of syncope, and both presence/extent of LGE and baseline NT-proBNP and troponin levels). GLS was not independently associated with sarcomere mutation status. To test the elastic net regression model, coefficients were estimated by elastic net regression from 75% of the data set (training) and used to predict GLS in the remaining 25% of the data set (testing). The mean absolute error for the model predictions was relatively small compared with the overall variability of GLS (Supplemental Figure 5).

FIGURE 3. Elastic Net Regression for Adjusted Predictors of GLS.

Elastic net regression for adjusted clinical and cardiac magnetic resonance predictors of global longitudinal strain. BAP = bone alkaline phosphatase; BMI = body mass index; CICP = C-terminal propeptide type 1 procollagen; ECV = extracellular volume fraction; Gal = Galectin; LA = left atrial; LAV = left atrial volume; LGE = late gadolinium enhancement; LVEF = left ventricular ejection fraction; MMP = matrix metalloproteinase; NT-proBNP = N-terminal pro-B-type natriuretic peptide; RV = right ventricular; RVEF = right ventricular ejection fraction; SAM = systolic anterior motion; ST2 = suppression of tumorigenicity 2; TIMP = tissue inhibitor metalloproteinase; other abbreviations as in Figures 1 and 2.

Holter monitor abnormalities occurred in 273 patients (196 nonsustained ventricular tachycardia, n = 196; ventricular tachycardia, n = 47; atrial fibrillation, n = 71; and multiple findings, n = 41), whereas ETT abnormalities occurred in 33 patients (ventricular tachycardia, n = 14; atrial fibrillation, n = 20; and multiple findings, n = 1). Univariable and multivariable predictors for abnormal Holter monitoring are shown in Table 6 (ETT results are presented in Supplemental Table 7). Both presence of LGE and GLS (odds ratio: 1.25; 95% CI: 1.09-1.43; P = 0.002) were independent predictors of abnormal Holter testing but neither reached significance for abnormal ETT. The proportion of patients with American Heart Association¹⁸ 2020 Hypertrophic Cardiomyopathy Guideline major risk factors for sudden cardiac death (LVEF <50%, sudden death in a first-degree relative, LV hypertrophy >30 mm, history of syncope, or LV apical aneurysm) were significantly higher for abnormal median GLS (39.3% vs 24.9%; P < 0.0001). Similarly, European Society of Cardiology¹⁹ risk scores were also higher in those with abnormal median GLS (3.03 ± 1.65 vs 2.45 ± 1.65; P < 0.0001), even among those patients without LGE (2.53 ± 2.30 vs 2.11 ± 1.95; P = 0.002). GLS had the highest correlation with calculated European Society of Cardiology risk score among the 3 global strain measures (GLS: r = 0.18; P < 0.0001; GCS: r = 0.15; P < 0.0001; GRS: r = −0.14; P < 0.0001).

TABLE 6.

Univariable and Multivariable Analyses for Predictors of Abnormal Holter Testing

	Univariable		Multivariable (AUC: 0.68)
	Odds Ratio (95% CI)	P Value	Odds Ratio (95% CI)	P Value
Age, y	1.03 (1.02-1.04)	<0.0001	1.04 (1.02-1.05)	<0.001
Female	0.90 (0.67-1.12)	0.46
Body mass index, kg/m2	1.02 (0.99-1.04)	0.14
History of syncope	0.74 (0.49-1.12)	0.15
LV outflow tract obstruction	0.84 (0.58-1.20)	0.33
Sarcomere mutation, positive	1.17 (0.89-1.53)	0.26
LV mass index, g/m2	1.01 (1.00-1.01)	0.02
Maximal wall thickness, mm	1.03 (1.00-1.05)	<0.05
LVEF	0.97 (0.96-0.99)	<0.001
LGE presence	2.51 (1.89-3.34)	<0.0001	1.92 (1.27-2.90)	0.002
Left atrial volume, mL	1.01 (1.00-1.01)	<0.0001
GLS, %	1.18 (1.11-1.23)	<0.0001	1.25 (1.09-1.43)	0.002
Global radial strain, %	0.99 (0.98-0.99)	<0.001
Global circumferential strain, %	1.14 (1.08-1.21)	<0.0001

Abnormal Holter testing was defined as the presence of atrial fibrillation and nonsustained or sustained ventricular tachycardia.

AUC = area under the curve; other abbreviations as in Table 1.

DISCUSSION

We evaluated the adjusted predictors of abnormal GLS measured from comprehensive 3D feature tracking analysis of routine cine CMR in 2,311 patients from HCMR (Central Illustration). We observed a number of important adjusted predictors, including female sex, presence of SAM, LVEF, presence/extent of LGE, baseline NT-proBNP/troponin levels, and apical/reverse curvature morphologic subtypes, among others. Evaluation of these predictors for abnormal myocardial contractile performance in HCM may provide insights into pathophysiological changes associated with higher risk phenotypes in HCM. Better understanding of these morphologic features in patients with HCM may aid differentiation from other hypertrophic phenotypes, such as athlete’s heart or hypertensive heart disease.²⁰ In addition, GLS was an independent predictor of abnormal Holter monitor testing, and both proportion of the American Heart Association major risk factors and calculated European Society of Cardiology risk scores for prediction of sudden cardiac death, inclusive of the subgroup without LGE, suggesting a potential role for prognostic evaluation in patients with HCM. We look forward to examination of relationships between LV strain and adverse cardiac outcomes in HCMR once follow-up of clinical events has been completed.

CENTRAL ILLUSTRATION. Comprehensive 3D Feature Tracking by CMR to Evaluate Adjusted Predictors of GLS from the NHLBI HCMR.

Comprehensive 3D feature tracking by CMR to evaluate adjusted predictors of GLS from the NHLBI (National Heart, Lung, and Blood Institute) HCMR (Hypertrophic Cardiomyopathy Registry). 3D LV myocardial strain from 2,311 cine CMR studies was used to evaluate adjusted clinical and functional/structural predictors of global longitudinal strain in the HCMR. 3D = 3-dimensional; BAP = bone alkaline phosphatase; BMI = body mass index; CICP = C-terminal propeptide type 1 procollagen; ECV = extracellular volume fraction; Gal = Galectin; GLS = global longitudinal strain; HCM = hypertrophic cardiomyopathy; LA = left atrial; LAV = left atrial volume; LGE = late gadolinium enhancement; LV = left ventricular; LVEF = left ventricular ejection fraction; MMP = matrix metalloproteinase; NT-proBNP = N-terminal pro-B-type natriuretic peptide; RV = right ventricular; RVEF = right ventricular ejection fraction; SAM = systolic anterior motion; ST2 = suppression of tumorigenicity 2; TIMP = tissue inhibitor metalloproteinase.

A number of studies have suggested independent prognostic utility of GLS derived by using speckle tracking echocardiography in HCM.^21,22 GLS was an independent predictor of sudden cardiac death in a cohort of 835 patients with HCM followed up for a median of 6.4 years,²² and Cui et al²¹ reported an association with all-cause mortality in a retrospective cohort of 857 patients with HCM and LVOT obstruction who underwent surgical myectomy followed up for a median of 8.3 years. Smaller studies have reported an independent association with GLS and heart failure in nonobstructive HCM²³ in addition to modest associations with NT-proBNP levels.²⁴ A recent systemic review evaluated 14 observational studies with >3,000 patients and reported that the majority of studies consistently observed an association between reduced GLS and risk of adverse cardiac events.²⁵ However, due to the heterogeneity in risk profile, varying duration of follow-up, and absence of several important predictors (ie, obstruction, apical aneurysm, presence/extent of LGE), a meta-analysis could not be conducted by the authors.

No clear cutoff threshold for GLS (ranging from −9.65% to −16%) has been incorporated into risk predictors, including the 2020 American Heart Association/American College of Cardiology HCM guidelines; therefore, the use of GLS for therapeutic decision-making remains exploratory.¹ LV strain, however, remains an attractive imaging biomarker for the potential diagnosis and risk stratification of patients with HCM as it may be routinely measured by using novel feature-tracking analytical tools from standard echocardiographic imaging and CMR (without contrast enhancement), is more sensitive and less affected by loading conditions than LVEF, and is quantitative and has been associated with the detection of both replacement and diffuse myocardial fibrosis.^26,27

Prior small cohort studies have reported an association between strain parameters and both LGE and native T1/ECV.^20,28 In this much larger cohort, we found much stronger direct relationships between strain parameters and the presence/extent of LGE compared with ECV or native T1. Moreover, we also found an association between GLS and reduced LA contractile function, an imaging biomarker associated with diastolic dysfunction in HCM and occurrence of atrial fibrillation in HCMR.^29–31 These results suggest that abnormal myocardial contractile performance and diastolic dysfunction may result to a greater extent from replacement fibrosis and/or regional hypertrophy compared with diffuse interstitial fibrosis or edema in HCM. Abnormal strain values were prevalent in certain HCM morphologies, including apical, apical aneurysm, and reserve septal curvature, and more strongly associated with LV mass index vs maximal wall thickness. This may provide a link between regional contractile performance and higher risk morphologies such as apical aneurysm.³² Abnormal median global strain parameters, particularly median GRS, were associated with a higher proportion of reverse curvature morphology compared with isolated basal septal hypertrophy, and may, therefore, play an important role in diagnosing HCM vs phenocopies (eg, hypertensive heart disease, athletic heart), and prognosis as these patients were younger, more likely sarcomere mutation positive, and represented the majority of cases with >10% LGE.⁶ Despite evidence of greater LV hypertrophy in patients with resting LV outflow tract obstruction, there were no differences in GLS. These results suggest that the development of regional hypertrophy, more diffuse myocardial disarray, and geometric alterations of the left ventricle may predispose to abnormal GLS as opposed to the presence of LVOT obstruction.³³ These findings are also supported by histopathologic studies exhibiting myocardial disarray inclusive of areas of normal wall thickness.^26,34

Women with HCM tend to present with more advanced disease, have greater symptom burden, are older at time of diagnosis, and are at higher risk of both heart failure and arrhythmic outcomes.³⁵ Women have smaller LV chamber size compared with men, which may lead to under recognition of LV hypertrophy with noninvasive imaging. In our analysis, female sex was a strong inverse predictor of abnormal GLS, suggesting potential sex-related differences in phenotypic manifestations or progression of disease. GLS had similar inverse relationship with maximal wall thickness for women and men, and women with <15 mm maximal wall thickness had abnormal GLS compared with published normative ranges.³⁶ Strain analysis provides a more sensitive evaluation of myocardial contractile performance than LVEF, and abnormal GLS may be an earlier marker for disease detection in women. From a prognostic standpoint, an evaluation of 2,226 women with HCM from SHaRe (Sarcomeric Human Cardiomyopathy Registry) reported that women with MYBPC3 sarcomere variants were 35% less likely to develop systolic dysfunction despite 50% higher mortality and greater burden of heart failure (controlling for presence of outflow tract obstruction).³⁵ Other smaller cohort studies have reported similar findings in women using diastolic echocardiographic evaluation.^18,37,38 Our observations also support the notion that diastolic dysfunction may play a more prominent role in the development of adverse outcomes in women with HCM. Future studies should continue to evaluate sex-related differences in prognostic imaging biomarkers of HCM.

Despite characterization of >2,000 variants in at least 11 genes encoding cardiac sarcomere proteins, a substantial proportion of patients with HCM have no identifiable sarcomere mutation.³⁹ Moreover, abnormalities in nonsarcomere components of the ventricle, such as apical displacement of the papillary muscles and elongation of the mitral valve leaflets, suggest a more complex pathobiological mechanism of disease as opposed to a monogenic disease model. Numerous mechanisms, inclusive of LV hypertrophy, myocardial disarray, valvular abnormalities, replacement fibrosis, and vasculopathy, all likely contribute in varying degrees to personalized phenotypic manifestations of disease in HCM. Smaller studies evaluating abnormalities of LV strain in sarcomere mutation-positive patients with no phenotypic manifestations of LV hypertrophy have reported conflicting findings.^40,41 Although we observed differences in all conventional global strain measures between sarcomere mutation-positive and -negative patients, and abnormal median GRS was associated with presence of sarcomere mutation status, we did not observe an independent association with abnormal GLS. Therefore, the impact of sarcomere mutation status on abnormalities of myocardial contractile performance is likely caused by complex epigenetic factors that require more comprehensive evaluation of various morphologic features, including LV hypertrophy, morphology, replacement/diffuse fibrosis, and microvascular disease, in addition to myocardial contractile performance.

Several serum biomarkers have been evaluated for risk stratification of adverse events in HCM. NT-proBNP, galectin-3, and troponin levels have all shown associations with adverse outcomes in HCM.^42–44 Elevations of serum NT-proBNP levels have been shown to correlate with diffuse subclinical fibrosis.⁴⁵ In this analysis, NT-proBNP and troponin levels were both predictors of adverse GLS, suggesting additional association with abnormal myocardial contractile performance. These results lend further consideration for clinical use of serum biomarkers for noninvasive serial evaluation regarding progression of disease or risk stratification of adverse events.

STUDY LIMITATIONS.

Due to the cross-sectional analysis of HCMR, causal relationships between various predictors and abnormal LV strain cannot be determined. In addition, there was a lower prevalence of patients with more severe hypertrophy and fibrosis (measured by LGE) than in some other published cohorts¹⁸; the results of this strain analysis may therefore not apply to HCM patients with more advanced disease. All CMR cine imaging was analyzed with a single well-validated feature-tracking post-processing tool. In current clinical practice, standardization is lacking across numerous different proprietary software tools that are used to derive strain, which can produce variation in derived values and agreed-upon cutoff values for therapeutic decision-making. In addition, although analysis of strain based on feature tracking allows both prospective and retrospective application to acquired CMR studies, it may have lower reproducibility compared with dedicated techniques such as strain-encoded imaging or tagging.²⁰ Finally, the absence of comparator groups limits analysis of the diagnostic performance of strain for identifying HCM from other hypertrophic phenotypes.

CONCLUSIONS

We evaluated independent predictors of abnormal GLS in the HCMR. These markers provide further insight into the utility of routine myocardial LV strain evaluation in the diagnosis and potential risk stratification of patients with HCM. Further analysis to evaluate independent associations of strain with adverse arrhythmic and heart failure outcomes from the HCMR is warranted to determine potential routine clinical utility and potential inclusion in risk scores.

TABLE 4.

Global Peak LV Systolic Strain Stratified According to Sex

	Male (n = 1,641)	Female (n = 669)	P Value
Longitudinal, %
Transmural	−15.07 ± 3.48	−16.45 ± 3.89	<0.0001
Endocardial	−7.08 ± 2.18	−7.53 ± 2.4	<0.0001
Epicardial	−10.55 ± 2.52	−11.55 ± 2.82	<0.0001
Circumferential, %
Transmural	−17.45 ± 3.48	−18.96 ± 3.59	<0.0001
Endocardial	−6.8 ± 1.78	−7.38 ± 1.89	<0.0001
Epicardial	−11.26 ± 2.36	−12.33 ± 2.47	<0.0001
Radial	32.76 ± 15.15	41.27 ± 19.99	<0.0001
Minimum principal, %
Transmural	−26.82 ± 3.98	−29.12 ± 4.01	<0.0001
Endocardial	−17.81 ± 3.26	−19.38 ± 3.43	<0.0001
Epicardial	−22.35 ± 3.59	−24.36 ± 3.72	<0.0001
Maximum principal, %	54.35 ± 18.52	69.85 ± 24.52	<0.0001

Values are mean ± SD.

Abbreviation as in Table 1.

PERSPECTIVES.

COMPETENCY IN MEDICAL KNOWLEDGE:

The determinants of abnormal GLS were evaluated in the HCMR using comprehensive 3D LV myocardial strain from cine CMR. The adjusted predictors of GLS included female sex, Black race, history of syncope, presence of systolic anterior motion of the mitral valve, reverse curvature and apical morphologies, LV ejection fraction, LV mass index, and both the presence and extent of LGE and baseline serum NT-proBNP and troponin levels. Strain parameters from routinely acquired CMR may provide further insights into the pathophysiology and determinants of higher risk phenotypes in HCM.

TRANSLATIONAL OUTLOOK:

The absence of comparator groups and association of strain parameters with adverse cardiac outcomes are the chief limitations. Comparison of strain parameters derived by similar feature-tracking CMR techniques are needed to help determine the diagnostic utility of strain in differentiating HCM from phenocopies (eg, hypertensive heart disease, athlete’s heart, Fabry/amyloid cardiomyopathy). Long-term follow-up of the HCMR will provide important data to determine the independent prognostic utility of strain in HCM for adverse cardiac events.

ABBREVIATIONS AND ACRONYMS

3D

3-dimensional

CMR

cardiac magnetic resonance

ECV

extracellular volume fraction

GCS

global circumferential strain

GLS

global longitudinal strain

GRS

global radial strain

HCM

hypertrophic cardiomyopathy

LGE

late gadolinium enhancement

LVEF

left ventricular ejection fraction

LVOT

left ventricular outflow tract

NT-proBNP

N-terminal pro-B-type natriuretic peptide

SAM

systolic anterior motion