Left Ventricular Hypertrophy With Strain and Aortic Stenosis
- Other version(s) of this article
You are viewing the most recent version of this article. Previous versions:
Abstract
Background—
ECG left ventricular hypertrophy with strain is associated with an adverse prognosis in aortic stenosis. We investigated the mechanisms and outcomes associated with ECG strain.
Methods and Results—
One hundred and two patients (age, 70 years [range, 63–75 years]; male, 66%; aortic valve area, 0.9 cm2 [range, 0.7–1.2 cm2]) underwent ECG, echocardiography, and cardiovascular magnetic resonance. They made up the mechanism cohort. Myocardial fibrosis was determined with late gadolinium enhancement (replacement fibrosis) and T1 mapping (diffuse fibrosis). The relationship between ECG strain and cardiovascular magnetic resonance was then assessed in an external validation cohort (n=64). The outcome cohort was made up of 140 patients from the Scottish Aortic Stenosis and Lipid Lowering Trial Impact on Regression (SALTIRE) study and was followed up for 10.6 years (1254 patient-years). Compared with those without left ventricular hypertrophy (n=51) and left ventricular hypertrophy without ECG strain (n=30), patients with ECG strain (n=21) had more severe aortic stenosis, increased left ventricular mass index, more myocardial injury (high-sensitivity plasma cardiac troponin I concentration, 4.3 ng/L [interquartile range, 2.5–7.3 ng/L] versus 7.3 ng/L [interquartile range, 3.2–20.8 ng/L] versus 18.6 ng/L [interquartile range, 9.0–45.2 ng/L], respectively; P<0.001) and increased diffuse fibrosis (extracellular volume fraction, 27.4±2.2% versus 27.2±2.9% versus 30.9±1.9%, respectively; P<0.001). All patients with ECG strain had midwall late gadolinium enhancement (positive and negative predictive values of 100% and 86%, respectively). Indeed, late gadolinium enhancement was independently associated with ECG strain (odds ratio, 1.73; 95% confidence interval, 1.08–2.77; P=0.02), a finding confirmed in the validation cohort. In the outcome cohort, ECG strain was an independent predictor of aortic valve replacement or cardiovascular death (hazard ratio, 2.67; 95% confidence interval, 1.35–5.27; P<0.01).
Conclusion—
ECG strain is a specific marker of midwall myocardial fibrosis and predicts adverse clinical outcomes in aortic stenosis.
Aortic stenosis is characterized by progressive valve narrowing of and secondary changes in the myocardium.1 In response to increased afterload, left ventricular hypertrophy (LVH) can develop to maintain wall stress and cardiac function. Although this process appears to be compensatory in the early stages, preclinical studies have suggested that cardiac performance can be preserved in the absence of hypertrophy.2,3 Moreover, the LVH response ultimately decompensates with progressive cell death and fibrosis, driving the transition to symptoms, heart failure, and adverse cardiovascular events.1,4,5 There is therefore considerable interest in identifying early, objective markers of this decompensation that might identify asymptomatic patients who would benefit from early valve replacement.
Clinical Perspective on p 1616
ECG strain is a well-recognized marker of LVH. However, the exact mechanism underlying the characteristic ST- and T-wave abnormalities associated with this pattern is uncertain. In this study, we hypothesized that ECG strain is a marker of left ventricular decompensation and investigated this association using cardiovascular magnetic resonance (CMR) to assess the degree of LVH and myocardial fibrosis and high-sensitivity plasma cardiac troponin I (cTnI) as a marker of myocardial injury. Moreover, we aimed to reassess the adverse prognosis previously associated with the ECG strain pattern in patients with aortic stenosis.6
Methods
Three cohorts were used for the study. In the mechanism cohort, we determined the pathophysiology underlying the ECG strain pattern using CMR and plasma cTnI concentration in patients recruited from the Edinburgh Heart Center. This was then independently validated in an external validation cohort from the Royal Brompton Hospital, London. Subsequently, in the outcome cohort, we examined the prognostic role of ECG strain in patients with aortic stenosis. The study was conducted in accordance with the Declaration of Helsinki and was approved by the local research ethics committee. Written informed consent was obtained from all participants.
Patient Populations
Mechanism Cohort
Patients with aortic stenosis (mild to severe) were recruited prospectively from the Edinburgh Heart Center. We excluded patients with other significant valvular heart disease (moderate or severe), contraindications to CMR, cardiomyopathies (acquired or inherited), left or right bundle-branch block, concurrent digoxin use, and impaired systolic function on CMR (ejection fraction <95th percentile for age and sex).7
Validation Cohort
Between 2011 and 2013, patients with moderate to severe aortic stenosis undergoing CMR were prospectively recruited from the Royal Brompton Hospital, London, with the use of similar exclusion criteria.
Outcome Cohort
Patients were initially recruited into the Scottish Aortic Stenosis and Lipid Lowering Trial, Impact on Regression (SALTIRE) study between March 2001 and April 2002, which comprised 155 patients with asymptomatic aortic stenosis who had been randomly assigned to either atorvastatin 80 mg or placebo once daily. Patients were excluded if already on a statin or if aortic valve replacement (AVR) was planned (because of either symptoms or impaired systolic function).8 For the purposes of this analysis, patients on digoxin or with uninterpretable ECGs or bundle-branch block patterns were excluded.
Electrocardiography
A standard 12-lead ECG was obtained for all participants, and interpretation of the ECG was performed independently by 2 observers who were blinded to the clinical data and imaging findings. LVH on ECG was diagnosed on the basis of the Romhilt-Estes point system (≥5 points),9 and ECG strain was defined as ≥1-mm concave down-sloping ST-segment depression with asymmetrical T-wave inversion in the lateral leads (Figure 1A).10
Figure 1. ECGs and myocardial biopsies in 2 patients with severe aortic stenosis. The ECG for patient A (A) demonstrated left ventricular hypertrophy and associated repolarization abnormalities (ST-segment depression and asymmetrical T-wave inversion in the lateral leads) characteristic of the ECG strain pattern, whereas the ECG for patient B (B) demonstrated left ventricular hypertrophy without the strain pattern. Compared with patient B, patient A had increased left ventricular mass index (169 vs 81 g/m2), increased plasma cardiac troponin I concentrations (8.4 vs 2.5 ng/L), and evidence of more extensive myocardial fibrosis on both cardiovascular magnetic resonance and histology (picrosirius red staining).
Echocardiography
Transthoracic echocardiography was performed in all participants in the mechanism and outcome cohorts. Maximum aortic valve jet velocity and mean pressure gradient were measured by velocity–time integral spectral tracing and the aortic valve area derived with the continuity equation. Multiple acoustic windows with the S51 and D2cwc probes (Philips Medical Systems, Best, the Netherlands) were assessed. The severity of aortic stenosis was classified according to the European Association of Echocardiography/American Society of Echocardiography guidelines.11 Transmitral early (E) and late (A) diastolic velocities and deceleration time of early filling velocity were measured at the tips of the mitral valve leaflets with pulsed-wave Doppler. Early (e′) diastolic velocities of the medial and lateral mitral annulus were measured with pulsed-wave tissue Doppler imaging. Diastolic function was determined from the E/A ratio, deceleration time, mean of medial and lateral e′, and E/e′. Midwall fractional shortening was estimated as an assessment of intrinsic myocardial contractility in the context of LVH.12
Cardiovascular Magnetic Resonance
CMR in the mechanism cohort was performed at 3 T (MAGNETOM Verio, Siemens AG, Healthcare Sector, Erlangen, Germany). In the validation cohort, CMR was performed at 1.5 T, as previously described.13 For the assessment of left ventricular function and mass, short-axis cine images from the mitral valve annulus to the apex were obtained by use of a balanced steady-state free-precession sequence (8-mm parallel slices with 2-mm spacing). The quantification of left ventricular function, volumes, and mass was assessed with dedicated software (Siemens AG Healthcare Sector, Erlangen, Germany) and indexed to body surface area. LVH on CMR was defined as a left ventricular mass index (LVMi) >95th percentile using age- and sex-specific normal ranges.7 Left ventricular longitudinal shortening was determined by measuring the difference in mitral annular displacement between end systole and end diastole. The mean value of the lateral and septal insertion sites (4-chamber view) and the anterior and inferior sites (2-chamber view) was used.
The assessment of focal replacement myocardial fibrosis was performed with late gadolinium enhancement (LGE) imaging, 15 minutes after administration of 0.1 mmol/kg gadobutrol (Gadovist/Gadavist, Bayer Pharma AG, Berlin, Germany). Two approaches were used: an inversion recovery fast gradient-echo sequence and a phase-sensitive inversion recovery sequence, performed in 2 phase-encoding directions to differentiate true late enhancement from artifact. The inversion time was optimized to achieve satisfactory nulling of the myocardium. Midwall LGE was determined visually by 2 independent operators who were blinded to the ECG findings. The amount of LGE was quantified with QMASS software (Medis Medical Imaging Systems, Leiden, the Netherlands) using a signal intensity threshold greater than twice the standard deviation above the mean value in a normal region of myocardium sampled on the same short-axis image. Areas thought to represent inversion artifact or contamination by blood pool or epicardial fat were manually excluded.
Myocardial extracellular volume fraction (ECV) has been demonstrated to act as a measure of diffuse myocardial fibrosis in a variety of cardiac conditions, including aortic stenosis.14–16 Recently, we have described a standardized approach to analyze myocardial ECV in patients with aortic stenosis, demonstrating excellent intraobserver, interobserver, and scan-rescan reproducibility of ±3%.17 In brief, myocardial T1 mapping was performed in the mechanism cohort using the modified look-locker inversion recovery sequence: flip angle, 35°; minimum TI, 100 milliseconds; TI increment, 80 milliseconds; and time delay, 150 milliseconds with a heartbeat acquisition scheme of 3-3-5.18 Regions of interest were drawn around the myocardium on the short-axis, precontrast, motion-corrected myocardial T1 maps and copied onto corresponding 20-minute postcontrast maps, with minor adjustments made to avoid partial volume effects and artifact (OsiriX version 4.1.1, Geneva, Switzerland). ECV was calculated according to the following formula: ECV=(ΔR1myocardium/ΔR1blood-pool)×(1−hematocrit), where ΔR1=(1/postcontrast T1−1/precontrast T1). Hematocrit was sampled at the time of CMR.
High-Sensitivity Plasma cTnI Assay
Plasma cTnI concentrations were determined in patients in the mechanism cohort as a marker of myocyte injury with the ARCHITECTSTAT high-sensitive troponin I assay (Abbott Laboratories, Abbott Park, IL). Previous data have shown that high-sensitivity plasma cTnI concentrations correspond to the magnitude of the hypertrophic response and extent of myocardial fibrosis in patients with aortic stenosis.19,20 The lower limit of detection for this assay is 1.2 ng/L, and the 99th percentile from a healthy reference population of 26 ng/L, with a 10% interassay coefficient of variation at 4.7 ng/L.21 Concentrations lower than the detection limit were assigned a value of 1.2 ng/L.
Calcium Scoring in the Outcome Cohort
ECG-gated noncontrast computed tomography scans of the coronary arteries and aortic valve were performed in all patients in the outcome cohort with a double-helix scanner (Twin II Flash, Philips Medical Systems). Coronary artery and aortic valve calcium scores were determined by a single operator using the Picker Cardiac Scoring software.8
Long-Term Follow-Up in the Outcome Cohort
Clinical outcomes were obtained in the outcome cohort and adjudicated by 2 independent investigators blinded to the clinical and ECG data. In-hospital and community deaths were captured from the General Register of Scotland. Cardiovascular death was defined as death resulting from myocardial infarction, sudden cardiac death, heart failure, or stroke; death related to cardiovascular procedures; and death resulting from other cardiovascular causes. Each death was classified by the 2 independent investigators, and any discrepancy was resolved by consensus. Furthermore, all events, including surgical AVR (no patients had transcatheter aortic valve implantation during follow-up), were confirmed by independent review of each patient’s healthcare record. All patients in the outcome cohort were managed in our tertiary cardiac center and reviewed at a multidisciplinary meeting before undergoing AVR. Only patients with established indications as per contemporary guidelines were referred for AVR.22,23
Statistical Analysis
Continuous variables were presented as mean±SD or median (interquartile range) as appropriate. The distribution of all continuous variables was assessed for normality with the Shapiro-Wilk test. Comparison of continuous variables was performed with the Student t test or 1-way ANOVA with post hoc Bonferroni adjustment when appropriate. The assumption of homogeneity of variances was tested with the Levene test. For nonparametric data, the Mann-Whitney U or Kruskal-Wallis test was used. Categorical variables were expressed as percentages and compared by use of the χ2 test for trend. All statistical analyses were performed with GraphPad Prism (GraphPad Software Inc, San Diego, CA), R version 2.15.2 (Vienna, Austria), and SPSS version 20.0.0 (IBM Corp, Armonk, NY). A 2-sided value of P<0.05 was considered statistically significant.
Mechanism Cohort
In the mechanism cohort, the association between ECG strain and left ventricular mass and aortic stenosis severity was assessed with multivariable linear regression analysis to adjust for potential confounders. Furthermore, we assessed determinants associated with ECG strain using univariate and multivariable logistic regression.
Outcome Cohort
In the outcome cohort, time-to-event curves in patients with and without ECG strain were estimated with the Kaplan–Meier method and compared by use of the log-rank test. To accommodate for competing risks, the association between time to AVR or cardiovascular death and ECG strain was modeled as a composite outcome in Cox proportional hazards models. The assumption for proportional hazards was assessed using the log (−log[survival]) versus log(survival time) plots and by examining the Schoenfeld residuals.
Results
One hundred and two patients with aortic stenosis (age, 70 years [interquartile range, 63–75 years]; male, 66%; aortic valve area, 0.9 cm2 [interquartile range, 0.7–1.2 cm2]) were recruited into the mechanism cohort, and an additional 64 patients were recruited into the validation cohort (age, 76 years [interquartile range, 69–84 years]; male, 69%; aortic valve area, 0.9 cm2 [interquartile range, 0.7–1.0 cm2]; Tables 1 and 2). After the exclusion of patients with uninterpretable ECGs or bundle-branch block and those receiving digoxin therapy (n=15), 140 patients from the SALTIRE study were analyzed as part of the outcome cohort (age, 69 years [interquartile range, 62–75] years; male, 70%; aortic valve area, 1.0 cm2 [interquartile range, 0.7–1.3 cm2]; Table 3). All patients in the mechanism and outcome cohorts were white. In the validation cohort, 92% were white and the remainder were South Asian. There were no observed racial differences with respect to the presence of LVH or strain on the ECG (P=0.95; Table 2).
| Patients With Aortic Stenosis | |||||
|---|---|---|---|---|---|
| All Patients (n=102) | No LVH (n=51) | LVH Without Strain (n=30) | ECG Strain (n=21) | P Value | |
| Clinical characteristics | |||||
| Age, y Male sex, n (%) Diabetes mellitus, n (%) Hypertension, n % Coronary artery disease, n (%) Systolic blood pressure, mm Hg Bicuspid aortic valve, n (%) NYHA class III and IV, n (%) | 70 (63–75)67 (66)8 (8)61 (60)28 (27)146±2133 (32)22 (22) | 70 (65–75)30 (59)5 (10)29 (58)11 (22)146±2213 (25)4 (8) | 70 (65–73)23 (77)2 (7)21 (70)9 (30)146±2012 (40)6 (20) | 69 (61–75)16 (76)2 (10)11 (52)8 (38)147±208 (38)12 (57) | 0.830.090.880.990.140.980.21<0.001 |
| ECG | |||||
| Aortic valve area, cm2 Aortic jet velocity, m/s MPG, mm Hg Dimensionless index | 0.9 (0.7–1.2)3.8±1.030 (20–39)0.29±0.10 | 1.0 (0.7–1.3)3.2±0.723 (14–32)0.31±0.10 | 0.9 (0.7–1.1)3.8±0.831 (22–41)0.28±0.09 | 0.7 (0.6–0.9)4.8±1.145 (37–64)0.25±0.09 | 0.02<0.001*†‡<0.0010.03‡ |
| Midwall fractional shortening, mm Mitral E/A ratio Deceleration time, ms Mean e′, cm/s E/e′ ratio | 10.9±1.71.0±0.4210±566.2±1.912.5 (10.0–16.7) | 11.2±1.90.9±0.4197±516.7±2.011.7 (9.4–14.9) | 10.9±1.51.0±0.4214±576.3±1.712.3 (9.3–15.4) | 9.9±1.20.9±0.4235±574.9±1.517.0 (13.0–23.0) | 0.01‡0.510.02<0.01‡<0.001†‡ |
| Cardiac MRI | |||||
| Indexed LV mass, g/m2 LV mass/EDV ratio, g/mL Indexed EDV, mL/m2 Indexed ESV, mL/m2 Indexed stroke volume, mL Ejection fraction, % Longitudinal shortening, mm Patients with midwall LGE, n (%) Amount of LGE, % ECV, % | 91±241.27±0.2669 (62–78)22 (18–27)49±1068±612.4±3.132 (31)0 (0–5.5)28.1±2.8 | 75±141.14±0.2267 (60–71)22 (17–25)45±868±513.1±2.74 (8)3.9 (1.8–7.0)27.4±2.2 | 99±181.33±0.2177 (69–88)25 (21–29)51±1067±612.9±3.17 (23)5.8 (5.0–7.6)27.2±2.9 | 118±221.51±0.2273 (65–86)23 (19–29)54±1268±89.9±2.721 (100)9.5 (7.5–14.2)30.9±1.9 | <0.001*†‡<0.001*†‡<0.0010.03<0.01*‡0.90<0.001†‡<0.001<0.01<0.001†‡ |
| Plasma high-sensitivity cTnI concentration, ng/L | 6.7(3.6–13.3) | 4.3(2.5–7.3) | 7.3(3.2–20.8) | 18.6(9.0–45.2) | <0.001 |
| Patients With Aortic Stenosis | |||||
|---|---|---|---|---|---|
| All Patients (n=64) | No LVH (n=48) | LVH Without Strain (n=5) | ECG Strain (n=11) | P Value | |
| Clinical characteristics | |||||
| Age, y Sex, males, n (%) Diabetes mellitus, n (%) Hypertension, n % Coronary artery disease, n (%) Systolic blood pressure, mm Hg Bicuspid aortic valve, n (%) NYHA class III and IV, n (%) Race, n (%) White South Asian | 76 (69–84)44 (69)16 (25)33 (52)26 (41)133 (119–142)14 (22)14 (22)59 (92)5 (8) | 78 (68–83)32 (67)13(27)26(54)21(44)134 (121–142)11 (23)11 (23)44 (92)4 (8) | 80 (63–89)4 (80)01 (20)1 (20)132 (124–158)005 (100)0 | 80 (73–87)8 (73)3 (27)6 (55)4 (36)123 (110–140)3 (27)3 (27)10 (91)1 (9) | 0.340.260.780.760.510.490.970.970.95 |
| Cardiac MRI | |||||
| Planimetered aortic valve area, cm2 Indexed LV mass, g/m2 LV mass/EDV ratio, g/mL Indexed EDV, mL/m2 Indexed ESV, mL/m2 Indexed stroke volume, mL Ejection fraction, % Patients with midwall LGE, n (%) | 0.9 (0.7–1.0)88 (74–113)1.17 (0.86–1.38)76 (67–104)28 (17–48)48 (38–55)64 (46,72)25 (39) | 0.8 (0.7–1.0)85 (72–107)1.18 (0.91–1.35)74 (64–95)27 (17–42)47 (40–54)65 (51,73)12 (25) | 1.0 (0.7–1.3)82 (74–113)1.05 (0.88–1.42)86 (67–97)29 (17–36)55 (47–65)68 (63–76)3 (60) | 0.7 (0.7–0.9)121 (102–133)1.13 (0.79–1.48)120 (75–153)65 (38–102)49 (36–52)44 (32–48)10 (91)* | 0.120.020.920.01<0.010.19<0.010.02 |
| Patients With Aortic Stenosis | ||||
|---|---|---|---|---|
| All Patients (n=140) | No ECG Strain (n=120) | ECG Strain (n=20) | P Value | |
| Clinical characteristics | ||||
| Age, y Male sex, n (%) Diabetes mellitus, n (%) Hypertension, n (%) Coronary artery disease, n (%) Systolic blood pressure, mm Hg | 69 (62–75)98 (70)4 (3)71 (50)24 (17)144±19 | 69 (61–75)82 (68)4 (3)58 (48)20 (17)144±20 | 75 (66–77)16 (80)013 (65)4 (20)142±16 | 0.050.430.590.230.750.65 |
| ECG | ||||
| Aortic valve area, cm2 Peak aortic jet velocity, m/s MPG, mm Hg Ejection fraction, % LVMi, g/m2 | 1.0 (0.7–1.3)3.4 (2.8–4.0)24 (18–35)69±10173 (142–205) | 1.0 (0.7–1.3)3.2 (2.8–3.9)22 (17–33)70±11164 (131–200) | 0.6 (0.4–0.8)3.9 (3.5–4.4)34 (26–44)69±9203 (177–223) | 0.03<0.01<0.0010.70<0.01 |
| Computed tomography | ||||
| Coronary calcium score, log AU Aortic valve calcium score, log AU | 1.6±1.23.6±0.6 | 1.6±1.33.6±0.6 | 1.7±1.24.0±0.4 | 0.690.01 |
| Plasma high-sensitivity cTnI concentration, ng/L | 7.5 (5.7–13.4) | 6.9 (5.3–11.4) | 17.3 (10.5–29.6) | <0.001 |
Mechanisms Underlying ECG Strain
Fifty-one patients in the mechanism cohort fulfilled the ECG criteria for LVH, demonstrating high predictive values for the presence of CMR-defined LVH (positive predictive value, 96%; negative predictive value, 89%). Of these, 21 patients had the strain pattern on their ECGs. These patients with ECG strain had the highest LVMi and most severe aortic stenosis compared with the other patient groups (those without LVH on their ECG and those with LVH but no ECG strain; Table 1), even after adjustment for age, sex, and systolic blood pressure (P<0.001 for both). Moreover, compared with other groups, these patients had increased end-diastolic volumes (P<0.01), worse diastolic function (P<0.001), and more severe symptoms (P<0.001; Table 1). Despite similar left ventricular ejection fractions, patients with LVH and ECG strain also had the worst longitudinal function (Figure 2) and midwall fractional shortening (Table 1).
Figure 2. Despite similar normal-range ejection fractions (A), patients with left ventricular hypertrophy (LVH) and ECG strain had the most impaired longitudinal shortening (B) and diastolic function (C). High-sensitivity plasma cardiac troponin I concentrations were 4-fold higher in patients with ECG strain compared with patients without LVH on ECG (D). Results are presented in box-and-whiskers plot (Tukey method).
Interestingly, all patients with LVH and ECG strain had focal midwall fibrosis (positive and negative predictive value, 100% and 86%, respectively; Figure 3B), strongly supporting ECG strain as a specific marker of replacement myocardial fibrosis. Moreover, these patients had more extensive diffuse myocardial fibrosis (ECV, 30.9±1.9% versus 27.2±2.9% in patients with LVH and no ECG strain versus 27.4±2.2% in patients without LVH; P<0.001; Figure 3A) and myocardial injury as assessed by high-sensitivity plasma cTnI. Indeed, plasma cTnI concentrations were >4-fold higher in patients with strain than in the other patient groups (18.6 ng/L [interquartile range, 9.0–45.2 ng/L] versus 7.3 ng/L [interquartile range, 3.2–20.8 ng/L] in patients with LVH and no ECG strain versus 4.3 ng/L [interquartile range, 2.5–7.3 ng/L] in patients without LVH; P<0.001; Figure 2D). Three patients with ECG strain had both an infarct and midwall pattern of fibrosis on LGE, and our findings remained unchanged even after their exclusion.
Figure 3. Patients with the strain pattern on the ECG had increased extracellular volume fractions, suggestive of increased diffuse myocardial fibrosis (A). Furthermore, all patients with ECG strain had a midwall pattern of late gadolinium enhancement (B). Of note, about a third of patients with midwall late gadolinium enhancement did not have ECG strain. The corresponding myocardial T1 map (A) and late gadolinium enhancement image (B) of a patient with ECG strain demonstrated evidence of focal myocardial fibrosis in the midcavity lateral wall. The extracellular volume fraction of the midcavity slice in this patient was 30.2%. LVH indicates left ventricular hypertrophy.
On univariate analysis, ECG strain was associated with an increased LVMi, more severe aortic stenosis, increased replacement and diffuse myocardial fibrosis, and diastolic dysfunction (all P<0.01; Table 4) but was not associated with the presence of coronary artery disease (odds ratio, 1.88; 95% confidence interval, 0.68–5.18; P=0.22). However, only increased myocardial fibrosis (either amount of LGE or ECV) and the severity of aortic stenosis maintained an independent association on multivariate analysis, with increased LVMi, increased myocardial injury, and diastolic dysfunction all dropping out of the model (models 3 and 4 in Table 4).
| Univariate Analysis | Multivariate Analysis(Model 1) | Multivariate Analysis(Model 2) | Multivariate Analysis(Model 3) | |||||
|---|---|---|---|---|---|---|---|---|
| OR (95% CI) | P Value | OR (95% Cl) | P Value | OR (95% Cl) | P Value | OR (95% Cl) | P Value | |
| Age, per 10 y | 0.91 (0.61–1.36) | 0.56 | 0.88 (0.38–2.03) | 0.76 | … | … | … | … |
| Male sex | 1.69 (0.56 -5.10) | 0.35 | 0.54 (0.07–3.93) | 0.54 | … | … | … | … |
| Coronary artery disease | 1.88 (0.68–5.18) | 0.22 | … | … | … | … | … | … |
| MPG, per 10 mm Hg | 1.80 (1.31–2.48) | <0.001 | 1.88 (1.02–1.13) | <0.01 | 1.93 (1.04–3.60) | 0.03 | 2.10 (1.22–3.60) | 0.01 |
| LVMi, per 10 g/m2 | 2.10 (1.49–2.95) | <0.001 | 1.95 (1.14–3.35) | <0.01 | 1.30 (0.63–2.66) | 0.47 | 1.77 (0.97–3.22) | 0.06 |
| Amount of LGE, % | 1.75 (1.35–2.27) | <0.001 | … | … | 1.73 (1.08–2.77) | 0.02 | … | … |
| ECV, % | 1.86 (1.38–2.47) | <0.001 | … | … | … | … | 1.55 (1.04–2.31) | 0.03 |
| High-sensitivity cTnI* | 3.14 (1.73–5.71) | <0.001 | 3.30 (1.24–8.80) | 0.02 | 3.18 (0.62–16.26) | 0.16 | 2.43 (0.83–7.10) | 0.11 |
| Mean e′ | 0.51 (0.34–0.75) | <0.01 | … | … | 1.71 (0.38–7.54) | 0.71 | 0.95 (0.46–1.94) | 0.88 |
Myocardial histology was available in 2 patients who underwent AVR and biopsy, supporting increased myocardial fibrosis in patients with LVH and ECG strain (Figure 1). However, not all patients with myocardial LGE had a strain pattern on the ECG. Indeed, of the 32 patients with myocardial LGE, 11 patients (34%) did not have any evidence of ECG repolarization abnormalities. Interestingly, these patients had ≈40% less replacement fibrosis on LGE compared with patients who had ECG strain (5.6% [interquartile range, 4.3%–7.5%] versus 9.5% [interquartile range, 7.5%–14.2%]; P=0.002), with no differences in the distribution of midwall LGE between these groups (P=0.78; see Distribution of Midwall Fibrosis in the online-only Data Supplement).
Validation Cohort
In the external validation cohort, similar findings were demonstrated (Table 2). There were 11 patients with ECG strain, of whom 10 had isolated midwall fibrosis and 1 had extensive fibrosis from a large myocardial infarct to explain the ECG changes. Conversely, 15 patients had midwall fibrosis but no ECG strain. In this cohort of patients with moderate to severe aortic stenosis, the positive and negative predictive values of LVH with ECG strain for midwall fibrosis were 91% and 72%, respectively. Patients with ECG strain were again observed to have an advanced hypertrophic response associated with increased LVMi and reduced myocardial performance (Table 2).
Prognostic Value of ECG Strain
In the outcome cohort, 20 patients (14%) had LVH with strain on ECG. Consistent with the mechanism cohort, patients with ECG strain had more severe aortic stenosis, increased LVMi, and elevated plasma cTnI concentrations compared with those without strain (Table 3). Of note, these elevated cTnI concentrations in patients with ECG strain were similar to those observed in the mechanism cohort (P=0.85). Over 10.6 years of follow-up (1254 patient-years), 63 patients had an AVR and 22 patients died of a cardiovascular cause of a total of 36 deaths. ECG strain was associated with a lower 10-year event-free survival rate for AVR or cardiovascular death (log-rank test, P<0.0001; Figure 4). Indeed, this association persisted even after adjustment for traditional markers of an adverse outcome, including systolic ejection fraction, severity of aortic stenosis, LVMi, and aortic valve calcium score (hazard ratio, 2.67; 95% confidence interval, 1.35–5.27; P<0.01; see Univariate and Adjusted Cox Models Predicting Time to Adverse Events in the online-only Data Supplement).
Figure 4. Kaplan–Meier estimates of time to event by status of ECG strain in the outcome cohort. Patients with ECG strain had significantly lower event-free survival compared with patients without ECG strain. LVH indicates left ventricular hypertrophy.
Discussion
This is the first CMR study to investigate the mechanisms underlying the ECG strain pattern in patients with aortic stenosis, demonstrating that it is a highly specific marker of midwall myocardial fibrosis. Moreover, ECG strain was associated with increased myocardial injury and impaired left ventricular performance and was an independent predictor of cardiovascular death or AVR. Our data therefore indicate that ECG strain is a powerful biomarker of left ventricular decompensation in aortic stenosis, with the ability to identify an at-risk population who may benefit from earlier valve replacement.
Currently, AVR is recommended in patients with severe aortic stenosis who have symptoms or evidence of left ventricular decompensation with an ejection fraction <50%.23 However, symptoms are often subjective, particularly in the elderly, and a reduced ejection fraction is frequently a late manifestation and is not necessarily reversible. There is therefore interest in exploring alternative, earlier, and more objective markers of ventricular decompensation in aortic stenosis.5
Previous echocardiographic studies have demonstrated that ECG strain is associated with an advanced hypertrophic response,24 and it has been hypothesized that the characteristic repolarization abnormalities relate to coronary perfusion abnormalities, even in the absence of coronary artery disease.25–28 Our study adds to these data, demonstrating a close association between ECG strain and myocardial injury and fibrosis. Indeed, across 2 independent cohorts, midwall myocardial fibrosis was present in 31 of the 32 patients with strain on their ECGs, and the remaining subject had an extensive infarct to explain the ECG changes. Moreover, patients with strain had evidence of higher plasma cTnI concentrations and worse myocardial function. It has been established that myocardial ischemia, cell death, and fibrosis are all key features that characterize the transition from hypertrophy to heart failure in aortic stenosis. Our study would therefore support ECG strain as a useful marker of left ventricular decompensation in patients with this condition.
In our outcome cohort, we have demonstrated that ECG strain acts as a strong independent predictor of AVR or cardiovascular death, over and above established prognostic markers such as systolic ejection fraction, severity of aortic stenosis, LVMi, and aortic valve calcium score. Indeed, patients with ECG strain had a >2-fold increase risk in adverse events compared with patients without ECG strain. This is in agreement with previous studies that have demonstrated an adverse prognosis associated with ECG strain.6,29,30 However, our study provides much longer periods of follow-up than have been described previously.
There are clear potential advantages of using ECG strain as a marker of left ventricular decompensation in aortic stenosis. A 12-lead ECG is readily available, cheap, and rapidly interpretable. However, although ECG strain is an extremely specific marker for myocardial fibrosis, it is less sensitive. Indeed, in our mechanism cohort, >30% of patients with replacement myocardial fibrosis did not have strain on the ECG. Importantly, these patients had 40% less replacement fibrosis compared with those with strain, suggesting that strain is a relatively late manifestation and that CMR offers even more sensitive detection of myocardial fibrosis and left ventricular decompensation.
Our data suggest that patients with ECG strain who are asymptomatic would derive long-term benefit from early AVR as a result of the prevention of progressive myocardial fibrosis and injury that would otherwise develop while the patient waited for the onset of symptoms. The stage is now set for randomized, controlled studies to investigate this strategy, examining the clinical utility of the ECG strain pattern in guiding early AVR alongside other novel and more sensitive markers of left ventricular decompensation, including high-sensitivity cTnI concentrations20 and midwall LGE.13
Limitations
In this study, separate cohorts were used to investigate the mechanism and prognosis of patients with ECG strain because CMR was not available in the original SALTIRE study. We therefore cannot directly confirm that ECG strain was similarly related to myocardial fibrosis in the outcome study. However, ECG strain in this population demonstrated the same associations with increased LVMi, aortic stenosis severity, and plasma cTnI concentrations, as observed in the mechanism cohort. Moreover, in our validation cohort, the same clear association between ECG strain and midwall LGE was also observed. We are therefore confident that ECG strain acts as a specific marker of midwall myocardial fibrosis and left ventricular decompensation in the predominantly white patients investigated in this study, although further studies are required for confirmation in different ethnic groups.
Conclusions
In patients with aortic stenosis, ECG strain is a specific marker of midwall myocardial fibrosis and an independent predictor of cardiovascular death or AVR. Future research should now examine whether the ECG strain should be used as a marker of left ventricular decompensation to guide early AVR in asymptomatic patients.
Acknowledgments
The Wellcome Trust Clinical Research Facility and the Clinical Research Imaging Centre, Edinburgh assisted with the conduct of the study. We thank Mary Stoddard and Edwin Carter for their assistance with this analysis and Abbott Laboratories for providing us with assay reagents.
Sources of Funding
Drs Shah, Dweck, Newby, and Mills are supported by the
Disclosures
Drs Mills and Shah received speaker fees from Abbott Laboratories, and Dr Mills has acted as a consultant for Beckman-Coulter. The other authors report no conflicts.
Footnotes
References
- 1.
Dweck MR, Boon NA, Newby DE . Calcific aortic stenosis: a disease of the valve and the myocardium.J Am Coll Cardiol. 2012; 60:1854–1863.CrossrefMedlineGoogle Scholar - 2.
Bartunek J, Weinberg EO, Tajima M, Rohrbach S, Katz SE, Douglas PS, Lorell BH . Chronic N(G)-nitro-L-arginine methyl ester-induced hypertension: novel molecular adaptation to systolic load in absence of hypertrophy.Circulation. 2000; 101:423–429.CrossrefMedlineGoogle Scholar - 3.
Esposito G, Rapacciuolo A, Naga Prasad SV, Takaoka H, Thomas SA, Koch WJ, Rockman HA . Genetic alterations that inhibit in vivo pressure-overload hypertrophy prevent cardiac dysfunction despite increased wall stress.Circulation. 2002; 105:85–92.CrossrefMedlineGoogle Scholar - 4.
Hein S, Arnon E, Kostin S, Schönburg M, Elsässer A, Polyakova V, Bauer EP, Klövekorn WP, Schaper J . Progression from compensated hypertrophy to failure in the pressure-overloaded human heart: structural deterioration and compensatory mechanisms.Circulation. 2003; 107:984–991.LinkGoogle Scholar - 5.
Chin CW, Vassiliou V, Jenkins WS, Prasad SK, Newby DE, Dweck MR . Markers of left ventricular decompensation in aortic stenosis.Expert Rev Cardiovasc Ther. 2014; 28:1–12.Google Scholar - 6.
Greve AM, Boman K, Gohlke-Baerwolf C, Kesäniemi YA, Nienaber C, Ray S, Egstrup K, Rossebø AB, Devereux RB, Køber L, Willenheimer R, Wachtell K . Clinical implications of electrocardiographic left ventricular strain and hypertrophy in asymptomatic patients with aortic stenosis: the Simvastatin and Ezetimibe in Aortic Stenosis study.Circulation. 2012; 125:346–353.LinkGoogle Scholar - 7.
Maceira AM, Prasad SK, Khan M, Pennell DJ . Normalized left ventricular systolic and diastolic function by steady state free precession cardiovascular magnetic resonance.J Cardiovasc Magn Reson. 2006; 8:417–426.CrossrefMedlineGoogle Scholar - 8.
Cowell SJ, Newby DE, Prescott RJ, Bloomfield P, Reid J, Northridge DB, Boon NA ; Scottish Aortic Stenosis and Lipid Lowering Trial, Impact on Regression (SALTIRE) Investigators. A randomized trial of intensive lipid-lowering therapy in calcific aortic stenosis.N Engl J Med. 2005; 352:2389–2397.CrossrefMedlineGoogle Scholar - 9.
Romhilt DW, Estes EH A point-score system for the ECG diagnosis of left ventricular hypertrophy.Am Heart J. 1968; 75:752–758.CrossrefMedlineGoogle Scholar - 10.
Hancock EW, Deal BJ, Mirvis DM, Okin P, Kligfield P, Gettes LS . AHA/ACCF/HRS recommendations for the standardization and interpretation of the electrocardiogram.J Am Coll Cardiol. 2009; 53:992–1002.CrossrefMedlineGoogle Scholar - 11.
Baumgartner H, Hung J, Bermejo J, Chambers JB, Evangelista A, Griffin BP, Iung B, Otto CM, Pellikka PA, Quiñones M ; American Society of Echocardiography; European Association of Echocardiography. Echocardiographic assessment of valve stenosis: EAE/ASE recommendations for clinical practice.J Am Soc Echocardiogr. 2009; 22:1–23.CrossrefMedlineGoogle Scholar - 12.
Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, Picard MH, Roman MJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MS, Stewart WJ ; Chamber Quantification Writing Group; American Society of Echocardiography’s Guidelines and Standards Committee; European Association of Echocardiography. Recommendations for chamber quantification: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology.J Am Soc Echocardiogr. 2005; 18:1440–1463.CrossrefMedlineGoogle Scholar - 13.
Dweck MR, Joshi S, Murigu T, Alpendurada F, Jabbour A, Melina G, Banya W, Gulati A, Roussin I, Raza S, Prasad NA, Wage R, Quarto C, Angeloni E, Refice S, Sheppard M, Cook SA, Kilner PJ, Pennell DJ, Newby DE, Mohiaddin RH, Pepper J, Prasad SK . Midwall fibrosis is an independent predictor of mortality in patients with aortic stenosis.J Am Coll Cardiol. 2011; 58:1271–1279.CrossrefMedlineGoogle Scholar - 14.
Flett AS, Hayward MP, Ashworth MT, Hansen MS, Taylor AM, Elliott PM, McGregor C, Moon JC . Equilibrium contrast cardiovascular magnetic resonance for the measurement of diffuse myocardial fibrosis: preliminary validation in humans.Circulation. 2010; 122:138–144.LinkGoogle Scholar - 15.
Fontana M, White SK, Banypersad SM, Sado DM, Maestrini V, Flett AS, Piechnik SK, Neubauer S, Roberts N, Moon JC . Comparison of T1 mapping techniques for ECV quantification: histological validation and reproducibility of ShMOLLI versus multibreath-hold T1 quantification equilibrium contrast CMR.J Cardiovasc Magn Reson. 2012; 14:88.CrossrefMedlineGoogle Scholar - 16.
Sado DM, Flett AS, Banypersad SM, White SK, Maestrini V, Quarta G, Lachmann RH, Murphy E, Mehta A, Hughes DA, McKenna WJ, Taylor AM, Hausenloy DJ, Hawkins PN, Elliott PM, Moon JC . Cardiovascular magnetic resonance measurement of myocardial extracellular volume in health and disease.Heart. 2012; 98:1436–1441.CrossrefMedlineGoogle Scholar - 17.
Chin CW, Semple S, Malley T, White AC, Mirsadraee S, Weale PJ, Prasad S, Newby DE, Dweck MR . Optimization and comparison of myocardial T1 techniques at 3T in patients with aortic stenosis.Eur Heart J Cardiovasc Imaging. 2014; 15:556–565.CrossrefMedlineGoogle Scholar - 18.
Messroghli DR, Greiser A, Fröhlich M, Dietz R, Schulz-Menger J . Optimization and validation of a fully-integrated pulse sequence for modified look-locker inversion-recovery (MOLLI) T1 mapping of the heart.J Magn Reson Imaging. 2007; 26:1081–1086.CrossrefMedlineGoogle Scholar - 19.
Røsjø H, Andreassen J, Edvardsen T, Omland T . Prognostic usefulness of circulating high-sensitivity troponin T in aortic stenosis and relation to echocardiographic indexes of cardiac function and anatomy.Am J Cardiol. 2011; 108:88–91.CrossrefMedlineGoogle Scholar - 20.
Chin CWL, Shah AS, McAllistair D, Cowell SJ, Alam S, Langrish JP, Strachan F, Hunter A, Choy AM, Lang CC, Walker S, Boon NA, Newby DE, Dweck MR . High sensitivity troponin I concentrations are a marker of an advanced hypertrophic response and adverse outcomes in patients with aortic stenosis.Eur Heart J. 2014; 35:2312–2321.CrossrefMedlineGoogle Scholar - 21. ARCHITECT STAT high sensitivity troponin I [package insert]. Abbott Park, IL: Abbott Laboratories; 2012.Google Scholar
- 22.
Vahanian A, Baumgartner H, Bax J, Butchart E, Dion R, Filippatos G, Flachskampf F, Hall R, Iung B, Kasprzak J, Nataf P, Tornos P, Torracca L, Wenink A ; Task Force on the Management of Valvular Hearth Disease of the European Society of Cardiology; ESC Committee for Practice Guidelines. Guidelines on the management of valvular heart disease: the Task Force on the Management of Valvular Heart Disease of the European Society of Cardiology.Eur Heart J. 2007; 28:230–268.CrossrefMedlineGoogle Scholar - 23.
Vahanian A, Alfieri O, Andreotti F, Antunes MJ, Baron-Esquivias G, Baumgartner H, Borger MA, Carrel TP, De Bonis M, Evangelista A, Falk V, Iung B, Lancellotti P, Pierard L, Price S, Schafers HJ, Schuler G, Stepinska J, Swedberg K, Takkenberg J, Von Oppell UO, Windecker S, Zamorano JL, Zambala M . Guidelines on the management of valvular heart disease (version 2012): the Joint Task Force on the Management of Valvular Heart Disease of the European Society of Cardiology (ESC) and the European Association for Cardio-Thoracic Surgery (EACTS).Eur Heart J. 2012; 33:2451–2496.CrossrefMedlineGoogle Scholar - 24.
Greve AM, Gerdts E, Boman K, Gohlke-Baerwolf C, Rossebø AB, Hammer-Hansen S, Køber L, Willenheimer R, Wachtell K . Differences in cardiovascular risk profile between electrocardiographic hypertrophy versus strain in asymptomatic patients with aortic stenosis (from SEAS data).Am J Cardiol. 2011; 108:541–547.CrossrefMedlineGoogle Scholar - 25.
Pringle SD, Macfarlane PW, McKillop JH, Lorimer AR, Dunn FG . Pathophysiologic assessment of left ventricular hypertrophy and strain in asymptomatic patients with essential hypertension.J Am Coll Cardiol. 1989; 13:1377–1381.CrossrefMedlineGoogle Scholar - 26.
Anderson HV, Stokes MJ, Leon M, Abu-Halawa SA, Stuart Y, Kirkeeide RL . Coronary artery flow velocity is related to lumen area and regional left ventricular mass.Circulation. 2000; 102:48–54.LinkGoogle Scholar - 27.
Villari B, Hess OM, Moccetti D, Vassalli G, Krayenbuehl HP . Effect of progression of left ventricular hypertrophy on coronary artery dimensions in aortic valve disease.J Am Coll Cardiol. 1992; 20:1073–1079.CrossrefMedlineGoogle Scholar - 28.
Okin PM, Devereux RB, Fabsitz RR, Lee ET, Galloway JM, Howard BV ; Strong Heart Study. Quantitative assessment of electrocardiographic strain predicts increased left ventricular mass: the Strong Heart Study.J Am Coll Cardiol. 2002; 40:1395–1400.CrossrefMedlineGoogle Scholar - 29.
Hering D, Piper C, Horstkotte D . Influence of atypical symptoms and electrocardiographic signs of left ventricular hypertrophy or ST-segment/T-wave abnormalities on the natural history of otherwise asymptomatic adults with moderate to severe aortic stenosis: preliminary communication.J Heart Valve Dis. 2004; 13:182–187.MedlineGoogle Scholar - 30.
Lund O, Nielsen TT, Emmertsen K, Flø C, Rasmussen B, Jensen FT, Pilegaard HK, Kristensen LH, Hansen OK . Mortality and worsening of prognostic profile during waiting time for valve replacement in aortic stenosis.Thorac Cardiovasc Surg. 1996; 44:289–295.CrossrefMedlineGoogle Scholar
CLINICAL PERSPECTIVE
Aortic stenosis is a condition that affects not only the valve but also the myocardium. The transition from compensatory left ventricular hypertrophy to heart failure appears to be the key factor in determining the development of symptoms and adverse events. Assessment of this transition is currently limited, and interest surrounds the development of novel biomarkers of left ventricular decompensation. In a cohort of 102 patients with aortic stenosis, we have demonstrated that ECG left ventricular hypertrophy with strain is a highly specific marker of left ventricular decompensation and, in particular, of replacement myocardial fibrosis as assessed with cardiovascular magnetic resonance, a finding validated in an independent external cohort of 64 patients. Moreover, we have confirmed ECG strain as an independent predictor of cardiovascular mortality or aortic valve replacement in 140 asymptomatic patients recruited from the Scottish Aortic Stenosis and Lipid Lowering Trial Impact on Regression (SALTIRE) study. Our data indicate that a 12-lead ECG, which is readily available, cheap, and rapidly interpretable, can identify high-risk patients with aortic stenosis who potentially might benefit from early valve replacement.
