The aim of this study was to evaluate whether left ventricular (LV) systolic strain in children and young adults with congenital aortic stenosis (AS) and preserved ejection fraction was different from normal subjects and to determine whether any alterations in strain were related to myocardial fibrosis. In this retrospective study, 29 patients with congenital AS with a median age of 15.3 years (range 1.7 to 23.7), highest lifetime AS peak Doppler gradient of 73 mm Hg (22 to 110), most recent AS peak Doppler gradient of 49 mm Hg (0 to 90), and ejection fraction of 65 (55 to 79) were included. Strain was measured using 2-dimensional speckle-tracking echocardiography. Cardiac magnetic resonance was used to identify focal fibrosis by late gadolinium enhancement (LGE) and diffuse fibrosis by calculating the extracellular volume fraction (ECV) from T1 measurements. Compared to age-matched controls (n = 29), patients with AS had reduced LV longitudinal (17.0 ± 3.0% vs 20.6 ± 2.2%, p <0.001) and radial strain (28.8 ± 8.6% vs 34.9 ± 8.5%, p = 0.01), and similar circumferential strain (26.2 ± 5.9% vs 26.4 ± 3.9%, p = 0.79). Median ECV in patients with AS was 0.27 (0.22 to 0.38) and was not significantly correlated with systolic strain. Patients with LGE (n = 7) had lower longitudinal strain than those without LGE (n = 21; 15.0 ± 2.2 vs 17.7 ± 3.1, p = 0.036). In conclusion, in this cohort of children and young adults with congenital AS and preserved ejection fraction, longitudinal and radial strain were reduced, and decreased longitudinal strain was associated with LGE but not ECV.
Decreased left ventricular (LV) global longitudinal strain by 2-dimensional speckle-tracking echocardiography has been found in children with congenital AS who have a normal ejection fraction, but its cause has not been established. One hypothesis is that pressure overload on the left ventricle may compromise myocardial perfusion and lead to ischemia and fibrosis. The development of cardiac magnetic resonance (CMR) techniques to noninvasively evaluate myocardial fibrosis provides an opportunity to explore this proposition. Using the late gadolinium enhancement (LGE) technique, focal regions of fibrosis can be detected. More recently, a CMR technique based on measurements of T1 relaxation times before and after contrast administration has been developed to determine the myocardial extracellular volume fraction (ECV) to assess diffuse fibrosis. The ECV reflects the fraction of myocardium that is extracellular and is higher when there is more fibrosis. We recently reported that in patients with congenital AS, both LGE prevalence and the ECV were increased compared to normal patients and associated with echocardiographic indexes of diastolic dysfunction. However, the association between these fibrosis measurements and systolic strain has not yet been explored. Therefore, the aim of this study was to evaluate whether LV myocardial strain in children and young adults with congenital AS and preserved LV ejection fraction is different from normal subjects and whether any alterations in strain are related to myocardial fibrosis.
Methods
A database search identified all patients at our institution who fulfilled the following criteria: (1) diagnosis of congenital AS with an aortic valve peak Doppler gradient ≥20 mm Hg at any time in their life, (2) age <30 years, (3) echocardiogram and CMR examinations with LGE and ECV measurement within 1 year of each other, and (4) an LV ejection fraction by CMR >54%. Subjects with a history of surgical or catheter interventions on the aortic valve were included. To establish a reference range for speckle-tracking echocardiography strain measurements, a database search identified age-matched subjects who had normal echocardiograms and no medical conditions that are associated with cardiac abnormalities. The study protocol was approved by the Committee on Clinical Investigation at Boston Children’s Hospital.
Echocardiograms were obtained using a Philips iE33 ultrasound system (Philips Healthcare, Andover, Massachusetts). Myocardial speckle-tracking strain quantification was performed offline on 2-dimensional B -mode (grayscale) images using Cardiac Performance Analysis software (TomTec Imaging Systems, Inc., Unterschleissheim, Germany) as previously described. LV longitudinal strain was measured in the apical 4-chamber view, and circumferential and radial strains were measured in the parasternal short-axis view at the level of the papillary muscles. The endocardial and epicardial contours were manually traced on a frame and the software tracking algorithm then propagated them to the other frames in the cardiac cycle. Contours were only accepted when visual inspection and the software indicated adequate tracking. The software automatically divided the myocardium on each view into 6 segments. Segmental strain was calculated as the average peak strain from 2 consecutive cardiac cycles. Global longitudinal, circumferential, and radial strain were calculated as the average peak strain from the 6 segments. Shortening, such as that normally seen in longitudinal and circumferential dimensions, yields negative strain values; however, for convenience, absolute values are reported. All strain measurements were performed by the same investigator to avoid interobserver variability. Measurements for shortening fraction were made on short-axis view images.
CMR examinations in AS subjects were performed on a 1.5T Philips Achieva scanner (Philips Healthcare, Best, The Netherlands). LV end-diastolic volume, end-systolic volume, ejection fraction, and mass were measured from a stack of cine steady-state free precession short-axis images. Diffuse myocardial fibrosis was assessed by calculating the LV myocardial ECV based on T1 measurements before and after gadolinium-based contrast agent administration. T1 measurements for ECV calculation were obtained using a previously described Look-Locker technique with bolus contrast administration. In brief, an electrocardiographic-gated breath-hold Look-Locker sequence with a segmented gradient echo cine acquisition was performed at a single midventricular short-axis slice, once before contrast administration, and 3 times after gadopentetate dimeglumine contrast administration (Magnevist; Bayer HealthCare Pharmaceuticals, Wayne, New Jersey). The partition coefficient for gadolinium (λ) was determined by fitting signal intensity versus time curves and plotting the resulting myocardial R1 (i.e., reciprocal of T1) against blood pool R1. Myocardium with LGE was excluded from the region of interest. The ECV was calculated by multiplying λ and 1 − hematocrit (expressed as a fraction). LGE imaging was performed in ventricular long- and short-axis planes, 15 minutes after contrast administration using a standard 2-dimensional breath-hold phase-sensitive inversion recovery sequence with the inversion time selected to null the myocardial signal.
Patient characteristics and clinical data were reported as medians and ranges for continuous variables and percentages for categorical variables. The Mann-Whitney test was used to compare strain in AS and normal subjects, regional strain variation between LV segments in AS and normal subjects, and to explore the association of strain with dichotomous parameters including clinical data, ECV, and LGE. The Spearman rank correlation was used to determine the relationship between strain and continuous parameters. All statistical tests were 2-sided. For the primary aims comparisons of strain in AS versus normal subjects, strain in AS subjects with elevated versus normal ECV, and strain in AS subjects with or without LGE, results were considered significant if p <0.05. When exploring the relationship of strain with various clinical parameters, significance was set at p <0.01 to account for multiple comparisons. For assessment of interobserver and intraobserver variability in strain analysis, an intraclass correlation coefficient was calculated. Analyses were performed using SPSS version 19.0.0 (SPSS Inc., Chicago, Illinois).
Results
Twenty-nine AS subjects met inclusion criteria and their age, gender, and clinical history are reported in Table 1 . All 29 subjects were in the study cohort from our previous report ; 6 subjects from our previous report were not included in this study because they failed to meet the inclusion criteria (3 had a CMR ejection fraction <54% and 3 subjects did not have an echo within 1 year from CMR). Twenty-nine age-matched normal subjects (median age 15.7 years [range 2.0 to 24.0], 69% men) with echocardiograms at Boston Children’s Hospital were identified for comparison of strain values.
| Age at echocardiogram (years) | 15.3 (1.7 – 23.7) |
| Males | 25 (86%) |
| Status-post balloon aortic valvuloplasty | 20 (69%) |
| Age at first balloon aortic valvuloplasty (years) | 0.3 (0-16.4) |
| Status-post aortic valve surgery | 12 (41%) |
| Age at first aortic valve surgery (years) | 3.3 (0-16.4) |
| Highest lifetime aortic valve peak Doppler gradient (mm Hg) | 73 (22-110) |
| Most recent aortic valve peak Doppler gradient (mm Hg) | 49 (0-90) |
| Left ventricular shortening fraction by echocardiography (%) | 38 (30-63) |
| Left ventricular ejection fraction by cardiac magnetic resonance (%) | 65 (55-79) |
| Left ventricular end-diastolic volume/body surface area by cardiac magnetic resonance (ml/m 2 ) | 104 (52-201) |
| Left ventricular mass/body surface area by cardiac magnetic resonance (g/m 2 ) | 74 (37-129) |
| Left ventricular mass/volume by cardiac magnetic resonance (g/cc) | 0.69 (0.45-1.0) |
| Aortic regurgitation fraction by cardiac magnetic resonance (%) | 25 (0-55) |
Figure 1 shows the mean and 95% confidence intervals for LV global longitudinal, circumferential, and radial strain for AS and normal subjects. Patients with AS had significantly lower global longitudinal strain (17.0 ± 3.0% vs 20.6 ± 2.2%, p <0.001) and lower global radial strain (28.8 ± 8.6% vs 34.9 ± 8.5%, p = 0.011) but no difference in global circumferential strain (26.2 ± 5.9% vs 26.4 ± 3.9%, p = 0.79). Intraobserver and interobserver agreements for global strain measurements were assessed in 10 randomly selected control subjects and are reported in Table 2 .
| Strain | Parameter | Intraobserver (n=10) | Interobserver (n=10) |
|---|---|---|---|
| Longitudinal | Mean difference | 0.08 ± 0.43 | 0.35 ± 0.65 |
| ICC | 0.78 (95% CI: 0.11, 0.95) | 0.76 (95% CI: 0.04, 0.94) | |
| Circumferential | Mean difference | 0.16 ± 0.80 | 0.20 ± 0.39 |
| ICC | 0.90 (95% CI: 0.58, 0.97) | 0.86 (95% CI: 0.44, 0.98) | |
| Radial | Mean difference | 0.58 ± 2.40 | 3.78 ± 1.82 |
| ICC | 0.70 (95% CI: -0.22, 0.93) | 0.70 (95%CI: -0.20, 0.93) |
Segmental values for LV longitudinal, circumferential, and radial strain in AS and normal subjects are summarized in Table 3 . Compared with normal subjects, patients with AS had significantly lower longitudinal strain in the basal (p <0.001) and mid (p <0.001) segments of the septum and in the basal (p = 0.009) and mid (p = 0.012) segments of the free wall. They had significantly lower radial strain in the midanterior (p = 0.028), midanteroseptal (p = 0.011), and midinferoseptal (p = 0.015) segments. The groups did not differ significantly in circumferential strain between any segments.
| Strain | Segment | AS subjects (n=29) | Normal subjects (n=29) | P |
|---|---|---|---|---|
| Longitudinal strain (%) | Lateral base | 15.9 ± 6.2 | 19.3 ± 3.8 | 0.009 |
| Lateral mid | 16.2 ± 7.4 | 19.6 ± 3.4 | 0.012 | |
| Lateral apex | 23.8 ± 8.9 | 25.1 ± 4.2 | 0.549 | |
| Septal base | 12.1 ± 5.5 | 17.4 ± 3.8 | <0.001 | |
| Septal mid | 11.0 ± 4.9 | 17.2 ± 3.0 | <0.001 | |
| Septal apex | 23.0 ± 8.9 | 25.1 ± 4.2 | 0.266 | |
| Global (average) | 17.0 ± 3.0 | 20.9 ± 2.4 | <0.001 | |
| Circumferential strain (%) | Mid-anterior | 26.5 ± 8.3 | 28.2 ± 5.9 | 0.396 |
| Mid-anteroseptal | 25.0 ± 8.7 | 24.5 ± 6.0 | 0.846 | |
| Mid-inferoseptal | 27.7 ± 8.6 | 25.7 ± 5.3 | 0.194 | |
| Mid-inferior | 23.5 ± 11.7 | 25.7 ± 5.4 | 0.560 | |
| Mid-inferolateral | 26.4 ± 8.7 | 23.9 ± 5.8 | 0.228 | |
| Mid-anterolateral | 28.3 ± 13.9 | 28.6 ± 6.3 | 0.608 | |
| Global (average) | 26.2 ± 5.9 | 26.4 ± 3.9 | 0.785 | |
| Radial strain (%) | Mid-anterior | 28.4 ± 14.1 | 36.4 ± 12.2 | 0.028 |
| Mid-anteroseptal | 29.9 ± 15.3 | 40.5 ± 15.5 | 0.011 | |
| Mid-inferoseptal | 27.2 ± 12.5 | 34.6 ± 10.4 | 0.015 | |
| Mid-inferior | 29.5 ± 13.2 | 34.1 ± 10.4 | 0.094 | |
| Mid-inferolateral | 29.7 ± 12.4 | 33.2 ± 14.9 | 0.351 | |
| Mid-anterolateral | 28.2 ± 14.3 | 30.9 ± 12.9 | 0.269 | |
| Global (average) | 28.8 ± 8.6 | 34.9 ± 8.5 | 0.011 |
CMR studies were performed at a median of 43 days (range 0 to 371) from the echocardiograms. The median myocardial ECV in AS subjects was 0.27 (range 0.22 to 0.38). Based on a prior study in our CMR unit, the upper limit of normal (2 standard deviation above the mean) for ECV in children and young adults is 0.28. Using this cut-off value, ECV was normal in 19 subjects and elevated in 10. There was no significant difference between longitudinal, circumferential, and radial strain in AS subjects with a normal versus elevated ECV ( Figure 2 ). Treating ECV as a continuous variable, there was no correlation between ECV and global longitudinal, circumferential, or radial strain.
LGE imaging was performed in 28 AS subjects and regions of LGE were found in 7 subjects (24%). A confluent layer of subendocardial LGE involving ≥12 segments was seen in 3 subjects. The following LGE distributions were found in the other 4 subjects: (1) posteromedial papillary muscle and an apical muscle band, (2) transmural in 4 small areas, (3) transmural in the anterior basal septum and endocardial surface of noncompacted myocardium, and (4) midwall in the basal–inferior segment. Patients with AS with LGE had significantly lower global longitudinal strain than those without LGE (15.0 ± 2.2 vs 17.7 ± 3.1, p = 0.036) ( Figures 3 and 4 ). There was a strong trend toward lower global longitudinal strain in patients with AS with confluent subendocardial LGE compared to those with LGE in other distributions (13.7 ± 1.9 vs 17.4 ± 3.0, p = 0.051). Global circumferential and radial strain were not different in those with versus those without LGE or in those with confluent subendocardial LGE versus those with LGE in other distributions.
