Cardiovascular magnetic resonance evidence of myocardial fibrosis and its clinical significance in adolescent and adult patients with Ebstein’s anomaly

Consecutive patients with unrepaired Ebstein’s anomaly referred to our institution for clinical evaluation gave their consent and were enrolled into a registry since June 2013. All patients received a comprehensive evaluation including a detailed medical history, physical examination, oxygen saturation test, routine blood examination (including serum creatine), electrocardiography, echocardiography, and CMR imaging. Patients in our registry database from June 2013 to Oct 2016 were reviewed and patients with complete data sets and those older than ten years of age were included in the study. Exclusion criteria were as follows: 1) Contraindications to contrast-enhanced CMR; 2) other coexisting cardiomyopathy, such as hypertrophic, restrictive, or infiltrative cardiomyopathy; 3) a history of known coronary heart disease or myocarditis; 4) other complex congenital cardiac anomalies (e.g., congenitally corrected transposition with Ebstein’s anomaly); 5) CMR image quality was too poor to analyze due to significant arrhythmia or difficulty with breath-holding. Patients were grouped into Carpentier’s classification of types A to D, based on the mobility of the tricuspid leaflet and the RV shape and contractility on echocardiography [22].

An equal number of age- and gender-matched subjects were included as controls. Healthy adults with no documented cardiovascular diseases or other major illnesses, and with normal echocardiograms, were recruited as volunteers. Subjects younger than 18 years old were selected for CMR examination to screen for myocarditis. Subjects were included as controls if all clinical data, routine blood work, electrocardiography, echocardiography, and CMR examination excluded cardiovascular disease. All patients and controls (or their parents if a subject was younger than 18 years old) gave written informed consent for their data to be included in this study.

CMR was performed on a 3 T scanner (Magnetom Tim Trio, Siemens Healthineers, Erlangen, Germany) with using a 32-channel phased-array cardiac coil. All images were acquired with electrocardiogram (ECG)-gated breath-hold technique. All patients were scanned according to the standard imaging protocol for Ebstein’s anomaly [23]. The protocol consisted of stacks of balanced steady-state free-precession (bSSFP) cine images acquired in the three long-axis (two-, three-, and four-chamber) views, consecutive short-axis views covering the LV from base to apex (TR/TE 3.4 ms /1.3 ms, flip angle 50°, FOV 300 × 340 mm, matrix size 256 × 144, slice thickness 8 mm), and consecutive axial cine bSSFP images covering the whole heart extending from the pulmonary bifurcation to just below the diaphragm. LGE images were acquired using a T1-weighted inversion recovery turboflash sequence 10 to 20 min after bolus contrast injection (Gadopentetate dimeglumine, 0.15 mmol/kg, Bayer HealthCare Pharmaceuticals, Wayne, New Jersey, USA) in the same planes. T1 measurements were performed by using a Modified Look-Locker Inversion Recovery sequence (MOLLI) at the LV basal-, mid- and apical- levels, with the following parameters: non-selective inversion pulse, bSSFP single shot readout with 35° flip angle, minimum inversion time 110 ms, inversion time increment 80 ms, TR/TE of 2.9/1.12 ms, FOV of 360 × 272 mm, matrix size of 256 × 144, Voxel size of 2.1 × 1.4 × 8 mm³ and slice thickness of 8 mm. The inversion time for each patient was chosen at the time of scanning from the TI scout images, with three heartbeats for recovery between each experiment. Native T1 measurements were acquired before injection of gadolinium, with an imaging scheme of 5(3)3), and post contrast T1 measurements were repeated in the same short-axis slices 10–15 min after the administration of gadolinium with a scheme of 4(1)3(1)2. The hematocrit was tested within 24 h of CMR scanning for ECV calculation.

CMR images were analyzed using a commercially available software package (QMass® 7.6, Medis, Leiden, The Netherlands). Biventricular volume and function were obtained by tracing epicardial and endocardial contours manually on both end-diastolic and end-systolic phases. The LV volume, function, and mass were measured based on the consecutive short-axis slices. RV volume and function were derived from the axial views, as described in previous studies [1, 24]. All volumetric parameters were indexed to body surface area (BSA). The tracing method for the chambers was similar to the previous description by Fratz et al. [2, 25]. A line along the tricuspid leaflets was used to segment the aRV and the functional right ventricle (fRV). The fRV is defined as the part of the RV distal to the tricuspid leaflets. The border demarcating the aRV and morphological right atrium (RA) is defined as the curved line along the presumed tricuspid valve annulus, from the free wall attachment of the anterior tricuspid valve leaflet to the point of presumed septal leaflet attachment [25]. The trabeculae and papillary muscles were included in the ventricular blood pool.

LGE images were visually assessed on phase sensitive inversion recovery (PSIR) images at the short- and long-axis views by two experienced CMR imagers (YCC, and JYS, both with > 300 cases/year experience in CMR). The myocardial T1 calculation was based on the LV short axis MOLLI images with motion correction. The endocardial and epicardial contours were traced manually on the pre-contrast T1-mapping images and post-contrast T1-mapping images. The T1 curve was fitted automatically using the software (QMass® 7.6, Medis), and the mean myocardial T1 was obtained. At the same time, the blood T1 was obtained by drawing an ROI in the blood pool within the LV cavity in the pre-contrast and post-contrast T1-mapping images, respectively. The ECV was calculated as [26]:

In our study, the severity index (SI) and total right/left-volume index (Total R/L-volume index) were utilized to quantify the severity of EA. SI, described by Fratz et al., was calculated using the area of the RA, aRV, fRV, left atrium (LA), and LV traced in end-diastole on the four-chamber view [1, 2, 25]. The total R/L-volume index was defined by Hösch et al. as the ratio of the total volume of the right heart and left heart, measured in the end-diastole on the axial views [27]. The equations for the two indices are shown as:

Total R/L-Volume Index = volume of \( \left(\frac{RA+ aRV+ fRV}{LA+ LV.}\right) \)

Statistical analysis was performed using SPSS 17.0 (SPSS Inc., International Business Machines, Armonk, New York, USA). Numerical data in normal distribution were expressed as the mean ± standard deviation (SD), or else were expressed as the median with the interquartile range. Continuous data were compared using a paired t-test. Categorical data were expressed as frequencies and compared using a χ² test or the Fisher’s exact test as appropriate. Correlations between variables were analyzed by using Pearson’s or Spearman correlation analyses. Univariable factors related to the ECV at p <  0.1 were included in the multivariate stepwise regression analysis. Statistical significance was defined as P <  0.05.

Ultimately, the Ebstein’s anomaly registry database comprised data from seventy-eight patients. Patients were excluded from the study because of: age < 10 years (n = 12), no contrast was used (n = 2), poor image quality caused by due to persistent atrial fibrillation or unable to breath hold (n = 10), and T1 mapping not performed (n = 10). The remaining 44 patients were included in the study. The mean age of the patients was 34.0 ± 16.2 years (range 10–60) years. Seven patients were between 10 and 18 years. Age, gender and BMI were well balanced between patients and the healthy control groups. Among the 44 patients, 29 (65.9%) had an atrial septal defect or patent foramen ovale. Tricuspid regurgitation was graded as severe in 32 (72.7%) and moderate in 12 (27.3%) patients (Table 1).

LV end diastolic volume (EDV) index was reduced in Ebstein’s anomaly patients (68.3 ± 16.8 vs. 79.0 ± 17.1 ml, P = 0.007) (Table 2). LV ejection fraction (EF) and LV mass index in Ebstein’s anomaly was decreased (52.8 ± 7.6%, vs. 60.6 ± 3.7%, P <  0.001; and 39.2 ± 9.1 g/m2 vs. 45.2 ± 11.1 g/m2, P = 0.004, respectively). LVEF was reduced (< 50%) in 34.1% (n = 15) of the Ebstein’s anomaly patients. The LV mass to volume ratio was not significantly different from the control group. Both total and fRV EDV indices and RV end systolic volume (ESV) index were significantly increased compared with those of the controls. The fRV EF in Ebstein’s anomaly was significantly reduced (P <  0.001).

LGE was seen in 10 out of 44 (22.7%) Ebstein’s anomaly patients. The typical LGE pattern was linear or patchy localized within the RV subendocardium of the basal septum (Fig. 1). Only one Ebstein’s anomaly patient showed focal transmural LGE in the basal septum and posteromedial papillary muscle (shown in Fig. 1). No remarkable LGE was seen in the LV free wall. In Ebstein’s anomaly patients with LGE, the indexed fRV EDV and ESV were significantly larger than those patients without LGE (both P value < 0.01). The fRV EF was lower in patients with LGE than without. The aRV volume was also significantly larger in the LGE-positive group than the LGE-negative group. Ebstein’s anomaly severity evaluated by the total R/L-volume index was more severe in patients with LGE than without LGE, while the SI was not significantly different between these two groups. Patients with LGE also had a higher NYHA functional class and lower oxygen saturation at rest than patients without LGE (P = 0.011 and 0.042, respectively). The LV volume and the LV EF were lower in patients with LGE; however, the difference was not significant. All data are shown in Table 3.

The native T1 was higher but not statistically significant in patients with Ebstein’s anomaly. The average myocardial ECV was significantly increased in patients with Ebstein’s anomaly (P < 0.001). The septal ECV was higher than that of the free ventricular wall in Ebstein’s anomaly patients and free wall ECV was significantly higher than that of the controls (P < 0.001). Twenty (45.5%) patients had an ECV higher than 30%, which is above the upper limits of normal. Except for one a-60-year old, all subjects in the control group had an ECV less than 30%. In patients with a higher ECV, the Ebstein’s anomaly severity index, NYHA class and tricuspid regurgitation were significantly worse than in patients with a normal ECV. The LVEF was much lower in Ebstein’s anomaly patients with a higher ECV than those with a normal ECV (Table 4). The hematocrit was observed to be lower in patients with higher ECV, while no linear correlation was present between the hematocrit and the ECV of the LV myocardium.

To explore the relationship between LGE and the ECV, we compared ECV between Ebstein’s anomaly patients with and without LGE. Figure 2 shows that the global ECV was higher in Ebstein’s anomaly patients with LGE than patients without LGE (P = 0.038). It is important to recognize that the ECV in patients without LGE was still significantly higher than that of the controls. The same trend was found (Fig. 2) when the ECV was analyzed separately in the septum versus the free wall. Individual global ECV values in the EA and control groups are shown in Fig. 3.

Among all variables, the ECV of the LV myocardium was significantly correlated with NYHA class, tricuspid regurgitation severity, the LV ESV index, the LV EF, the fRV EDV index, the fRV ESV index, and the fRV EF (Table 5). Moreover, we found that the ECV of the LV myocardium had a moderate correlation with the Ebstein’s anomaly severity (severity index: r = 0.392, P = 0.009; Total R/L-volume index: r = 0.612, P < 0.001). Multivariate correlation analysis showed that ECV was independently correlated with LVEF.

There are several limitations to our study. First, the present study consisted of both adolescent and adult patients, and does not represent the entire Ebstein’s anomaly spectrum. Second, the ECV of the RV myocardium was not quantified because the RV free wall was too thin in Ebstein’s anomaly for accurate assessment. Third, the cross-sectional study design limited the ability to determine the impact of myocardial fibrosis on the long-term prognosis of Ebstein’s anomaly. Our study indicated that further longitudinal observation is needed to clarify the impact of myocardial fibrosis and remodeling in Ebstein’s anomaly patients.