Patients with systemic right ventricles frequently experience progressive heart failure and conduction abnormalities leading to abnormal ventricular activation. Activation delay–induced mechanical dyssynchrony can contribute to ventricular failure and is identified by a classic strain pattern of paradoxical opposing wall motion that is an excellent predictor of response to cardiac resynchronization therapy in adults with left bundle branch block. The specific aims of this study were to compare right ventricular (RV) mechanics in an adult systemic right ventricle population versus control subjects, evaluate the feasibility of this RV strain pattern analysis, and determine the frequency of the classic pattern.
Young adults ( n = 25) with d-transposition of the great arteries, status post Mustard or Senning palliation (TGA-MS), were ambispectively enrolled and compared with healthy young adults ( n = 30) who were prospectively enrolled. All subjects were imaged using novel three–apical view (18-segment) RV longitudinal speckle-tracking strain analysis (EchoPAC) and electrocardiographic data.
Patients with TGA-MS had diminished RV global peak systolic strain compared with control subjects (−12.0 ± 4.0% vs −23.3 ± 2.3%, P < .001). Most patients with TGA-MS had intrinsic or left ventricular paced right bundle branch block. A classic pattern was present in 11 of 25 subjects (44%), but this pattern would have been missed in four of 11 based only on the RV four-chamber (six-segment) model. Only three subjects underwent cardiac resynchronization therapy. Both subjects who had the classic pattern responded to cardiac resynchronization therapy, whereas the one nonresponder did not have the classic pattern.
Systemic right ventricles demonstrated decreased function and increased mechanical dyssynchrony. The classic pattern of activation delay–induced mechanical dyssynchrony was frequently seen in this TGA-MS population and associated with activation delays. This comprehensive RV approach demonstrated incremental value.
Progressive right ventricular (RV) failure is a nearly universal problem for adults with systemic right ventricles. Understanding how ventricular activation abnormalities contribute to RV dysfunction and the potential for cardiac resynchronization therapy (CRT) may avoid transplantation or prevent death in properly selected patients. Before the arterial switch procedure became the gold standard repair for patients born with d-transposition of the great arteries, the Mustard and Senning palliations were the surgical procedures of choice. They both effectively baffle the systemic venous return to the left heart and the pulmonary venous return to the right heart, which solved the problem of the parallel left and right heart circulations but maintained the morphologic right ventricle as the systemic pumping ventricle. Most institutions transitioned to the arterial switch procedure >20 years ago, so patients with d-transposition of the great arteries, status post Mustard or Senning palliation (TGA-MS), now constitute an adult congenital population in their 20s to 40s. In this aging population, there is almost universal progressive RV failure in adults that is due partly to a lifetime systemic pressure load. However, abnormal ventricular activation causing mechanical dyssynchrony may also contribute to this progression, which has been demonstrated in the left ventricles of patients with normal anatomy with left bundle branch block (LBBB). Patients with TGA-MS frequently have activation abnormalities of their systemic right ventricles, such as intrinsic right bundle branch block (RBBB) or left ventricular (LV) pacing–induced RBBB. RBBB often leads to progressively mechanical RV dyssynchrony and dysfunction, and studies have shown that a subgroup of patients with TGA-MS benefit from CRT. However, the high CRT nonresponse rate is a major clinical problem and highlights the need for a specific predictive marker for CRT response in the systemic right ventricle.
In adults with LV cardiomyopathy, predicting CRT response by traditional time to peak (TTP) echocardiographic strain indices or prolonged QRS duration on electrocardiography leads to an unacceptably high nonresponse rate. Recently, a classic pattern of dyssynchrony has been identified using regional strain pattern analysis that more strongly predicts both short- and long-term CRT response and is associated with LBBB criteria on electrocardiography. This pattern analysis identifies the physiology of opposing wall motion caused by a significant activation delay, referred to as activation delay–induced mechanical dyssynchrony (ADI-MD). The paradoxical wall motion of ADI-MD consists of early septal contraction opposed by early stretch in the activation-delayed RV free wall, followed by late free wall contraction causing early termination of septal contraction. Independent groups have also reported on a similar strain pattern analysis with strong CRT response predictive characteristics. This pattern has also been identified in both the left and right ventricles of various patient populations, including case reports in the systemic right ventricle.
In the past, RV longitudinal strain analysis has used only a right ventricle–centered four-chamber apical (RV4) view despite potential heterogeneities in the systemic right ventricle. Recently, a comprehensive 18-segment (18S) strain analysis model, using three apical RV images, has been introduced in a normal adult population to more comprehensively evaluate complex regional RV mechanics. The aims of this study were to (1) compare RV mechanics in a young adult systemic right ventricle population versus control subjects and (2) evaluate for the presence of the classic pattern of ADI-MD in the right ventricle in this systemic right ventricle population. We hypothesized that the classic pattern would be present in a subgroup of patients with systemic right ventricles and would be associated with electrical activation abnormalities.
The TGA-MS population consisted of all the adult subjects who were identified for strain-protocol echocardiography with the three apical RV views during routine adult congenital heart clinic visits from June 2010 to September 2012 at Duke University Medical Center. The patients with TGA-MS for this study were ambispectively identified, with subjects imaged before January 2012 retrospectively identified and those after that date prospectively enrolled. However, all subjects with TGA-MS in this study were imaged with the same echocardiographic protocol, including comprehensive RV apical acquisitions on the GE Vivid E9 (GE Vingmed Ultrasound AS, Horten, Norway) optimized for strain analysis, because the protocol was implemented in early 2010 for all patients with systemic right ventricles. This is believed to represent a consecutive capture of all subjects with TGA-MS imaged during this period. No subjects were excluded because of inadequate image quality. Health information and electrocardiograms were obtained from the medical record. A small subset of this population was previously described as pilot data.
The control population was prospectively enrolled from a population of healthy young adults recruited for a study to assess echocardiographic predictors of pulmonary edema when diving underwater. In the normal population, feasibility of the 18S RV strain analysis, normal ranges, and reproducibility were previously presented. A normal cohort ( n = 30) was age- and sex-matched to the TGA-MS population. All subjects underwent comprehensive screening echocardiography with the comprehensive apical RV views with strain analysis. Inclusion criteria for control subjects were age ≤18 years, no history of cardiac abnormalities, and normal echocardiographic findings, including anatomy, LV ejection fraction, fractional area change (FAC), and tricuspid annular plane systolic excursion (TAPSE). Exclusion criteria for control subjects were any abnormal echocardiographic findings, prolonged QRS duration, and serious systemic disease. This study was approved by the institutional review board at Duke, and all prospectively enrolled subjects provided informed consent.
Echocardiography with Comprehensive Apical RV Views
Echocardiography was performed with grayscale images optimized for longitudinal speckle-tracking strain analysis (50–90 frames/sec). Three apical RV views were obtained with the subject in the standard left lateral recumbent position. The three apical RV views have equivalent imaging planes to the four-, two-, and three-chamber LV apical views with the transducer angled rightward ( Figure 1 ), as previously published. To name the apical RV views, the view with the outflow tract (transducer notch at 1 o’clock) was named the RV outflow view. The view isolating the RV inflow and apical regions (transducer notch at 11 o’clock) was named the RV inflow view. The RV four-chamber view name was maintained (transducer notch at 3 o’clock). Optimizing these views often required repositioning of the transducer a few centimeters toward the axillary line from the typical LV-focused apical view and angling the imaging plane anteriorly and rightward. All echocardiographic studies were acquired with a Vivid E9 using a 3.5-MHz ultrasound probe. RV systolic function was also assessed by FAC and TAPSE.
In the planning stage of this study, efforts were made to obtain high-quality short-axis views of the right ventricle for circumferential strain assessment. However, imaging of the anterior right ventricle was rarely adequate, and the decision was made to focus only on longitudinal strain in this study. Furthermore, the criteria for the classic pattern of ADI-MD were created using longitudinal strain and have never been studied with circumferential or short-axis strain analysis, even in the left ventricle.
Two-Dimensional Speckle-Tracking Strain Analysis
Offline strain analysis was performed using EchoPAC PC version BT11 (GE Vingmed Ultrasound AS). The reference point was placed at QRS onset to ensure that early systolic events were not missed. The systemic semilunar valve closure timing relative to the QRS complex was defined on spectral Doppler from the parasternal short-axis or apical view. The endocardial border was manually traced in end-systole for speckle-tracking. The region of interest was adjusted to exclude the pericardium. The integrity of speckle-tracking was automatically detected and visually confirmed. In case of poor tracking, tracing was readjusted or the shape of the region of interest changed. Segments with persistent inadequate tracking were excluded from analysis. Feasibility was determined as the percentage of all segments with adequate tracking. Segmental peak longitudinal systolic strain (PLS) was measured in six segments in each of the three RV views as peak strain between QRS onset and semilunar valve closure. These six segments included the basal, mid, and apical segments on the septal and RV walls, with a total of 18 segments from the three apical views.
The software calculated a segmental PLS score for each segment, and these were averaged from all 18 segments to produce global PLS. Segmental PLS was also averaged by walls (apical, septal, and free wall) to produce regional PLS. This software has been validated for the determination of global PLS in LV strain from the four-, two-, and three-chamber apical LV views, and those imaging places were maintained with our comprehensive RV views.
Because longitudinal peak strain is by convention a negative value (shortening in systole), the nomenclature of this report will refer to greater strain as a larger negative number (e.g., peak strain of −15% is greater than −5%, or peak strain −15% > −5%).
TTP Dyssynchrony Analysis
TTP intervals were measured as the duration from the onset of the QRS complex to the first distinct peak in a strain curve. One TTP measure of RV mechanical synchrony was the SD of this duration in all 18 segments (TTP SD ). Another TTP measure was the maximum opposing wall interval (TTP op ), measured as the maximal interval between the opposing walls in the mid or basal segments. TTP SD ≥ 60 msec and TTP op ≥ 130 msec were used as traditional predictors of CRT response in adults with LBBB cardiomyopathy.
Regional Strain Pattern Analysis
The methodology for LV regional strain pattern analysis has been published in a population with LBBB heart failure to identify a classic dyssynchronous contraction pattern reflective of a significant ventricular activation delay. The three criteria to define the classic pattern are (1) early peak contraction with early termination of shortening in at least one septal or apical segment and early stretching in at least one opposing wall segment, (2) early peak contraction occurring in the first 70% of the systolic ejection phase, and (3) late contraction after semilunar valve closure in the segment with early stretch. If these criteria are met in a single view, the classic pattern is diagnosed. Examples of this RV classic pattern in patients with TGA-MS are provided in Figure 2 .
CRT was performed in three subjects with TGA-MS with FAC ≤ 25% who were dyssynchronous by TTP and had prolonged QRS durations on electrocardiography. Echocardiograms including 18S RV regional strain analysis were obtained before CRT and after CRT pacer optimization. Patients 1 and 2 had transvenous, dual-chamber pacemakers with LV septal leads before the addition of epicardial, unipolar leads placed in the basal region of the lateral RV free wall. Patient 3 had the same pre-CRT pacing system and was upgraded with the same epicardial RV lead, but it was placed in the basal region of the anterolateral free wall.
Continuous variables are reported as mean ± SD and were compared using paired or unpaired t tests after visualizing the data to confirm normality. Categorical variables (reported as percentages) were tested for differences using χ 2 tests. Univariate logistic regression was used to model predictors of the classic pattern. The r 2 statistic was calculated to determine correlations between two linear variables using the Spearman rank correlation test. P values <.05 were considered to indicate statistical significance.
Interobserver variability of regional pattern analysis using the same cardiac cycle was performed on all 30 control subjects and the first 19 patients with TGA-MS by a second experienced reader blinded to the primary analysis. The strain curves were redrawn by the second reader de novo. In case of disagreements, subgroup analysis was based on the primary read. Statistical analyses were performed using Stata version 12 (StataCorp LP, College Station, Texas).
Subject characteristics for 30 control subjects and 25 patients with TGA-MS are reported in Table 1 . Control subjects were in the normal range for all electrocardiographic and echocardiographic measures. Patients with TGA-MS had at least moderate dilation or hypertrophy of the systemic right ventricle and prolonged QRS duration due to paced RBBB, intrinsic RBBB, or incomplete RBBB (RBBB morphology with QRS duration of 100–120 msec). Dysrhythmias were also frequently seen in the TGA-MS population, including ventricular tachycardia (16%), atrial fibrillation or flutter (64%), and other supraventricular tachycardias (24%). These dysrhythmias led to ablation procedures in 12 subjects (48%). Most subjects with TGA-MS were in New York Heart Association (NYHA) class II or III.
|Control subjects ( n = 30)||Patients with TGA-MS ( n = 25)|
|Age (y)||31 ± 9 (20–52)||34 ± 7 (22–49)|
|Women||16 (53%)||17 (68%)|
|QRS duration (msec)||90 ± 11 (73–111)||133 ± 34 (86–192)|
|RBBB: left ventricular pacing||0||13 (52%)|
|RBBB: no pacing||0||3 (12%)|
|Incomplete RBBB||0||7 (28%)|
|NYHA class I||0||4 (16%)|
|NYHA class II||0||13 (52%)|
|NYHA class III||0||8 (32%)|
|>2 CHF medications||0||11 (44%)|
|Moderate or greater TR||0||4 (16%)|
|TV replacement||0||2 (8%)|
|LV-PA shunt||0||2 (8%)|
Functional Strain Analysis
Strain measurements are reported in Table 2 (global and regional) and Figure 3 (segmental). RV global PLS was significantly lower in the TGA-MS population, with 22 of 25 subjects (88%) demonstrating abnormal global PLS (<−17%). The regional differences in PLS between the free wall, apex, and septum seen in the control subjects ( P < .001) were not present in the TGA-MS population ( P > .50). Analyzing all 55 subjects, global PLS correlated well with TAPSE ( r 2 = −0.80, P < .001) and FAC ( r 2 = −0.75, P < .001) across the entire range of systolic RV function. Evaluating only the TGA-MS population ( n = 25), the correlation between global PLS and TAPSE ( r 2 = −0.68, P < .001) or FAC ( r 2 = −0.69, P < .001) is slightly lower but remains good.
|Control subjects ( n = 30)||Patients with TGA-MS ( n = 25)||P|
|RV FAC (%)||42.2 ± 5.1 (35 to 57)||27.4 ± 7.4 (10 to 45)||<.001|
|RV TAPSE (mm)||21.7 ± 3.5 (16 to 30)||10.2 ± 3.1 (3 to 16)||<.001|
|RV global PLS||−23.3 ± 2.3% (−18 to −27)||−12.0 ± 4.0% (−3 to −20)||<.001|
|Septal PLS||−19.5 ± 2.0% (−15 to −24) ∗||−12 ± 4.1% (−1 to −24) †||<.001|
|Apical PLS||−21.9 ± 3.6% (−15 to −30) ∗||−11.6 ± 4.7% (−3 to −23) †||<.001|
|Free wall PLS||−27.7 ± 3.8% (−20 to −36) ∗||−12.0 ± 4.7% (0 to −22) †||<.001|
|TTP SD (msec)||25.2 ± 7.9 (10 to 42)||80.5 ± 34.1 (28 to 140)||<.001|
|TTP SD > 60 msec||0||15 (60%)||<.001|
|TTP op ≥ 130 msec||0||19 (76%)||<.001|
|Classic pattern||0||11 (44%)||<.001|