Background
Previous studies have suggested the presence of dyssynchrony in the functionally single ventricle. The aim of this study was to investigate the presence of classic-pattern dyssynchrony (CPD), characterized by typical early and late deformation of opposite walls, and its relation to QRS duration and myocardial function in patients with single-ventricle physiology after Fontan palliation.
Methods
In a retrospective cross-sectional study, 101 adolescent and adult patients with single-ventricle physiology after the Fontan procedure were investigated. Strain curves were visually assessed for the presence of CPD. Systolic and diastolic function were assessed using echocardiography.
Results
One hundred one patients were included, with varying anatomic morphology: two sizable ventricular components ( n = 21), right dominant ( n = 21), left dominant ( n = 49), and undefined anatomy ( n = 10). Fifteen of 101 Fontan patients had CPD. Forty-three percent of patients with two sizable ventricular masses displayed CPD, mostly with prolonged QRS, while the number of patients with CPD with right-dominant (9%) and left-dominant (6%) morphology was significantly lower ( P = .016). Those with CPD displayed significantly ( P < .05) larger QRS widths (142 ± 22 vs 112 ± 24 msec), lower ejection fractions (31 ± 14% vs 45 ± 14%), lower global early diastolic strain rates (0.7 ± 0.5 vs 1.2 ± 0.8 sec −1 ), and global systolic circumferential (−10 ± 5% vs −16 ± 7%) and longitudinal (−9 ± 5% vs −14 ± 5%) strain, respectively.
Conclusions
CPD is present in a proportion of adolescent and adult patients after Fontan palliation. The presence of CPD is associated with reduced systolic and diastolic function compared with Fontan patients without CPD. Because the presence of CPD appears to be a promising predictor for response to cardiac resynchronization therapy in patients with biventricular circulation, these findings may have important potential for prospective evaluation of cardiac resynchronization therapy in patients with univentricular circulation.
Highlights
- •
CPD can be found in a substantial number of adolescent and adult patients with Fontan circulation.
- •
In particular, patients with two sizable ventricular components display broad QRS width and CPD.
- •
Cardiac function is reduced in Fontan patients with CPD compared to those without.
- •
These findings might have important implications for prospective evaluation of CRT in patients with Fontan circulation.
Surgical and medical advances have improved life expectancy for patients with functionally single ventricles and Fontan palliation. However, patients with functionally single ventricles often experience progressive heart failure and arrhythmias, leading to transplantation or death.
Previous studies in children with functionally single ventricles have suggested a high frequency of ventricular dyssynchrony, especially in hypoplastic left heart syndrome (HLHS). Cardiac resynchronization therapy (CRT) has been shown to benefit adults with left ventricular dysfunction due to intraventricular conduction delay. Three retrospective reviews of pediatric CRT subjects with primarily congenital cardiac disease and a case report of a patient with HLHS with dyssynchrony showed promise for improvement of function in the majority of patients (75%–87%) undergoing CRT. Prior echocardiographic studies in patients with functionally single ventricles defined dyssynchrony as increased time to peak dispersion of segmental velocity or deformation. However, in the Predictors of Response to Cardiac Resynchronization Therapy study in adult patients with heart failure, these parameters correlated poorly with CRT response.
In contrast, studies in adults with heart failure have reported reproducible deformation patterns of mechanical dyssynchrony resulting from electrical conduction delays in the presence of left bundle branch block (LBBB). Classic-pattern dyssynchrony (CPD) can be seen by parasternal M-mode or strain or strain rate (SR) echocardiography and typically shows early short-lived contraction of the septum (“flash”) opposed by early stretch and late contraction of the delayed lateral wall. CPD was highly associated with the presence of LBBB by electrocardiography in adults and was highly predictive of response to CRT.
In this study we aimed to investigate the presence of CPD-like contraction patterns in the functionally single ventricle using segmental strain analysis. We hypothesized that CPD, as a specific marker for electromechanical dyssynchrony, is associated with worse myocardial function in the functionally single ventricle. By investigating the presence of CPD, we aimed to specifically identify possible CRT responders among adult patients with functionally single ventricles.
Methods
Study Subjects
In a retrospective cohort study, we identified patients after Fontan palliation from the Adult Congenital Heart Program at Stanford University. From the hospital’s registry, patients were identified by the diagnostic code for functionally single ventricle status post Fontan palliation. All patients who underwent Fontan surgery with echocardiograms available for review were included. Type of anatomy and surgery were retrieved from medical records. We used both the echocardiographic images and medical records to define the morphology of the functionally single ventricle as having two sizable ventricular components (usually with an unrestrictive interventricular communication), dominant right ventricle, dominant left ventricle, or other. In cases of undefined morphology or when two sizable ventricular components were present, the echocardiogram was reassessed. Hearts not assigned to either of the three groups were labeled “undefined” anatomy. The most recent echocardiograms were chosen for reanalysis. Clinical data from the medical records, including electrocardiograms, blood pressure, height, weight, and New York Heart Association class were obtained from the same or the closest available to the date of echocardiography.
Echocardiograms and clinical data from 99 healthy control subjects matched by age and sex were retrieved from a population-based study (HUNT) conducted in Trondheim, Norway. Strain and conventional echocardiographic measurements were performed in the same way and by the same observer as in Fontan patients.
The study was conducted in accordance with institutional human subjects committee guidelines and was approved by the institutional review board at Stanford University.
Echocardiography
Images were acquired between January 2001 and May 2015 using a Philips iE33 ultrasound scanner (Philips Medical Systems, Andover, MA) at Stanford Hospital or a Siemens Acuson 512 or SC 2000 ultrasound scanner (Siemens Healthcare, Erlangen, Germany) at Lucile Packard Children’s Hospital. Images in the HUNT study were acquired between January 2007 and June 2008 with a Vivid 7 ultrasound scanner (GE Healthcare, Little Chalfont, United Kingdom). At least one apical and one midventricular short-axis (SAX) view were available in the majority of studies. In these two projections, sagittal, transverse, and longitudinal ventricular end-systolic and end-diastolic diameters were measured using grayscale recordings. Wall thickness was measured from SAX images.
Ejection fraction and ventricular volumes were estimated using the monoplane Simpson method in the apical projection.
Ventricular diastolic function was assessed from atrioventricular valve (AVV) pulsed-wave Doppler recordings measuring the maximal velocity of early filling (E) and atrial contraction (A), AVV inflow deceleration time, and E/A ratio. We further calculated a parameter similar to E/e′ by using e′ (tissue velocity E) from displacement-derived tissue velocities from the basal segments in apical views. The degree of valvular regurgitation was based on assessment of the color Doppler images. When regurgitation was greater than mild to moderate, the severity of the regurgitation was based on further echocardiographic criteria according to the guidelines.
Two-Dimensional Speckle-Tracking Analysis for Strain and SR
Strain and SR were analyzed offline in an apical projection equivalent to a four-chamber view and a SAX midventricular view using 2D Cardiac Performance Analysis version 1.1 (TomTec Imaging Systems, Unterschleissheim, Germany) and Syngo VVI software (Siemens Medical Solutions USA, Mountain View, CA). In these releases, Syngo and 2D Cardiac Performance Analysis use identical speckle-tracking algorithms and thus produce equivalent values. The images used were acquired at a frame rate of 40 ± 12 frames/sec (range, 30–93 frames/sec). For strain and SR analysis, endocardial longitudinal and circumferential curves were used. Ventricular systolic and diastolic volumes and ejection fraction were calculated using the Simpson method, derived from the endocardial trace of one long-axis two-dimensional loop. Global or segmental strain curves from missing endocardial segments or visually incorrect tracking due to subjectively low-quality images were discarded. The onset of the QRS complex was used as the onset of the cardiac cycle and thus the 0% time point of strain. As shown in Figure 1 , tissue velocities were derived from displacement registrations of the basal segments of the image equivalent to the apical four-chamber view and expressed as the peak velocities from the medial and lateral wall as well as the mean of both values. The definition of peak strain and timing of those in systole and over the entire cardiac cycle are demonstrated in Figure 2 . By extracting original segmented strain and SR curves, measurements on curves from both software vendors were performed in the same way. Total strain was defined as the difference between the first positive peak and the following negative peak strain during one cardiac cycle. Average longitudinal and circumferential strain and SR were calculated from the deformation of an unsegmented endocardial region of interest. Postsystolic shortening (PSS) expresses the difference between diastolic strain and peak systolic strain. PSS is present when diastolic strain is higher than peak systolic strain. Each projection of the ventricle was divided into six segments. Strain and SR were measured in the pressure-generating “outer” walls of the functionally single ventricle. When an unrestrictive interventricular communication was present, the free lateral walls of both ventricular masses were analyzed. When the septum was intact or a restrictive interventricular communication was present, septal strain was included. In the absence of a septum or when the rudimentary (nondominant) ventricle was out of plane, we analyzed the free lateral walls of the dominant ventricle.
The timing of aortic valve opening, and aortic valve closure (AVC), AVV opening, E-wave termination, and AVV closure was measured from Doppler flow sampled in the ventricular outflow or inflow tract. On the basis of Doppler-derived time measurements, using an excel macro, segmental values for peak positive and peak negative strain were extracted during systole (QRS onset to AVC) or from the entire the cardiac cycle, while diastolic segmental SR and basal velocities were defined as peak values during E and A filling.
Definition of CPD
Strain curves from the long-axis and SAX projections were visually assessed. Figure 3 displays different patterns of segmental contraction. Figure 3 B shows a variant that can typically be seen when regionally reduced strain is present. Here, some segments display delayed contraction in the absence of conduction delay. Time-to-peak measurements can vary solely because of regionally reduced segmental function, where peak systolic strain in the late contracting segment is often lower compared with the earlier contracting segments. CPD ( Figure 3 C) is the typical dyssynchronous pattern described in patients with LBBB. CPD segments with early electrical activation show early contraction during the isovolumic contraction period, followed by segmental stretch during early systole (also termed “flash”), resulting in low end-systolic strain values of the early-activated segment. Segments with conduction delay display early stretch during isovolumic contraction, followed by delayed contraction. In CPD, segments with delayed activation typically reach higher end-systolic strain values compared with early-activated segments. Figure 3 D shows a deformation pattern when conduction delay is present, but systolic strain peaks are simultaneous, not dyssynchronous. This pattern is often seen in patients with pacemakers or LBBB at the presence of normal ventricular function. In contrast to CPD, early-activated segments contract during the entire duration of systole, up to but not past AVC. Patients were grouped into CPD and non-CPD groups. The non-CPD group included normal function ( Figure 3 A), regionally reduced function ( Figure 3 B), and conduction delay with simultaneous peaks ( Figure 3 D).
Statistical Analyses
Continuous variables are reported as mean ± SD or as median and range. Frequencies and presence of dyssynchrony patterns in right-dominant, left-dominant, anatomy with two sizable ventricular masses were compared using χ 2 tests. Differences between patient groups and between patients and healthy control subjects were tested using univariate analysis of variance. Contrast analysis was used to test differences between groups. In all tests, two-way P values < .05 were considered to indicate statistical significance. Intraobserver reproducibility for global and segmental two-dimensional strain using the same cardiac cycle were performed on 15 randomly chosen subjects for SAX and long-axis images, with repeated analysis ≥2 weeks apart, blinded to the previous interpretation. Reproducibility was expressed with intraclass correlation coefficients. Statistical analyses were performed using SPSS version 23.0 (IBM, Armonk, NY).
Results
Patient Characteristics
Of 107 Fontan patients, a total of 101 adolescent and adult patients after Fontan palliation and 99 age-matched healthy control subjects were identified. Six patients were excluded from the study because of missing echocardiograms. Image quality was not sufficient for strain analysis in one patient, and strain analysis could not be performed because of technical problems in two patients. Eighty-five of 1,212 segments (7.5%) were excluded from analysis in the patient cohort.
Table 1 displays intraobserver variabilities. Table 2 shows the baseline characteristics of the study population. The average age was 21 ± 10 years, and 60% were male. Eighty percent were in New York Heart Association class I or II, and all patients were in sinus rhythm at the time of echocardiography. There were significant differences in height, weight, and systolic blood pressure between the Fontan and control groups. One patient had severe ventricular outflow tract stenosis, and three patients had moderate to severe AVV regurgitation.
| Variable | Intraclass correlation coefficient | |
|---|---|---|
| Global | Segmental | |
| Circumferential strain | 93.5 | 76.9 |
| Longitudinal strain | 92.1 | 73.4 |
| Circumferential SR S | 88.0 | 82.5 |
| Longitudinal SR S | 93.0 | 70.1 |
| Circumferential SR E | 89.3 | 83.2 |
| Longitudinal SR E | 89.0 | 71.7 |
| Longitudinal velocity E | 84.3 | |
| Diagnosis | Unbalanced AVSD | HLHS | PA IVS | TA | DILV | Others ‡ | All Fontan patients | Healthy control subjects |
|---|---|---|---|---|---|---|---|---|
| ( n = 19) | ( n = 15) | ( n = 4) | ( n = 23) | ( n = 21) | ( n = 19) | ( N = 101) | ( n = 99) | |
| Male | 16 (84) ∗ , † | 7 (46) | 2 (50) | 12 (52) | 17 (81) ∗ , † | 6 (32) | 60 (60) | 53 (54) |
| Age (y) | 20 ± 9 ∗ | 21 ± 7 | 25 ± 5 | 28 ± 9 | 25 ± 7 | 26 ± 10 | 21 ± 10 | 25 ± 8 |
| Age at Fontan (y) | 5.4 ± 4 | 6.0 ± 5 | 13 ± 12 | 9.1 ± 8 | 6.6 ± 5 | 10 ± 9 | 6 ± 6 | |
| Height (cm) | 159 ± 21 ∗ | 166 ± 17 ∗ | 173 ± 14 | 164 ± 22 ∗ | 163 ± 15 ∗ | 162 ± 13 | 152 ± 25 ∗ | 176 ± 8 |
| Weight (kg) | 57 ± 20 ∗ | 62 ± 16 ∗ | 65 ± 13 | 63 ± 16 ∗ | 62 ± 18 ∗ | 62 ± 14 | 52 ± 23 ∗ | 79 ± 17 |
| Systolic BP (mm Hg) | 113 ± 12 ∗ | 106 ± 10 ∗ , † | 111 ± 18 | 118 ± 19 ∗ | 117 ± 13 ∗ | 108 ± 12 ∗ | 113 ± 14 ∗ | 125 ± 14 |
| Diastolic BP (mm Hg) | 69 ± 10 | 64 ± 9 | 66 ± 4 | 70 ± 12 | 69 ± 11 | 67 ± 12 | 68 ± 11 | 69 ± 10 |
| HR (beats/min) | 84 ± 15 ∗ , † | 72 ± 12 | 78 ± 14 ∗ | 70 ± 13 † | 79 ± 18 ∗ | 76 ± 11 ∗ | 76 ± 15 ∗ | 64 ± 13 |
| Type of Fontan | ||||||||
| Classic | 3 (16) | 1 (7) | 0 (0) | 17 (74) | 7 (33) | 5 (26) | 36 (36) | |
| Lateral tunnel | 3 (16) | 7 (47) | 2 (50) | 5 (22) | 7 (33) | 10 (53) | 33 (33) | |
| Extracardiac tunnel | 13 (68) | 7 (47) | 2 (50) | 1 (4) | 7 (33) | 4 (21) | 32 (32) | |
| Moderate to severe AVV regurgitation | 1 (5) | 1 (7) | 1 (5) | 1 (10) | 4 (4) | |||
| NYHA class | ||||||||
| I | 3 (17) ∗ | 1 (8) ∗ | 1 (33) | 4 (18) ∗ | 1 (6) ∗ | 5 (31) ∗ | 15 (16) ∗ | 99 (100) |
| II | 8 (44) | 10 (77) | 1 (33) | 17 (77) | 14 (33) | 9 (56) | 59 (64) | |
| III | 5 (28) | 2 (15) | 1 (33) | 1 (5) | 3 (17) | 2 (13) | 16 (17) | |
| IV | 2 (11) | 0 (0) | 0 (0) | 0 (0) | 0 (0) | 0 (0) | 2 (2) |
∗ P < .05, comparison between each group and healthy control subjects.
† P < .05, comparison between each group and all other Fontan patients.
‡ Ten patients with single ventricles, six with double-outlet right ventricles, two patients with transposition of the great arteries and AVC, and one patient with transposition of the great arteries and hypoplastic right ventricle.
Ventricular Function in Patients with and without CPD
Only one of the patients with tricuspid atresia (TA) in our cohort displayed a nonrestrictive interventricular communication, where strain was measured in the right ventricular free wall, which consisted of one basal segment. Among patients with atrioventricular septal defect (AVSDs), only one patient displayed a small interventricular communication and a rudimentary right ventricle, where strain was measured in the left ventricular free wall and septum. All other patients with AVSDs displayed a rudimentary septum, an unrestrictive interventricular communication, and two sizable ventricular masses. In this instance, strain was measured in the free walls of the left- and right-sided ventricles. Three patients with double-outlet right ventricles had an unrestrictive interventricular communication, where we measured strain in the lateral wall of the left-sided ventricle instead of the septum. All other patients had a restrictive interventricular communication. Table 3 displays the presence of CPD in the different anatomic variants. Overall, CPD was present in 15% of the Fontan patients. The highest prevalence was found in patients with two sizable ventricular components (nine of 21 [43%]), and this was significantly higher than patients with dominant left or right ventricles. Most patients with dominant left ventricular morphology did not have CPD. Thus, none of the patients with TA or pulmonary atresia with intact ventricular septum had CPD. Furthermore, 18 of the 21 patients with double-inlet left ventricles and 18 of 19 patients with HLHS did not have CPD. Three of 1 ventricular paced hearts displayed CPD, two with two sizable ventricles. Among all patients with CPD without ventricular pacing, seven patients with AVSDs and one patient with HLHS displayed right bundle branch block. LBBB was seen in one patient with TA. Two patients, one with double-outlet right ventricle and one with single ventricle, could not be classified as LBBB or right bundle branch block. Of the four patients with CPD under ventricular pacing, one with an AVSD, one with undefined anatomy, and one with double-inlet left ventricle displayed LBBB, while one patient with TA displayed right bundle branch block. Three patients in the CPD group with ventricular pacing were atrially sensed, and one was atrially paced. All 14 patients with pacemakers had atrial and ventricular leads implanted, but one patient had an atrial lead only. In the non-CPD group, two of the nine patients with ventricular pacing were atrially paced, and seven were atrially sensed. Three patients were paced in multiple sites of the ventricle, one of whom displayed CPD.
| Anatomy | Total | CPD | No CPD | Ventricular paced with CPD | Ventricular paced without CPD |
|---|---|---|---|---|---|
| AV canal (two sizable ventricles) | 19 | 8 (42) | 11 (58) | 1 (5) | 2 (11) |
| HLHS (RV dominant) | 15 | 1 (7) | 14 (93) | 0 | 0 |
| PA IVS (LV dominant) | 4 | 0 | 4 | 0 | 1 |
| TA (LV dominant) | 23 | 0 | 23 (100) | 0 | 0 |
| DILV (LV dominant) | 21 | 3 (14) | 18 (86) | 1 (5) | 4 (19) |
| Others | 19 | 3 (16) | 16 (84) | 1 (5) | 1 (5) |
| Two sizable ventricles | 21 | 9 (43) ∗ | 12 (57) | 2 (10) | 1 (5) |
| RV dominant | 21 | 2 (9) | 19 (91) | 0 | 1 (5) |
| LV dominant | 49 | 3 (6) | 46 (94) | 1 (2) | 5 (10) |
| SV | 10 | 1 (10) | 9 (90) | 0 | 1 (10) |
| All patients | 101 | 15 (15) | 86 (85) | 3 (3) | 8 (8) |
∗ P = .016 for χ 2 test between patients with two sizable ventricles and right or left ventricular dominance for presence of true dyssynchrony.
Table 4 displays systolic and diastolic function in patients with CPD compared with those without CPD. Values from healthy control subjects illustrate normal values in this age group. In the presence of CPD, QRS width was increased compared with absence of CPD, with a trend toward delayed aortic valve opening and shortened ejection time. With CPD, all functional parameters were decreased, indicating lower systolic function, as reflected by higher end-systolic volumes, lower ejection fraction, average longitudinal and circumferential strain, and impaired relaxation as expressed by reduced SR E. Figure 4 shows a typical example of a patient with two sizable ventricles with prolonged QRS complex, typical posterior displacement of the atrioventricular node, and left anterior bundle branch block on electrocardiography, CPD with early contraction of the inferior wall, and delayed contraction of the anterior wall. QRS duration correlated significantly with both longitudinal ( R = 0.57) and circumferential ( R = 0.47) strain, longitudinal ( R = 0.48) and circumferential ( R = 0.49) SR, and ventricular volumes ( R = 0.52). All patients with CPD had prolonged QRS durations (mean, 142 msec; range, 104–195 msec), whereas not all patients with prolonged QRS durations had CPD.
