Background
In studies of adult patients, increased QRS duration and mechanical dyssynchrony have been associated with decreased ventricular function. The aim of this study was to test the hypothesis that similar findings would be present in a population of patients with hypoplastic left heart syndrome (HLHS) after the Fontan procedure.
Methods
A retrospective cross-sectional study was conducted. All patients with HLHS after the Fontan procedure were eligible. QRS duration was measured using 12-lead electrocardiography. Echocardiographic measurements of mechanical dyssynchrony included Doppler tissue imaging (DTI) QRS to onset of s′ wave difference between the left ventricle and the right ventricle, time to peak strain, time to peak systolic strain rate (SRs), the standard deviation of time to peak strain rate (modified Yu strain), and the standard deviation of time to peak SRs (modified Yu SRs). Right ventricular (RV) functional measurements included DTI s′ wave, DTI RV myocardial performance index, global strain, global SRs, and RV fractional area change. Pearson’s correlations were performed between the variables.
Results
Thirty-one echocardiographic studies were performed on 26 patients. The median age was 5.3 years (range, 2.5–15.4 years). QRS duration was correlated significantly with global SRs ( r = 0.42). Time to peak SRs was correlated significantly with DTI s′ wave ( r = −0.48) and global SRs ( r = 0.37). Modified Yu SRs was correlated significantly with global strain ( r = 0.35) and RV fractional area change ( r = −0.35).
Conclusions
Both QRS duration and mechanical dyssynchrony were correlated with RV function, albeit weakly. The clinical significance of these findings is intriguing, but only larger studies will determine if these measurements are reliable in guiding treatment options for this complex patient population.
Because of surgical and medical advances, life expectancy is increasing for patients with hypoplastic left heart syndrome (HLHS). A major issue these patients face as they age is progressive heart failure, with a reported incidence as high as 40%. One treatment that has been established to improve morbidity and mortality in the adult heart failure population is cardiac resynchronization therapy (CRT). One criterion used to determine if CRT is an appropriate treatment option is QRS duration. Various mechanical dyssynchrony measurements, measured via echocardiographic techniques, have also been used to determine if patients would respond to CRT, but results have been mixed.
CRT has been used in the congenital heart population for heart failure as well, but indications are usually determined on a case-by-case basis. Increased QRS duration in patients with HLHS compared with historical controls with normal hearts has been previously noted. In addition, increased mechanical dyssynchrony via echocardiographic evaluation has been documented in patients with single right ventricles. The relationship of QRS duration and mechanical dyssynchrony to right ventricular (RV) function, however, has not been clearly defined in patients with single RV physiology.
We hypothesized that QRS duration and mechanical dyssynchrony measurements in patients with HLHS after the Fontan procedure would correlate with measurements of ventricular function, similar to findings in adult studies.
Methods
The institutional review board approved this cross-sectional, retrospective study. All patients with HLHS who had undergone the Fontan procedure were included. Patients also needed to have adequate echocardiographic images with frame rates of 60 to 110 frames/sec performed obtained using a Vivid I or Vivid 7 machine (GE Healthcare, Wauwatosa, WI). An electrocardiogram obtained within 3 months of the echocardiogram was also required. The offline echocardiographic database was searched for these specific patients. Patients needed to be in stable sinus rhythm or in an atrial-derived rhythm. Patients whose intrinsic rhythm lacked atrioventricular synchrony and patients who had ventricular paced rhythms were excluded. Baseline demographics included age, weight, cardiac diagnosis, previous surgeries, and oxygen saturation on room air.
Electrocardiography
Rhythm and QRS duration were determined on 12-lead electrocardiography. QRS duration was measured in lead V 1 for consistency by a single observer (C.L.C.) blinded to clinical and echocardiographic data. Measurements were made offline using the Muse system (GE Healthcare).
Echocardiography
All echocardiographic studies were performed on a Vivid I or Vivid 7 machine. All measurements were made in triplicate by a single observer (J.G.). Although the observer who measured the echocardiographic variables could in fact see the electrocardiographic tracings while obtaining the echocardiographic measurements, the observer did not measure or have knowledge of QRS width. Postprocessing of all images was completed offline using EchoPAC version 7 (GE Healthcare).
Standard Echocardiography
Views equivalent to an apical four-chamber view were obtained ( Figure 1 ). Images were optimized for the visualization of the epicardial and endocardial borders of the single right ventricle. RV fractional area change ([RV diastolic area − RV systolic area]/RV diastolic area) was obtained per published guidelines.
Doppler Tissue Imaging (DTI) Measurements
Spectral DTI measurements of the right free wall and left free wall at the level of the atrioventricular valve annulus were obtained in the apical four-chamber view. DTI peak systolic annular velocity (s′) of the RV free wall was measured as one marker of RV function. The myocardial performance index (MPI) was calculated ([isovolumetric contraction time + isovolumetric relaxation time]/ejection time) for the right ventricle from the DTI values as another marker of RV function ( Figure 2 ).
QRS onset to the onset of the s′ wave was measured from DTI images of the free walls using a simultaneous electrocardiogram. The absolute difference between the RV and left ventricular free wall intervals was determined and used as a marker of mechanical dyssynchrony.
Strain Measurements
The endocardial border of the single right ventricle in an apical four-chamber view was traced from the septal-atrioventricular annular hinge point to the apical septum and then to the RV lateral wall at the lateral-atrioventricular annular hinge point. The automated epicardial-to-endocardial computer-generated border, or region of interest, was adjusted to include the epicardium. The borders were accepted only if visual inspection showed adequate tracking and if the software indicated adequate tracking for all segments (green). Patients whose segments did not track well because of artifacts or inadequate visualization of the lateral borders of the right ventricle were excluded. Measurements of global RV strain and global RV systolic strain rate (SRs) were obtained as markers for RV function ( Figure 3 ).
QRS onset to time to peak strain and SRs was measured for all six segments. The standard deviations of the time to peak strain and the time to peak SRs were then calculated to give modified Yu indices for strain and SRs, respectively. QRS onset to time to peak global strain and global SRs was also obtained as other values for mechanical dyssynchrony.
Correlations
RV fractional area change, RV free wall DTI s′ wave, RV MPI, RV global strain, and RV global SRs were used as measurements of RV function. QRS duration was used as a marker of electrical dyssynchrony. QRS onset to s′ wave difference between the RV and left ventricular free walls, QRS onset to time to peak global strain and SRs, modified Yu strain, and modified Yu strain rate were used as values for mechanical dyssynchrony.
Pearson’s correlations were performed between RV function measurements and QRS duration as well as RV function measurements and RV mechanical dyssynchrony values. P values ≤.05 were considered significant. To examine interobserver and intraobserver variability, three measurements were taken for each parameter on a random sample of 10 subjects and completed by one observer. An additional observer completed a single set of measurements on the same 10 subjects on all parameters. Intraclass correlation coefficients (ICCs) were used to measure the levels of intraobserver and interobserver variability. Agreement was considered excellent (>0.75), good (0.60–0.74), or poor (<0.40) as defined by the ICC. All statistical analyses were performed using SAS version 9.2 (SAS Institute, Inc., Cary, NC).
Results
A total of 30 patients with HLHS were identified, with 35 studies performed. Four studies were excluded, one because of an atrial arrhythmia, one because of the lack of an electrocardiographic tracing at the time of imaging, and two because of inadequate imaging and/or speckle-tracking analysis. A total of 31 studies from 26 patients were thus available for analysis.
The median age of the patients at the time of echocardiography was 5.3 years (range, 2.5–15.4 years), the median time from Fontan procedure was 38.7 months (range, 2.8–134.3 months), the median weight was 10.7 kg (range, 11.2–60.8 kg), and median oxygen saturation on room air was 95% (range, 80%–99%). Twenty patients underwent hybrid procedures and six patients Norwood procedures with modified Blalock-Taussig shunts as their initial surgical procedures. Fifteen patients underwent extracardiac nonfenestrated Fontan procedures, four underwent extracardiac fenestrated Fontan procedures, and seven patients underwent lateral tunnel fenestrated Fontan procedures as their last surgical cardiac procedures.
The mean QRS duration was 96 ± 18 msec. Twenty-two of the 31 electrocardiographic studies were performed on the same day as the echocardiographic studies, with the rest performed within 3 months of echocardiography.
RV mechanical dyssynchrony measurements are presented in Table 1 . RV functional measurements are presented in Table 2 . Normative data, previously published in other studies, were available for a few variables and are shown in the far right column for comparative purposes in both tables. The normative data are from pediatric patient populations with structurally normal hearts.
| Variable | Fontan study group | Normative data |
|---|---|---|
| DTI QRS to onset of s′ wave difference (msec) | 28.5 ± 48.8 | 17.7 ± 11.8 |
| Time to global peak strain (msec) | 355 ± 48 | 320.1 ± 28.6 |
| Time to global peak SRs (msec) | 195 ± 39.4 | |
| Modified Yu strain (msec) | 52.6 ± 22.8 | 21.7 ± 7.3 |
| Modified Yu SR (msec) | 120 ± 53.1 |
| Variable | Fontan study group | Normative data |
|---|---|---|
| DTI s′ wave for RV free wall (cm/sec) | 5.5 ± 1.3 | 13.2 ± 0.9 |
| DTI RV free wall MPI | 0.4 ± 0.1 | 0.3 ± 0.1 |
| Global strain (%) | −17.6 ± 2.9 | −31.3 ± 3.1 |
| Global SRs (s −1 ) | −1.08 ± 0.2 | −2.01 ± 0.3 |
| RV fractional area change (%) | 30.6 ± 4.9 | 54.6 ± 10.5 |
Correlations between mechanical dyssynchrony variables and QRS duration are shown in Table 3 . Two of the five mechanical dyssynchrony measurements had correlations with QRS duration.
| Correlation coefficients with QRS duration | r | P |
|---|---|---|
| Time to peak strain | 0.42 | .02 |
| Time to peak SRs | 0.42 | .02 |
| Modified Yu strain | −0.09 | .62 |
| Modified Yu SRs | 0.07 | .70 |
| DTI QRS to onset of s′ wave difference | 0.07 | .70 |
Correlations between QRS duration and mechanical dyssynchrony values with RV measures of function are highlighted in Table 4 . Numerical values in the table represent the significant r values. QRS duration was positively correlated with global SRs. Qualitatively, this means that a longer QRS duration corresponds to less negative strain rate values and, theoretically, decreased RV function. Time to peak global SRs was inversely correlated with DTI s′ wave and positively correlated with global SRs. This suggests that increasing mechanical dyssynchrony portends worse RV function. Modified Yu strain rate was positively correlated with strain and negatively correlated with fractional area change. There was a weak association between the modified Yu strain rate and global SRs that did not reach significance ( r = 0.30, P = .09). Again, this implies that increasing mechanical dyssynchrony is associated with worsening RV functional measurements. A subset analysis was performed on the 22 patients who underwent electrocardiography and echocardiography on the same day. The correlations for this subset of patients were similar to those found in the original group.
