Patients undergoing staged repair for single right ventricular (RV) anomalies are at high risk for adverse outcomes, with reported mortality rates of 17% to 34% during the first year of life. The Pediatric Heart Network Single Ventricle Reconstruction (SVR) trial compared outcomes in 549 infants undergoing Norwood procedures randomized to either a modified Blalock-Taussig shunts (MBTS) or right ventricle–pulmonary artery shunts (RVPAS) at 15 North American centers. The primary result of the trial found better 1-year transplantation-free survival in subjects who received RVPAS compared with those who had MBTS. This advantage, however, appears to diminish over time. At 3 years of age, transplantation-free survival was similar between the groups, because of worse transplantation-free survival in the RVPAS group between 1 and 3 years, and RV systolic function as assessed by echocardiography was worse for the survivors in the RVPAS group. The long-term effect of the ventriculotomy required for RVPAS placement at the time of Norwood procedure has been suggested as a potential cause of this deterioration in RV function and increase in the hazard risk for late adverse outcome, but a direct correlation has not been demonstrated.

Qualitative assessment of systemic RV function has shown significant intraobserver variability and poor correlation with cardiac magnetic resonance imaging. Because of the complex geometry of the right ventricle, quantitative two-dimensional (2D) and Doppler measures by echocardiography are difficult to perform, correlate modestly at best with cardiac magnetic resonance estimates of ejection fraction, and are not routinely used in clinical practice. Prior analysis of this SVR cohort has shown that standard echocardiographic measures of RV systolic and diastolic function are similar between shunt groups at 14 months.

Emerging technologies have shown more promise in the evaluation of RV function. Speckle-tracking echocardiography has provided reliable assessment of global and regional myocardial deformation without dependence on the angle of interrogation. It has been applied to the systemic right ventricle and has shown good correlation with cardiac magnetic resonance–derived RV ejection fraction. The SVR cohort presents a unique opportunity to evaluate a large group of patients with single right ventricles randomly assigned to the initial surgical shunt type. We sought to compare global and regional RV deformation indices in SVR survivors at 14 months of age, after the stage 2 procedure, to assess for shunt-related differences in RV function.

Methods

Study Population

The SVR trial enrolled patients with single morphologic RV lesions undergoing the Norwood procedure at 15 centers in the United States and Canada between May 2005 and July 2008. Inclusion and exclusion criteria have been previously described. Protocol echocardiographic data were collected at specific intervals during the first year as well as at 14 months of age. All protocol echocardiograms were maintained at the echocardiography core laboratory at the Medical College of Wisconsin, and the 14-month echocardiograms were selected for this analysis after removal of the initial shunt as part of the stage 2 procedure. Those with inadequate imaging for speckle-tracking were excluded.

Echocardiographic Analysis

Digital Imaging and Communications in Medicine image clips from the parasternal short-axis (SAX) view below the level of the atrioventricular valve(s) at the midventricular level and from the apical four-chamber (A4C) view at the cardiac crux were selected for analysis when the endocardium of the right ventricle was displayed throughout the cardiac cycle. These Digital Imaging and Communications in Medicine images were stored at a frame rate of 30 frames/sec. Analysis was initially performed with manual tracing of ventricular subendocardium by a single blinded reviewer using commercially available offline echocardiographic speckle-tracking software (TomTec Imaging Systems GmbH, Unterschleissheim, Germany). The subendocardial border was automatically tracked through the cardiac cycle by the software and the overlaid map of the subendocardial tracking was then reviewed by two senior pediatric cardiologists with experience in speckle-tracking to confirm the accuracy of the tracking and adjust as necessary for a consensus best tracing. Once confirmed, the software automatically calculates global circumferential (from the SAX view) and longitudinal (from the A4C view) strain and strain rate as well as regional strain and strain rate from six predetermined segments of the right ventricle ( Figure 1 ). In addition, the software automatically calculates the fractional area change from the SAX view and the single-plane Simpson’s ejection fraction from the A4C tracing. For the purposes of this study, we elected not to use the apical longitudinal segments, because of poor reliability.

Regional segments from the (A) parasternal short-axis and (B) apical four-chamber imaging planes.

Statistical Analysis

Descriptive data are presented as mean ± SD, median with range, or number with percentage of total. Comparisons between groups were made using Student t tests or Wilcoxon rank sum tests on the basis of distribution using SAS OnDemand 4.3 (SAS Institute, Cary, NC). Because the presence of a ventriculotomy and its effect on RV function was the primary objective, the groups were compared on the basis of shunt in place at the completion of the Norwood procedure (non-intention-to-treat). For all differences, P values < .05 were considered to indicate statistical significance.