Background
Novel echocardiographic techniques based on myocardial deformation have not been extensively evaluated to assess right ventricular (RV) and left ventricular (LV) response after pulmonary valve replacement (PVR) in patients with repaired tetralogy of Fallot.
Methods
Between 2003 and 2012, 133 patients undergoing first-time PVR after tetralogy of Fallot repair underwent echocardiographic assessment at Mayo Clinic. The last echocardiogram before PVR and 1 year after surgery were retrospectively analyzed with Velocity Vector Imaging.
Results
Mean age at PVR was 35.5 ± 16.2 years (54% women). Longitudinal peak systolic strain and strain rate before PVR were low: for the left ventricle, −14.8 ± 3.5% and −0.8 ± 0.2 sec −1 , and for the right ventricle, −16.2 ± 4.1% and −0.9 ± 0.3 sec −1 , respectively. There was no significant change in either parameter after surgery. A close correlation between LV and RV deformational parameters was found before PVR and was maintained after surgery. In the multivariate analysis, patients with better LV and RV peak systolic strain preoperatively were found to have better LV and RV peak systolic strain after surgery ( P = .004 and P = .006, respectively). However, patients with the most improvement in deformation were those with worse RV function preoperatively ( P = .002). Mean New York Heart Association class at early follow-up improved from 2.2 ± 0.8 to 1.2 ± 0.6 ( P < .0001); RV peak systolic strain was the only factor associated with symptomatic improvement.
Conclusion
LV and RV systolic and diastolic deformational parameters were decreased in patients with repaired tetralogy of Fallot undergoing PVR, and there was no significant change after surgery. However, preoperative systolic deformational parameters were predictive of postoperative ventricular function and New York Heart Association class after PVR and may be helpful to identify optimal timing for surgical intervention in this cohort.
Tetralogy of Fallot (TOF) is the most common type of cyanotic congenital heart disease. Severe pulmonary regurgitation is common in patients with repaired TOF, often necessitating reconstruction of the right ventricular (RV) outflow tract (RVOT) and pulmonary valve replacement (PVR). However, identifying the best timing for PVR continues to be challenging in clinical practice, because it may affect the recovery of RV function after surgery. The advent and use of new echocardiographic techniques based on myocardial deformation may help refine the decision-making process in these patients. These novel echocardiographic techniques have been used to assess RV and left ventricular (LV) dysfunction and ventricular interaction after TOF repair. If found to provide an accurate measure of ventricular function and functional class improvement after PVR, deformation imaging could be used as an alternative diagnostic tool in patients with repaired TOF, offering quantitative ventricular functional assessment to all patients irrespective of their clinical status or concomitant device status. However, there are few data and no consensus on the use of these new echocardiographic techniques to help determine the timing of operative intervention in patients with TOF. In addition, RV and LV response after PVR has not been extensively evaluated using these new techniques.
The primary aims of this study were (1) to evaluate whether patients with repaired TOF with good ventricular function before PVR had better ventricular function, functional class, and survival at early follow-up than patients with decreased ventricular function measured by novel deformational echocardiographic techniques and (2) to evaluate if the quantitative measures obtained with myocardial deformation are equivalent to ventricular functional assessment with cardiac magnetic resonance imaging (MRI) in this patient cohort.
Methods
Study Population
We performed a retrospective review of all patients with repaired TOF who underwent first-time PVR at our institution between 2003 and 2012 ( n = 146). To be included in this study, patients were required to have undergone at least one preoperative echocardiographic examination at our institution. Patients with associated pulmonary atresia, TOF with absent pulmonary valve, and/or concomitant atrioventricular canal defects were excluded. Patients who refused research authorization were also excluded. Overall, 133 of 146 patients (91%) met the inclusion criteria to be analyzed.
The study was approved by the Mayo Clinic Institutional Review Board. Medical histories, as well as perioperative and follow-up data, were collected using all available records and postoperative surveys that are sent on a routine scheduled basis. In addition, the Social Security Death Index was reviewed.
Echocardiography
We selected the last echocardiogram before surgery and that closest to 1 year after surgery (range, 3.6 months to 2.7 years) for each patient.
Strain Analysis
Digital echocardiographic images were transferred to a dedicated workstation for offline analysis. Images were analyzed with Velocity Vector Imaging software (Siemens Medical Solutions USA, Inc, Mountain View, CA). Images were analyzed only when overall quality was good and enabled visualization of the entire RV free wall. Mean frame rates on all views were between 38 and 41 Hz. Strain analysis was performed by a single observer blinded to the clinical data. The endocardium was manually traced, and the region of interest was manually adjusted to the wall thickness. Adequate tracking was visually assessed, and strain curves were accepted only if tracking appeared appropriate. The apical four-chamber view was used for the LV analysis (septum and free wall), and the RV free wall was used for the RV analysis.
Two-Dimensional (2D) Echocardiography and Doppler Analysis
Categorical variables such as chamber size, ventricular function, and valve regurgitation were codified using a numeric scale for analysis purposes: −1 = small, 0 = normal or none, 1 = borderline or trivial, 2 = mild, 3 = moderate, and 4 = severe.
Intraobserver and Interobserver Reliability Analysis
Intraobserver and interobserver reliability analyses of LV and RV peak systolic longitudinal strain and strain rate (SR) and diastolic SR were performed in 10 random studies. Curves were traced anew, and strain curves generated anew, on the same cardiac cycle by the same observer and independently by a second observer for intraobserver and interobserver reliability, respectively. Absolute difference divided by the mean of the repeated observations and expressed as a percentage was used to measure variability.
Outcomes
Three different outcomes were chosen: (1) LV and RV peak systolic longitudinal strain at follow-up; (2) functional class at follow-up, defined in terms of the New York Heart Association (NYHA) classification; and (3) death of any cause.
MRI
Cardiac MRI before PVR was available for 49 patients. Correlations comparing MRI with strain imaging and to the traditional 2D echocardiography were performed. All MRI studies were performed on a 1.5-T system (Signa; GE Healthcare, Waukesha, WI) using an eight-element phased-array cardiac coil. After obtaining initial scout images, short-axis cine balanced steady-state free precession images were obtained from the atrioventricular ring to the apex, and axial steady-state free precession images were obtained. The RV and LV volumes and ejection fraction were obtained by manual tracing of endocardial borders from the short-axis images at end-diastole and end-systole. Beginning in 2008, RV volumes and ejection fraction from the axial images were also routinely obtained. The following measurements were included: RV and LV end-diastolic volume indices, RV and LV end-systolic volume indices, and RV and LV ejection fractions.
Statistical Analysis
Descriptive statistics are reported as proportions for discrete data and as mean ± SD for continuous data, except for variables that were not normally distributed, in which case medians and interquartile ranges are used, unless specified differently. Student t tests were used to compare continuous variables. Chi-square tests of independence were used to compare categorical variables, except for cells with percentage predicted <5%, for which Fisher exact tests were used. Comparisons of patient and echocardiographic data before and after surgery were performed using t -test pairs or Wilcoxon tests accordingly. For quantifying correlations between two variables, the Spearman correlation test was applied. Stepwise multivariate analysis included only variables that had small numbers of missing data and P values ≤ .20 in the univariate analysis. P values < .05 were considered statistically significant.
Results
Patient Characteristics
A total of 133 patients with TOF underwent PVR, at a mean age of 35.5 ± 16.2 years, a mean of 29.4 ± 12.4 years after the initial repair. Characteristics of the population are summarized in Table 1 . Almost half of the cohort had three or more previous cardiac interventions. These included palliation procedures such as the Brock procedure or systemic–to–pulmonary artery shunts and closure of residual ventricular septal defects. The majority of patients were symptomatic at the time of surgery. On cardiac MRI, their mean indexed RV end-systolic volume was 100 ± 30 mL/m 2 , RV end-diastolic volume 172 ± 44 mL/m 2 , and RV ejection fraction 42 ± 8%. Their mean indexed LV end-systolic volume was 28 ± 12 mL/m 2 , LV end-diastolic volume 69 ± 19 mL/m 2 , and LV ejection fraction 58 ± 9%. A total of 41 patients underwent tricuspid valvuloplasty, and two underwent tricuspid replacement at the time of PVR. Patients did not undergo specific RVOT patch resection with PVR. Some patients had partial RVOT patch resection during valve insertion but not as a method of RVOT enlargement.
| Variable | Value |
|---|---|
| Women | 72 (54.1%) |
| BSA (m 2 ) | 1.8 ± 0.4 |
| BMI (kg/m 2 ) | 25.7 ± 6.4 |
| Age at initial repair (y) | 3.3 (0.2–47.2) |
| Previous palliation procedure | 47 (35.6%) |
| Previous cardiac surgery ≥ 3 | 59 (44.7%) |
| Age at PVR (y) | 35.0 (3.6–64.9) |
| QRS at PVR (msec) | 154 ± 29 |
| QRS after PVR (msec) | 148 ± 30 |
| NYHA class before PVR ≥ II | 103 (77.4%) |
| NYHA class after PVR ≥ II | 12 (12.4%) |
| Follow-up time (y) | 3.0 ± 2.7 |
| Death | 5 (3.9%) |
Echocardiography before and after PVR
The preoperative echocardiographic study was performed 56 ± 87 days before PVR, with a median of 37 days. The postoperative echocardiographic study was performed 390 ± 181 days after PVR, with a median of 1 year.
Strain Imaging
Table 2 shows LV and RV peak systolic and diastolic deformation parameters both before and after PVR. Peak systolic deformation parameters were decreased for both the left and right ventricles preoperatively and did not change significantly after surgery. RV diastolic SR worsened slightly after surgery.
| Variable | Echocardiography before PVR | Echocardiography after PVR | Matched difference | P | |||
|---|---|---|---|---|---|---|---|
| n | Value | n | Value | n | Value | ||
| Left ventricle | |||||||
| Strain (%) | 76 | −14.8 ± 3.5 | 52 | −14.5 ± 3.3 | 35 | 0.6 ± 0.5 | .30 |
| Systolic SR (sec −1 ) | 76 | −0.8 ± 0.2 | 52 | −0.8 ± 0.2 | 35 | 0.02 ± 0.03 | .50 |
| Diastolic SR (sec −1 ) | 76 | 0.9 ± 0.3 | 52 | 0.9 ± 0.3 | 35 | −0.04 ± 0.4 | .40 |
| RV free wall | |||||||
| Strain (%) | 108 | −16.2 ± 4.1 | 55 | −15.8 ± 4.4 | 46 | 0.2 ± 0.7 | .80 |
| Systolic SR (sec −1 ) | 108 | −0.9 ± 0.3 | 55 | −0.9 ± 0.3 | 46 | 0.007 ± 0.05 | .90 |
| Diastolic SR (sec −1 ) | 108 | 1.0 ± 0.3 | 55 | 0.9 ± 0.5 | 46 | −0.2 ± 0.06 | .02 |
| 2D echocardiography and Doppler | |||||||
| LV EF (%) | 131 | 57 ± 8 | 59 | 59 ± 6 | 57 | 2.8 ± 1.0 | .006 |
| Mitral E/A ratio | 114 | 1.8 ± 0.7 | 53 | 1.6 ± 0.6 | 46 | −0.15 ± 0.1 | .20 |
| Lateral mitral E/e′ ratio | 101 | 9.4 ± 4.0 | 47 | 10.0 ± 4.2 | 34 | 1.9 ± 0.9 | .03 |
| RV MPI | 119 | 0.3 ± 0.2 | 15 | 0.3 ± 0.07 | 14 | −0.02 ± 0.05 | .70 |
| RV pressure (mm Hg) | 129 | 47 ± 17 | 56 | 40 ± 13 | 55 | −7.2 ± 2.4 | .004 |
| LA size | 123 | 0 (0 to 2) | 41 | 0 (0 to 2) | 37 | 0 (0 to 0) | .20 |
| LV size | 131 | 0 (0 to 0) | 64 | 0 (0 to 0) | 63 | 0 (0 to 0) | .50 |
| LV systolic function | 129 | 0 (0 to 0) | 60 | 0 (0 to 0) | 58 | 0 (0 to 0) | .048 |
| PR | 131 | 4 (4 to 4) | 63 | 1 (0 to 1) | 63 | −3 (−4 to −3) | <.001 |
| RA size | 124 | 3 (2 to 4) | 38 | 2 (2 to 3) | 36 | 0 (−2 to 0) | .02 |
| RV hypertrophy | 103 | 3 (2 to 3) | 15 | 3 (2 to 3) | 14 | 0 (−1 to 1) | .90 |
| RV size | 132 | 4 (3 to 4) | 65 | 3 (2 to 3) | 65 | −1 (−2 to 0) | <.001 |
| RV systolic function | 131 | 2 (1 to 3) | 63 | 2 (0 to 3) | 62 | 0 (−1 to 0) | <.001 |
| TR | 132 | 2 (2 to 3) | 65 | 2 (1 to 2) | 65 | −1 (−2 to 0) | <.001 |
Intraobserver and interobserver variability for LV and RV peak systolic and diastolic deformation parameters is presented in Table 3 .
| Variable | Intraobserver variability | Interobserver variability | ||
|---|---|---|---|---|
| Mean ± SD | % | Mean ± SD | % | |
| LV strain | −17.7 ± 2.8 | 4 | −18.2 ± 3.2 | 9 |
| LV systolic SR | −0.9 ± 0.1 | 10 | −1.0 ± 0.2 | 18 |
| LV diastolic SR | 1.1 ± 0.3 | 11 | 1.2 ± 0.4 | 22 |
| RV strain | −18.3 ± 2.0 | 6 | −18.5 ± 2.2 | 11 |
| RV systolic SR | −0.9 ± 0.1 | 13 | −1.0 ± 0.1 | 22 |
| RV diastolic SR | 1.0 ± 0.3 | 16 | 1.1 ± 0.2 | 37 |
Two-Dimensional and Doppler Imaging
Table 2 shows detailed information on 2D and Doppler parameters before and after PVR. Both LV ejection fraction and qualitative LV function improved after surgery, although the changes were not clinically relevant. As expected, pulmonary regurgitation decreased significantly, and thus right-sided chambers remodeled and RV systolic function improved.
Ventricular-Ventricular Interactions before and after PVR
A close correlation between LV and RV peak systolic strain and SR and diastolic SR was found before PVR and was maintained after surgery. However, no correlation was found between chamber size and contralateral ventricular function either before or after PVR. Patients with higher degrees of pulmonary regurgitation had larger right ventricles both before and after surgery ( r = 0.30, P = .003, and r = 0.20, P = .06), although this was not correlated with LV size or function.
Outcomes
Patients were followed for a mean duration of 3.0 ± 2.7 years; six patients were lost to follow-up (4.5%).
Factors Associated with LV and RV Function at Follow-Up
Univariate and multivariate analyses for LV and RV peak systolic strain after PVR are depicted in Table 4 . In the multivariate analysis, better LV peak systolic strain preoperatively was associated with better LV peak systolic strain after surgery. Similarly, better RV peak systolic strain preoperatively was independently associated with better RV peak systolic strain after surgery, although the patients who had the most improvement were those with worse function preoperatively (0.5 ± 0.2%, P = .002), as shown in Figure 1 .
