Right ventricular (RV) dysfunction is a significant cause of morbidity and mortality after surgical correction of tetralogy of Fallot (TOF). Transatrial/transpulmonary repair avoids a ventriculotomy (in contrast to the transventricular approach) in order to preserve the structure and function of the right ventricle. We performed a pilot prospective randomized controlled trial in infants with TOF undergoing primary repair.
A pilot prospective controlled clinical trial was conducted in infants with TOF undergoing primary repair between January 2008 and December 2009. One hundred and six patients were recruited in the trial and divided into a transatrial–transpulmonary approach group (Group A; n = 53) and a transventricular approach group (Group B; n = 53), depending on the different surgical techniques used.
Preoperative patient characteristics and procedure-related variables were similar. There were no deaths in Group A, while two patients died in Group B. There were significant differences in cardiopulmonary bypass time (95.02 ± 23.8 vs. 85.23 ± 22.63 minutes, p = 0.032), cross-clamp time (69.4 ± 10.36 vs. 61.17 ± 9.38 minutes, p = 0.035), inotropic support (1.63 ± 0.97 vs. 2.1 ± 1.09 days, p = 0.02), intubation time (26.62 ± 12.48 vs. 33.02 ± 17.55 hours, p = 0.033), duration of stay in the intensive care unit (ICU) (2.25 ± 1.28 vs. 2.85 ± 1.46 days, p = 0.026), and the incidence of arrhythmia [3 patients (5.7%) vs. 10 patients (18.9%), p = 0.038]. No significant differences in right/left ventricular pressure ratio and hospital stay were observed.
Transatrial/transpulmonary repair of TOF is associated with excellent surgical results and immediately follow-up.
infancy;tetralogy of Fallot;transatrial/transpulmonary;transventricular
Tetralogy of Fallot (TOF) is the most common cyanotic congenital heart disease1 characterized by large ventricular defect (VSD), right ventricular hypertrophy, right ventricular outflow tract (RVOT) obstruction, and an over-riding aorta. Since the first successful repair of TOF by Lillehei and associates in 1954 in a 10-year-old boy,2 this most common cyanotic heart disease has been studied extensively; however, controversies still exist regarding whether an early primary repair or a two-stage approach is the technique of choice.3 An early primary repair is preferred for several reasons, including reduction of Right ventricular (RV) pressure overload and elimination of hypoxemia, cyanosis, and secondary organ damage. The optimal age and surgical approach of TOF have been debated for several decades. Currently, most centers favor primary elective repair in infancy.4; 5; 6 ; 7
Traditionally, TOF was repaired through an RV incision providing an excellent exposure for closure of the VSD and relief of RVOT obstruction. However, RV function is impaired after ventriculotomy owing to a reduction in the regional wall motion around the incision,8 which may increase the incidence of ventricular arrythmias and sudden death. The right atrial approach (combined with a transpulmonary approach) to repair of TOF was reported by Hudspeth and coworkers in 1963.9 Subsequently, it has been shown that TOF can be repaired reliably through a transatrial/transpulmonary approach,10; 11; 12 ; 13 which avoids the impairment of RV function caused by ventriculotomy. However, no perspective data are available to the surgeon to compare the two approaches and adopt the optimal one.
The principal aim of our study randomized controlled trial was to compare between the transatrial/transpulmonary and transventricular approaches for the primary repair of TOF in infancy by evaluating their advantages and disadvantages and find the optimal approach.
This study was approved by the Institutional Review Board of Third Military Medical University, and was in compliance with the Health Insurance Portability and Accountability Act Regulations and the Declaration of Helsinki. The Institutional Review Board waived the need for individual patient consent. We used the consolidated standards of reporting trials checklist for design and conduct of this study.
In total, 106 patients were recruited in this study between January 2008 and December 2009. An informed patient consent was obtained before operation. For this study, patients with TOF but without important complicated features were selected. Such patients were considered to be those with classic TOF who underwent repair without using a valved extracardiac conduit or an orthotopically inserted pulmonary valve substitute. All patients were diagnosed by echocardiography and cardiac computed tomography. The McGoon ratio and left ventricular volume index of diastolic end (LVEDVI) were calculated. The patients with a McGoon ratio of <1.2 and an LVEDVI of <30 mL/m2 were excluded. Patients undergoing a two-stage operation and the abnormal coronary artery course (double-left anterior descending branches, preventing a transventricular approach to repair TOF) were also excluded.
Randomization was undertaken with sequential closed envelopes containing the treatment strategies assigned to each patient. In each case, the surgeon opened the envelope on the day of surgery to assign the treatment strategy to each patient.
All complete repairs were performed after standard median sternotomy using a cardiopulmonary bypass with moderate systemic hypothermia (25–28°C). The patent ductus arteriosus was repaired before cross-clamping the ascending aorta to prevent lung perfusion. Antegrade cold crystalloid and cold blood cardioplegia were used for myocardial protection. The transatrial/transpulmonary surgical approach was adopted in Group A and transventricular surgical approach in Group B.
In Group A, visualization of the VSD was usually adequate through the tricuspid valve and was even easier after division and resection of the obstructing muscle bundles. An autologous pericardial patch (washed in 0.6% glutaraldehyde for 10 minutes) was cut to the appropriate size, usually equal to the diameter of the mid ascending aorta. It was then sutured to the right side of the septum with a continuous suture technique, with 6-0 Prolene ACC. Suturing was usually started on the anterior limb of the septal band, as close as possible to the aortic valve, and continued inferiorly to the angle between the anterior and posterior limbs of the septal band, past the medial papillary muscle and under any chordae tendinae from the septal leaflet, until the posteroinferior rim of the defect was reached. At this point, suturing must be done approximately 5 mm away from the crest of the VSD itself and only on the RV side, to avoid injury to the atrioventricular node or bundle of His. This initial arm of the suture was then brought to the right atrium passing the needle through the septal leaflet annulus of the tricuspid valve, away from the coronary sinus orifice and the triangle of Koch. The other needle was brought superiorly over the infundibular septum and the aortic valve annulus, without injuring the aortic valve leaflets. The root of the aorta was pressed when it difficult to expose the aortic valve annulus. The sutures were kept close to the aortic valve annulus to avoid residual interventricular shunts. The needle was then brought to the atrium through the anterior leaflet of the tricuspid valve at its junction with the ventriculoinfundibular fold. Finally, the patch was sutured under the septal leaflet of the tricuspid valve, over the penetrating bundle, down to the needle of the first arm of the suture. Pericardial pledgets might be used to reinforce the suturing at the septal leaflet.
In Group B, a longitudinal incision was made in the right ventricle, a few millimeters away from the pulmonary valve. The obstructing muscles in the RVOT were excised until the VSD could be detected clearly. The VSD was closed with an autologous pericardial patch. After pulling the patch down, the sutures were continued in a clockwise fashion, taking care to keep it on the right side of the septum. In the region of the bundle of His, it was safer to place the sutures in the base of the tricuspid valve leaflet. Particularly in case of a very thin tissue, extra pledgeted sutures could be used. The other end of the continuous suture was used to run round in an anticlockwise manner and a knot was tied at the top. If the annulus was not enough, the incision was carried upward through the pulmonary valve to the bifurcation of the main pulmonary artery. In case of a proximal narrowing in one of the individual pulmonary arteries, the incision should be continued across the stenotic part. When making the incision through the pulmonary valvular annulus, it is best to try to take the incision between cusps so as to preserve the cusps. To assess the adequacy of the RVOT, Hegars dilators were introduced through RVOT to assess adequacy based on normalized sizes according to for body surface area.14 In addition, peak RV/LV pressure ratio was measured by direct RV puncture to rule out any significant residual outflow tract obstruction. For a postrepair ratio of <0.7, the RVOT appeared to be reconstructed adequately. Pulmonary annulus was enlarged by a glutaraldehyde-treated autologous pericardial patch when necessary. Transesophageal echocardiography was used to evaluate the surgical result immediately.
Associated cardiac anomalies, such as patent ductus arteriosus, or patent foramen ovale were treated simultaneously. The major aortopulmonary collateral arteries were evaluated by the doctor and coil embolization was performed.
The data were managed and analyzed using SPSS13.0. All continuous variables were expressed as mean ± standard deviation (M ± SD). Continuous variables were compared by standard t tests. Categorized variables were analyzed with chi-square tests. All statistical tests were two tailed. A p value of <0.05 was considered statistically significant.
Patient characteristics and procedure-related variables were similar. Fifty-three patients (26 boys and 27 girls, aged 11.53 ± 5.58 months) were assigned to Group A (transatrial–transpulmonary approach group) and 53 patients (28 boys and 25 girls, aged 10.43 ± 4.7 months) to Group B (transventricular approach group). There was no significant difference in the McGoon ratio and LVEDVI between the groups. Both groups were comparable for patient characteristics and procedure-related parameters (Table 1 ; Table 2).
|Variables||Group A (n = 53)||Group B (n = 53)||t/χ2||p|
|Mean age (D ± SD)||11.53 ± 5.58||10.43 ± 4.7||1.09||0.278|
|Sex ratio (M/F)||26/27||28/25||0.151||0.698|
|Weight (kg)||7.61 ± 2.27||7.06 ± 2.05||1.301||0.196|
|Body surface area||0.52 ± 0.14||0.49 ± 0.13||0.986||0.327|
|Hematocrit (%)||47.28 ± 8.16||48.73 ± 8.61||0.889||0.376|
|Cyanotic spell history (%)||8 (15.1%)||11 (20.8%)||0.577||0.613|
|Arrhythmias||4 (7.5%)||3 (5.7%)||0.153||>0.99|
|RV-PA mean gradient (mmHg)||84.6 ± 11.25||81.64 ± 12.63||0.382||0.703|
|NYHA class III-IV||6 (11.3%)||10 (18.9%)||1.18||0.416|
|McGoon ratio||1.46 ± 0.23||1.49 ± 0.26||0.794||0.429|
|LVEDVI (mL/m2)||38.37 ± 8.66||41.36 ± 11.47||1.512||0.134|
Types of arrhythmia identified were atrial fibrillation in three and nonsustained ventricular tachycardia in four.
Group A = transatrial–transpulmonary approach; Group B = transventricular approach; LVEDVI = left ventricular volume index of diastolic end; NYHA = New York Heart Association; RV = right ventricular; SD = standard deviation.
|Variables||Group A (n = 53)||Group B (n = 53)||t/χ2||p|
|Patent foramen ovale (%)||17 (32.1%)||7 (13.2%)||5.386||0.02|
|Patent ductus arteriosus at time of complete repair (%)||7 (13.2%)||4 (7.5%)||0.913||0.339|
|Atrial septal defect (%)||11 (20.8%)||9 (17%)||0.247||0.62|
|Right aortic arch (%)||3 (5.7%)||0||3.087||0.079|
|MAPCAs (%)||8 (15.1%)||6 (11.3%)||0.329||0.566|
|Left superior vena cava (%)||2 (3.8%)||8 (15.1%)||3.975||0.046|
MAPCAs = major aortopulmonary collateral arteries.
Two patients died in Group B compared with no death in Group A (Table 3). Overall, two patients died in the hospital: one patient for serious low cardiac output syndrome and multiple organ failure, and the other for malignant arrhythmia of ventricular fibrillation and ventricular flutter.
|Variables||Group A (n = 53)||Group B (n = 53)||t/χ2||p|
|CPB time (min)||95.02 ± 23.8||85.23 ± 22.63||2.17||0.032|
|Cross-clamp time (min)||69.4 ± 20.36||61.17 ± 19.38||2.13||0.035|
|RV/LV pressure ratio||0.45 ± 0.13||0.41 ± 0.1||1.71||0.091|
|Inotropic support (d)||1.63 ± 0.97||2.1 ± 1.09||2.36||0.02|
|Intubation time (h)||26.62 ± 12.48||33.02 ± 17.55||2.16||0.033|
|ICU stay (d)||2.25 ± 1.28||2.85 ± 1.46||2.27||0.026|
|Hospital stay (d)||10.8 ± 2.03||10.6 ± 2.37||0.466||0.642|
|Presence of low cardiac output (no. patients)||1 (1.9%)||5 (9.4%)||2.827||0.093|
|Renal insufficiency (newly diagnosed)||1 (1.9%)||2 (3.8%)||0.343||0.558|
|Arrhythmiasa||3 (5.7%)||10 (18.9%)||4.3||0.038|
CPB = cardiopulmonary bypass; Group A = transatrial–transpulmonary approach; Group B = transventricular approach; ICU = intensive care unit; LV = left ventricular; RV = right ventricular.
a. Types of arrhythmia identified were atrial fibrillation, supraventricular tachycardia, and nonsustained ventricular tachycardia.
There were significant differences in cardiopulmonary bypass time (95.02 ± 23.8 vs. 85.23 ± 22.63 minutes, p = 0.032), cross-clamp time (69.4 ± 10.36 vs. 61.17 ± 9.38 minutes, p = 0.035), inotropic support (1.63 ± 0.97 vs. 2.1 ± 1.09 days, p = 0.02), intubation time (26.62 ± 12.48 vs. 33.02 ± 17.55 hours, p = 0.033), duration of stay in the intensive care unit (ICU) (2.25 ± 1.28 vs. 2.85 ± 1.46 days, p = 0.026), and the incidence of arrhythmia [3 patients (5.7%) vs. 10 patients (18.9%), p = 0.038] between the two groups. No difference was observed in RV/LV pressure ratio (0.45 ± 0.13 vs. 0.41 ± 0.1, p = 0.091), hospital stay (10.8 ± 2.03 vs. 10.6 ± 2.37 days, p = 0.642), reoperation [3 patients (5.7%) vs. 0 patients, p = 0.079], and newly diagnosed renal insufficiency (1, 1.9% vs. 2, 3.8%, p = 0.558) between two groups ( Table 3). In Group A, two infants were reoperated for bleeding and one for a twist of the left pulmonary artery, using a transannular patch (TAP).
Echocardiography did not reveal residual VSD in any patient before discharge. The RVOT pressure gradient was low without significant hemodynamic effects. Most patients in both groups had mild pulmonary insufficiency. Only 5.7% in Group A and 11.8% in Group B had moderate–severe pulmonary insufficiency. Mild–moderate impairment of RV function occurred only in a few patients (3.8% in Group A and 7.84% in Group B). Tricuspid valve function was well preserved and most patients had mild Tricuspid valve insufficiency (TI). Overall, RV and LV functions were well preserved. Predischarge echocardiography data are summarized in Table 4.
|Variables||Group A (n = 53)||Group B (n = 51)||t/χ2||p|
|Right ventricular outflow tract pressure gradient||14.46 ± 6.15||13.47 ± 6.9||0.768||0.444|
|Pulmonary insufficiencya||3 (5.7%)||6 (11.8%)||1.23||0.268|
|Tricuspid regurgitationa||5 (9.4%)||4 (7.84%)||0.083||0.773|
|Right ventricular functionb||2 (3.8%)||4 (7.84%)||0.792||0.374|
a. Moderate or severe insufficiency.
b. Mild–moderate impairment.
The mean follow-up period was 39.6 ± 5.59 months. No significant residual lesion has been detected and no reoperation has been necessary. One died from severe lung infection with ventricular arrhythmia during late follow-up. Detailed echocardiographic late follow-up data are summarized in Table 5. Specifically, during this follow-up period, the RVOT remained free of significant obstruction, and pulmonary insufficiency or tricuspid valve insufficiency did not progress. Thus, mean RVOT gradient remained low and no patient has developed significant RVOT obstruction. More patients who underwent transventricular repaired developed arrhythmia, especially nonsustained ventricular tachycardia, than those who underwent transatrial/transpulmonary surgery (1.96% vs. 14%, p = 0.025). RV function remained stable during this follow-up period in patients in whom the transatrial/transpulmonary approach was used, while it was impaired later in those in whom the transventricular approach was used (0 vs. 8%, p = 0.039).
|Variables||Group A (n = 53)||Group B (n = 50)||t/χ2||p|
|NYHA class I–II||51 (96.2%)||46 (92%)||1.516||0.22|
|Arrhythmiasa||1 (1.89%)||7 (14%)||5.13||0.024|
|Right ventricular outflow tract pressure gradient||12.16 ± 5.56||11.6 ± 6.84||0.648||0.56|
|Pulmonary insufficiencyb||3 (5.66%)||9 (20%)||3.659||0.056|
|Tricuspid regurgitationb||5 (9.43%)||4 (8%)||0.066||0.797|
|Right ventricular functionc||0||4 (8%)||4.411||0.036|
NYHA = New York Heart Association.
a. Atrial fibrillation, supraventricular tachycardia, and nonsustained ventricular tachycardia.
b. Moderate or severe insufficiency.
c. Mild–moderate impairment.
TOF has been known to exist for more than 100 years and the typical cardiac abnormalities have been known to surgeons. The operative correction of TOF has been performed for more than 40 years.15 The varied features of this defect and sequelae after repair have been understood better in the past two decades. The timing of surgical correction of TOF is controversial. The predominant trend in the timing of surgical procedure has been toward an earlier intervention with elective repair within the first 4–6 months of life,3 although others have advocated earlier repairs in symptomatic neonates.16 ; 17 According to the results of many studies,18; 19; 20 ; 21 the surgical procedure can be performed with relatively low mortality (0–7%). In our center, the correction of TOF was performed in infants usually older than 6 months.
Generally, it is accepted that early correction of TOF minimizes secondary damages to the heart, lungs, and nervous system and that functional results after elective primary repair early in infancy are superior to either two-stage repair or repair later in life. However, primary repair during infancy may need more frequent use of a TAP, but the association between the use of a TAP and the long-term outcome is uncertain. The myriad reports showed that TAP could lead to long-term pulmonary valve insufficiency (PI), although this remains a controversial issue. Several studies have demonstrated good long-term outcomes after repair of TOF with a TAP. Bacha and colleagues22 reported that the functional status, as measured by the New York Heart Association class, was similar to TAP and valve-sparing. Kirklins group23 found that the risk of reoperation for PI at 20 years was 7%; they ascribed this low rate of reoperation, despite a significant incidence of PI, to the adaptive properties of the right ventricle. These findings are contradicted by other studies showing the deleterious effects of pulmonary insufficiency on the right ventricle. Cardiac magnetic resonance cine examinations of TOF patients late after repair have demonstrated that PI is closely associated with TAP and results in significant RV dysfunction, even in asymptomatic patients.24 ; 25 Gatzoulis and associates26 showed that PI and TAP are associated with the development of ventricular tachycardia and sudden death late after TOF repair. The group at the Great Ormond Street found that a total of 124 patients with TAPs were highly associated with PI and PI was associated with RV and LV dysfunctions.27 The late effects of TAP and PI on both RV and patient functional status are further documented by the growing number of reports on the ventricular improvement and symptomatic relief achieved with prosthetic pulmonary valve insertion late after repair.28; 29 ; 30
In our study, slicing technique of the RVOT was a crucial step in the transatrial–transpulmonary repair of TOF, which could relieve the obstructive RVOT.31 In addition, we made great efforts to avoid a TAP in order to preserve the annulus and pulmonary valve, which may improve the postoperative RV function and reduce long-term complications including arrhythmias and PI. In those cases, an infundibular patch was required; we think that two patches above and below the annulus are preferable to a single patch crossing the annulus. This is consistent with the recommendation of the Northwestern University group: avoidance of a TAP with preservation of the annulus of the pulmonary valve.30 Minimal incision in the right ventricle was another way to preserve RV function.
RV diastolic dysfunction often occurs in patients following repair of TOF. The postoperative course in those patients is characterized by low cardiac output, longer requirement of inotropics, longer stay in the ICU, and longer chest drainage time. Therefore, assessment of RV diastolic function is important in patients undergoing correction of TOF, particularly in those in whom a transventricular approach was used. Index of myocardial performance (IMP) is a new parameter, which has been validated as the measure of both systolic and diastolic functions of the right ventricle. Sachdevand colleagues reported32 the lower the IMP, the more the possibility for the surgical approach to have complications during the postoperative course with a longer duration of inotropic support, ventilation time and ICU stay, higher central filling pressures, prolonged chest tube drainage, and a higher dose of diuretics. However, as time goes by, RV dysfunction and pulmonary valve insufficiency slowly develop in almost all patients to some degree.33 Exercise capacity decreases in patients with RV dilation, but most patients remain asymptomatic. Pulmonary valve replacement (PVR) may be a frequent consideration for patients with fatigue and exercise intolerance. PVR unexplained by factors other than RV dilatation and dysfunction would be a strong indication for valvular replacement. It is associated with improvements in exercise performance, as well as in RV size and function.30
Transmural myocardial scarring of the right ventricle is an important factor in the development of malignant ventricular arrhythmias. The QRS prolongation (>180 ms) is the most sensitive predictor of development of malignant ventricular arrhythmias.33 Transatrial–pulmonary correction of TOF may prevent the incision in RV and reduce the incidence of ventricular arrhythmias. The smaller the incision, the more the possibility of avoiding malignant ventricular arrhythmias if ventriculotomy is necessary. The relation between RV dilatation and ventricular arrhythmias (and sudden death) is controversial.
A limitation of the present study is its relatively short period of follow-up, which should be long enough to obtain more detailed information on the RV function by magnetic resonance imaging or IMP.
In conclusion, we assume that TOF can be corrected safely in infancy. This can also be accomplished safely in all patients and should be attempted despite symptoms, minimizing the injury on the RV muscular structures, and preserving the structure and function of the right ventricle for a better long-term outcome of this growing population.