Author + information
- Received December 9, 2004
- Revision received June 10, 2005
- Accepted June 21, 2005
- Published online October 4, 2005.
- Marco Pocar, MD, PhD⁎,‡,
- Emmanuel Villa, MD⁎,‡,
- Alexandra Degandt, MD⁎,
- Philippe Mauriat, MD†,
- Philippe Pouard, MD† and
- Pascal R. Vouhé, MD, PhD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Pr. Pascal R. Vouhé, Service de Chirurgie Cardiaque, Hôpital Necker-Enfants Malades, 149, Rue de Sèvres, 75743 Paris Cedex 15, France
Objectives The study was designed to evaluate perioperative and late results after primary, single-stage arterial switch operation (ASO) associated with aortic arch obstruction repair. Outcome of patients with more than five years of follow-up were analyzed.
Background The treatment of patients with transposition of the great arteries, or other forms of ventriculoarterial discordance suitable for an ASO, with coexisting arch obstruction is a difficult task. Single-stage repair has become the treatment of choice at many institutions, but large series with long-term results are seldom reported.
Methods Between 1990 and 1998, a primary operation including aortic arch repair through a midline sternotomy was performed in 38 patients. The relief of arch obstruction was accomplished during a period of hypothermic circulatory arrest, employing a wide pericardial patch to enlarge the inner curvature of the entire arch in most patients.
Results There were nine (24%) hospital deaths. None could be directly related to aortic arch repair, but additional risk factors for an ASO were common (right ventricular hypoplasia, complex coronary anatomy, uncommon relationship between the great vessels or severe pulmonary hypertension). There were no late deaths. Four patients required cardiac reoperation, whereas three underwent successful treatment of recurrent coarctation with balloon angioplasty.
Conclusions Infants with ventriculoarterial discordance and aortic arch obstruction represent a high-risk subgroup of candidates for an ASO. Despite a non-negligible operative mortality, single-stage primary repair represents the treatment of choice, and follow-up of operative survivors is favorable. Pericardial patch enlargement is a reliable technique for arch obstruction repair.
The association between transposition of the great arteries (TGA) and aortic arch obstruction is relatively infrequent, especially in the absence of a ventricular septal defect (VSD). When this combination occurs, the natural history is very poor and the treatment constitutes a difficult task (1–5). Hypoplasia of the aortic arch is more commonly encountered in specific forms of ventriculoarterial discordance, being reported in up to 50% of cases in some subgroups of double outlet right ventricle (RV), especially in the presence of a subpulmonary VSD (Taussig-Bing anomaly) (6). Historically, the approach to such complex malformations favored a two-stage treatment, including primary coarctation repair with pulmonary artery banding, followed by intracardiac repair at a later date (5–7). During the past two decades, a one-stage procedure has gradually become the treatment of choice to avoid the detrimental effects of pulmonary artery banding on the left ventricle and neoaortic root and valve (8,9).
Our group has adopted this approach since 1990. This retrospective study includes all infants who underwent single-stage anatomic repair of TGA complexes with aortic arch obstruction at the Hôpital Laënnec in Paris between July 1990 and June 1998, and describes perioperative and medium-to-long-term results.
Between July 1990 and June 1998, 38 consecutive patients underwent an arterial switch operation (ASO) with concomitant repair of aortic arch obstruction. Most patients (31 of 38, 82%) underwent operation during the first month of life. Intravenous prostaglandin infusion to maintain patency of the ductus arteriosus and percutaneous atrial septostomy (Rashkind maneuver) were necessary in most neonates to prevent arterial desaturation. Except for two patients with type B interrupted aortic arch, diffuse arch hypoplasia was invariably present, with the area of most severe narrowing located immediately after the origin of the left subclavian artery. A localized coarctation “ridge” at the aortic isthmus was found in about three-quarters of the cases. Patients were divided into three subgroups, according to coronary anatomy: a first group included infants with Yacoub’s type A anatomy (normal pattern) (10), a second group comprised type D and E coronary arteries (circumflex artery originating from the right coronary ostium, and right coronary artery originating from the left ostium, respectively), and a third group included all other more complex patterns. The presence of specific high-risk situations, namely, a single or paracommissural coronary ostium, an intramural coronary, or a course between the great vessels of either coronary artery, was also analyzed separately. The two latter conditions were distinguished because the proximal segment of two of four coronary branches coursing between the great vessels was tangential to the aortic wall but not truly intramural (4,11,12). The majority of patients had an associated VSD that was often related to a variable degree of conal septal malalignment. Consequently, some degree of right ventricular outflow tract obstruction (RVOTO) was present in about a quarter of the cases, whereas a perimembranous VSD was more common in TGA with near-anteroposterior great vessels. Relative hypoplasia of the ascending aorta was always present when compared with the pulmonary trunk and determined an important mismatch between the two great vessels, the pulmonary artery diameter being 25% to 30% (or more) greater than the aorta. Also, the relationship between the great vessels varied widely, with almost one-half of the patients showing a non-anteroposterior anatomy. An unbalanced anatomy with a smaller than normal RV is more common in TGA with arch obstruction (2,4,13). In such instances, RV hypoplasia was retrospectively defined when the tricuspid valve annulus diameter measured two standard deviations below normal values. Demographic and preoperative characteristics are listed in Table 1.
All patients underwent surgery through a median sternotomy. After the institution of cardiopulmonary bypass, always using a single arterial cannula in the ascending aorta, the aorta was cross-clamped at 28°C core temperature during systemic cooling and the myocardium protected according to a modified Buckberg protocol (antegrade warm induction, cold infusion every 20 min, and controlled warm aortic root reperfusion). The ascending aorta and pulmonary trunk were then transected and the Lecompte maneuver performed in all cases.
After reaching 18°C, extracorporeal perfusion was stopped and the aortic arch reconstructed under hypothermic circulatory arrest (HCA). Since November 1996, all patients but one underwent selective antegrade cerebral perfusion through the innominate artery during HCA. An extended resection and end-to-end anastomosis was performed in nine patients, mainly in the early period. In five of nine instances, however, associated patch enlargement of the ascending aorta and proximal arch to overcome the mismatch between the arterial trunks was unavoidable. As a consequence, patch enlargement became the technique of choice for aortic arch repair, extending from below the insertion of the ductus arteriosus to the transection line in the ascending aorta. This also applied to infants with interrupted aortic arch, who were the only patients to undergo resection and end-to-end anastomosis after 1995. Twenty-nine patients underwent patch repair alone, employing glutaraldehyde-treated autologous pericardium in most cases.
After completion of the aortic arch repair, systemic arterial perfusion was resumed, and the ASO performed (14). The VSD patches were sutured in place either through the right atrium and tricuspid valve or through the great arteries, or both, always avoiding a right ventriculotomy. The conal septum was excised in three infants with double outlet RV and RVOTO (subaortic stenosis). The Rashkind atrial septal defect was only loosely closed with one or two separate stitches, with the aim to leave a restrictive interatrial communication in patients likely to develop perioperative pulmonary hypertension (PHT). A patch of fresh autologous pericardium was used to reconstruct the defect in the neopulmonary root after detachment of the coronary ostia; a second patch of bovine pericardium was employed to widen the pulmonary artery anteriorly, in case of important mismatch between the great vessels. The neopulmonary root was then anastomosed to the pulmonary trunk during controlled aortic root reperfusion before aortic declamping. Operative data are summarized in Table 2.
Reviews of clinical records and contacts with the referring cardiologists served for collection of hospital and follow-up data. Echocardiograms were available for all patients, whereas cardiac catheterization was performed in case of symptoms or electrocardiographic/echocardiographic findings suggesting myocardial ischemia or before school age for the purpose of a prospective study (15). Patients were selected for analysis if surgery had been performed before the preceding five years at the time the study was conceived. Duration of follow-up among hospital survivors was 99.2 ± 43.7 months (longest follow-up: 13.7 years). Three patients referred to our hospital, two from Algeria and one from French Guyana, were lost to long-term follow-up. Excluding such patients, follow-up was 110.4 ± 29.5 months. Patients were also divided into groups to analyze mortality trends over time. Reintervention refers to open cardiovascular surgery (reoperation) and interventional catheterization (balloon angioplasty for recurrent coarctation), excluding delayed sternal closure, revision for postoperative bleeding, infection or chylothorax, and pacemaker implantation or replacement.
Chi-square test (or Fisher exact test when 2 × 2 tables had a cell with an expected frequency of <5; in practice, all significant p values of 2×2 table variables refer to Fisher exact test), and stepwise logistic regression were used for univariate and multivariate analysis of variables thought to affect adverse outcome. Values of p < 0.05 were considered statistically significant. Variables found to be significant or that approached significance (p < 0.1), according to univariate analysis, were subsequently evaluated in the multivariate analyses. The Kaplan-Meier method was used to estimate the probability of survival and freedom from adverse outcome events. All 95% confidence intervals (CIs) were calculated as ± 2 SEs. Difference in probability estimates was calculated with the log-rank test. The Cox proportional hazard model was used for multivariate analysis of time-dependent variables. Continuous variables are expressed as mean ± SD. Data were analyzed with SPSS for Windows (version 11.5.1, SPSS Inc., Chicago, Illinois).
Actuarial 10-year survival was 76% (95% CI 62% to 90.1%) (Fig. 1).There were nine hospital deaths (24%), all during the first postoperative month, and no late deaths. Survival probability remained constant after the perioperative period and, in practice, was determined by operative mortality alone.
Among the nine patients who died, RV hypoplasia (five of nine) and coronary malperfusion (four of nine), alone or in association, were the most common primary causes of death. When occurring in the presence of a hypoplastic RV, coronary malperfusion was related to overdistension of the right heart and tearing of a relocated coronary artery (an unsuccessful attempt to place a RV-to-pulmonary artery conduit was made in one patient). In both such instances, this complication occurred in the presence of side-by-side great vessels and a single coronary artery (one common left ostium with the right coronary between the aorta and pulmonary artery, and one single right coronary with an intramural left anterior descending branch). The interrelationship between coronary complications, RV failure, and great vessels’ anatomy is reflected by the fact that four of five infants with a hypoplastic RV showed complex coronary artery patterns. Similarly, the two patients with coronary malperfusion alone as a leading cause of death had non-anteroposterior great vessels (one showed side-by-side great arteries, and one had a leftward, anteriorly located aorta). Infants with a hypoplastic RV included the two patients cases with situs viscerum inversus and the patient with tricuspid valve dysplasia. Both neonates who died because of uncontrollable PHT had an associated VSD. Finally, no death seemed directly related to aortic arch repair.
The analysis of the correlation between outcome and the time interval between the first operation in the series (June 15, 1990) and surgery failed to show a linear “learning curve effect”. There were, however, 5 of 11 (45%) deaths during the first two years’ experience and 4 of 27 (15%) during the following 6 years.
Complete heart block with subsequent implantation of a permanent pacemaker, left phrenic nerve paralysis requiring diaphragm plication, deep wound infection, and left recurrent laryngeal nerve impairment occurred in one instance each, whereas postoperative chylothorax developed in two patients (one of whom necessitated surgical ligation of the thoracic duct). No neurologic complications with permanent sequelae were observed.
Reintervention and other late events
Actuarial freedom from overall reintervention, reoperation, and recoarctation requiring surgery or percutaneous angioplasty among operative survivors was 89% (95% CI 77% to 100%), 96.3% (95% CI 89% to 100%), and 92.6% (95% CI 82% to 100%) at 1 year; 77% (95% CI 61% to 93.6%), 89% (95% CI 76% to 100%), and 89% (95% CI 76% to 100%) at 5 years; and 72% (95% CI 54% to 90.3%), 83% (95% CI 66% to 98.9%), and 89% (95% CI 76% to 100%) at 10 years, respectively (Fig. 2).Recoarctation was amenable to percutaneous treatment and never required open surgery. Reoperations were performed for RVOTO or left main coronary stenosis or both. One patient who underwent RVOTO repair required a RV-to-pulmonary artery conduit because of the coronary anatomy. Patch enlargement for left main coronary artery stenosis was accomplished with fresh autologous tissue, such as a portion of saphenous vein or innominate vein (16). No patient required late neoaortic valve or root surgery.
During late follow-up, four patients showed mild RVOTO. Trivial neoaortic valve regurgitation was documented in three cases, and a mildly dilated neoaortic root in two. Two children with known coronary stenoses (one left, and one right coronary artery stenosis, respectively, both in type A anatomy) and no signs of inducible myocardial ischemia were treated conservatively and underwent strict controls. No patient developed true or false aortic arch aneurysm after patch reconstruction, in spite of the generous use of fixed autologous pericardium.
Finally, no major neurologic complications were observed perioperatively. Although cognitive function tests were not available, long-term neurodevelopmental outcome was satisfactory and seems comparable to that of a neonatal ASO combined with HCA or low-flow cardiopulmonary bypass. In this respect, HCA was limited to the aortic arch repair and is not adopted for isolated ASO at our institution.
Despite the improvements in surgical techniques and perioperative intensive care management, the ASO in the presence of aortic arch obstruction remains a surgical challenge, with coarctation reported as a risk factor for late death after a successful operation (17,18). Although long-term mortality and freedom from reoperation rates among survivors were, in some series, similar to patients who underwent the ASO for “uncomplicated” classical TGA, several aspects seem to pertain specifically to this subset of surgical candidates. No death could be related to the aortic arch repair, per se, but complex anatomy is more common and determines a higher complexity of the surgical strategy and, consequently, an increased operative mortality (2,14,19,20).
The aspects discussed in the following sections are subdivided according to their different impact on mortality and morbidity.
RV hypoplasia and dysfunction
Right ventricular failure is a known cause of operative death after the ASO (5,14). A hypoplastic RV represented a leading cause of death in five of nine instances, either alone or in combination with coronary malperfusion. Perioperative RV dysfunction was the strongest predictor of death and the only variable that reached statistical significance at multivariate analysis. The association between aortic arch obstruction and ventriculoarterial discordance encompasses a wide spectrum of anatomic variants, and the rate of RV hypoplasia is higher than for isolated TGA (2,4,13). Conceptually, the resulting malformation at the extreme end could be called, with a provocative term, “hypoplastic right heart syndrome.” Indeed, the known association of hypoplastic left ventricular structures and aortic coarctaction is reflected, in TGA, with hypoplasia of the aortic connection and valve, RV infundibulum (RVOTO, see subsequent text), and tricuspid valve orifice, especially in the absence of a VSD. Interestingly, 4 of 8 infants (50%) with an intact ventricular septum but only 5 of 30 (17%) with an associated VSD had a hypoplastic RV. On the other hand, every effort should be made to avoid a right ventriculotomy in patients requiring VSD closure.
Although no patient had a straddling or markedly overriding tricuspid valve in this series, these conditions might render insufficient the definition of a hypoplastic RV solely on the basis of tricuspid orifice dimensions. Echocardiographic data concerning the tricuspid valve were not always available, especially in earlier records and in case of balanced ventricular morphology. Consequently, a correlation between the tricuspid valve orifice and outcome could not be analyzed; however, measurements were always specified in case of a small RV. Thus, a subgroup of patients with relative hypoplasia could be retrospectively outlined among those with an unbalanced anatomy.
Finally, no patient had a severely diminutive RV, but primary biventricular repair yielded a high mortality rate among patients with RV hypoplasia (5 of 9, 56%; 3 of 5, 60% with an associated VSD; 2 of 4, 50% with no VSD). Thus, the opportunity of a staged procedure and, eventually, of a single or “one-and-a-half” ventricle approach should be considered. Residual shunting (see below) might further reduce early mortality and seems to enhance RV growth (13). The lower prevalence of unsuccessful outcomes in later operations might reflect an improved selection of patients for primary repair.
Coronary and great vessels’ anatomy
Coronary malperfusion is a major issue in patients undergoing an ASO: in our experience, this was the primary cause of death in four of nine cases, either alone (two patients) or associated with RV failure (two patients, see previous). Except for patients with an intact ventricular septum, non-type A coronary patterns were common (see following); similar considerations can be drawn for the relationship between the great arteries, which was often non-anteroposterior. Coronary anatomy tended toward a higher degree of complexity in relation to rightward displacement of the aorta and side-by-side great vessels; interestingly, coronaries were type A in the instance of a left-ward anterior aorta. Furthermore, a pronounced difference in size between the great vessels has been well documented in TGA with associated cardiac malformations, in particular, VSD, coarctation, and Taussig-Bing anomaly (21), and a dilated pulmonary artery has been a constant observation in this series. All these aspects determine a technically higher-risk coronary transfer in many instances.
Supravalvular aortic stenosis might result from the anastomosis of a dilated neoaortic root with a diminutive ascending aorta and has been reported as a cause of operative mortality after an ASO (18). This complication has not been encountered, possibly because wide patch enlargement of the entire arch was performed to overcome the mismatch in size with the neoaortic root.
Relocated coronaries might be compressed along their proximal course by dilated pulmonary arteries, and this can be observed late after the operation (one patient required reoperation for obstruction of a compressed left coronary ostium). The advantages of the Lecompte maneuver have been questioned, particularly in case of side-by-side great arteries (4,11,22). Our policy has been to translocate the pulmonary arteries anteriorly, except in case of a right posterior aorta, which was never encountered in this series. Although the opportunity not to perform the Lecompte maneuver might be considered selectively in case of side-by-side great vessels with complex coronary patterns to avoid coronary compression by dilated (left) pulmonary arteries, the neoaorta assumes a smoother course when reconstructed posteriorly (especially with enlarged pulmonary arteries) and, in the present experience, required no reoperation (see the following text). Furthermore, reoperative surgery for RVOTO is technically simpler in case of an anteriorly translocated pulmonary outflow. Interestingly, the use of fresh autologous pericardium during primary neopulmonary root reconstruction might lessen the probability of recurrent RVOTO, because pericardial tissue has been shown to potentially differentiate into arterial vascular wall (23).
Associated VSD and PHT
The majority of patients showed a coexisting VSD (30 of 38, 79%); however, despite a high incidence of RV hypoplasia in case of intact ventricular septum (4 of 8, 50%), mortality was not increased in this subgroup (2 of 8, 25%), possibly because of a lower impact of PHT, but also because type A coronary anatomy was encountered in all cases except one.
Both patients who died because of intractable PHT had an associated VSD and were refractory to intensive treatment including inhaled nitric oxide, which was first used at our institution in 1992 (24). One patient also developed superior vena caval thrombosis with infection of a central venous catheter. Interestingly, VSD repair and the presence of RVOTO were risk factors for a prolonged stay in the intensive care unit: both might be correlated to PHT, because only two of ten patients with RVOTO had an intact ventricular septum. The possible role of incomplete atrial partitioning in determining a smoother course in patients likely to develop perioperative PHT remains to be ascertained, but the implantation of a fenestrated atrial or ventricular septal patch has been advocated in patients with a hypoplastic RV undergoing an ASO (13). A one-way valved atrial septal patch has also been suggested in case of defects characterized by RV hypoplasia or PHT (25). Moreover, the latter approach and a temporary central aorto-pulmonary shunt have been successfully used as rescue measures to wean infants with PHT from cardiopulmonary bypass (25,26). The potential for left-to-right shunting, with a consequent decrease in cardiac output if the left ventricle is dysfunctional after repair, represents a drawback that should be kept in mind in case of a non-valved interatrial communication. Finally, although it was not employed in our two patients, extracorporeal membrane oxygenation should be considered in patients with refractory PHT and, ideally, should be initiated in a controlled fashion to restore adequate perfusion before organ damage (27).
RVOTO and the fate of the neoaortic root
Although other authors have stressed the importance of subaortic stenosis and have been more aggressive in this respect (3,28,29), the conal septum was excised in only three cases. One operative death is partially attributable to residual subpulmonary stenosis in a patient with a Taussig-Bing heart, RV hypoplasia, and the left anterior descending coronary artery originating from the right coronary ostium, in whom an attempt to place a RV-to-pulmonary trunk conduit was unsuccessful. Conversely, the role of subaortic stenosis at time of the ASO is more critical in infants for whom a two-stage approach is selected, because RVOTO is likely to develop after pulmonary artery banding when this is performed together with coarctation repair as a primary operation (19). In the present series, only three patients underwent reoperation for the relief of RVOTO (one underwent associated left main coronary patch enlargement), whereas four others showed mild RVOTO at long-term follow-up. In the multi-institutional study on TGA conducted by the Congenital Heart Surgeon’s Society, the coexisting coarctaction of the aorta is an incremental risk factor for proximal right-sided outflow obstruction (30). One reoperated patient had the Taussig-Bing anomaly and, in a recent series from Germany, RVOTO was reported to often complicate the postoperative course of Taussig-Bing hearts and to be the main indication for reintervention (31).
The fate of the neoaortic root and valve is of particular concern after the ASO (32). The probability of neoaortic root dilation and need for reoperation in the long term is likely to be higher when primary enlargement of the pulmonary trunk is part of the congenital anomaly; however, no neoaortic root reoperation was necessary in this series, whereas three patients showed trivial neoaortic valve insufficiency and two others had a mildly enlarged neoaortic root. Apparently, this does not significantly differ from other patients after an ASO but might be influenced by the extensive patch enlargement for the relief of aortic obstruction (33). In this respect, a two-stage strategy contemplating pulmonary artery banding is probably deleterious, if not contraindicated.
Recoarctation has been observed in a minority of cases (3 of 29 survivors) and never required open surgery. The one-stage primary repair strategy probably correlates with this low rate, because aortic arch obstruction is related to diffuse arch hypoplasia rather than true coarctation localized at the aortic isthmus. Arch anatomy and the mismatch between the arterial trunks make patch augmentation extended to the entire arch and ascending aorta an almost indispensable step of the operation. In this respect, a left thoracotomy, which represents the standard approach for coarctation repair, is clearly less than ideal from a strictly surgical standpoint. Furthermore, no aneurysm formation was observed during follow-up, confirming the usefulness and reliability of fixed autologous pericardium for aortic reconstruction.
The study includes 38 consecutive patients with an infrequent complex congenital cardiovascular abnormality. This represents a relatively small population for statistical analysis and includes few late events. As a result, perioperative RV failure was the only variable to reach statistical significance at multivariate analysis among the predictors of death at univariate analysis, and no specific risk factors for adverse events during follow-up could further be outlined.
Among four patients who were referred from abroad, three could not be traced for long-term follow-up. These patients were less critically ill than average and were somewhat “naturally selected,” because they underwent surgical repair at age 87, 108, and 251 days, respectively. None of these infants died, but none had RV hypoplasia, a high-risk coronary pattern, or side-by-side great vessels. Because there were no late deaths and few reoperations in the whole series, it might be speculated that these patients are unlikely to be at increased risk for late death or reoperation: on the contrary, they probably have a reduced risk, and their inclusion in the study should not bias the estimates in this respect. Obviously, nothing is known about the fate of the aortic arch repair.
Finally, the operation is technically demanding, and the results outlined above were obtained with all procedures performed by the same operating surgeon (P.V.). The difference in operative mortality between early and late operations imposes caution in performing one-stage repair in non-experienced hospitals: in this setting, a simpler palliative operation performed as a life-saving “bridge-to-anatomic repair” might be preferable.
Long-term outcome after primary one-stage repair of TGA complexes and aortic arch obstruction seems similar to the overall late course of unselected patients after the ASO. The incidence of recurrent coarctation is low and could always be treated without open surgery; however, these infants are a higher-risk subgroup for the ASO. Right ventricular hypoplasia, complex coronary patterns associated with non-anteroposterior great arteries, and severe PHT represent common additional risk factors for operative mortality. Extensive patch enlargement of the inner curvature of the entire arch with autologous pericardium is, in our experience, the technique of choice for the relief of aortic arch obstruction. The concerns for future neoaortic root dilation and RVOTO constitute, at least, relative contraindications for pulmonary artery banding, whereas resection of the conal septum seems justified only in case of severe malalignment and RVOTO. With the possible exception of infants with unbalanced ventricles and a relatively hypoplastic RV, these considerations render a one-stage operation the preferred approach for the treatment of these critically ill newborns.
The authors are grateful to Ms. Corinne Pasquet for her kind secretarial assistance.
- Abbreviations and Acronyms
- arterial switch operation
- confidence interval
- hypothermic circulatory arrest
- pulmonary hypertension
- right ventricle/ventricular
- right ventricular outflow tract obstruction
- transposition of the great arteries
- ventricular septal defect
- Received December 9, 2004.
- Revision received June 10, 2005.
- Accepted June 21, 2005.
- American College of Cardiology Foundation
- Milanesi O.,
- Thiene G.,
- Bini R.M.,
- Pellegrino P.A.
- Moene R.J.,
- Ottenkamp J.,
- Oppenheimer-Dekker A.,
- Bartelings M.M.
- Kouchoukos N.T.,
- Blackstone E.H.,
- Doty D.B.,
- Hanley F.L.,
- Karp R.B.
- Blume E.D.,
- Altmann K.,
- Mayer J.E.,
- Colan S.D.,
- Gauvreau K.,
- Geva T.
- Serraf A.,
- Lacour-Gayet F.,
- Bruniaux J.,
- et al.
- Yacoub M.H.,
- Radley-Smith R.
- Castaneda A.R.,
- Jonas R.A.,
- Mayer J.E.,
- Hanley F.L.
- Li J.,
- Tulloh R.M.,
- Cook A.,
- Schneider M.,
- Ho S.Y.,
- Anderson R.H.
- Bonhoeffer P.,
- Bonnet D.,
- Piechaud J.F.,
- et al.
- Williams W.G.,
- McCrindle B.W.,
- Ashburn D.A.,
- Jonas R.A.,
- Mavroudis C.,
- Blackstone E.H.,
- Congenital Heart Surgeons’s Society
- Hutter P.A.,
- Thomeer B.J.M.,
- Jansen P.,
- et al.
- Katsumata T.,
- Shilfeld A.,
- Westaby S.
- Chaturvedi R.R.,
- Macrae D.,
- Brown K.L.,
- et al.
- Mohammadi S.,
- Serraf A.,
- Belli E.,
- et al.