In conventional repair of the associated malformations

Intermediate Results of the Double-Switch Operations for Atrioventricular Discordance Masahiro Koh, MD, Toshikatsu Yagihara, MD, Hideki Uemura, MD, Koji Kagisaki, MD, Ikuo Hagino, MD, Toru Ishizaka, MD, and Soichiro Kitamura, MD Department of Cardiovascular Surgery, National Cardiovascular Center, Osaka, Japan, and Department of Cardiothoracic Surgery, Royal Brompton Hospital, London, United Kingdom Background. Since 1987, anatomic biventricular repair using the double-switch operations has been our principal choice for patients with atrioventricular discordance. These alternative procedures have the theoretical advantage of using the anatomic left ventricle to support the systemic circulation. Methods. A total of 45 patients underwent the doubleswitch operation. Their ages ranged from 6 months to 21 years. Associated malformations included pulmonary atresia in 27, pulmonary stenosis in 11, and Ebstein s malformation in 5. An atrial switch plus an arterial switch procedure was performed in 7, and an atrial switch plus a Rastelli-type ventriculoarterial switch procedure in 38. Follow-up ranged from 6 months to 15 years. Results. Early mortality was 8.9% (n 4). In the latter half of the series (n 23, since 1994), there was no early death. Six patients died late. Actuarial survival at 5 and 10 years was 83.6% and 77.6%, respectively. Six patients required conduit replacement, and 2 required revision of an intraatrial baffle for pulmonary venous channel obstruction and infection, respectively. Freedom from reoperation was 95.3% at 5 years and 76.2% at 10 years. Freedom from arrhythmia was 88.8% at 5 years and 78.4% at 10 years. The systemic ventricular ejection fraction was 0.568 0.103 at 1 year (n 39), 0.555 0.105 at 5 years (n 17), and 0.539 0.098 at 10 years (n 12). Conclusions. The surgical results of the double-switch operations have been improving. Intermediate follow-up suggests that these alternative procedures are a reasonable option for patients with atrioventricular discordance. (Ann Thorac Surg 2006;81:671 7) 2006 by The Society of Thoracic Surgeons Accepted for publication Aug 15, 2005. Presented at the Forty-first Annual Meeting of The Society of Thoracic Surgeons, Tampa, FL, Jan 24 26, 2005. Address correspondence to Dr Yagihara, Department of Cardiovascular Surgery, National Cardiovascular Center, 5-7-1 Fujishiro-dai, Suita, Osaka, 565-8565, Japan; e-mail address: yagihara@hsp.ncvc.go.jp. In conventional repair of the associated malformations in patients with atrioventricular discordance (AVD), the morphologic right ventricle supplies blood to the systemic circulation. This is a significant problem, because with or without repair of the associated cardiac defects, regurgitation of the tricuspid valve and decreased right ventricular function tend to develop unpredictably with time [1 4]. Since 1987, anatomic biventricular repair using the double-switch operations (DSO) has been our principal choice for patients with AVD [5]. The DSO includes two combinations of procedures. A combined atrial and arterial switch procedure can be performed in those patients in whom the pulmonary valve is normal and left ventricular pressure has been maintained at systemic levels by virtue of a large ventricular septal defect or previous pulmonary artery banding. In those patients with pulmonary valve stenosis or atresia associated with a large ventricular septal defect, a combined atrial switch and a Rastelli-type ventriculoarterial switch procedure can be performed. These alternative procedures result in support of the systemic circulation by the morphologic left ventricle and the mitral valve. The tricuspid valve, even with some morphologic abnormalities, is placed within the low-pressure pulmonary circulation and is likely to function reasonably. This approach should improve the long-term outcome of patients with AVD [6 10]. However, the DSO might result in future complications, such as atrial arrhythmia or venous pathway obstruction related to the surgical maneuver of intraatrial redirection of blood, or obstruction of the conduit. It has not yet been determined whether the potential advantages outweigh the risks. In this study, we evaluated the intermediate results of the DSO to clarify the optimal surgical strategy in this setting. Patients and Methods Patient Selection From September 1987 to September 2004, 62 patients with AVD and two balanced ventricles, including 6 with AVD-like heterotaxy, were referred to us (Fig 1). Patients with a hypoplastic ventricle or straddling atrioventricular valve were excluded. Fifty patients had pulmonary valve stenosis or atresia, and 12 had a normal pulmonary valve. In those 50 with pulmonary valve stenosis or atresia, a ventricular septal defect was considered to be suitable for intraventricular rerouting in 44. Thirty-eight of them 2006 by The Society of Thoracic Surgeons 0003-4975/06/$32.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2005.08.014

672 KOH ET AL Ann Thorac Surg DOUBLE-SWITCH OPERATION 2006;81:671 7 Fig 1. Patient selection flow-chart. (AVD atrioventricular discordance; LVOTR left ventricular outflow tract reconstruction; LVP left ventricular pressure; PA/PS pulmonary atresia or pulmonary valve stenosis; PV pulmonary valve; TR tricuspid regurgitation; TVR tricuspid valve replacement; VSD ventricular septal defect.) underwent the DSO including an atrial switch plus a Rastelli-type procedure. We chose conventional repair including closure of the ventricular septal defect and reconstruction of the left ventricular outflow tract in the remaining 6, because of advanced age in 4 (27, 36, 44, and 47 years), a rather small volume of the left ventricle in 1, and the presence of significant mitral valvular regurgitation in 1. In the other 6 patients with pulmonary valve stenosis or atresia, the ventricular septal defect was unsuitable for rerouting and we chose conventional repair. The defect was small and remote from the aortic orifice in 3, the tricuspid valve attachment disturbed rerouting in 2, and a crisscross heart arrangement was found in 1. In 12 patients with normal pulmonary valves, the left ventricular pressure was maintained at a systemic level in 3 owing to a large amount of shunt associated with the ventricular septal defect. In the other 9 patients, left ventricular pressure was depressed and tricuspid regurgitation was present. We performed pulmonary artery banding in 4; 3 were banded for congestive heart failure, and 1 for left ventricular training, followed by a combined atrial and arterial switch procedure. The remaining 5 were older patients who developed significant tricuspid regurgitation and congestive heart failure (4, 14, 21, 37, and 52 years). We carried out tricuspid valve replacement in these 5 patients [11]. A total of 45 patients underwent the DSO; a combined atrial and arterial switch procedure in 7, and a combined atrial switch and a Rastelli-type ventriculoarterial switch procedure in 38. The age at time of operation ranged from 6 months to 21 years (Fig 2). Morphology Of the 45 patients who underwent the DSO, atrial situs was solitus in 32, inversus in 7, left atrial isomerism in 5, and right atrial isomerism in 1. Regarding ventriculoarterial connections, a solitary outlet of the aorta arising from the morphologic right ventricle with pulmonary atresia was present in 27, a double-outlet right ventricle with pulmonary stenosis in 7, VAD with pulmonary stenosis in 4, and VAD with a normal pulmonary valve in 7. All but 1 patient had a ventricular septal defect. Other associated malformations were Ebstein s anomaly in 5, Wolff-Parkinson-White syndrome in 2, major aortopulmonary collateral arteries in 2, clefted mitral valve in 1, and coarctation of the aorta in 1. Previous Surgery Four patients underwent pulmonary artery banding as stated above. One of them had coarctation of the aorta, and extended direct anastomosis of the aorta concomitant with pulmonary artery banding was required in his early infancy. A total of 36 systemic to pulmonary artery shunts were created in 21 patients. Six with hypoplasia of the pulmonary artery, probably related to the previous shunting operation, required pulmonary artery reconstruction using a heterologous pericardial roll by means of a lateral thoracotomy [12]. Two patients had a major aortopulmonary collateral artery and arborization anomaly. They underwent staged unifocalization of the pulmonary artery [13]. Surgical Procedure We previously described the details of our surgical procedure [5]. Standard aortobicaval cardiopulmonary bypass with moderate hypothermia (28 C) and cold antegrade crystalloid cardioplegia were used. Atrial Switch Procedure The Mustard procedure was performed in 36 patients and the Senning procedure in 9. In the Mustard procedure, a glutaraldehyde-treated heterologous pericardial patch was used for an intraatrial baffle in 32 and an expanded polytetrafluoroethylene patch in 4. To supplement the pulmonary venous chamber in the Mustard procedure, as is occasionally required in the Senning procedure, we sutured an additional patch to the old right atrium. Arterial Switch Procedure Seven patients with a normal pulmonary valve underwent an arterial switch procedure. A standard Lecompte maneu- Fig 2. Age at operation. Gray columns represent patients having a combined atrial switch and a Rastelli-type procedure. White columns represent those having a combined atrial switch and arterial switch procedure.

Ann Thorac Surg KOH ET AL 2006;81:671 7 DOUBLE-SWITCH OPERATION 673 ver was applied in 4. The original Jatene procedure was applied in the remaining 3, who had a relatively side-byside relationship of the great arteries. For reconstruction of the pulmonary artery, an autologous or heterologous pericardial patch was generously inserted in all. Rastelli-Type Ventriculoarterial Switch Procedure Thirty-eight patients with pulmonary valve stenosis or atresia underwent a Rastelli-type procedure for ventriculoarterial switching. The internal conduit was patched to create a channel between the morphologic left ventricle and the aortic orifice. The intraventricular communication was enlarged to create an unobstructed left ventricular outflow tract in 5. In 4 patients, who had relatively small ventricular septal defects, an additional aortopulmonary anastomosis was constructed [5, 14, 15]. The right ventricular outflow tract was reconstructed with an external conduit in 26; a handmade trileaflet conduit made of glutaraldehyde-treated heterologous pericardium (what we call a valved pericardial roll ) was used in all [16]. Nonconduit repair, minimizing the use of prosthetic material, was carried out in 12. A direct anastomosis between the right ventricle and the pulmonary artery was created in 4; an REV technique was applied to facilitate direct anastomosis in 2 [17, 18]. A piece of autologous pericardial patch was inserted posteriorly between the right ventriculotomy and the pulmonary artery in 8, creating continuity of autologous tissue. For the anterior aspect of the right ventricular outflow tract, an expanded polytetrafluoroethylene patch was used in 5, a heterologous pericardium in 3, and an autologous pericardium in 4. In 11 of 12 patients, we avoided the use of a valved conduit; a monocusp made of autologous pericardium or expanded polytetrafluoroethylene sheet was applied. Data Analysis All surviving patients had periodic follow-up at our institution. Patients data were retrospectively reviewed. The follow-up period ranged from 6 months to 15 years, with a mean of 6.4 4.8 years. Routine postoperative catheterization studies were carried out in the majority at approximately 1, 5, and 10 years after the operation. Data are presented as the mean the standard deviation. Cumulative survival estimates were made by the Kaplan-Meier method. The log-rank test was applied for comparison of survival curves. Differences in continuous variables were assessed by Student s t test. Values of p less than 0.05 were considered significant. All statistical tests were conducted with JMP 5.1.1 software (SAS Institute Inc, Cary, NC). Results Survival There were four early deaths, and overall early mortality was 8.9%. All 4 patients underwent a combined atrial switch and a Rastelli-type procedure. The cause of early mortality was mediastinitis in 1, ventricular failure in 1, and bleeding in the airway in 2. Six patients died late. A Fig 3. Survival. (A) Overall survival. (B) Survival differences with era: 1987 1993 (dotted line) versus 1994 2004 (solid line). 3-year-old girl with AVD-like right atrial isomerism developed infective endocarditis 1 month after the Mustard plus a Rastelli-type procedure. She tolerated a reoperation in which the infected intraatrial baffle was exchanged; however, she eventually succumbed to pneumonia 1 year later. A 3-year-old boy who had an atrial plus arterial switch procedure died of coronary artery obstruction 17 months after surgery. The remaining 4 late mortalities were sudden deaths. Two of them had a history of atrial tachyarrhythmia, and 1 had recurrent epileptic seizures. The other patient was doing quite well, and the cause of death was unknown. Overall survival at 5 and 10 years was 83.6% 5.7% and 77.6% 7.8%, respectively (Fig 3). In the latter half of the series (n 23, since 1994), there was no early mortality. Actuarial survival in the latter half was 95.2% 4.7% at 5 years. Reinterventions Six patients required reoperation because of conduit obstruction after the atrial switch plus Rastelli-type procedure. The conduit was replaced with a new, valved pericardial roll in 4, and with anterior patching using autologous tissue in 2 [16, 19]. The other 5 patients who developed conduit obstruction were treated with balloon dilatation. Therefore, 11 of 26 implanted external conduits required reintervention. In 12 patients who had nonconduit repair, none required reintervention for the right ventricular outflow tract (Fig 4). Regarding venous pathway obstruction, 1 patient who had the Senning plus an arterial switch procedure developed progressive pulmonary venous obstruction and required revision of the intraatrial baffle 2 months later. Among 36 patients who had the Mustard procedure, 10

674 KOH ET AL Ann Thorac Surg DOUBLE-SWITCH OPERATION 2006;81:671 7 Fig 5. Freedom from reoperation. after in 17, and 8 to 13 years after in 12; the majority of patients had favorable hemodynamics (Fig 7). One patient after combined atrial and arterial switch procedure experienced progressive left ventricular enlargement after complete atrioventricular block. Fig 4. Freedom from reintervention. (A) Freedom from reintervention for right ventricular outflow tract obstruction (RVOTO): conduit repair (dotted line) versus nonconduit repair (solid line). (B) Freedom from reintervention for venous pathway obstruction: Senning (solid line) versus Mustard (dotted line). required reintervention for venous obstruction caused by calcification of the heterologous pericardium used as an intraatrial baffle (Fig 4). Balloon dilatation was carried out in 9 patients; 8 in the superior vena cava and 1 in the inferior vena cava. One patient underwent revision of the Mustard baffle concomitant with conduit replacement. No patient developed obvious left ventricular outflow tract obstruction after a Rastelli-type ventriculoarterial switch procedure. Among those 7 patients who underwent the arterial switch procedure, none have developed significant aortic regurgitation, despite 4 who required pulmonary artery banding before the DSO. A total of 8 patients have needed reoperation thus far (Fig 5). Arrhythmia Documented arrhythmia occurred in 8 patients (Fig 6). Two suffered from complete atrioventricular block immediately after the surgery, probably caused by closing the ventricular septal defect in one, and enlarging the intraventricular communication in another. Another patient had complete atrioventricular block spontaneously 7 years later. Atrial tachyarrhythmia was documented in 5 patients. One of them had Wolff-Parkinson-White syndrome, and unsuccessful catheter ablation was attempted before the DSO. Atrial tachyarrhythmia recurred 17 months after the surgery. All were controlled with medical treatment; however, 2 died suddenly, probably as a result of arrhythmia. Catheterization Data Routine postoperative cardiac catheterization was carried out 1 year after the operation in 39 patients, 3 to 7 years Functional Status Thirty-four of the surviving patients were in New York Heart Association functional class I at the most recent follow-up. The remaining patient exhibited proteinlosing enteropathy and was in class II. Medications except for antiplatelet agents were not required in 76.2% at 5 years and 83.3% at 10 years. Comment Since 1987, we have used the DSO as our principal surgical procedure for patients with AVD [5]. The DSO includes the following two procedures: a combined atrial switch and arterial switch procedure for patients with a normal pulmonary valve and a combined atrial switch and a Rastelli-type procedure for those with pulmonary atresia or stenosis. Although these alternative procedures have the potential advantage of supporting the systemic circulation with the morphologic left ventricle, they are rather extensive and might have future complications. The optimal surgical strategy is still controversial. Selection of candidates is essential to accomplish the DSO successfully. The DSO combined with an atrial and arterial switch procedure is attractive in patients with a normal pulmonary valve. However, particularly in older patients, the left ventricle placed in the low-pressure Fig 6. Freedom from arrhythmia.

Ann Thorac Surg KOH ET AL 2006;81:671 7 DOUBLE-SWITCH OPERATION 675 Fig 7. Catheterization data. Black circles represent patients having a combined atrial switch and a Rastelli-type procedure. White circles represent patients having a combined atrial switch and an arterial switch procedure. (LVEDV left ventricular end-diastolic volume; LVEF left ventricular ejection fraction.) circulation for a prolonged period is not prepared to support the systemic circulation. Pulmonary artery banding appears to be an option for providing adequate left ventricular training when done at an early age; however, it is not always indicated for older patients. In this series, only 1 young patient (5 years old) with depressed left ventricular pressure underwent pulmonary artery banding with the aim of left ventricular training followed by a combined atrial and arterial switch procedure 4 months later. The other 3 patients required pulmonary artery banding in their early infancy mainly because of congestive heart failure associated with tricuspid regurgitation. Considering the difficulties with left ventricular training, we abandoned the idea of the DSO for the 4 older patients, although they had a reasonable morphology for this alternative procedure. The optimal timing of the DSO is also crucial. For patients with a normal pulmonary valve, we prefer surgery in late infancy to minimize the deterioration of the left ventricular musculature and the deleterious effect of banding on the native pulmonary valve that is to be placed in the aortic position. We have no experience with this procedure in the neonatal period or early infancy. When we encounter a symptomatic neonate or baby in early infancy, we consider pulmonary artery banding primarily, followed by the DSO several months later. We believe that a good indication for the DSO combined with an atrial and a Rastelli-type ventriculoarterial switch procedure was pulmonary stenosis or atresia with a ventricular septal defect that was subaortic in position and suitable for rerouting with an intraventricular baffle from the left ventricle to the aorta. Occasionally, particularly in patients with mild pulmonary stenosis, the ventricular septal defect was unsuitable for rerouting with respect to size or location. This situation might be treated by direct enlargement of the ventricular septal defect or creation of an additional aortopulmonary anastomosis [5, 14, 15]. With these maneuvers, no patient has developed significant left ventricular outflow tract obstruction after ventriculoarterial switching. We previously reported that excessive enlargement of the intraventricular communication might induce ventricular dysfunction, particularly when the ventricular septal defect is a perimembranous inlet type [5, 20]. The presence of such a restrictive ventricular septal defect might be better treated by construction of an additional aortopulmonary anastomosis. An aortic translocation or a truncal switching might be a possible option to establish ventriculoarterial switching in such a situation [21, 22]. Regarding the timing of surgery for patients with pulmonary valve stenosis or atresia, we prefer to perform the DSO somewhat later. Operation in infancy in this setting might be disadvantageous because the ventricular volume is often too small at this stage of life to construct an unobstructed intraventricular channel. However, in older patients, there is another problem of palliative systemic to pulmonary shunting that might cause a volume load on the heart. Moreover, an acquired systemic to pulmonary collateral artery might develop in patients with prolonged cyanosis, which is unequivocally a risk of the operation [23]. It might cause both undesirable volume loading of the ventricle and unexpected airway or intestinal bleeding during the postoperative period. In the early era of this series, 3 patients experienced sudden bleeding in the airway during the postoperative period. These 3 older patients, 8, 9, and 21 years of age, underwent multiple systemic to pulmonary shunting operations as palliative procedures. Sophistication of the surgical strategy should contribute to improvement of early results of the DSO. There is no doubt that better results were also related to technical improvement [5, 20]. The perfusion time was shorter in the latter half (317 95 versus 259 47 minutes; p 0.014), as well as the cross-clamping time (166 32 versus 147 34 minutes; p 0.058). Reinterventions in the intermediate term are common, as prosthetic material is often required to accomplish the DSO. In particular, valved conduits inevitably deteriorate and obstruct within the intermediate term. We initially preferred the use of an external conduit for right ventricular outflow tract reconstruction in a Rastelli-type procedure. In this setting, we always placed the conduit to avoid crossing the midline to prevent compression of the conduit between the heart and the sternum [5]. The conduit tended to be long and winding, and early obstruction was a concern. Moreover, in the reoperation, dissection of the conduit placed far from the midline might be complicated. However, in our experience, conduit life was reasonable in comparison with other kinds of conduit repair, and there was no particular problem with conduit replacement [16]. Since 1992, nonconduit repair using autologous tissue has been our primary technique for infants and young children. This technique might reduce the requirement for reoperation, although right ventricular dysfunction owing to pulmonary regurgitation might adversely influence the outcome [24]. It is encouraging that all patients in this series having nonconduit repair have been free from reintervention. Currently, we reserve the use of conduits for older children

676 KOH ET AL Ann Thorac Surg DOUBLE-SWITCH OPERATION 2006;81:671 7 and adults, particularly for patients with elevated pulmonary vascular resistance or ventricular dysfunction who might be helped by competent valve function [16]. In the early era, we primarily chose the Mustard procedure for atrial switching. This technique is easily applicable to various situations, including apicocaval juxtaposition or anomalous venoatrial connection associated with heterotaxy [5, 25]. However, calcification of the heterologous pericardium, placed in the atrium in the Mustard procedure, occasionally caused venous pathway obstruction. In our series, 10 patients required balloon dilatation of the narrowed venous pathway. We generally decided on early catheter intervention, even when the narrowing was still mild and subclinical, considering that this technique was often ineffective once the baffle had severely calcified. Actually, only 3 of these 10 patients had symptoms of venous pathway obstruction. Currently, we prefer the Senning procedure, in which whole autologous tissue is used for the intraatrial rerouting. We expect this technique might minimize long-term venous pathway obstruction. Aortic regurgitation after an arterial switch procedure is a major concern over the long term, because pulmonary artery banding is common in this setting. In our series, the aortic valves functioned well in all, including 4 patients who required pulmonary artery banding before the DSO. Atrial tachyarrhythmia is the most frequent morbid problem. It is an inherent complication of the atrial switching procedure, although this particular anomaly is often accompanied by atrial rhythm disturbances before surgery [1 3]. An effort to avoid atrial arrhythmia is mandatory, including meticulous design of the baffle and careful attention to the suturing technique. Catheterization data showed that function of the left ventricle placed in the systemic circulation has been reasonably maintained in the majority. We previously reported that the coronary veins of the morphologic left ventricle commonly drain into the pectinated floor of the right atrium directly, not to the coronary sinus [26]. Such independent openings of the small coronary veins might be injured at the time of atrial switching, resulting in impairment of ventricular function during the early postoperative period; however, we could not find any such deleterious effects in the intermediate term. In conclusion, the surgical results of the DSO have been improving. The majority of survivors is doing well with respect to New York Heart Association functional class, and is free from medications. We believe that the DSO is able to minimize deleterious complications associated with AVD. References 1. McGrath LB, Kirklin JW, Blackstone EH, Pacifico AD, Kirklin JK, Bargeron LM Jr. Death and other events after cardiac repair in discordant atrioventricular connection. J Thorac Cardiovasc Surg 1985;90:711 28. 2. Szufladowicz M, Horvath P, de Leval M, Elliott M, Wyse R, Stark J. Intracardiac repair of lesions associated with atrioventricular discordance. Eur J Cardiothorac Surg 1996;10:443 8. 3. Yeh T Jr, Connelly MS, Coles JG, et al. Atrioventricular discordance: results of repair in 127 patients. J Thorac Cardiovasc Surg 1999;117:1190 203. 4. Koh M, Yagihara T, Uemura H, Kagisaki K, Kitamura S. Functional biventricular repair using left ventricle pulmonary artery conduit in patients with discordant atrioventricular connections and pulmonary outflow tract obstruction. Does conduit obstruction maintain tricuspid valve function? Eur J Cardiothorac Surg 2004;26:767 72. 5. Yagihara T, Kishimoto H, Isobe F, et al. Double switch operation in cardiac anomalies with atrioventricular and ventriculoarterial discordance. J Thorac Cardiovasc Surg 1994;107:351 8. 6. Ilbawi MN, DeLeon SY, Backer CL, et al. An alternative approach to the surgical management of physiologically corrected transposition with ventricular septal defect and pulmonary stenosis or atresia. J Thorac Cardiovasc Surg 1990;100:410 5. 7. Yamagishi M, Imai Y, Hoshino S, et al. Anatomic correction of atrioventricular discordance. J Thorac Cardiovasc Surg 1993;105:1067 76. 8. Imai Y, Sawatari K, Hoshino S, Ishihara K, Nakazawa M, Momma K. Ventricular function after anatomic repair in patients with atrioventricular discordance. J Thorac Cardiovasc Surg 1994;107:1272 83. 9. Imamura M, Drummond-Webb JJ, Murphy DJ Jr, et al. Results of the double switch operation in the current era. Ann Thorac Surg 2000;70:100 5. 10. Langley SM, Winlaw DS, Stumper O, et al. Midterm results after restoration of the morphologically left ventricle to the systemic circulation in patients with congenitally corrected transposition of the great arteries. J Thorac Cardiovasc Surg 2003;125:1229 41. 11. Numata S, Uemura H, Yagihara T, Kawahira Y, Yoshikawa Y, Kitamura S. Replacement of the morphologically tricuspid valve in children with discordant atrioventricular connections. J Heart Valve Dis 1999;8:649 54. 12. Uemura H, Yagihara T, Kawashima Y, Yamashita K, Kamiya T. Intrapulmonary reconstruction of pulmonary arteries using a heterologous pericardial roll. Ann Thorac Surg 1995; 59:1464 9. 13. Yagihara T, Yamamoto F, Nishigaki K, et al. Unifocalization for pulmonary atresia with ventricular septal defect and major aortopulmonary collateral arteries. J Thorac Cardiovasc Surg 1996;112:392 402. 14. Kawashima Y, Matsuda H, Taniguchi K, Kobayashi J. Additional aortopulmonary anastomosis for subaortic obstruction in the Rastelli-type repair for the Taussig-Bing malformation. Ann Thorac Surg 1987;44:662 4. 15. Uemura H, Yagihara T, Kawashima Y, Kamiya T. Regurgitation through the morphologically pulmonary valve after additional aortopulmonary anastomosis. J Heart Valve Dis 1995;4:176 81. 16. Koh M, Yagihara T, Uemura H, et al. Long-term outcome of right ventricular outflow tract reconstruction using a handmade tri-leaflet conduit. Eur J Cardiothorac Surg 2005;27: 807 14. 17. Lecompte Y, Neveux JY, Leca F, et al. Reconstruction of the pulmonary outflow tract without prosthetic conduit. J Thorac Cardiovasc Surg 1982;84:727 33. 18. Nomura F, Kadoba K, Ichikawa H, Akedo H, Matsuda H. Direct anastomosis of pulmonary artery-to-right ventricular outflow for correction of tetralogy of Fallot with pulmonary atresia. Ann Thorac Surg 1996;62:1199 202. 19. Danielson GK, Downing TP, Schaff HV, Puga FJ, DiDonato RM, Ritter DG. Replacement of obstructed extracardiac conduits with autogenous tissue reconstructions. J Thorac Cardiovasc Surg 1987;93:555 9. 20. Uemura H, Yagihara T. Anatomic biventricular repair by intraventricular and intraatrial rerouting in patients with discordant atrioventricular connections. In: Redington AN,

Ann Thorac Surg KOH ET AL 2006;81:671 7 DOUBLE-SWITCH OPERATION 677 Brawn WJ, Deanfield JE, Anderson RH, eds. The right heart in congenital heart disease. London: Greenwich Medical Media, 1998:237 42. 21. Nikaidoh H. Aortic translocation and biventricular outflow tract reconstruction. J Thorac Cardiovasc Surg 1984;88:365 72. 22. Yamagishi M, Shuntoh K, Matsushita T, et al. Half-turned truncal switch operation for complete transposition of the great arteries with ventricular septal defect and pulmonary stenosis. J Thorac Cardiovasc Surg 2003;125:966 8. 23. Ichikawa H, Yagihara T, Kishimoto H, et al. Extent of aortopulmonary collateral blood flow as a risk factor for Fontan operations. Ann Thorac Surg 1995;59:433 7. 24. Lacour-Gayet F, Serraf A, Komiya T, et al. Truncus arteriosus repair: influence of techniques of right ventricular outflow tract reconstruction. J Thorac Cardiovasc Surg 1996;111:849 56. 25. Uemura H, Yagihara T, Kawahira Y, Yoshikawa Y. Anatomic biventricular repair by intraatrial and intraventricular rerouting in patients with left isomerism. Cardiol Young 2001;11:12 6. 26. Uemura H, Ho SY, Anderson RH, et al. Surgical anatomy of the coronary circulation in hearts with discordant atrioventricular connections. Eur J Cardiothorac Surg 1996;10:194 200. DISCUSSION DR CHRISTO I. TCHERVENKOV (Montreal, Quebec, Canada): I congratulate you on your large and excellent experience with the double-switch operation. I am particularly interested in the patients with dextrocardia, to give us a little bit of an insight into the two technical aspects. One is the Mustard/Senning part of the operation and the other one the placement of the conduit. In our small experience with patients with dextrocardia is that the performance of the Mustard operation is quite difficult due to the small size of the right atrium. In these patients we have performed a modified Senning operation by using the in situ pericardial well technique. My last question is, based on your experience, are you ready to offer a double-switch operation in a neonate with corrected transposition and intact ventricular septum, since left ventricular preparedness is not an issue in the first few weeks of life. DR KOH: Your first question is regarding the atrial switching in the setting of dextrocardia. In the earlier era, we preferred the Mustard procedure in the patient with apicocaval juxtaposition. Those patients have a rather small atrial wall, as you mentioned, and we think the Mustard procedure is easily applicable. The Senning procedure is rather difficult in this setting. Recently we preferred the Senning procedure in order to minimize future venous pathway obstruction. In the Senning procedure, generally we augment the functional left atrium with a Gore-Tex patch. And what is your next question? DR TCHERVENKOV: Is the placement of the conduit in patients with dextrocardia to the right or the left of the aorta? DR KOH: We always place the conduit not to cross the midline, on the opposite side of the caval veins. We intend to avoid compression between the heart mass and the sternum as well as turbulence at the anastomosis. This design generally makes the conduit long and winding; however, our data showed conduit life was satisfactory. DR TCHERVENKOV: And my last question, what would you recommend for the 3-week-old neonate with corrected transposition and intact ventricular septum? Would you recommend the double-switch procedure? DR CARL L. BACKER (Chicago, Illinois): He doesn t have enough time to answer that one, Christo. DR TCHERVENKOV: All I need is a yes or a no. DR KOH: No. Initially we will place banding, and 6 or 5 months later we will carry out the double-switch operation. DR GLEN S. VAN ARSDELL (Toronto, Ontario, Canada): This is a fascinating paper with a beautiful, detailed analysis, but I think it doesn t completely address a very important question that we, as a society, still have to deal with. First off, I think there s a lot of interest in the VSD (ventricular septal defect). How do you objectively evaluate your VSD to insure that it s of adequate size? I believe your group has earlier shown that if you enlarge the VSD the ventricular function is worse. So how do you deal with the VSD? And are you following that? What type of gradients do you have? Have you had to reoperate for VSDs? DR KOH: In the patient with pulmonary atresia or pulmonary stenosis, VSD is always located at the subaortic position and is large enough. When pulmonary stenosis is mild, VSD may be small. In such cases, we carried out VSD enlargement or Damus-Kaye-Stanzel anastomosis, what we call an additional aortopulmonary anastomosis, to create an unobstructed ventricular channel. We think excessive enlargement of VSD may impair ventricular function due to damage of the septal branches. One patient in our early series developed ventricular dysfunction after VSD enlargement, as you mentioned; however, no patient among the other 4 who required VSD enlargement in this study had similar results. For the patients with small VSD, additional aortopulmonary anastomosis is an effective alternative to create an unobstructed channel. There is no case that developed LVOTO (left ventricular outflow tract obstruction) in our series. DR VAN ARSDELL: And maybe I should just ask one rhetorical question. Yesterday we talked about Rastellis, and the outcome for Rastellis, in the best-case scenario, is era-independent and 60% survival at 20 years. I would submit, as a rhetorical question, our Fontan survivals at 20 years are probably about 85% or 90%. Should we, as a society, be exploring doing Fontans on these patients instead of pursing a biventricular repair? DR BACKER: That sounds like a good topic for next year s debates, and we ll put you as one of the protagonists. DR VAN ARSDELL: I don t have the answer. DR BACKER: Well, that s why we re going to have you talk.