Outcomes After the Palliative Arterial Switch Operation in Neonates With Single-Ventricle Anatomy

Outcomes After the Palliative Arterial Switch Operation in Neonates With Single-Ventricle Anatomy Jeffrey S. Heinle, MD, Kathleen E. Carberry, MPH, RN, E. Dean McKenzie, MD, Aimee Liou, MD, Paul A. Katigbak, BS, and Charles D. Fraser, Jr, MD Michael E. DeBakey Department of Surgery, Division of Congenital Heart Surgery, Baylor College of Medicine; Department of Congenital Heart Surgery, Surgical Outcomes Center, Texas Children s Hospital; and Department of Pediatrics, the Lillie Frank Abercrombie Section of Pediatric Cardiology, Baylor College of Medicine, Houston, Texas Background. Newborns with single-ventricle anatomy, transposition of the great arteries, and systemic outflow obstruction are challenging patients most often managed with a Norwood or Damus-Kaye-Stansel (DKS) procedure. The palliative arterial switch operation (paso) offers the theoretical advantage of avoiding a systemicto-pulmonary artery shunt physiology and posterior entrapment of the left pulmonary artery while aligning the single left ventricle with the posterior semilunar valve. Limited outcome data exist for the paso. The purpose of this study was to examine the clinical course of patients after this operation. Methods. We conducted a retrospective review of all neonates undergoing a paso at our institution from July 1995 to June 2011. Results. Fourteen patients underwent paso at a median age of 7 days (2 16 days). Primary diagnoses included double-inlet left ventricle (6 patients [43%]), tricuspid atresia (TA) (4 patients [29%]), and other (4 patients [29%]). Concomitant procedures at initial operation included aortic arch reconstruction (13 patients [93%]), pulmonary artery banding (6 patients [43%]), and placement of a systemic-to-pulmonary artery shunt (1 patient [7%]). Median intensive care unit stay and hospital length of stay were 11 days (4 60 days) and 21 days (9 84), respectively. There were no deaths. All patients had advanced to a bidirectional cavopulmonary shunt (BCPS). Four patients required intervention between the paso and BCPS procedures: systemic-to-pulmonary artery shunt (3 patients [21%]) and pulmonary artery banding revision (1 patient [7%]). Eleven patients had Fontan completion; the remaining 3 patients are candidates for this procedure. At the time of the Fontan operation, 7 (64%) patients required pulmonary artery augmentation. All patients were alive at last follow-up (median, 5 years [3 months 14 years]). At last echocardiographic follow-up (median, 5 years [3 months 10 years]), all patients had normal ventricular function, trivial to mild neoaortic insufficiency, and no left ventricular outflow obstruction. Conclusions. The paso should be considered in neonates with single ventricle, transposition, and systemic outflow obstruction because it affords a favorable anatomic arrangement for long-term palliation, with excellent survival and preserved ventricular function. (Ann Thorac Surg 2013;95:212 9) 2013 by The Society of Thoracic Surgeons Surgical palliation of newborns with single-ventricle anatomy is achieved through staged reconstruction with the ultimate goal of a Fontan circulation. Patients with single ventricle and transposed great vessels and systemic outflow obstruction represent a challenging subgroup. This substrate is typically seen in patients with double-inlet left ventricle (DILV) or TA with transposed great vessels (Fig 1A). Systemic outflow obstruction occurs at the level of the bulboventricular foramen (BVF) or subaortic area or at the aortic valve (Fig 1B). This is frequently associated with aortic arch hypoplasia, coarctation, or interruption. Furthermore, there is typically Accepted for publication Sept 11, 2012. Presented at the Fifty-eighth Annual Meeting of the Southern Thoracic Surgical Association, San Antonio, TX, Nov 9 12, 2011. Address correspondence to Dr Heinle, Department of Congenital Heart Surgery, Texas Children s Hospital, 6621 Fannin St, MC 19345H, Houston, TX 77030; e-mail: jsheinle@texaschildrens.org. unrestricted pulmonary blood flow through a large posteriorly located semilunar valve. Significant systemic outflow obstruction precludes a successful Fontan operation and must be addressed. Management decisions made in the early neonatal period affect the patient s ultimate candidacy for Fontan palliation. Surgical options for initial newborn palliation include enlargement of the obstructed BVF [1], pulmonary artery banding with arch reconstruction [2 6], or more commonly modifications of the DKS or Norwood procedure [4, 7 10]. The palliative arterial switch operation (paso) [11 13] is an alternative approach with the theoretical advantages of aligning the single systemic ventricle with the unobstructed posterior semilunar valve while avoiding systemic-to-pulmonary artery shunt physiology and posterior entrapment of the left pulmonary artery. Limited outcome data exist for the paso applied in this setting. The purpose of this study was to examine the clinical course and outcomes of patients with single- 2013 by The Society of Thoracic Surgeons 0003-4975/$36.00 Published by Elsevier Inc http://dx.doi.org/10.1016/j.athoracsur.2012.09.028

Ann Thorac Surg HEINLE ET AL 2013;95:212 9 PALLIATIVE ARTERIAL SWITCH OPERATION 213 Fig 1. Double-inlet left ventricle with transposed great arteries and hypoplastic subaortic outflow chamber. (A) Typically there is large size discrepancy between posterior dilatated main pulmonary artery and ascending aorta. Associated aortic arch obstruction is common. (B) Systemic outflow obstruction can occur at level of bulboventricular foramen, subaortic area, and/or aortic valve. Pulmonary blood flow through large posterior pulmonary valve is unrestricted. ventricle anatomy who underwent palliation with the arterial switch operation (ASO) in the neonatal period at our institution. Patients and Methods Patient Population The Texas Children s Hospital Congenital Heart Surgery database was queried for all neonates (age 30 days) with single-ventricle anatomy who underwent an ASO. The medical records of these patients were retrospectively reviewed after approval by the Baylor College of Medicine Institutional Review Board. Given the retrospective nature of the study, permission was given for a waiver of consent. All patients undergoing the paso operation from July 1995 until June 2011 (N 14) constitute the study cohort for this analysis. Records were reviewed for demographic, perioperative, and follow-up information. All pre-fontan catheterization results were reviewed by a single interventional cardiologist. Patient Demographics and Presentation Fourteen neonates underwent a paso at a median age of 7 days (2 16 days) and a median weight of 3.3 kg (2.6 4.2 kg). All patients were boys. Primary diagnoses included DILV in 6 (43%) patients, TA in 4 (29%) patients, and other in 4 (29%) patients. Of the 4 patients with diagnosis of other, 1 had congenitally corrected transposition of the great arteries with right atrioventricular valve atresia; 1 had congenitally corrected transposition of the great arteries with a ventricular septal defect, double-outlet right ventricle with right ventricular hypoplasia, and severe subaortic stenosis; 1 had d-transposition of the great arteries, a right ventricular dominant unbalanced atrioventricular canal, and subaortic obstruction; and 1 had doubleoutlet right ventricle, right atrioventricular valve atresia, and severe subaortic stenosis. All patients had transposed great vessels (9 d-transpositions, 5 l-transpositions). Thirteen patients had aortic arch obstruction: 10 (71%) had aortic arch hypoplasia or coarctation and 3 (21%) had complete arch interruption. Twelve (86%) patients presented with definite subaortic stenosis and/or a restrictive BVF. We believe the other 2 patients had the substrate for subsequent subaortic obstruction. Surgical Technique All operations were performed through a median sternotomy with cardiopulmonary bypass support. Arch reconstruction was facilitated using antegrade cerebral perfusion in 12 of 13 patients. After cardioplegic arrest, the ascending aorta was transected several millimeters distal to the aortic valve commissures (Fig 2A). In 12 of 13 patients, direct arch anastomosis was achieved without the use of patch material (Fig 2B). Coronary transfer was accomplished in a standard manner. A LeCompte maneuver was performed in all patients. Six patients required patch augmentation of the ascending aorta to account for the size discrepancy of the neoaortic root and the ascending aorta (Fig 2C). The pulmonary sinuses were reconstructed with autologous pericardial patches (Fig 2D). After weaning from bypass, an assessment was made of relative pulmonary blood flow. In 6 (43%) patients, a pulmonary artery band was placed at the initial operation to control excessive pulmonary blood flow, and in 1 (7%) patient a right subclavian-topulmonary artery shunt was constructed to supplement pulmonary blood flow. Statistical Analysis Standard descriptive statistical analyses were performed. Means, medians, standard deviations, and ranges were used to describe continuous variables. Frequencies and percentages were used for categorical data. All analyses

214 HEINLE ET AL Ann Thorac Surg PALLIATIVE ARTERIAL SWITCH OPERATION 2013;95:212 9 Fig 2. Palliative arterial switch operation. (A) Transection of ascending aorta and main pulmonary artery. Division of descending thoracic aorta in preparation for arch advancement. (B) Autologous tissue repair of arch obstruction avoiding need for patch material. (C) LeCompte maneuver bringing pulmonary artery anterior to reconstructed aorta and avoiding left pulmonary artery compression. Coronary artery translocation to neoaortic root. (D) Autologous pericardial patch reconstruction of neopulmonary sinuses. Neopulmonary artery anastomosis may need to be moved onto branch pulmonary artery to avoid compression of translocated coronary artery. were conducted with Microsoft Excel (Microsoft Corp, Redmond, Washington). Results Operative Data Median cardiopulmonary bypass, aortic cross clamp, and circulatory arrest times were 266 (189 327), 175 (11 226), and 6 (0 44) minutes, respectively. Median antegrade cerebral perfusion time was 53 minutes (20 113 minutes). Concomitant procedures included aortic arch reconstruction in 13 (93%) patients and atrial septectomy in 10 (71%) patients. One (7%) patient underwent neoaortic valve repair because of pulmonary valve regurgitation noted on preoperative echocardiography. A pulmonary artery band was placed in 6 (43%) patients and a systemic-topulmonary artery shunt was used in 1 (7%) patient at the time of the paso. Hospital Course Median intensive care unit and hospital lengths of stay were 11 days (4 60 days) and 21 days (9 84 days), respectively. There were no hospital deaths. Four (29%) patients underwent unplanned reoperations before hospital discharge. Two (14%) patients required placement of a systemic-to-pulmonary artery shunt for desaturation at 3 and 28 days, respectively, after the paso. A third patient required loosening of the pulmonary artery band at 23 days after the paso. The fourth patient required an early bidirectional cavopulmonary shunt (BCPS) and also underwent arch repair for recurrent arch obstruction at 77 days after the paso. One (7%) patient had postoperative bleeding requiring reoperation. Follow-Up There have been no late deaths. All 14 patients have advanced to a BCPS. Median age and weight at BCPS was 5 months (2 12 months) and 6.4 kg (5.2 8.3 kg), respectively. One patient required placement of a systemic-topulmonary artery shunt after hospital discharge and before BCPS at 52 days after the initial operation. Eleven patients (79%) have proceeded to Fontan completion, and the remaining 3 (27%) patients are candidates for such a procedure. Pre-Fontan cardiac catheterization demonstrated that all patients were favorable candidates for a Fontan operation, with low pulmonary pressures and transpulmonary gradient, low ventricular filling pressures, excellent cardiac output, and adequate development of the pulmonary arteries (Table 1; Fig 3). Ten of the 11 patients underwent total cavopulmonary connection with an extracardiac conduit, and 2 (14%) patients had fenestrations created. The median age and weight at Fontan completion were 3.6 years (2.4 7 years) and 14.7 kg (13.4 18.2 kg), respectively. At the time of Fontan completion, 7 (64%) patients required patch aug- Table 1. Pre-Fontan Cardiac Catheterization Data Variable Results Age a 3 y (2 6 y) Weight a 13.9 kg (13.1 16.9 kg) Pulmonary artery pressure (mean) 10.6 2.8 mm Hg LAP (mean) 6.9 2.7 mm Hg TPG 3.7 0.9 mm Hg PVR (WU) 1.5 0.5 EDP 7.3 2.3 mm Hg Qs 4.2 1.2 L/min/m 2 MVo 2 68.7% 5.6% McGoon ratio 1.7 0.4 Nakata index 263.4 68.8 mm 2 /BSA a Median with range, all others mean/standard deviation. BSA body surface area; EDP end-diastolic pressure; LAP left atrial pressure; MVo 2 mixed venous oxygen saturation; PVR (WU) pulmonary arteriolar resistance (Woods units); TPG transpulmonary gradient.

Ann Thorac Surg HEINLE ET AL 2013;95:212 9 PALLIATIVE ARTERIAL SWITCH OPERATION 215 Fig 3. Pulmonary angiogram obtained after palliative arterial switch at time of pre-fontan catheterization showing good development of branch pulmonary arteries. with arch repair [2 6], and modifications of the DKS or Norwood procedure [4, 7 10]. Enlargement of the BVF can be technically quite difficult in the young infant or neonate [1], risking injury to the conduction system or aortic valve. The presence of an obstructed aortic arch and performing the operation in the neonatal period has been shown to negatively impact survival in this patient group [1], suggesting that this may be a less than optimal approach in neonates. Although pulmonary artery banding has been advocated as initial palliation in the neonatal period to protect the pulmonary vasculature in patients with functional single ventricle and unobstructed pulmonary blood flow, there is a recognized high incidence of subaortic obstruction developing in single-ventricle hearts with transposition and a rudimentary subaortic chamber after pulmonary artery banding [14, 19]. This is particularly true in those patients with aortic arch obstruction [22]. Subaortic obstruction may complicate eventual Fontan palliation by inducing ventricular hypertrophy [23], leading to subendocardial ischemia and diastolic dysfunction and has led some groups to abandon pulmonary artery banding as palliation when subaortic stenosis is present or anticipated [22, 24]. mentation of the pulmonary artery. All patients were alive at last follow-up (median, 5 years [3 months 14 years]). At last echocardiographic follow-up (median, 5 years [3 months 10 years]) all patients had normal ventricular function, mild or a lesser amount of neoaortic insufficiency, and no left ventricular outflow obstruction. Comment Successful palliation of the patient with single ventricle and systemic outflow obstruction requires a management strategy that leads to a durable Fontan circulation. Decisions made in the neonatal period ultimately affect candidacy for eventual Fontan completion, and in this group of patients must address the outflow obstruction and unrestricted pulmonary blood flow at the time of initial palliation. Criteria for a successful Fontan circulation are well known. The importance of preserving ventricular function (both systolic and diastolic) has been realized since the mid-1980s. Subaortic obstruction [14 17] and diastolic dysfunction because of ventricular hypertrophy [18 20] have been recognized as risk factors for poor outcome after a Fontan operation. In a recent series from Birmingham, England, Hosein et al [21] identified 2 risk factors in the current era that influence outcome in a Fontan procedure: impaired ventricular function and elevated pulmonary artery pressures, again stressing the need for early protection of the pulmonary vascular bed and relief of the outflow obstruction in this group. Management Options for Single Ventricle and Systemic Outflow Obstruction Various management options to address these issues in this group of patients with single ventricle have included enlargement of the BVF [1], pulmonary artery banding Fig 4. Palliative arterial switch. After arterial switch, systemic ventricle is aligned directly with large semilunar valve and with systemic circulation in an undistorted orientation. Systemic outflow obstruction is relieved. Subaortic obstruction is converted to subpulmonary obstruction restricting pulmonary blood flow and resulting in a banded circulation.

216 HEINLE ET AL Ann Thorac Surg PALLIATIVE ARTERIAL SWITCH OPERATION 2013;95:212 9 Fig 5. Ventriculogram obtained after palliative arterial switch at time of pre-fontan catheterization showing alignment of systemic ventricle with systemic circulation. (A) Anteroposterior view. (B) Lateral view. The DKS procedure [25 27], originally described for biventricular repair of transposition of the great arteries, has been extended for use in single-ventricle palliation. Variations of this technique [4, 8 10] and the Norwood [7] operation have been described that address the subaortic obstruction in the subset of patients with single ventricle and transposed great arteries. These techniques were extended to neonates and infants as primary palliation [28 31] and have allowed for successful staged palliation to the Fontan operation [28, 29, 32]. Palliative Arterial Switch Option The ASO as a palliative procedure was first described by Freedom and colleagues [33] in Toronto and subsequently promoted by Karl and colleagues [11] at the Royal Children s Hospital in Melbourne, Australia. This group was the first to report the use of the AOS as initial palliation in 6 neonates with univentricular hearts (DILV or TA), transposition of the great arteries, and a small BVF. Five of the 6 patients also required arch repair as part of the initial palliation. There was 1 operative death (17%) and 1 late death in this series, which compared favorably to the poor results seen in newborns who underwent palliation with a DKS procedure. Subsequent small series from Lacour-Gayet and colleagues [12] and Ceresnak and associates [13] suggested favorable early results, with the challenge of balancing pulmonary blood flow being notable in both reports. In the latter series, 6 of 9 original patients were successfully transitioned to a completion Fontan operation. Critics of this approach cite the more technically demanding operation, particularly with the major size discrepancy between the aortic and pulmonary roots [2, 34, 35], and the fact that pulmonary blood flow may be unpredictable [3, 28, 35]. Palliative Arterial Switch Operation: Texas Children s Hospital Experience The current report represents the largest series of neonates with single-ventricle anatomy undergoing the paso. In this study, there was no operative, interstage, or late mortality. All patients were able to be successfully palliated to a BCPS, and all have either proceeded to a Fontan operation or are considered to be good candidates for such a procedure. These results clearly compare favorably to contemporary published series of either Norwood or DKS palliation [36]. The goal of the initial palliation strategy is to optimize the anatomy and physiology to allow for durable palliation with a Fontan operation. In the setting of DILV or TA with transposition of the great arteries and a hypoplastic subaortic outlet chamber, there is typically an enormous size discrepancy between the aorta and the main pulmonary artery. In this setting, the pulmonary valve is typically the most suitable semilunar valve and the subpulmonary area is unobstructed. By performing an arterial switch and aligning the systemic ventricle with the systemic outflow tract (Figs 4, 5), the subaortic obstruction is effectively relieved. There has been no evidence of recurrent obstruction. This thereby avoids a stimulus for ventricular hypertrophy and the potential for diastolic dysfunction. Pre-Fontan hemodynamics (Table 1) were excellent in all patients in our series. In the setting of straight anterior-posterior great vessels or when the ascending aorta is leftward of the main pulmonary artery, creating a DKS or Norwood type of reconstruction pulls the main pulmonary artery to the patient s left. This may distort the pulmonary valve and result in iatrogenic pulmonary valve insufficiency (neoaortic insufficiency). This anatomy sets up nicely for the arterial switch and leaves the orientation of the pulmonary root intact, minimizing the potential for distortion of the root. We have not seen significant neoaortic insufficiency in any patient in this study. By switching the subaortic obstruction to subpulmonary obstruction, a banded circulation is created in the neonatal period (Fig 4). This early protection of the pulmonary vascular bed is evidenced by low pulmonary artery pressures, transpulmonary gradient, and pulmonary vascular resistance at the time of pre-fontan catheterization. In the setting of the large posteriorly located main pulmonary artery, if a DKS or Norwood connection is created, the posterior space for the left pulmonary artery is compromised. The LeCompte maneuver at the time of the arterial switch brings the branch pulmonary arteries anterior to the neoaorta, optimizing pulmonary artery growth potential (Fig 3). Furthermore this allows for easier access to the pulmonary arteries at subsequent

Ann Thorac Surg HEINLE ET AL 2013;95:212 9 PALLIATIVE ARTERIAL SWITCH OPERATION 217 stages of palliation and facilitates construction of an extracardiac Fontan conduit. The importance of an effective long-term strategy for palliation at the outset cannot be overemphasized. We believe that the paos is an attractive alternative for the initial palliation of the specific morphologic substrate in patients with single ventricle and transposition and recognized or potential systemic outflow obstruction (particularly in the setting of aortic arch obstruction). The dominant semilunar valve is aligned with the systemic circulation in an undistorted orientation. Typically, a banded circulation is achievable, which is more easily managed than a shunted circulation. The branch pulmonary arteries are brought anterior to the neoaorta and thus can be readily accessed if the need for repair arises. In most cases, an arch reconstruction using all autologous tissue is achievable. There is no need to attempt to enlarge the BVF and risk complete heart block or recurrent obstruction. The resulting anatomy and physiology sets these patients up extremely well for a subsequent Fontan procedure. We remain encouraged by the observations of no operative mortality or late attrition, with all patients considered to be good candidates for Fontan operations. We are also encouraged by the lack of interstage mortality, an issue that has plagued patients undergoing Norwood palliation. Limitations This study is limited by its retrospective nature. There is inherent selection bias in the way the patients were chosen (ie, those patients who had been selected for the paso at the time of operation). No real direct comparison can be made regarding this approach with other management options. Follow-up of these patients is limited to midterm follow-up. The long-term fate of the neoaortic valve, the translocated coronary arteries, development of the pulmonary arteries, and durable success of the Fontan has not yet been determined. Conclusions The paso should be considered in neonates with single ventricle, transposition, and systemic outflow obstruction because it affords favorable anatomic arrangement for long-term palliation with excellent survival and preserved ventricular function. Results from this study support an early aggressive approach to avoid potential late complications precluding successful palliation. The neonatal palliative arterial switch should be viewed as an additional important tool in the surgical armamentarium in the management of these complicated patients. References 1. Jahangiri M, Shinebourne EA, Ross DB, Anderson RH, Lincoln C. Long-term results of relief of subaortic stenosis in univentricular atrioventricular connection with discordant ventriculoarterial connections. Ann Thorac Surg 2001;71: 907 10. 2. Webber SA, LeBlanc JG, Keeton BR, et al. Pulmonary artery banding is not contraindicated in double inlet left ventricle with transposition and aortic arch obstruction. Eur J Cardiothorac Surg 1995;9:515 20. 3. Odim JN, Laks H, Drinkwater DC, Jr, George BL, Yun J, Salem M, Allada V. Staged surgical approach to neonates with aortic obstruction and single-ventricle physiology. Ann Thorac Surg 1999;68:962 7; discussion 968. 4. Clarke AJ, Kasahara S, Andrews DR, et al. Mid-term results for double inlet left ventricle and similar morphologies: timing of Damus-Kaye-Stansel. Ann Thorac Surg 2004;78: 650 7; discussion 657. 5. Rodefeld MD, Ruzmetov M, Schamberger MS, Girod DA, Turrentine MW, Brown JW. Staged surgical repair of functional single ventricle in infants with unobstructed pulmonary blood flow. Eur J Cardiothorac Surg 2005;27:949 55. 6. Kajihara N, Asou T, Takeda Y, et al. Staged surgical approach in neonates with a functionally single ventricle and arch obstruction: pulmonary artery banding and aortic arch reconstruction before placement of a bidirectional cavopulmonary shunt in infants. Pediatr Cardiol 2010;31:33 9. 7. Jonas RA, Castaneda AR, Lang P. Single ventricle (single- or double-inlet) complicated by subaortic stenosis: surgical options in infancy. Ann Thorac Surg 1985;39:361 6. 8. Lamberti JJ, Mainwaring RD, Waldman JD, et al. The Damus-Fontan procedure. Ann Thorac Surg 1991;52:676 9. 9. Waldman JD, Lamberti JJ, George L, et al. Experience with Damus procedure. Circulation 1988;78:III32 9. 10. Fiore AC, Rodefeld M, Vijay P, et al. Subaortic obstruction in univentricular heart: results using the double barrel Damus-Kaye Stansel operation. Eur J Cardiothorac Surg 2009;35:141 6. 11. Karl TR, Watterson KG, Sano S, Mee RB. Operations for subaortic stenosis in univentricular hearts. Ann Thorac Surg 1991;52:420 7; discussion 427 8. 12. Lacour-Gayet F, Serraf A, Fermont L, et al. Early palliation of univentricular hearts with subaortic stenosis and ventriculoarterial discordance. The arterial switch option. J Thorac Cardiovasc Surg 1992;104:1238 45. 13. Ceresnak SR, Quaegebeur JM, Pass RH, Hordof AJ, Liberman L. The palliative arterial switch procedure for single ventricles: are these patients suitable Fontan candidates? Ann Thorac Surg 2008;86:583 7. 14. Freedom RM, Williams WG, Trusler GA, Rowe RD. Penkoske pulmonary artery. Surgical palliation of subaortic stenosis in the univentricular heart. J Thorac Cardiovasc Surg 1984;87:767 81. 15. DeLeon SY, Ilbawi MN, Idriss FS, et al. Fontan type operation for complex lesions. Surgical considerations to improve survival. J Thorac Cardiovasc Surg 1986;92:1029 37. 16. Coles JG, Kielmanowicz S, Freedom RM, et al. Surgical experience with the modified Fontan procedure. Circulation 1987;76:III61 6. 17. Myers JL, Waldhausen JA, Weber HS, et al. A reconsideration of risk factors for the Fontan operation. Ann Surg 1990;211:738 43; discussion 744. 18. Kirklin JK, Blackstone EH, Kirklin JW, Pacifico AD, Bargeron LM, Jr. The Fontan operation. Ventricular hypertrophy, age, and date of operation as risk factors. J Thorac Cardiovasc Surg 1986;92:1049 64. 19. Freedom RM, Benson LN, Smallhorn JF, Williams WG, Trusler GA, Rowe RD. Subaortic stenosis, the univentricular heart, and banding of the pulmonary artery: an analysis of the courses of 43 patients with univentricular heart palliated by pulmonary artery banding. Circulation 1986;73:758 64. 20. Seliem M, Muster AJ, Paul MH, Benson DW, Jr. Relation between preoperative left ventricular muscle mass and outcome of the Fontan procedure in patients with tricuspid atresia. J Am Coll Cardiol 1989;14:750 5. 21. Hosein RB, Clarke AJ, McGuirk SP, et al. Factors influencing early and late outcome following the Fontan procedure in the current era. The Two Commandments? Eur J Cardiothorac Surg 2007;31:344 52; discussion 353. 22. Franklin RC, Sullivan ID, Anderson RH, Shinebourne EA, Deanfield JE. Is banding of the pulmonary trunk obsolete for infants with tricuspid atresia and double inlet ventricle with

218 HEINLE ET AL Ann Thorac Surg PALLIATIVE ARTERIAL SWITCH OPERATION 2013;95:212 9 a discordant ventriculoarterial connection? Role of aortic arch obstruction and subaortic stenosis. J Am Coll Cardiol 1990;16:1455 64. 23. Freed MD, Rosenthal A, Plauth WH, Jr, Nadas AS. Development of subaortic stenosis after pulmonary artery banding. Circulation 1973;48:III7 10. 24. Rothman A, Lang P, Lock JE, Jonas RA, Mayer JE, Castaneda AR. Surgical management of subaortic obstruction in single left ventricle and tricuspid atresia. J Am Coll Cardiol 1987; 10:421 6. 25. Damus PS, Thomson NB, Jr, McLoughlin TG. Arterial repair without coronary relocation for complete transposition of the great vessels with ventricular septal defect. Report of a case. J Thorac Cardiovasc Surg 1982;83:316 8. 26. Kaye MP. Anatomic correction of transposition of great arteries. Mayo Clin Proc 1975;50:638 40. 27. Stansel HC, Jr. A new operation for d-loop transposition of the great vessels. Ann Thorac Surg 1975;19:565 7. 28. Lotto AA, Hosein R, Jones TJ, Barron DJ, Brawn WJ. Outcome of the Norwood procedure in the setting of transposition of the great arteries and functional single left ventricle. Eur J Cardiothorac Surg 2009;35:149 55; discussion 155. 29. Mosca RS, Hennein HA, Kulik TJ, et al. Modified Norwood operation for single left ventricle and ventriculoarterial discordance: an improved surgical technique. Ann Thorac Surg 1997;64:1126 32. 30. Tchervenkov CI, Beland MJ, Latter DA, Dobell AR. Norwood operation for univentricular heart with subaortic stenosis in the neonate. Ann Thorac Surg 1990;50:822 5. 31. van Son JA, Reddy VM, Haas GS, Hanley FL. Modified surgical techniques for relief of aortic obstruction in [S,L,L] hearts with rudimentary right ventricle and restrictive bulboventricular foramen. J Thorac Cardiovasc Surg 1995;110: 909 15. 32. McElhinney DB, Reddy VM, Silverman NH, Hanley FL. Modified Damus-Kaye-Stansel procedure for single ventricle, subaortic stenosis, and arch obstruction in neonates and infants: midterm results and techniques for avoiding circulatory arrest. J Thorac Cardiovasc Surg 1997;114:718 25; discussion 725 6. 33. Freedom RM, Williams WG, Fowler RS, Trusler GA, Rowe RD. Tricuspid atresia, transposition of the great arteries, and banded pulmonary artery. Repair by arterial switch, coronary artery reimplantation, and right atrioventricular valved conduit. J Thorac Cardiovasc Surg 1980;80:621 8. 34. Mee RB. Neonatal palliative switch for complex univentricular heart. Semin Thorac Cardiovasc Surg 1994;6:39 40. 35. Serraf A, Conte S, Lacour-Gayet F, et al. Systemic obstruction in univentricular hearts: surgical options for neonates. Ann Thorac Surg 1995;60:970 6; discussion 976 7. 36. Ohye RG, Sleeper LA, Mahony L, et al. Comparison of shunt types in the Norwood procedure for single-ventricle lesions. N Engl J Med 2010;362:1980 92. DISCUSSION DR JAMES A. QUINTESSENZA (St. Petersburg, FL): Congratulations, Dr Hienle, on your presentation and thank you for sending the manuscript in advance. Dr Heinle and associates from Texas presented their experience with 14 patients undergoing a palliative arterial switch procedure as an initial procedure for patients with functional single ventricle, transposed great vessels, aortic stenosis, and subaortic obstruction. This concept was initially reported by Freedom in 1980 to relieve outlet obstruction in an older Fontan patient. Subsequent reports by Karl and Lacour-Gayet described the approach in neonates in an effort to create an unobstructed ventricular outlet, preserve semilunar valve function, control pulmonary blood flow through a naturally restrictive outflow tract, and include a pulmonary artery band or shunt when needed. Additionally, a LeCompte maneuver was added to minimize left pulmonary artery narrowing, which was felt to be a problem in this series. The results in this series are excellent, with no mortality, and all patients were suitable for a Fontan at a mean of 6 years follow-up, with 11 of 14 having had a completed Fontan. I have a few questions. This procedure is not frequently performed. Can you expand on the specific selection criteria you used in choosing this option? This must be a very select group from your institution. I have a second question. Our goal in all of this should be to determine the best initial palliation to make the most ideal Fontan candidate. Do you think a trial comparing palliative arterial switch to Damus-Kay-Stansel (DKS) patients in matched groups would be valuable to really tease out the differences in both the short- and long-term outcomes regarding the optimal approach? For example, it might be of interest to study issues like aortic insufficiency where there may be competing risk factors for AI in the different subgroups. In the DKS patients, surgical reconstruction may cause semilunar valve distortion and insufficiency compared to the arterial switch patients, where there is arterial trunk size mismatch and enlargement, and valvar insufficiency may develop. Overall, I enjoyed the paper and commend you on excellent results in this complex group of patients. DR HEINLE: Thank you very much for your comments, appreciate it. I agree that these are not frequently performed operations, and I believe that s for a variety of reasons. One is that the lesions we are talking about are not that common. In our series, and in others, this is a relatively small patient group. As for the selection criteria, we do not have absolute selection criteria. I do think that much of the decision is following intraoperative inspection. Our concern when we are looking at these patients is just what you said, what is the best way to palliate these kids for the long term? So how are the pulmonary arteries going to develop, how are you going to align the outflow tract? In our opinion, you align the outflow tract very well with this operation. We have been concerned that in some patients who undergo the modified DKS, particularly when you have the large posterior root, that you get entrapment of the left pulmonary artery and poor development of the LPA. It is hard to address this problem. In the palliative arterial switch, by pulling the arteries anterior with a LeCompte maneuver, you now have access to the branch pulmonary arteries if they require augmentation. I think the anatomy probably does not set up well if you have side-by-side great arteries or if you have difficult coronary anatomy arrangement. Most of our patients had a Yacoub A or normal coronary pattern; only 1 patient had type B. I am sure that has to do with some selection bias. As far as a trial goes, I think that it is probably always best, but it would be very difficult to get adequate patient numbers. Our study is over more than 10 years, and it s only 14 patients. It may be hard to have a large enough cohort to randomize and to have enough power to identify differences. DR JAMES S. TWEDDELL (Milwaukee, WI): That was an excellent presentation. I have always considered the palliative switch to be a procedure looking for an indication, and I am not

Ann Thorac Surg HEINLE ET AL 2013;95:212 9 PALLIATIVE ARTERIAL SWITCH OPERATION 219 entirely certain that I have been persuaded otherwise by your presentation here today. It seems to combine all of the difficulties of an arterial switch (prolonged cross-clamp time, potential for coronary compromise and pulmonary arteries that have been subjected to the LeCompte maneuver) with problems of singleventricle palliation related to adjusting pulmonary blood flow. Fifty percent of your patients did require some adjustment of the pulmonary blood flow at the initial hospitalization; 4 required banding, 1 required a shunt; in addition some patients required a shunt in the interstage period. It seems to me that the goal for neonatal palliation of these single-ventricle lesions ought to be an operation that doesn t require any additional intervention until you get to the second stage. I am unconvinced that the alignment of the outflow tract is any different or superior to that achieved with a Norwood operation. Were there patients with double-inlet left ventricle and arch hypoplasia or tricuspid atresia with transposition that had a more conventional Norwood style operation, and have you considered comparing the outcomes of those patients to the ones you presented here today? Specifically have you looked at the pulmonary artery anatomy and the potential for coronary insufficiency? Have you had any patients in this group that have ventricular dysfunction, because with a very small aortic root, it can be challenging to excise and transfer buttons into a very large pulmonary artery. Thank you for an interesting presentation. DR HEINLE: I appreciate the comments. I think all excellent points and certainly things we have thought about. In my mind, the ultimate palliation is not to get them through the neonatal period but to set them up for a long-term palliation, and I don t think you want to complicate the neonatal period. Really, in our opinion, it is a complex operation, but these patients, for us, are easier to manage than some of the Norwoods, because they have mostly banded physiology when you are coming out of the operating room. So you have a long bypass run, you have a long ischemic period, but you have banded physiology, not shunt physiology, coming out of the operating room, and I think that simplifies their early management. There is no question that you cannot just send these kids home and bring them back for their Glenn. The bulboventricular foramen, as we know, even if you band these kids, is going to get small. You need to watch for that, and it is true that more than about half of these patients will need some adjustment of their pulmonary blood flow, but those operations are fairly well tolerated, also, because you are out of the early postoperative period, and we don t put in a shunt that we intend to get them to Fontan. This is a shunt that is going to get them to Glenn that we still do relatively early in infancy, so even if you get shunted, we are not waiting for you to be a year or 2; so they do get some volume load, but it is not a long-standing volume load. And your other question about the coronaries, I do think it can be difficult when there is a large size mismatch. At least 1 surgeon in our group prefers to put the aorta back together before translocating the coronaries so that you get a better idea of the translocation site. However, we have not had any coronary insufficiency or had to address coronary problems in these patients. DR EDWARD L. BOVE: (Ann Arbor, MI): You sort of alluded to it, but I was a little surprised not to hear about any follow-up on the coronaries. Obviously these patients must have had caths, but I don t think you mentioned that unless I missed it. DR HEINLE: No, you are correct. DR BOVE: In addition, to the best of my knowledge, I am unaware of any long-term follow-up of the arterial switch that doesn t have some incidence of coronary problems, often asymptomatic, so I think you have added an element of concern with this operation which you wouldn t have otherwise. Finally, it is probably a bit duplicitous to say that these patients have a long operation and a long clamp time, but they are banded. The reason it is a long operation and a long clamp time is because you are doing an arterial switch. It seems to be circuitous logic. I probably would side with Jim, as I think it is an operation looking for an indication. DR HEINLE: I think the coronary follow-up is a good point. Certainly we can go back and look at the angiograms. All I can really comment on is the ventricular function that no patient had depressed function. That does not certainly address the actual coronary anatomy. We will certainly need to follow this as these patients get farther out. Maybe I misstated my earlier point. I think that these patients are easier to manage because they have a banded physiology and not a shunt physiology. I think that makes a big difference in taking care of these. It is a small group, but they all made it through the initial palliation and there was no interim attrition. You can make the argument that now that we are doing Sano s, the Norwood physiology is more stable. However, if you look at Bill Brawn s group, this is 1 group where they do not do a Sano. They still use a systemic-to-pa shunt because of the orientation of the ventricles. So again, compared to a modified Norwood, we feel this physiology is easier to manage in the postoperative period.