The successful application of the Fontan operation for

Transcription

1 Modified Norwood Operation for Single Left Ventricle and Ventriculoarterial Discordance: An Improved Surgical Technique Ralph S. Mosca, MD, Hani A. Hennein, MD, Thomas J. Kulik, MD, Dennis C. Crowley, MD, Erik C. Michelfelder, MD, Achi Ludomirsky, MD, and Edward L. Bove, MD Section of Thoracic Surgery, Department of Surgery, and Division of Pediatric Cardiology, Department of Pediatrics, The University of Michigan School of Medicine, Ann Arbor, Michigan Background. Patients with univentricular hearts and ventriculoarterial discordance with potentially obstructed systemic blood flow continue to pose difficult management problems. The goals of neonatal palliative operations are to control pulmonary blood flow while avoiding pulmonary artery distortion, to relieve systemic outflow tract obstruction, and to avoid heart block. Methods. Between January 1987 and December 1996, 38 patients with either tricuspid atresia or a double-inlet left ventricle and ventriculoarterial discordance underwent a modified Norwood procedure. Their mean age was 15 days, and their mean weight was 3.4 kg. Aortic arch anomalies were present in 92% of the patients. Morbidity and mortality statistics, intraoperative data, and postoperative echocardiograms were reviewed. Results. There were 3 early deaths (7.8%) and 5 late deaths (13.1%). The actuarial survival rates at 1 month, 1 The successful application of the Fontan operation for virtually all forms of single-ventricle lesions has highlighted the importance of avoiding the many risk factors now known to reduce the likelihood of an optimum long-term outcome. Although it is difficult to quantify, diastolic function, or compliance, of the systemic ventricle has been found to be an important predictor of this outcome. Hearts characterized by an outlet chamber supporting the aortic valve usually are dependent on an entirely muscular outlet, the bulboventricular foramen (BVF), for systemic blood flow (Fig 1). The tendency for this outlet to become increasingly restricted with time, resulting in progressive obstruction to systemic blood flow, ventricular hypertrophy, and decreased ventricular compliance, is well documented [1]. Because the usual presentation of infants with this condition also includes unrestricted pulmonary blood flow, early efforts to protect the pulmonary vascular bed by reducing pulmonary flow and pressure with pulmonary artery banding generally have been used. However, this procedure may Presented at the Thirty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Feb 3 5, Address reprint requests to Dr Mosca, Pediatric Cardiovascular Surgery, Michigan Congenital Heart Center, F7830 Mott Hospital, Box 0223, 1500 E Medical Center Dr, Ann Arbor, MI ( rmosca@umich.edu). year, and 5 years were 89%, 82%, and 71%, respectively. Follow-up was complete in all children at a mean interval of 30 9 months. None of the patients had significant neoaortic valve insufficiency, and 1 patient required therapy for residual aortic arch obstruction. Nine patients (30% of the survivors) have undergone the hemi-fontan procedure, and 18 patients (60%) successfully have undergone the Fontan procedure. Conclusions. In this patient population, we recommend the modified Norwood procedure as the neonatal palliative treatment of choice. It can be performed with acceptable early morbidity and mortality, and it improves suitability for the Fontan procedure. It reliably relieves all levels of systemic outflow tract obstruction, controls pulmonary blood flow, and avoids heart block. (Ann Thorac Surg 1997;64: ) 1997 by The Society of Thoracic Surgeons accelerate the progression of subaortic stenosis by stimulating spontaneous reduction in BVF size and increasing myocardial hypertrophy [2]. To avoid the development of subaortic stenosis and progressive ventricular hypertrophy, we have not performed pulmonary artery banding, but instead have used a modification of the Norwood procedure in all patients with single-ventricle lesions who have actual or potential systemic outflow tract obstruction. This report reviews our experience in the treatment of neonates and infants with ventriculoarterial discordance and either a doubleinlet left ventricle or tricuspid atresia. The early and mid-term outcome with respect to mortality, relief of systemic outflow tract obstruction, and suitability for the Fontan procedure, are reviewed. The use of a modified surgical approach is described in detail. Material and Methods Patients The records of all 38 patients with a single left ventricle and ventriculoarterial discordance who underwent surgical palliation using a modified Norwood procedure between January 1987 and December 1996 were reviewed. The diagnosis was established by two-dimensional Doppler echocardiography in all patients. There were 10 patients 1997 by The Society of Thoracic Surgeons /97/$17.00 Published by Elsevier Science Inc PII S (97)

2 Ann Thorac Surg MOSCA ET AL 1997;64: MODIFIED NORWOOD PROCEDURE 1127 with tricuspid atresia and transposition of the great arteries and 28 patients with a double-inlet left ventricle and transposition of the great arteries. Additional pertinent information recorded included the relation of the great arteries, the size of the pulmonary arteries, the size of the BVF, and the presence of aortic arch hypoplasia, coarctation, or interruption. Echocardiographic studies on all 38 patients were available for review. The initial and most recent studies were analyzed for the presence of neoaortic valve insufficiency, ventricular outflow tract obstruction, and obstruction along the neoaortic arch. The size of the BVF also was measured in the initial and latest studies. Image quality limited serial assessment of the BVF size to 20 patients. The dimensions of the BVF were measured in orthogonal planes from either the parasternal (long and short axis) or the subcostal (sagittal and coronal axis) window. The cross-sectional area of the BVF (BVFA) was calculated by using the formula for a standard ellipse: BVFA 3.14 (A/2) (B/2), where A equals the longitudinal dimension and B equals the transverse dimension of the ellipse. The BVFA was indexed to the body surface area by dividing the former by the latter. The mean rate of growth of the BVF was expressed as the mean of the individual changes in the BVFA index between the initial and follow-up studies divided by the time between the studies. The median patient age at the time of operation was 15 days (range, 2 days to 13 months) and the median weight was 3.3 kg (range, 2.1 to 9.8 kg). Only 9 patients were older than 1 month. Twenty patients (54%) were receiving mechanical ventilation before operation, 11 (30%) were receiving inotropic support, and 20 (53%) were receiving an infusion of intravenous prostaglandin to maintain ductal patency. Intraoperative data collected included pulmonary blood flow type, shunt size, cardiopulmonary bypass time, cross-clamp time, and circulatory arrest time. Complete postoperative Doppler echocardiography was performed in each patient, with particular attention paid to assessing the size of the BVF and the presence of neoaortic insufficiency or residual aortic arch obstruction. Postoperative systemic oxygen saturation levels were recorded just before hospital discharge. Fig 1. Single left ventricle with ventriculoarterial discordance. The aorta is positioned anterior and slightly to the right of the pulmonary artery. The systemic flow to the hypoplastic ascending aorta (arrow) is by virtue of a muscular, potentially restrictive bulboventricular foramen. Surgical Technique After routine midline sternotomy and partial thymectomy, the ascending aorta, arch vessels, proximal descending thoracic aorta, and pulmonary arteries were mobilized. Cardiopulmonary bypass was established with cannulation of the main pulmonary trunk in patients with a patent ductus arteriosus. The branch pulmonary arteries were occluded, and cooling in preparation for circulatory arrest was begun, according to standard techniques. When the ductus arteriosus was not patent, cannulation of the ascending aorta was used. After at least 20 minutes of cooling to a nasopharyngeal temperature of 20 C, the circulation was arrested, the perfusion cannulas were removed, and the heart was arrested with cold blood cardioplegia. If it was restrictive, the atrial septum was excised completely, working through the right atrial appendage cannulation site. Both great vessels were divided a few millimeters distal to the sinotubular ridge, and the distal pulmonary artery bifurcation was closed using a polytetrafluoroethylene patch (Gore-Tex; W.L. Gore and Associates, Flagstaff, AZ). The ductal tissue was excised completely and the ascending aorta was opened along its inner curvature to a point 10 to 15 mm distal to the ductal insertion (Fig 2). A patch was configured from a large-sized pulmonary allograft and was used to reconstruct the entire ascending aorta, transverse arch, and proximal descending aorta. Fig 2. Single left ventricle with ventriculoarterial discordance. The atrial septum has been excised. Both great vessels are divided just beyond the sinotubular ridge. The ductal tissue is excised completely and the aortic arch is reconstructed with an allograft patch.

3 1128 MOSCA ET AL Ann Thorac Surg MODIFIED NORWOOD PROCEDURE 1997;64: Table 1. Intraoperative Data a Parameter Time (min) Range (min) Cardiopulmonary bypass Cross-clamp Circulatory arrest a Data are expressed as means standard deviation. Fig 3. Single left ventricle with ventriculoarterial discordance. The augmented distal aorta is anastomosed end-to-end to the divided main pulmonary artery, incorporating the proximal ascending aorta. The augmented distal aorta was anastomosed in an end-to-end fashion to the proximal divided main pulmonary artery, incorporating the proximal aorta in the anastomosis (Fig 3). In this fashion, unobstructed systemic blood flow is provided from the left ventricle and coronary blood flow is provided from the proximal ascending aorta (Fig 4). Pulmonary blood flow was provided by a modified innominate artery to the pulmonary artery through a polytetrafluoroethylene shunt in neonatal patients (Fig 5) and through a bidirectional superior cavopulmonary anastomosis in older patients. In general, a 3.5-mm shunt was used in patients who weighed 3 to 4 kg and a 4-mm shunt was used in those who weighed more than 4 kg. The optimum initial systemic arterial oxygen saturation is approximately 75% to 80% (Po 2 28 to 35 mm Hg). When the systemic arterial oxygen saturation is higher, the inspired oxygen level is decreased to the level of room air (inspired oxygen fraction 0.21) if necessary. The addition of carbon dioxide was not necessary in the treatment of any of the patients in this series. When the systemic arterial oxygen saturation is too low ( 65%), it is important to determine whether the cause is elevated pulmonary vascular resistance or a technical problem with the shunt itself. Temporarily elevated pulmonary vascular resistance will improve with time or perhaps the addition of inhaled nitric oxide, whereas a technical problem with the shunt will require revision. Results Thirty-eight patients underwent palliative procedures. Follow-up was complete at a mean interval of 30 9 months. Cardiopulmonary bypass data, expressed as the mean times the standard deviations, were a cardiopulmonary bypass time of minutes, a cross-clamp time of minutes, and a circulatory arrest time of minutes (Table 1). Before operation, the mean Fig 4. Right anterior oblique angiogram of a patient with a doubleinlet left ventricle and transposition of the great arteries before a hemi-fontan procedure. An unobstructed systemic outflow exists through the pulmonary valve (neoaorta) and reconstructed aortic arch. Fig 5. Single left ventricle with ventriculoarterial discordance. The completed repair. Pulmonary blood flow is provided by an innominate-to-pulmonary artery shunt. The pulmonary artery bifurcation is closed with a patch.

4 Ann Thorac Surg MOSCA ET AL 1997;64: MODIFIED NORWOOD PROCEDURE 1129 Table 2. Postoperative Pulmonary Insufficiency a Level of Severity No. of Patients a As determined by transthoracic echocardiography at a mean follow-up of 26 4 months. Fig 6. Actuarial survival curve. Survival rates at 1 month, 1 year, and 5 years were 89%, 82%, and 71%, respectively. initial BVF area index was 1.73 cm 2 /m 2. Thirty-five patients (92%) had associated aortic arch hypoplasia, coarctation, or interruption requiring repair. There were 3 early deaths and 5 late deaths. Actuarial survival rates at 1 month, 1 year, and 5 years were 89%, 82%, and 71%, respectively (Fig 6). Concerning the early deaths, 1 patient who was critically ill before operation, with metabolic acidosis resulting from systemic hypoperfusion and pulmonary overcirculation, died on postoperative day 20 of fungal sepsis and multisystem organ failure. The second and third patients initially did well after the procedure but died suddenly after unexplained bradycardic events. Of the 5 late deaths, 1 patient died at 2 months of age after an episode of Klebsiella sepsis, the second died suddenly at 3.5 months (autopsy revealed muscular changes within the peripheral pulmonary arteries consistent with vascular obstructive disease), the third died during a catheterization for evaluation for the Fontan procedure, the fourth died after the Fontan procedure as a result of low cardiac output, and the fifth died at home at 2 months of unexplained causes. Complications included three postoperative wound infections, two significant pleural effusions requiring drainage, and three pericardial effusions, one of which required subxiphoid drainage. The mean hospital stay for the first-stage palliative procedure was days. The mean postoperative systemic oxygen saturation, measured just before hospital discharge, was 79% 8%. Currently, 9 patients (24%) have undergone successful conversion to the second-stage palliative procedure, and 18 patients (47%) successfully have completed all three stages. Echocardiographic hemodynamic data were complete in all the patients at a mean follow-up of 26 4 months. All patients had no or trace neoaortic (pulmonary valve) insufficiency (Table 2), 5 patients had mild residual gradients (peak instantaneous pressure gradient, 15 to 30 mm Hg), and 1 patient had moderate residual arch obstruction (peak instantaneous pressure gradient, 30 mm Hg) (Table 3). The significant residual arch obstruction (48 mm Hg) in the last patient, confirmed at cardiac catheterization, was corrected at the second-stage (hemi- Fontan) procedure, with good results. Complete data regarding the BVF were available in 20 patients. Although the absolute BVF increased in approximately 50% of the patients, when indexed to body surface area, there was an overall decrease with time (Table 4). Comment Hearts with a univentricular atrioventricular connection are characterized by a wide variation in morphology. The relative volumes of pulmonary and systemic blood flow are determined in large part by the type of ventriculoarterial connection (concordant or discordant) and the presence and severity of subaortic or subpulmonary obstruction. In those hearts with a discordant ventriculoarterial connection (eg, either a double-inlet left ventricle or tricuspid atresia, with transposition of the great arteries), systemic blood flow must traverse a BVF, and associated obstruction at the level of the aortic arch frequently is present. The degree of subaortic obstruction has been shown to have a significant effect on survival [1]. The BVF may be restricted at birth, or it may narrow insidiously over time. Matitiau and colleagues [3] examined the progression of obstruction of the BVF in 28 infants with a double-inlet left ventricle or tricuspid atresia with transposition of the great arteries, and found that the most important determinant of late BVF obstruction was its initial size. In that report, the mean initial BVF area index was significantly smaller in those patients with associated arch obstruction. Although the BVF appeared to grow, the growth did not parallel somatic growth and the BVF tended to become obstructed with time [3]. Patients with univentricular hearts whose systemic flow is dependent on a BVF and subaortic outflow cham- Table 3. Postoperative Aortic Arch Obstruction a Level of Severity (mm Hg) No. of Patients a Reported as the peak instantaneous pressure gradient as determined by transthoracic echocardiography.

5 1130 MOSCA ET AL Ann Thorac Surg MODIFIED NORWOOD PROCEDURE 1997;64: Table 4. Rate of Change in Size of Bulboventricular Foramen a Patient No. BVFA 1 (cm 2 ) BVFAI 1 (cm 2 /m 2 ) BVFA 2 (cm 2 ) BVFAI 2 (cm 2 /m 2 ) dbvfai (cm 2 m 2 mo 1 ) Mean a Although the absolute bulboventricular foramen size did increase in approximately 50% of patients, this was not predictable, and when indexed to body surface area, there was an overall decrease with time. BVFA 1 preoperative area; BVFAI 1 indexed preoperative area; BVFA 2 postoperative area; BVFAI 2 indexed postoperative area; dbvfai change in indexed area over time. ber continue to pose difficult management problems [1, 4]. Survival before definitive repair in a group of patients with a double-inlet left ventricle was reviewed by Franklin and associates [5]. In that report, the prognosis was determined predominantly by the specific morphology. Patients with naturally occurring pulmonary outflow tract stenosis and restricted pulmonary blood flow had predicted survival rates of 96% at 1 year and 79% at 10 years. However, similar patients with unobstructed pulmonary blood flow had lower predicted survival rates: 79% at 1 year and 60% at 10 years. In addition, those patients with unobstructed pulmonary blood flow associated with obstruction to systemic blood flow at any level had an even more guarded prognosis: 36% at 1 year and 11% at 10 years. This study underscores the important influence of systemic outflow tract obstruction on the overall outcome of this group of patients. The influence of palliative operations on survival was reviewed by Franklin and co-workers [6]. Although surgical intervention increased the relative risk of early death, all three procedures analyzed (systemic-to-pulmonary artery shunt, banding of the pulmonary artery, and pulmonary artery banding with associated coarctation repair) resulted in improved survival at 6 months when compared with unoperated patients. The subset of patients who required pulmonary artery banding with or without aortic arch repair had much higher perioperative risk and no real long-term benefit beyond 6 months, however. The Fontan operation increasingly is being applied to patients with all forms of univentricular hearts. Several preoperative factors have been recognized as important predictors of long-term survival, including significant ventricular systolic or diastolic dysfunction. Although it is difficult to quantify, diastolic dysfunction has emerged as particularly important for early and late outcome [7]. Staged reconstruction for many types of univentricular hearts has been adopted by most centers in an attempt to eliminate risk factors in anticipation of the Fontan procedure. Specifically, removing ventricular volume and pressure overload as early in life as possible is likely to improve outcome and protect long-term ventricular function. Those hearts whose systemic blood flow is dependent on the BVF are at constant risk of decreasing ventricular compliance from progressive subaortic stenosis, ventricular hypertrophy, and subendocardial ischemia [8]. Thus, it is imperative that optimum methods of palliation are chosen in infancy to avoid these known risk factors. Early pulmonary vascular obstructive disease will develop in infants with unrestricted pulmonary blood flow without surgical intervention. In addition, the resulting ventricular volume overload may result in ventricular dilatation and impaired ventricular function. Although pulmonary artery banding may provide effective control of pulmonary blood flow and pressure, it may lead to rapid diminution in the size of the BVF as a result of the immediate reduction in ventricular diastolic volume as well as subsequent myocardial hypertrophy [9 12]. Pulmonary artery banding also risks distortion of the branch right or left pulmonary arteries and the pulmonary valve [13]. Direct surgical enlargement of the BVF may be complicated by heart block, impaired ventricular func-

6 Ann Thorac Surg MOSCA ET AL 1997;64: MODIFIED NORWOOD PROCEDURE 1131 tion, recurrent or residual obstruction, and ventricular aneurysm formation [14]. This approach is limited further by the fact that the obstruction to systemic outflow may occur within the cavity of the outflow chamber as well as from adjacent atrioventricular valve tissue, in addition to the BVF itself [15, 16]. These problems are not addressed by enlargement of the BVF and may be difficult to relieve. Banding of the main pulmonary artery and creation of an aortopulmonary window proximal to the band was devised in an attempt to provide unobstructed aortic outflow through the pulmonary artery [17]. This procedure, however, results in increased pressure and volume work of the systemic ventricle and often inexact control of the pulmonary blood flow. Division of the main pulmonary artery and end-to-side anastomosis with the ascending aorta (Damus-Kaye-Stansel procedure) has been used with variable success, but it may lead to distortion of the great vessels and neoaortic insufficiency. This approach does not address the aortic arch obstruction that often is associated with these lesions [18 21]. The arterial switch procedure has been used, trading subaortic obstruction for subpulmonary obstruction [22]. Pulmonary blood flow is unpredictable, however, and a subsequent systemic-to-pulmonary artery shunt or a pulmonary artery band may be required. Finally, left ventricular apical-to-aortic conduits have been used, but these are limited by conduit failure and the inevitable need for a technically complex reoperation [23]. The method described in the current report affords an entirely extracardiac means of providing an unobstructed systemic outflow tract regardless of the anatomic nature of the subaortic obstruction and the fate of the BVF, and it is similar to the bivalve approach described by Lamberti and colleagues [24]. Although the BVF appears to grow in a significant number of patients, it frequently does not parallel somatic growth and is likely to become restricted over time. In a study reviewing the course of hospital survivors of the Fontan procedure by Finta and associates [25], systemic ventricular outflow tract obstruction developed in a significant number of patients (12%), even after definitive repair. The subgroup of patients who relied on a ventricular septal defect or BVF for systemic outflow experienced obstruction significantly more frequently (21%). In that study, even small outflow gradients progressed over time, emphasizing that the development of subaortic stenosis is a constant risk. These findings confirm those of earlier studies [3], and they serve to emphasize that predicting which patients will experience obstruction of the BVF is unreliable, at best. The technique described in this report also stresses the use of small systemic-to-pulmonary shunts to control pulmonary blood flow and systemic hypoperfusion. Restricting pulmonary blood flow in a predictable manner has improved survival among neonates undergoing arch reconstruction in association with a shunt-dependent pulmonary circulation [26]. However, increased early stability after neonatal palliation must be weighed against the frequent need for an earlier hemi-fontan procedure as the pulmonary blood flow becomes inadequate as a result of the patient s growth and increased activity. However, use of the hemi-fontan procedure early in infancy has been well tolerated. In a series of 85 infants undergoing the hemi-fontan operation within the first 6 months of life at the University of Michigan, the hospital mortality rate was 6%, supporting the notion that this approach is suitable even for these complex patients [27]. The ultimate goal of neonatal palliative operations is to provide the optimum anatomic and physiologic conditions for a Fontan procedure. In view of the fact that 92% of the patients in this study required aortic arch repair in addition to control of pulmonary blood flow, they otherwise would have required banding of the pulmonary trunk and arch repair as the initial palliative procedure. However, using the method we have described, all the survivors have undergone or are suitable candidates for a Fontan operation (79%), which stands in striking contrast to the 1 (8%) of 12 patients who were alive and suitable candidates for definitive repair in the series by Franklin and colleagues [6] of patients treated with initial banding of the pulmonary artery and repair of the aortic arch. These results support the continued use of this approach for patients with univentricular hearts and actual or potential obstruction to systemic outflow through the BVF. The technique described is preferable to other methods because it controls pulmonary blood flow while avoiding pulmonary artery distortion, obviates the need for incisions in the ventricle and the possible development of complete heart block, and permanently relieves all levels of systemic outflow tract obstruction. In addition, this procedure avoids extracardiac conduits and is technically easier to perform than the traditional Damus- Kaye-Stansel procedure combined with aortic arch reconstruction. This approach helps to eliminate the development of ventricular hypertrophy from early in life and should provide for optimum ventricular function for the Fontan procedure. References 1. Somerville J, Becu L, Ross D. Common ventricle with acquired subaortic stenosis. Am J Cardiol 1974;34: Barber G, Hagler DJ, Edwards WD, et al. Surgical repair of the univentricular heart (double-inlet left ventricle) with obstructed anterior outlet chamber. J Am Coll Cardiol 1984; 4: Matitiau A, Geva T, Colan SD, et al. Bulboventricular foramen size in infants with double-inlet left ventricle or tricuspid atresia with transposed great arteries: influence on initial palliative operation and rate of growth. J Am Coll Cardiol 1992;19: Moodie DS, Ritter DG, Tajik AJ, et al. Long term follow-up in the unoperated univentricular heart. Am J Cardiol 1984;53: Franklin CG, Spiegelhalter DJ, Anderson RH, et al. Doubleinlet ventricle presenting in infancy I. Survival without definitive repair. J Thorac Cardiovasc Surg 1991;101: Franklin RCG, Spiegelhalter DJ, Anderson RH, et al. Double-inlet ventricle presenting in infancy II. Results of palliative operations. J Thorac Cardiovasc Surg 1991;101: Kirklin JK, Blackstone EH, Kirklin JW, et al. The Fontan operation: ventricular hypertrophy, age and date of operation as risk factor. J Thorac Cardiovasc Surg 1986;92: Freedom RM, Sondheimer H, Dische R, et al. Development

7 1132 MOSCA ET AL Ann Thorac Surg MODIFIED NORWOOD PROCEDURE 1997;64: of subaortic stenosis after pulmonary artery banding for common ventricle. Am J Cardiol 1977;39: Rychik J, Jacobs ML, Norwood WI. Acute changes in left ventricular geometry after volume reduction operation. Ann Thorac Surg 1995;60: Rychik J, Donofrio MT, Jacobs ML, Norwood WI. Early changes in ventricular septal defect size and ventricular geometry in the single left ventricle after volume-unloading surgery. J Am Coll Cardiol 1995;26: Jensen RA, Williams RG, Laks J, et al. Usefulness of banding of the pulmonary trunk with single ventricle physiology at risk for subaortic obstruction. Am J Cardiol 1996;77: 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: Freedom RM, Benson LN, Smallhorn JF, et al. Subaortic stenosis, the univentricular heart and banding of the pulmonary artery: an analysis of the course of 43 patients with univentricular heart palliated by pulmonary artery banding. Circulation 1986;73: O Leary PU, Driscoll DJ, Connor AR, et al. Subaortic obstruction in hearts with a univentricular connection to a dominant left ventricle and an anterior subaortic outlet chamber: results of a staged approach. J Thorac Cardiovasc Surg 1992;106: Cheung HC, Lincoln C, Anderson RH, et al. Options for surgical repair in hearts with univentricular atrioventricular connection and subaortic stenosis. J Thorac Cardiovasc Surg 1990;100: Freedom RM, Dische MR, Rowe RD. Pathologic anatomy of subaortic stenosis and atresia in the first year of life. Am J Cardiol 1977;39: Neches WH, Park SC, Lenox CC, et al. Tricuspid atresia with transposition of the great arteries and closing ventricular septal defect: successful palliation by banding of the pulmonary artery and creation of a pulmonary window. J Thorac Cardiovasc Surg 1973;65: Damus PS. Letter to the editor. Ann Thorac Surg 1975;20: Kaye MP. Anatomic correction for transposition of the great arteries. Mayo Clin Proc 1975;50: Lin AE, Laks H, Barber G, et al. Subaortic obstruction in complex congenital heart disease: management by proximal pulmonary artery to ascending aorta end to side anastomosis. J Am Coll Cardiol 1986;7: Stansel HC. A new operation for d-loop transposition of the great arteries. Ann Thorac Surg 1975;19: Lacour-Gayet F, Serraf A, Fermont L, et al. Early palliation of univentricular hearts with subaortic stenosis and ventriculoarterial discordance. J Thorac Cardiovasc Surg 1992;109: Frommelt DC, Rocchini AD, Bove EL. Natural history of apical left ventricle to aortic conduits in pediatric patients. Circulation 1991;89(Suppl 3): Lamberti JJ, Mainwaring RD, Waldman JD, et al. The Damus-Fontan procedure. Ann Thorac Surg 1991;52: Finta KM, Beekman RH, Lupinetti FM, Bove EL. Ventricular outflow obstruction progresses after the Fontan operation. Ann Thorac Surg 1994;58: Iannettoni MD, Bove EL, Mosca RS, et al. Improving results with first stage palliation for hypoplastic left heart syndrome. J Thorac Cardiovasc Surg 1994;107: Bradley SM, Mosca RM, Hennen HA, et al. Bidirectional superior cavopulmonary connection in young infants. Circulation 1996;94(Suppl 2):5 11. DISCUSSION DR BABULAL SETHIA (Birmingham, England): I enjoyed your paper very much, and we have had similar experience with these types of patients, but I have two questions. You said that you cannulate the pulmonary artery electively. However, in most of these patients, the ascending aorta is of good size and can be cannulated to undertake the operation, reducing some of the circulatory arrest time. Do you have any comments on that? DR MOSCA: The vast majority of these patients have small ascending aortas and concomitant aortic arch hypoplasia or coarctation. Thus, it can be difficult to cannulate the ascending aorta directly. Cannulation of the main pulmonary artery in a sinus just above the pulmonary valve is simple, safe, and reliable. Manipulation of the arterial cannulas is reduced greatly during mobilization of the head vessels and proximal descending thoracic aorta. In our experience, the circulatory arrest times have averaged approximately 50 minutes, which is well tolerated as long as uniform cooling has been performed for more than 20 minutes before the arrest portion of the procedure and the cardiac output after the repair is good. If the circulatory arrest time becomes an issue, the distal ascending aorta can be cannulated after repair of the aortic arch and reinstitution of low-flow cardiopulmonary bypass. In a patient with a larger aorta ( 6 mm), the option of direct aortic cannulation also is feasible. DR SETHIA: I think one can save a little bit of clamp or arrest time in that way. I have a second question. I note that you still use a pulmonary or other type of homograft to reconstruct the aortic arch. In our series of more than 100 Norwood procedures, we had had to use a homograft in only about 5 patients. Do you think that it is really necessary to continue to use a homograft or some form of augmentation of the transverse aortic arch? DR MOSCA: Are you referring to the technique where you bring the pulmonary artery right up under the aortic arch? DR SETHIA: You do not have to bring the pulmonary valve right up under the aortic arch, but you have to mobilize the aortic arch quite extensively with the head and neck vessels and the descending aorta. And you can undertake a primary anastomosis. DR MOSCA: This technique, popularized by you and Dr Brahm, also is applicable. However, in our opinion, the technique you described does not relieve the aortic arch obstruction as reliably, and it may produce tension at the arch anastomosis by virtue of bringing the divided pulmonary artery up under the aortic arch. We believe that the incidence of neocoronary obstruction and neoaortic valve insufficiency is less using our technique, which eliminates any tension on the aorta or distortion of the pulmonary valve. DR SETHIA: So you have used the other technique? DR MOSCA: We used it before 1987, but have switched to the technique I just presented, and our results have improved. DR SETHIA: Thank you.