Pediatric heart transplantation after operations involving the pulmonary arteries

J

THORAC CARDIOVASC SURG

1991;102:386-95

Pediatric heart transplantation after operations involving the pulmonary arteries A prohibitive perioperative mortality has been previously ascribed to pediatric heart transplantation after palliative operatiees for congenital heart disease involving the pulmonary arteries. Of 46 children who have undergone heart transplantation at our institution between June 1984 and February 1990, 7 (15%; mean age 8 ± 3 years; range 1 to 18 years) have previously undergone such operations: right ventricle to pulmonary artery conduit/homograft for levo-transposition of the great arteries (2~ Waterston shunt for tricuspid and pulmonary atresia (1~ pulmonary artery banding for single ventricle (1~ Fontan procedure for single ventricle (1~ first-stage Norwood procedure for hypoplastic left heart syndrome (1~ and classic right Blalock-Taussig shunt for atrioventricular canal with pulmonic stenosis (1~ Three categories of pulmonary artery anatomy that require different approaches to reconstruction at the time of transplantation are recognized: abnormalities of position, pulmonary outflow obstruction, and previous systemic- or atrial-pulmonary coenecnons, At operation, individualized pulmonary arterial reconsnucnon was employed, including use of previously created right ventricular-pulmonary artery conduits/homografts and angioplasty (with and without pericardial patches). Transplantation was successful in aU patients. Posttransplant right ventricular-pulmonary artery pressure gradients and pulmonary vascular resistance indices were acceptable, with a tendency to decrease with time. Two patients had critical right ventricular failure postoperatively; one of them required support with extracorporeal membrane oxygenation. There was no perioperative mortality, with three deaths occurring from 5 to 39 months after transplantation. All surviving patients are in New York Heart Association functional class I. Techniques borrowed from the repair of congenital cardiac lesions can be applied to subgroups of children undergoing heart transplantation. Additional length of donor aorta and pulmonary artery should be harvested for possible use in designing pulmonary artery connections. Previous palliative operations. involving the pulmonary arteries with associated complex pulmonary artery anatomy are not of themselves an insurmountable obstacle to successful heart transplantation.

Matthew M. Cooper, MD (by invitation)," Laszlo Fuzesi, MD (by invitation)," Linda J. Addonizio, MD (by invitation),"Daphne T. Hsu, MD (by invitation)," Craig R. Smith, MD (by invitation)," and Eric A. Rose, MD,a New York, N.Y.

A t most centers patients with congenitalheart disease constitute a minority of pediatric recipients of cardiac transplants.!" As experience has increased with such patients, there has beena broadeningof the complexity of From the Departments of Surgery' and Pediatrics," Columbia-Presbyterian Medical Center, Columbia University College of Physicians and Surgeons, New York, N.Y. Read at the Seventieth Annual Meeting of The American Association for Thoracic Surgery, Toronto, Ontario, Canada, May 7-9, 1990. Address for reprints: Matthew M. Cooper, MD, Department of Surgery, Columbia-Presbyterian Medical Center, 630 West l68th St., New York, NY 10032.

12/6/28121

386

congenitalproblemsthat may be treated successfully with transplantation.v" Until recently, successful transplantation in patients with "complex" pulmonary artery anatomy or previous palliative operationsinvolving the pulmonaryarteries has been performed infrequently and with increasedperioperative mortality.l" Our experience at Columbia-Presbyterian Medical Center with such patients has prompted this report. Patients and methods Ofa total of 46 patients ranging in agefrom 4days to 18 years who received cardiac allografts at our institution between June 1984and February 1990, seven (15%; six male) have undergone

Volume 102

Pediatric heart transplantation 3 8 7

Number 3 September 1991

Table I. Patient characteristics Patient

Age (yr)

Anatomy

Previous operations

13.8

L-TGA, pulmonary stenosis, VSD

I. Valved RV-PA conduit, ASD, VSD closure 2. Systemic A V valve replacement, pacemaker I. Bilateral modified B-T shunts 2. Valved RV-PA homograft, VSD closure, takedown B-T shunts Waterston shunt

L-TGA, pulmonary atresia

2

3.8

3

17.7

4

4.6

5

15.8

Tricuspid and pulmonary atresia, ASD,VSD Single ventricle, mitral atresia, L-TGA Single ventricle, L-TGA

6 7

1.2 2.1

Hypoplastic left heart syndrome AV, canal, pulmonic stenosis

Atrial septectomy, PA banding

I. PA banding 2. Fontan Stage I, Norwood procedure B-T shunt

L-TGA. Leva-transposition of the great arteries; VSD. ventricular septal defect; RV. right ventricle; PA, pulmonary artery; ASD. atrial septal defect; AV, atrioventricular; B-T, Blalock-Taussig.

previous palliative operations involvingthe pulmonary arteries. The specific anatomic abnormalities included in this select group of patients and their previous operations are shown in Table I. Preoperative hemodynamic parameters pertinent to the pulmonary.circulation are shown in Table II. Transplantation was done in these patients at a mean age of 8.4 ± 3 years (mean ± standard error of the mean), with a range of 1 to 18 years. All patients were in New York Heart Association (NYHA) functional class IV, and five of seven (71%) were dependent on inotropic agents before transplantation. Recipients and donors are matched by weight and ABO compatibility. Donor hearts are arrested with an iced crystalloid cardioplegic solution and maintained in an iced slush solution until implantation. Our immunosuppressive regimen has evolved and now includes "triple" therapy comprising cyclosporine, azathioprine (2 to 4 mg/kg), and steroids (methylprednisolone, 30 mg/kg intraoperatively, followedby prednisone, 0.5 mg/kg/day tapered to -0.1 mg/kg/day). Cyclosporine (6 mg/kg) is usually administered orally before transplantation and then adjusted to maintain whole blood levels of approximately 200 ng/rnl. However, three of the seven patients in this group received only cyclosporine and steroids because their transplantations were performed before our switch to triple therapy in October 1986. Induction was accomplished in one patient with a 14-day course of OKT3, a murine monoclonal antibody that recognizes the CD3 antigen (Ortho Diagnostics, Raritan, N.J.), with cyclosporine started on posttransplantation day 4. We choose to follow up our pediatric patients with routine endomyocardial biopsies, as we do our adult patients. Operative strategy. We recognize three categories of pulmonary artery anatomy that necessitate different approaches to reconstruction during transplantation. These include abnormalities of position, pulmonary outflow obstruction, and previous systemic-pulmonary or atrial-pulmonary connections. At operation individualized pulmonary arterial reconstruction was performed. The average ischemic time of the donor hearts was 177 ± 24 minutes (range, 80 to 267 minutes). Abnormalities of position. Two patients had levo-transposition of the great arteries (L-TGA) and had undergone re-

Table II. Preoperative hemodynamics PVRI

PAS/PAD (mmHg)

PA (mmHg)

Inotropic support

7.2 4.2

70/42 75/12 36/24

50 34 32

+ + +

2.4 5.4 2.5

22/10

16

28/20

24

Patient

(RU)

I 2 3 4 5 6 7

+ +

PVRI, Pulmonary vascular resistance index; PAS, pulmonary artery systolic pressure; PAD, pulmonary artery diastolic pressure; PA, mean pulmonary artery pressure.

constructions for which either a ventricular-pulmonary artery conduit or a homograft was used before transplantation. Transplantation was necessitated in both patients because of refractory ventricular or biventricular failure. The first patient initially had placement of a valved right ventricular-pulmonary artery conduit and closure of atrial and ventricular septal defects for palliation of L-TGA with pulmonic stenosis (Fig. 1). Subsequently replacement of the systemic atrioventricular valve and pacemaker placement were required before transplantation at the age of 14 years. At the initial operation the bifurcation of the main pulmonary artery was opened bilaterally across the stenotic right and left branch orifices,and the valved conduit was sutured to their confluence. Therefore, at the time of transplantation, the valved conduit was transected just distal to the valve prosthesis, and the donor pulmonary artery was anastomosed end to end to its remaining portion. Because of the concomitant mesocardia, the pulmonary artery anastomosis was placed most comfortably posterior to the aortic anastomosis. Our second patient with L-TGA and pulmonary atresia had previously undergone creation of bilateral modified BlalockTaussig shunts in infancy followed by reconstruction with a right ventricular-pulmonary artery valved homograft, closure

3 8 8 Cooper et al.

The Journal of Thoracic and Cardiovascular Surgery

Fig. I. Patient 1 after reconstruction with a valved right ventricular-pulmonary artery conduit for L-TGA The conduit had been sutured to the confluence of the stenotic right and left branch pulmonary artery orifices. At transplantation the conduit was transected just distal to the valve prosthesis, and the donor pulmonary artery was anastomosed end to end to its remaining portion. The concomitant mesocardia allowed the pulmonary artery anastomosis to lie posterior to the aortic anastomosis. of the ventricular septal defect, and takedown of the BlalockTaussig shunts at 3 years of age (Fig. 2). At the time of transplantation the ventricular-pulmonary artery homograft was transected just distal to its valve, and the donor pulmonary artery was anastomosed end to end to the remaining portion of the homograft. In this patient, in contrast to the previous patient with L-TGA, the pulmonary artery anastomosis was positioned anterior to the aortic anastomosis, as shown in Fig. 2. Pulmonary outflow obstruction. Our first patient with obstructive pulmonary artery anatomy was referred for transplantation at the age of 18 years after hemodynamic conditions were found to be unsatisfactory for a proposed Fontan procedure. A Waterston shunt had previously been created for palliation of tricuspid and pulmonary atresia at the age of 4 months (Fig. 3). There was partial obstruction of the right pulmonary artery, with minimal flow to the left pulmonary artery, because of "kinking" at the site of the Waterston shunt. Severe left ventricular dysfunction with coexisting aortic valve insufficiency and mitral valve regurgitation precluded additional palliation. At operation the aorta was transected at the superior rim of the shunt site, and the pulmonary artery was opened longitudinally through the same. It was elected to reconstruct both pulmonary arteries with a long pericardial patch on the Waterston shunt site. Alternatively, a long spatulated anastomosis with the donor pulmonary artery could have been employed. A tight Mersilene band (Ethicon, Inc., Somerville, N.J.) was present around the proximal main pulmonary artery in our next patient. This patient had previously undergone atrial septectomy and pulmonary artery banding for palliation of a single ventricle with mitral atresia (Fig. 4). Subsequent ventricular failure prompted transplantation. At operation, an incision was made in the lateral aspect of the main pulmonary artery,

extending across the Mersilene band and into the proximally constricted left pulmonary artery in its intrapericardial portion (Fig. 5). The donor pulmonary artery was then appropriately fashioned and the pulmonary artery anastomosis was performed, effectively patching the area with the donor's pulmonary artery bifurcation. This patient also had L-TGA. However, pulmonary anatomy at the time of transplantation posed a problem with respect to pulmonary outflow rather than an abnormality of position, given the absence of a previously placed ventricular-pulmonary artery connection. Systemic-pulmonary or atrial-pulmonary connection. The first patient in this category was born with a univentricular heart (double-inlet left ventricle; L-TGA) and had previously undergone pulmonary artery banding followed later by a Fontan procedure. Before transplantation several attempts were made at pharmacologic sclerosis of the pleural spaces to help manage pleural effusions that resulted in a severe restrictive pulmonary functional defect. At operation the main pulmonary artery was found to be of good size. The right atrial-pulmonary artery connection extended onto the right main pulmonary artery (Fig. 6). The right atrium was hemisected and the Fontan connection taken down, leaving an extremely wide opening at the junction of the right and main pulmonary arteries. The right-sided portion of this pulmonary artery ostium was oversewn for a distance of 2 em, leaving a right pulmonary artery orifice of approximately 10 mm in diameter and a main pulmonary artery orifice of good size match for end-to-end anastomosis with the donor pulmonary artery. During implantation of the donor heart, left atrial decompression via a stab wound in the fossa ovalis was required after completion of the left atrial anastomosis because of the extensive pulmonary collateral circulation. The sixth patient was 14 months old at the time of trans plan-

Volume 102

Pediatric heart transplantation 3 8 9

Number 3 September 1991

Fig. 2. Reconstruction for L-TGA in patient 2. An right ventricular-pulmonary artery valved homograft was employed at the time of pretransplant reconstruction. In an approach similar to that employed in patient 1, the ventricular-pulmonary artery homograft was transected distal to its valve, and end-to-end pulmonary artery anastomosis was performed to the remaining homograft. In contrast to the previous patient, however, the pulmonary artery anastomosis was positioned anterior to the aortic anastomosis.

tation. The patient had previously undergone a first-stage Norwood procedure for hypoplastic left heart syndrome (Fig. 7) but had unsuitable hemodynamics, including tricuspid regurgitation and poor right ventricular function, precluding a subsequent Fontan procedure. At operation the pulmonary arteries were confluent, although the main pulmonary artery was small. The branch pulmonary arteries admitted 6 mm diameter Hegar dilators. The pulmonary arteriotomy was extended approximately 2 cm distal to the bifurcation onto each of the branch pulmonary arteries. Identification of the previously placed central polytetrafluoroethylene* shunt was difficult because of its short length (2 mm), and incision of the descending aorta was necessary for its visualization and excision. The arch vessels were then crossclamped and circulatory arrest was instituted at a minimum temperature of 20 0 C for 23 minutes, allowing the defect in the descending aorta to be patched with a piece of donor neoaorta, Cardiopulmonary bypass was reinstituted, and the atrial anastomoses were performed and tailored in such a way as to create a new atrial septum because the atrial septum was surgically absent. An additional 45 minutes of circulatory arrest was then employed to facilitate anastomosis of the donor pulmonary artery to the confluence of the main pulmonary artery, with extension out onto the branch pulmonary arteries bilaterally. The donor aorta was then anastomosed to the remainder of the neoascending aorta. The final patient in this group had a complete atrioventricular canal and pulmonic stenosis for which a classic right Blalock-Taussig shunt was performed. The Blalock-Taussig shunt was doubly ligated at the time of transplantation, and a *Gore-Tex shunt manufactured by W. L. Gore & Associates, Inc., Elkton, Md.

standard end-to-end pulmonary artery anastomosis was performed. Poor left ventricular function precluded additional palliative intervention. Pathologic analysis of the explanted heart, in fact, demonstrated a cardiomyopathic ventricle with extensive fibrosis.

Posttransplantation course

Right ventricu/aroutftow. All patients underwent right heart catheterization within 1 month of transplantation with measurement of right ventricular-pulmonary artery pressure gradients. Five of seven patients had gradients in the 0 to I I mm Hg range, and two had gradients greater than 30 mm Hg, which decreased during the ensuing months. Of these two latter patients, one was the patient who required postoperative extracorporeal membrane oxygenation (ECMO) and had an initial right ventricular-pulmonary artery pressure gradient of 46 mm Hg that decreased as time progressed, as discussed in the next section. The pulmonary vascular resistance index (PVRI) averaged 3.4 ± 0.8 RU* (range 0.5 to 5.7 RU) aftertransplantation. The right ventricular-pulmonary artery pressure gradients and PVRIs in these patients showed some tendency to decrease with time. Right ventricular failure. Two patients had severe perioperative right ventricular failure after transplantation. One of these was the patient with L-TGA status after placement of a right ventricular-pulmonary artery valved conduit. Preoperatively the PVRI was 7.2 RU on a dosage regimen of Dobutarnine, decreasing to 5.5 RU with large doses of nitroprusside in the catheterization laboratory. Pulmonary artery pressures were 70/42 mm Hg, with a mean of 50 mm Hg, dropping to 60/30 *RU (resistance unit)

= mm

Hg/Lyrnin/rn.?

3 9 0 Cooper et al.

The Journal of Thoracic and Cardiovascular Surgery

Fig. 3. Transplantation after Waterston shunt for palliation of tricuspid and pulmonary atresia. "Kinking" of right pulmonary artery was present at site of Waterston shunt with minimal flow to left pulmonary artery. Reconstruction after takedown of Waterston shunt employed a long pericardial patch at shunt site.

mm Hg, with a mean of 40 mm Hg with nitroprusside. These parameters placed this patient in a somewhat higher risk group," and he was inotrope dependent as well. After transplantation, nitroprusside was required for right ventricular afterload reduction and pulmonary vasodilatation. A multiple gated acquisition scan obtained postoperatively indicated a right ventricular ejection fraction of - 30% compared with a left ventricular ejection fraction of 85%. The right ventricular-pulmonary artery gradient averaged 10 mm Hg, and he subsequently recovered from his right ventricular failure. Profound right ventricular failure developed postoperatively in the patient receiving a transplant after stage I palliation for hypoplastic left heart syndrome. On postoperative day I the right ventricular failure necessitated institution of ECMO support via the right common carotid artery and the right atrium directly. The pretransplantation PVRI was elevated at 5.4 RU. The patient was successfully weaned from ECMO after 72 hours and ultimately from ventilatory support entirely. Unfortunately, the patient had a severe right hemispheric infarct resulting from ligation of the right common carotid artery at the time of cannulation for ECMO, as well as a unilateral phrenic nerve palsy. Her somewhat tenuous respiratory status was further compromised by a parainfluenza respiratory infection that ultimately led to a spiraling downhill course, including respiratory arrest and death 9 months after transplantation. During this latter period cardiac function remained normal. In fact, serial postoperative catheterizations demonstrated good right ventricular function with persistent but decreasing gradients (from 46 mm Hg to a low of 15 mm Hg) between the main pulmonary artery and the branch pulmonary arteries. Some degree

of stenosis at the branch anastomoses was demonstrated angiographically. Rejection. All patients had at least one rejection episode. Two patients required courses of OKT3 for refractory rejection, and one required a course of antithymocyte globulin as well. Both patients recovered from these episodes with excellent function, although one ultimately required placement of a permanent DDD pacemaker for sinus arrest. Additional morbidity. Patient 4 did well initially and improved to NYHA functional class I status. Three years after transplantation, however, a diagnosis of stage IA Hodgkin's disease with lymphocyte predominance was made. The disease was treated with mechlorethamine, Oncovin (vincristine), procarbazine, and prednisone chemotherapy. In February 1989, almost 5 years after his transplantation, right-sided failure developed, and a mass in the right atrium, thought to be consistent with residual lymphoma, was found. He therefore underwent retransplantation. Pathologic examination demonstrated only organizing thrombus and chronic rejection. He has since done well and returned to school and full activity despite requiring placement of a permanent D DD pacemaker because of sinus bradycardia. The early postoperative course of an additional patient (patient 3) was complicated by constrictive hemodynamics for which pericardiectomy and evacuation of organizing hematoma from the pericardial space were ultimately required 2 months after transplantation. He has subsequently done well and is in NYHA functional class I 3 years after transplantation. Mortality and survival. In this challenging group of patients transplantation was performed without perioperative mortality.

Volume 102 Number 3 September 1991

Pediatric heart transplantation 3 9 I

Fig. 4. Pretransplant anatomy after pulmonary artery banding for univentricular heart. A tightMersilene bandwas present around the main pulmonary artery. One patient died suddenly of acute myocardial infarction resulting from atherosclerotic occlusion of the left circumflex coronary artery5 months postoperatively. Therewas onedeath at 9 months resulting from chronic respiratory failure superimposed on neurologic deficit. Sudden deathoccurred in another patient at 39 months from acute rejection resulting from noncompliance with medications. The four surviving patients are alive and well 9 to 77 months after transplantation, and all, including the patient who underwent retransplantation 5 years afterinitial transplantation, are in NYHA functional class 1. A curve of actuarial survival (approximately 67% at 5 years) is shown in Fig. 8.Strikingly, the patient witha severe restrictive pulmonary defect who underwent transplantation with takedown ofhis Fontan connection isnow in NYHA functional class I andisagain playing basketball. Chronic pleural effusions have completely resolved. Discussion The ability to perform successful transplantation in patients who have had previous palliative operations involving the pulmonary arteries is in essencethe problem of achieving acceptable pulmonary vascular resistance (i.e.,minimal or no anatomic obstruction and pulmonary vasoconstriction). Acceptable is difficult to define but is intended to suggest a range of values that will allow continued effective function of the donor right ventricle after transplantation. Anatomic reconstruction must be sufficient to prevent the imposition of an additional afterload

Fig. 5. Transplantation and pulmonary artery reconstruction after removal of pulmonary artery band. Incision was madein lateral aspect ofmain pulmonary arteryandextended across the Mersilene bandandonto the proximally constricted leftpulmonaryartery. Donor pulmonary artery wasfashioned so that its bifurcation effectively patched thedefect inrecipient pulmonary artery. on the new heart above the preexisting pulmonary vascular resistance of the recipient. The use of donor hearts from somewhat larger donors (donor weight 25% greater than recipient weight) for patients with an elevated PVRI preoperatively has been suggested to reduce the difficulties associated with pulmonary vasoconstriction. In fact, preliminary studiessuggestthat hearts from infant donors with body weight at least 75% greater than recipient weight tend to have better systolic function in the early postoperative period," However, Addonizio and colleagues? have recently examined a possible role for such larger donor hearts in patients with borderline or elevated PVRI before transplantation. In a series of 23 pediatric patients with PVRI greater than or equal to 6 RU who received a cardiac transplant, there was no significant differencein the prevalenceof right ventricular failure or death from right ventricular failure postoperativelywhen patients receivinga heart from donors at least 25% larger

The Journal of Thoracic and Cardiovascular

3 9 2 Cooper et al.

Surgery

l

(

Fig. 6. Transplantationafter Fontanprocedure. Right atrial-pulmonary artery connection extendedontorightmain pulmonaryartery. Orifice of right main pulmonaryartery waspartiallyoversewn, as wasthe main pulmonaryartery, creatinga main pulmonaryartery orificeofgoodsizematch forend-to-end anastomosis withdonorpulmonary artery.

than recipients were compared with recipients receiving a heart from smaller donors. The ability to reconstruct pulmonary anatomy satisfactorily at the time of transplantation (i.e., to accomplish satisfactory right ventricular and pulmonary artery outflow) is in part also dependent on the location of the abnormalities. In general, the more distal the lesions the less amenable they are to satisfactory reconstruction. Trento and colleagues' reported that four of the six deaths in their patients with congenital heart disease undergoing transplantation were attributable to acute failure of the donor right ventricle. Included in this group were two patients who required pulmonary artery reconstruction for stenoses resulting from previous systemic-pulmonary artery shunts (one patient with a Potts shunt and left pul-

monary artery stenosis; one patient after bilateral Blalock-Taussig shunts) and one patient who had had bilateral Blalock-Taussig shunts performed after a first-stage Norwood procedure and had distortion of the branch pulmonary arteries at the hila bilaterally. More recently, Mayer and coworkers" reported a series of seven patients with congenital heart disease who underwent orthotopic heart transplantation. Of these patients, four had undergone previous palliative procedures. These procedures involvedthe pulmonary arteries, including first-stage palliation for hypoplastic left heart syndrome (two patients), Fontan procedure (one patient), and pulmonary artery banding combined with additional palliation of a univentricular heart (one patient). Perioperative right ventricular failure occurred in one patient, and one patient

Volume 102 Number 3 September 1991

Pediatric heart transplantation 3 9 3

Fig. 7. Transplantation after first-stage Norwood procedure for palliation of hypoplastic left heart syndrome. Identification of central polytetrafluorethylene shunt was difficult because of its short length (2 mm) and required incision of descending aorta for its visualization and excision. Circulatory arrest was employed to allow the defect in the descending aorta to be patched with a piece of donor neoaorta and to facilitate anastomosis of donor pulmonary artery to confluence of main pulmonary artery with extension out onto the branch pulmonary arteries bilaterally. Donor aorta was then anatomosed to remainder of neoascending aorta.

required prolonged inotropic and ventilator support. There was no perioperative mortality among these four patients. Severeperioperative right ventricularfailureoccurred intwoofourpatients.In onethe pretransplantation PVRI was7.2RU and in the other, 5.4 RU. Bothpatientswere alreadyreceiving infusions of dobutamineat the time of thesedeterminations. Reduction in the PVRI occurredin response to nitroprusside in the first patient, decreasing the valueto 5.5 RU. Althoughthis patient wasclearlyin a higherrisk group, the laboratory response to a pulmonaryvasodilator portends a morefavorable prognosis.' In this patient,acute right ventricularfailure was managed successfully postoperatively withnitroprusside. However, the second patient required support with ECMO for 3 days. This patientwassuccessfully weaned from ECMO and maintained satisfactory right ventricular function. Complications arisingfromthe methodofcannulation for ECMO, which included ligation of the right common carotid artery, ultimately contributed to her death. We would not hesitate to consider ECMO in the future for temporary support of acute right ventricularfailure but would look more favorably on central cannulation to achieve this end.'? With regard to pretransplantation

100 III

>

80

::J

-... en

60

GI

l:

40

GI Q.

20

...>

U

l

I

0 0

20

40

60

80

Months After Transplantation Fig. 8. Actuarial survival for patients undergoing transplantation after palliative procedures involving pulmonary arteries (Kaplan-Meier). N = 3 at 36 months after transplantation.

selection and preparationof potentialrecipients, Addonizioand coworkers7 havesuggested that a courseof dobutamine or amrinone may be of benefit in lowering the PVRI sufficiently to allow successful transplantation, particularly in thosepatientswhose PVRI shows minimal response to pharmacologic vasodilation.

The Journal of Thoracic and Cardiovascular Surgery

3 9 4 Cooper et al.

Abnormalities of pulmonaryartery position have previously been addressed at the time of transplantation.!"!' Harvest of additional donor aorta and pulmonary artery is stressed. In addition, Doty'' stresses complete mobilization of the aorta to the pericardial reflection, as wellas dissection from the pulmonary artery, to facilitateaortic and pulmonaryartery connections. In our patients with L-TGA, previously placedventricular-pulmonary artery connections were utilized for the pulmonary artery anastomosis. Chartrand.lf- 15 Menkis," and their colleagues have further extended the realm of transplantation by successfully addressing atrial abnormalities as well as anomalous pulmonary and systemic venous return at the time of transplantation. Modified preparation of the donor heart in conjunction with multiple techniques of atrial reconstruction have been employed." There was no morbidity or mortality in our patients directly related to the preexisting anatomic substrate or subsequent transplantationwithpulmonaryartery reconstruction.Mortality related to graft coronaryatherosclerosisand rejection has previously plaguedseries of pediatric transplant recipients.' Reductions in these factors contributingto late mortalitymustawaitthe development of new, effective, and more specific immunosuppressive modalities. The complexity of pulmonary artery reconstruction requiredat the time of transplantationin patientssuchas these is variableand depends on the anatomic substrate, which may include abnormal position, pulmonaryoutflow obstruction, or previous systemic-pulmonary or atrial-pulmonary connections. Techniques borrowed from the repair of congenital cardiac lesions can be applied to such subgroups of children undergoing heart transplantation. Additionallengthsof donor aorta and pulmonary artery shouldbe harvested for possible use in pulmonary artery reconstruction. Each patient tends to present a uniqueset of problems, and surgicalimprovisation based on anatomic findings at the time of transplantationmay be necessary. Certainly,difficult pulmonaryartery anatomycan be addressedat the timeof transplantation, and, therefore,should not exclude such patients from consideration for cardiac transplantation.

REFERENCES 1. Heck CF, Shumway Sl, Kaye MP. The registry of the International Society for Heart Transplantation: sixth official report-1989. 1 Heart Transplant 1989;8:271-6. 2. Starnes VA, Bernstein D, Oyer PE, et al. Heart transplantation in children. 1 Heart Transplant 1989;8:20-6.

3. Trento A, Griffith BP, Fricker Fl, Kormos RL, Armitage 1, Hardesty RL. Lessons learned in pediatric heart transplantation. Ann Thorac Surg 1989;48:617-23. 4. Mayer IE Jr, Perry S, O'Brien P, et al. Orthotopic heart transplantation for complex congenital heart disease. 1 THORAC CARDIOVASC SURG 1990;99:484-92. 5. Mavroudis C, Harrison H, Klein JB, et al. Infant orthotopic cardiac transplantation. 1 THORAC CARDIOVASC SURG 1988;96:912-24. 6. Bailey L, Concepcion W, Shattuck H, Huang L. Method of heart transplantation for treatment of hypoplastic left heart syndrome. 1 THORAC CARDIOVASC SURG 1986;92: 1-5. 7. Addonizio LJ, Gersony WM, Robbins RC, et al. Elevated pulmonary vascular resistance and cardiac transplantation. Circulation 1987:76(Pt 2):V52-5. 8. Kanakriyeh MS, Mathis CM, Boucek MM, McCormack 1, Gundry S, Bailey LL. Effect of donor size on graft function post infant heart transplantation [Abstract]. J Heart Transplant 1990;9:77. 9. Addonizio LJ, Hsu DT, Gersony WM, Smith CR, Rose EA. Elevated pulmonary vascular resistance and cardiac transplantation: Are larger donors better? [Abstract]. Presented, American College of Cardiology, Fortieth Annual Scientific Session, Atlanta, Ga., March 3-7, 1991. 10. Galantowicz ME, Stolar CIH. Extracorporeal membrane oxygenation for perioperative support in pediatric heart transplantation [Abstract]. 1 THORAC CARDIOVASC SURG [In press]. II. Reitz BA, Jamieson SW, Gaudiani VA, Oyer PE, Stinson EB. Method for cardiac transplantation in corrected transposition of the great arteries. J Cardiovasc' Surg 1982; 23:293-6. 12. Harjula ALJ, Heikkila LJ, Nieminen MS, Kupari M, Keto P, Mattila SP. Heart transplantation in repaired transposition of the great arteries. Ann Thorac Surg 1988;46:611-4. 13. Doty DB. Cardiac transplantation: transposition of the great arteries. In: Cardiac surgery [A looseleaf workbook and update service]. St. Louis: Mosby-Year Book, 1988: 24-5. 14. Chartrand C, Dumont L, Stanley P. Pediatric cardiac transplantation. Transplant Proc 1989;21:3349-50. 15. Chartrand C, Guerin R, Kanagh M, Stanley P. Pediatric heart transplantation: surgical considerations for congenital heart diseases. 1 Heart Transplant 1990;9:608-17. 16. Menkis A, McKenzie FN, Novick RJ, et al. Special considerations for heart transplantation in congenital heart disease. J Heart Transplant 1990;9:602-7.

Discussion Dr. John E. Mayer, Jr. (Boston. Mass.). We have a similar experience in Boston but with only 16 total transplants. Ten of the patients, however, had congenital heart disease as their primary etiology, and eight of them had had prior operations. Three have had atrial level repairs of transposition; four, palliation for single ventricle, including two who had had a stage I

Volume 102 Number 3 September 1991

hypoplastic repair; and one had had a Fontan operation. All of those patients survived the operation, and there has been only one late death in that group from chronic rejection at 2Y2 years. The one problem that we have seen is particularly prevalent in the older patients who have had chronic cyanosis for many years. We had initial problems with renal failure in the perioperative period in two such teenaged patients, and both of those patients actually had to stop receiving cyclosporine and institute antilymphocyte therapy. Have you had any similar experience or have you altered your immunosuppressive protocol for those groups of patients? Dr. Cooper. We have not had that problem, but I should mention that although our immunosuppressive protocol now includes cyciosporine, three of the patients in this group underwent transplantation before our switch to triple therapy. No, we have not encountered such renal failure. Dr. William A. Gay (Salt Lake City, Utah). Have you had to use any prosthetic graft material or any xenograft material to gain length on your great vessels in any of these cases? Dr. Cooper. No, we have not had to employ such material. That was among the advantages of using portions of the existing conduits in the two cases of transposition. If harvest of sufficientadditional donor material is performed in continuity, that

Pediatric heart transplantation

395

does not pose a problem, at least it has not in our experience. Dr. Gay. I have a perplexing problem of a child who has had a bidirectional Glenn shunt done and needs a transplant. The superior vena cava has been interrupted. I would appreciate your advice. Dr. John E. Mayer, Jr. (Boston, Mass.). We discussed a similar case at the 1989 Annual Meeting of The Western Thoracic Surgical Association (J THORAC CARDIOVASC SURG 1990;99:484-92). Not only by harvesting enough great artery tissue but also by harvesting extended segments of the venous anatomy of the donor, one can overcome anomalies of venous position as well. In two patients who have had prior atrial level repair of transposition, rather than trying to reestablish an atrial septum, I have used the atrial tissue that was there for left atrial anastomosis and then have done two cavo-cavo anastomoses. I think that's a much simpler way of dealing with the problem. I certainly will do all of my post-Senning transplants that way. I also want to emphasize that it was important to think about harvesting not only extended great vessels but also extended venous segments, although you may have to fight with the liver specialists over the lower venous segments.