Repair of truncus arteriosus in the neonate and young infant

Repair of Truncus Arteriosus in the Neonate and Young Infant Edward L. Bove, MD, Robert H. Beekman, MD, A. Rebecca Snider, MD, Louise B. Callow, RN, MSN, David J. Underhill, MD, Albert P. Rocchini, MD, Macdonald Dick 11, MD, and Amnon Rosenthal, MD Divisions of Thoracic Surnerv , and Pediatric Cardioloev. C.S. Mott Children’s Hospital, The University of Michigan Medical Center, Ann Arbor, Michigan
I,,

Infants with truncus arteriosus present a difficult management issue. Because of the high operative mortality, repair is often delayed beyond the first 3 to 6 months of age. We reviewed our experience with 11 neonates and young infants with truncus arteriosus undergoing repair (median age, 21 days). Five patients also had major truncal valve insufficiency, and 2 required valve replacement. Right ventricle-pulmonary artery continuity was established with a porcine valved conduit in 3 patients and an aortic or pulmonary homograft in 8. There was 1 operative death (9%;70% confidence limits, 39622%)and 1 late death over a mean follow-up of 21 months (range,

T

runcus arteriosus accounts for approximately 2% to 4% of all congenital cardiac malformations [l-31. It is characterized by a single arterial trunk and semilunar valve leaving the heart [4, 51. This vessel gives rise to the systemic, pulmonary, and coronary circulations. The severe hemodynamic burden imposed by the high flow and elevated pulmonary artery pressure leads to the development of congestive heart failure early in life. Low diastolic pressure d u e to runoff from the truncal artery into the low-resistance pulmonary circuit can result in decreased coronary perfusion and diminished ventricular performance. The presence of substantial truncal valve regurgitation further compounds the clinical picture. Without treatment, survival for more than 1 year is rare. Furthermore, the development of pulmonary vascular occlusive disease occurs early and can lead to a significantly increased operative risk or even inoperability by 3 to 6 months of age (61. Following the landmark contribution of Ebert and colleagues [7], several reports [8, 91 have demonstrated improvement in the outcome of patients undergoing complete repair of truncus arteriosus. However, survival after repair in the first month of life remains uncommon, and some groups [8, 10, 111 continue to recommend postponing operation until the patient is 3 to 6 months of age. At The University of Michigan, we have pioneered the Presented at the Twenty-fourth Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Sep 26-28, 1988. Address reprint requests to Dr Bove, Section of Thoracic Surgery, The University of Michigan Medical Center, 1500 E Medical Center Dr, 2120 Taubman Center, Box 0344, Ann Arbor, MI 48109.

0 1989 by The Society of Thoracic Surgeons

4 to 32 months). Eight of the 9 late survivors are growing

normally. Echocardiographic examination revealed normal ventricular function in all patients (mean shortening fraction, 39%). Doppler assessment demonstrated trivial prosthetic or homograft valve regurgitation in 7 patients and mild to moderate obstruction in 5 patients. This recent experience with repair of truncus arteriosus indicates that the operative risk is low even in the neonate. Repair in the first month of life should be recommended before the development of critical congestive heart failure or irreversible pulmonary vascular disease. (Ann Tliorac Surg 1989;47:499-506)

surgical treatment of truncus arteriosus since the first successful repair was performed at this institution in 1962 [12]. Although early repair with a nonvalved conduit was initially preferred [13, 141, more recent experience has indicated that insertion of a valve in the pulmonary position may provide a substantial hemodynamic advantage in these generally very ill patients. Beginning in 1986 at The University of Michigan, all infants with truncus arteriosus have undergone complete repair with a valved conduit as soon as the patient is referred for initial diagnosis, regardless of age or size. Because the majority of these patients are seen soon after birth, most of our repairs have been performed in neonates. We, therefore, reviewed our entire recent experience with the surgical treatment of truncus arteriosus to assess the results of this practice.

Material and Methods Between January 1986 and May 1988, a total of 13 patients with the diagnosis of truncus arteriosus were seen at C.S. Mott Children’s Hospital of The University of Michigan. One 8-year-old boy underwent repair after pulmonary artery banding performed in infancy at another institution, and is excluded from further review. One neonate sustained a cardiac arrest during catheterization and died within a few hours before any surgical intervention. The remaining 11 patients underwent primary repair in infancy and form the basis of this report. There were 4 male and 7 female infants (Table 1). Age ranged from 4 days to 4 months (median age, 21 days) with only 2 patients older than 3 months. Weight ranged from 1.9 to 4.1 kg (mean weight, 3.3 kg). 0003-4975/89/$3.50

500

BOVEETAL TRUNCUS ARTERIOSUS IN NEONATES

Ann Thorac Surg 1989;47499-506

Table 1 . Clinical Characteristics of Pntieiits With Triiizciis Arteriosiis Pa tien t No. 1

Weight (kg1

Sex

Associated Diagnoses

4 mo

3.8

F

4 days

2.8

2 mo

3.5

3.7 mo

4.1

21 days

3.5

Preoperative respiratory arrest, CNS damage, partial DiGeorge syndrome None None None Left SVC, absent right SVC, restrictive VSD None Nonconfluent pulmonary arteries None Partial DiGeorge syndrome Cleft lip and palate, DiGeorge syndrome None

Age

6

28 days

1.9

7

12 days

3.8

F F

8

13 days

3.4

9

10 days

2.2

10

21 days

3.4

M M M

11

6 days

3.9

M

congestive heart failure;

CNS

CHF

=

=

central nervous system;

SVC

The clinical characteristics and associated diagnoses are also listed in Table 1. One patient (patient 1) sustained central nervous system damage when an endotracheal tube became dislodged during air transport prior to surgical repair. Three patients had either partial (2) or complete (1)DiGeorge syndrome. One patient (patient 5) had a restrictive ventricular septal defect and a n absent right superior vena cava, and 1 (patient 7) had nonconfluent pulmonary arteries with the left pulmonary artery supplied by a ductus arteriosus. Preoperative diagnosis was established in each case by two-dimensional echocardiography/Doppler examination. Because of suspected associated lesions, preoperative cardiac catheterization was

=

superior vena cava;

Outcome and Length of Follow-up CNS damage, no cardiac symptoms; 32 mo Well; moderate CHF; 31 mo Well; 30 mo Well; 27 mo Operative death Well; 19 mo Well; 18 mo Well; 14 mo Late truncal valve replacement; 14 mo Late death Well; 4 mo VSD

=

ventricular septal defect

performed in 3 patients. Aortography only was performed in 1 patienl to better quantify the degree of truncal valve regurgitation. Using the classification of Collett and Edwards [15], 1 patient had type i anatomy and 5 patients had type 2 (Table 2). Four patiirnts had an intermediate form between types 1 and 2 with both pulmonary arteries arising from the left posterior w d l of the truncal vessel. The remaining patient had nonconfluent pulmonary arteries. The right pulmonary artery a rose from the left posterolateral aspect of the truncal vessel, and the left pulmonary artery arose from the ductus arteriosus. The ventricular septal defect was a typical outlet defect in 9 patients. In 1, it extended

Table 2. Operative Details Patient No.

Anatomical Classification

1

2

2

1

3

2

4 5

2 1-2

6

1-2

7

1"

8

2

9

2

10 11

1-2

1-2

VSD Anatomy Perimembranous outlet Outlet Outlet Outlet Restrictive outlet Outlet Outlet Outlet Outlet Outlet Outlet

Truncal Regurgitation 0

3+ 0

2+

0 0 4+h

0 3+

0 4+b

Associated Procedures None None None None VSD enlargement None TVR,' reconstruction of nonconfluent pulmonary arteries None None None TVR

This patient had truncal valve stenosis (36-mm Hg gradient). a The right pulmonary artery arose only from the truncal vessel. replacement was performed ten days after repair. VSD = ventricular septal defect T V R = truncal valve replacement;

Truncal valve

BOVEETAL TRUNCUS ARTERIOSUS IN NEONATES

Ann Thorac Surg 1989;47499-506

501

Table 3 . Right VentriclelPulrnonary Artery Conduits Patient No. 1 2 3

4

5 6

to the tricuspid valve annulus (perimembranous), and in the remaining patient, the defect was moderately restrictive. Truncal valve regurgitation was absent in 6 patients, 2+ in 1 patient, 3+ in 2 patients, and 4+ in 2 patients. In these last 2 patients, the valves were quadricuspid with thickened, immobile leaflets that did not close in diastole (Fig 1).Both patients had systolic gradients of 36 mm Hg across the valve by Doppler examination; none of the other patients had evidence of outflow tract gradients. Operative repair was performed using bicaval cannulation with low-flow cardiopulmonary bypass at systemic temperatures of 20°C. The pulmonary arteries were snared at the onset of bypass, and dilute blood cardiople-

Type

14 12 12

Carpentier-Edwards Hancock Hancock Aortic homograft Aortic homograft Aortic homograft Aortic homograft Pulmonary homograft Pulmonary homograft Pulmonary homograft Pulmonary homograft

13 13 8

7

10

8

14

9

11 14 14

10

Fig 1 . (Patient 7.) Parasternal long-axis view in an infant with truncus arteriosus. The truncal ualve is thickened, rolled, and nryxornatous. This still frame 7uas taken during diastole and dernonstrates a large regurgitant orifice. By Doppler examination, there ulas se71ere truncal valve regurgitation and a 36-rmn H g peak instantaneous pressure gradient across that valve. (A = anterior; I = inferior; LA = left atrium; LV = left ventricle; RV = right ventricle; TR = truncus arteriosus.)

Size (mm)

11

gia (4°C) was delivered through the truncal vessel after placement of the cross-clamp. Cross-clamp times ranged from 45 to 95 minutes (mean time, 76 minutes). The pulmonary arteries were excised from within the truncal vessel after the bifurcation was exposed through a short, vertical anterior incision (Fig 2). This allowed accurate removal of the pulmonary bifurcation without injury to the truncal valve or left coronary artery. The resultant defect was closed with a patch in each instance to prevent distortion of the truncal valve. Right ventricle-pulmonary artery continuity was established with a valved conduit in each patient (Table 3). In the first 3 patients in the series, a heterograft (14-mm Carpentier-Edwards in 1 patient, 12-mm Hancock in 2 patients) was implanted. Cryopreserved aortic homografts (CryoLife, Inc, Marietta, GA) were used in 4 patients. The last 4 patients received cryopreserved pulmonary homografts, which are currently preferred because of their thinner walls and greater ease of handling.

Fig 2 . Znitial step in the repair of truncus arteriosus. A n incision is made in the anterior wall of the truncal vessel to allow accurate rernoval of the pulmonary artery bifurcation without damaging the valve or left coronary artery.

502

BOVEETAL TRUNCUS ARTERIOSUS IN NEONATES

n

Fig 3 . The cryopreserved honiograft is stented over a dilator and extended yroxinially with a knitted Dacron conduit equal to or a feeril rnillirneters srnaller than the diameter of the homograft.

Size selection was generally determined by homograft availability and patient size; the 8-mm aortic homograft was selected for the patient weighing only 1.9 kg. Despite the greater technical ease of implanting homografts than heterografts, "oversizing" was avoided. Homografts were extended proximally with knitted Dacron conduits equal to or one size smaller than the homograft (Fig 3). The Dacron extension was then sharply beveled so that only a 2- or 3-mm rim of Dacron remained posteriorly (Fig 4). In the patient with nonconfluent pulmonary arteries (patient 7), the transected end of an aortic homograft was sutured to the divided left pulmonary artery in the hilum, and the right pulmonary artery was then connected to the side of the homograft (Fig 5). In the 2 patients who required truncal valve replacement, the vertical incision used to expose the pulmonary orifices was extended through the valve annulus into the ventricular septal defect. The ventricular septum was then divided further

Fig 4. The honiograft is attached to the pulnionary artery bifurcation distally. The proximal Dacron extension is beveled to leave only a short rini of Dacron posteriorly. The ventricular septal defect is closed in the standard fashion.

Ann Thorac Surg 1989:47499-506

as described for aortoventriculoplasty to accommodate a large patch, which widened the annulus and permitted insertion of a 17-mm Omniscience disc prosthesis. Truncal valve replacement was performed ten days after the initial repair in I patient (patient 7) because of severe regurgitation and congestive heart failure. This patient also had evidence of peripheral embolization presumed to originate from thrombus on the ventricular septal defect patch where it contacted the regurgitant jet (echocardiographic diagnosis), although no evidence of intracardiac thrombus was found at reoperation. Valve replacement was performed at the same time as the truncus repair in another patient (patient 11). In the presence of substantial untreated truncal valve regurgitation, the clamp was left in position longer, generally until systemic temperatures reached at least 32" to 35°C to allow immediate effective ventricular contractions to eject the regurgitant volume.

Results There was 1 operative death (9%; 70% confidence limits, 3%-22%). The patient who died was a 3-week-old infant (patient 5) in whom Ventricular fibrillation developed soon after induction of anesthesia. Multiple attempts at defibrillation were unsuccessful, and cardiopulmonary bypass was rapidly established. Although repair was completed uneventfully, the patient could not be weaned from bypass. One late death occurred 3 months after operation. The infant (patient 10) had DiGeorge syndrome and probable graft versus host disease, and was readmitted because of gastroenteritis and dehydration. An unrecognized perforation of the subclavian vein occurring during central venous line insertion caused the patient's death. The remaining 9 patients have been followed for a mean of 21 months (range, 4 to 32 months). One patient (patient 9) required truncal valve replacement 6 months after

BOVEETAL TRUNCUS ARTERIOSUS IN NEONATES

Ann Thorac Surg 1989;47:499-506

503

Fig 5. (Patient 7.) The repair it7 the patient with noriconfluent pulmonary arteries was accomplished by first attaching the divided left pulmonary artery to the distal end of a 10-rnrn aortic hoinpyraft. The mobilized right pulmonary artery z c m then iinplanted into the side of the hoinograft.

repair for persistent congestive heart failure and failure to thrive. No other patient has required reoperation to date. The 2 additional patients with substantial untreated truncal valve regurgitation (patients 2, 4) have continued to tolerate the regurgitation well and are growing, although l (patient 2) has moderate cardiomegaly and congestive heart failure. No patient has complete heart block. Nine patients underwent detailed two-dimensional echocardiographic/Doppler studies 1 month to 29 months after repair (mean interval, 15 months). The findings are shown in Table 4. Left ventricular function was normal in each patient, with fractional shortening ranging from 29% to 56% (mean, 39%; normal, 228%).Truncal valve regurgitation was absent in 6 patients, trivial to mild in 1, and moderate in 2 patients. Conduit valve regurgitation was common (7/9), particularly in patients with homografts,

Table 4. Two-Dimensional EchocardiographylDoppler Study at F o ~ ~ o w - u ~ Patient No. 1 2 3 4 6 7 8 10 11

LV Fractional Conduit RV-PA Valve Gradient Shortening Truncal Valve @) Regurgitation Regurgitation (mm Hg) 36 34 34 39 43 47 29 56 36

0

2+ 0 2+ 0 0 1+ 0 0

1+ 1+

0 29 23 52” 35

1+

14

1+

0 0 0

0 1+

1+

0 1+

This patient had a 17-mm Hg gradient across the conduit and an additional 35-mm Hg gradient across a stenosis at the origin of the left pulmonary artery.

a

LV = left ventricular;

PA = pulmonary artery;

RV

=

right ventricle.

but was never more than trivial in any patient. Doppler estimates of right ventricle-pulmonary artery peak instantaneous gradients ranged from 0 to 52 mm Hg. In the patient with a 52-mm Hg gradient, the majority of the obstruction was located at the origin of the left pulmonary artery (35 mm Hg) with an additional 17 mm Hg across the conduit itself. Postoperative cardiac catheterization was performed in 2 patients. In 1 (patient l), catheterization was done 12 months after operation and revealed normal pulmonary artery pressure and resistance. There was no truncal valve regurgitation or residual shunt. The other patient (patient 9) was studied elsewhere 5 months postoperatively because of persistent congestive heart failure believed secondary to severe truncal valve regurgitation. Truncal valve replacement was subsequently performed at another institution. An additional study was obtained after valve replacement because the infant could not be weaned from ventilator support. There was no residual shunt, truncal (prosthetic) valve regurgitation, or left ventricular dysfunction. Pulmonary artery pressure and resistance were normal. Ventilator dependency was thought to be secondary to phrenic nerve palsy. This patient was eventually weaned from the ventilator and discharged from the hospital. Neither patient had evidence of conduit stenosis at recatheterization.

Comment It is now clear that infants with truncus arteriosus are best treated by early complete repair. Palliation with pulmonary artery banding offers no advantage and, even with modern surgical techniques, is unlikely to improve the mortality rates that can now be achieved with complete physiological repair. However, the exact timing of operation, particularly in neonates who have minimal or wellcompensated congestive heart failure, remains in doubt. A number of authors [8, 10, 11, 161 continue to believe that the best results are obtained if repair is delayed beyond 2 to 3 months of age. It is important to note,

504

BOVEETAL TRUNCUS ARTERIOSUS IN NEONATES

A

Ann Thorac Surg 1989;47:499-506

B

Fig 6 . (Patient 2.) ( A ) Posferoanterior and ( B ) lateral chest roentgenogram nrnde after repair roith a 12-rnwi Hancock conduit. Proper positioning of the valve away froni the sternum is demonstrated.

however, that medical treatment often gains very little and may further complicate the surgical risk if failure to thrive, infection, diminished ventricular function, or pulmonary vascular occlusive disease intervene. For a number of years, it has been the practice at The University of Michigan to perform complete repair a s early as possible for patients with truncus arteriosus. Although repair with a nonvalved conduit was advocated in the past because of the technical difficulties encountered when inserting valved porcine prostheses in small infants [13, 141, we now believe that the hemodynamic advantage of a valved conduit outweighs any potential disadvantages. With a few technical modifications, it is nearly always possible to place a 12-mm valved Dacron conduit even in tiny neonates (Fig 6). Furthermore, the current use of cryopreserved homografts offers a substantial technical improvement and the availability of conduits smaller than 12 mm if necessary. This policy of early repair with valved conduits in all patients with truncus arteriosus has resulted in a unique experience of repair mostly in neonates. Although small in number, our series still allows some meaningful conclusions regarding complete repair in this potentially high-risk age group. Of the 11 patients, 8 were 4 weeks of age or less and 5 had moderate to severe truncal valve regurgitation. Two of these patients underwent successful truncal valve replacement at 22 days and 6 days of age. The former patient also required simultaneous repair of nonconfluent pulmonary arteries. Despite these factors, there was only 1 death (9%) in our series. The cause of death is not readily apparent, and no explanation could be found for the ventricular fibrillation that occurred before

repair. The single late death was clearly noncardiac in cause, a s this patient had an entirely uneventful recovery after operation and evidence of an excellent hemodynamic repair (see Table 4, patient 10). Our method of repair differs little from previous reports from other centers. We prefer bicaval cannulation with low-flow cardiopulmonary bypass at 20°C. This has been possible in all patients, including the infant weighing 1.9 kg. A single dose of dilute blood cardioplegia is delivered, as this provides additional protection during the period of aortic cross-clamping [17, 181. No special precautions are taken in patients with major truncal valve regurgitation when replacement is not planned except to assure prompt initiation of cardiac action on reperfusion to eject the regurgitation volume of blood reaching the left ventricle. Truncal valve replacement should be performed at the time of complete reuair when the valve is severely regurgitant, particularly if stenosis is also present, regardless of patient age or size. In our experience, attempts to postpone valve replacement in this setting have been unsuccessful. Division of the truncal valve annulus with patch enlargement of the valve ring as described for aortoventriculoplasty was readily accomplished in 2 patients and allowed insertion oi a relatively large disc prosthesis. Possibly, insertion of an aortic homograft, either freehand or as an extended root replacement, may offer a substantial advantage over a mechanical valve [19]. However, the longevity of homogi-afts in the aortic position in the neonate remains unknown. Use of fresh and cryopreserved homografts to provide right ventricle-pulmonary artery continuity is a more tested procedure wii h great advantages over Dacron

Ann Thorac Surg 1989;47499-506

conduits containing porcine valves [20, 211. Avoidance of rigid prostheses and a greater technical ease of suturing with improved hemostasis are extremely desirable, particularly in neonates. Although mild regurgitation is common, in none of our patients has this been hemodynamically significant. Long-term durability is limited by valve calcification and neointimal proliferation, particularly when woven Dacron is used to lengthen the homograft [22]. Our use of knitted materials with little posterior rim to minimize the circumferential amount of Dacron may improve longevity. The residual gradients found in the homograft group (see Table 4) were thought to be secondary to the small size of the conduits themselves and, in one instance, to stenosis of a branch of the left pulmonary artery. No evidence of isolated proximal narrowing at the Dacron extension was found. In conclusion, this experience supports the notion that complete repair of truncus arteriosus should be performed as early as possible after diagnosis. Optimally, this would occur within the first month of life, regardless of symptoms. When the diagnosis is made at birth or even in utero and the neonatal pulmonary vascular resistance remains elevated, postponing repair for a few weeks to allow resistance to fall is beneficial. Cardiac catheterization may not be necessary if two-dimensional echocardiography/Doppler examination can identify both pulmonary arteries, demonstrate the anatomy of the aortic arch, and define the function of the truncal valve. It might also be necessary if specific quantification of pulmonary vascular resistance is required. Truncal valve replacement should be performed when severe regurgitation is present, especially when associated with truncal valve stenosis. The truncal valve should be left alone when regurgitation is less than severe. Right ventricle-pulmonary artery continuity is best established with a cryopreserved pulmonary or aortic homograft, although porcine heterograft conduits remain a satisfactory alternative 1231. This practice has resulted in an acceptable operative mortality (9%) despite the high percentage of neonates and the complexity of associated lesions.

Addendum Since the final preparation of this manuscript, 2 additional neonates have undergone successful repair of truncus arteriosus, both at 1 day of age. One patient also required replacement of the truncal valve because of severe insufficiency.

References 1. Calder L, Van Praagh R, Van Praagh 5, et al. Truncus arteriosus communis. Clinical, angiographic, and pathologic findings in 100 patients. Am Heart J 1976;92:22-38. 2. Tandon R, Hanck AJ, Nadas AS. Persistent truncus arteriosus: a clinical, hemodynamic, and autopsy study of 19 cases. Circulation 1963;28:105040. 3. Van Praagh R, Van Praagh S. The anatomy of common aorticopulmonary trunk (truncus arteriosus communis) and its embryologic implications. A study of 57 necropsy cases. Am J Cardiol 1965;16:40&25.

BOVEETAL TRUNCUS ARTERIOSUS IN NEONATES

505

4. Crupi G, Macartney FJ, Anderson RH. Persistent truncus arteriosus. A study of 66 autopsy cases with special reference to definition and morphogenesis. Am J Cardiol 1977;40:56978. 5. Ceballos R, Soto 8 , Kirklin JW, Bargeron LM Jr. Truncus arteriosus: an anatomical-angiographic study. Br Heart J 1983;49:589-99. 6. Juaneda E, Haworth SG. Pulmonary vascular disease in children with truncus arteriosus. Am J Cardiol 1984;54:1314LU.

7. Ebert PA, Turley K, Stanger P, Hoffman JIE, Heymann MA, Rudolph AM. Surgical treatment of truncus arteriosus in the first 6 months of life. Ann Surg 1984;200:4514. 8. Sharma AK, Brawn WJ, Mee RBB. Truncus arteriosus: surgical approach. J Thorac Cardiovasc Surg 1985;90:45-9. 9. Stegmann T, Oelert H, Luhmer I, Kallfelz HC, Borst HG. Surgical correction of truncus arteriosus type 1. Thorac Cardiovasc Surg 1982;30:163-6. 10. DiDonato RM, Fyfe DA, Puga FJ, et al. Fifteen-year experience with surgical repair of truncus arteriosus. J Thorac Cardiovasc Surg 1985;89:414-22. 1. Ziemer G, Luhmer I, Siclari F, Kallfelz HC. Truncus arteriosus type A3: complex repair with cryopreserved pulmonary homograft. Eur J Cardio-thorac Surg 1987;l:llO-5. 2. Behrendt DM, Kirsh MM, Stern A, Sigmann J, Perry B, Sloan H. The surgical therapy for pulmonary artery-right ventricular discontinuity. Ann Thorac Surg 1974;18:122-37. 3. Peetz DJ, Spicer RL, Crowley DC, Sloan H, Behrendt DM. Correction of truncus arteriosus in the neonate using a nonvalved conduit. J Thorac Cardiovasc Surg 1982;83:74%6. 4. Spicer RL, Behrendt D, Crowley DC, et al. Repair of truncus arteriosus in neonates with the use of a valveless conduit. Circulation 1984;7O(Suppl 1):2&9. 5. Collett RW, Edwards JE. Persistent truncus arteriosus. A classification according to anatomic types. Surg Clin North Am 1949;29:1245-70. 6. Kirklin JW, Barratt-Boyes BG. Truncus arteriosus. In: Cardiac surgery. New York: John Wiley & Sons, 1986:911-31. 17. Bove EL, Stammers AH, Gallagher KP. Protection of the neonatal myocardium during hypothermic ischemia: effect of cardioplegia on left ventricular function in the rabbit. J Thorac Cardiovasc Surg 1987;94:115-23. 18. Lynch MJ, Bove EL, Zweng TN, Fox MH, Boiling SF, Gallagher KP. Protection of the neonatal heart following normothermic ischemia: a comparison of oxygenated saline and oxygenated versus nonoxygenated cardioplegia. Ann Thorac Surg 1988;45:650-5. 19. Clarke DR. Extended aortic root replacement for treatment of left ventricular outflow tract obstruction. J Cardiac Surg 1987; 2(Suppl):121-8. 20. Kirklin JW, Blackstone EH, Maehara T, et al. Intermediateterm fate of cryopreserved allograft and xenograft valved conduits. Ann Thorac Surg 1987;44:59-06. 21. Kay PH, Ross DN. Fifteen years' experience with the aortic homograft: the conduit of choice for right ventricular outflow tract reconstruction. Ann Thorac Surg 1985;40:360-4. 22. Bull C, Macartney FJ, Horvath P, et al. Evaluation of longterm results of homograft and heterograft valves in extracardiac conduits. J Thorac Cardiovasc Surg 1987;94:12-9. 23. Boyce SW, Turley K, Yee ES, Verrier ED, Ebert PA. The fate of the 12 mm porcine valved conduit from the right ventricle to the pulmonary artery. J Thorac Cardiovasc Surg 1988;95: 201-7.

506

BOVEETAL TRUNCUS ARTERIOSUS IN NEONATES

Ann Thorac Surg 1989;47499-506

DISCUSSION DR KEVIN TURLEY (San Francisco, CA): I thank The Society for the privilege of discussing this paper and the authors for forwarding authors for the excellent results in this series of difficult patients. At the University of California, San Francisco, from 1975 through 1988, 217 patients less than 6 months of age underwent operation. The time frame for operation, that is, our recommended age for optimal elective repair, has decreased during that period from less than 6 months prior to 1980 to less than 3 months through 1984 to less than 6 weeks in the current era. The reason we thought operation was difficult in infants less than 6 weeks or a month old was the mismatch that we often see between a heterograft conduit (the smallest available being a 12-mm conduit) and the right ventricle of these small infants. The use of allografts has changed this situation substantially, and early repair in the first month of life is both feasible and desirable. In reviewing our heterograft series, we have found that obstructions occurred in the Dacron conduits at multiple levels, with pseudointimal formation, valve degeneration, and involvement of the proximal conduit in the obstructing process. In one instance, dissection of the pseudointima of the proximal conduit resulted in obstruction. For this reason, our allografts are placed with the muscle ridge posteriorly, leaving the mitral leaflet to form part of the arterial patch. A Teflon felt bolster is placed posteriorly, the suture is taken down the ventricular muscle, about one third of the way through the ventricular septa1 defect patch, and then back and u p through the muscle, placing this muscle within the ventricular cavity. The mitral valve is then augmented with a piece of pericardium or allograft arch tissue, placing the allograft valve perpendicular to the ventricular wall, thereby eliminating major valvular regurgitation while avoiding the need for any Dacron extension with the obstructive consequences. I ask the authors, first, about this concern of the placement of the circumferential Dacron graft proximal to the allograft. Second, I ask that they expound on the case of graft versus host reaction 3 months postoperatively in the patient with DiGeorge syndrome. We have not encountered this problem, and it is one

of great concern. Finally, it is stated in the manuscript that the authors would operate in early infancy when the patient is first seen. We see a number of patients with pulmonary hypertension in the first week of life. Would the authors operate at that point or wait for the development of congestive heart failure and high flow to initiate an operative approach? DR BOVE: I thank Dr Turley for his comments. I will reply to his questions beginning with the Dacron extension. We, too, have had concerns about the use of Dacron to extend these conduits. 'We prefer the pulmonary homograft over the aortic homograft, which, of course, does not give us the option of using the anterior leaflet of the mitral valve to extend it over the ventriculotomy. What we have utilized, not only in the patients with truncus but in all of our patients who have had homografts, are proximal extensions of the conduit with knitted Dacron baked in the autoclave either with autologous blood or albumin. We have not had hemorrhage problems with that method, although it remains to be seen if the use of knitted material versus woven material will decrease the incidence of peel formation and late obstruction. Regarding the case of graft versus host disease in the 1 infant with DiGeorge syndrome, I have to say that the diagnosis was never firmly established. The infant was admitted to the cardiology service and died as a result of the mishap with central venous line placement before the diagnosis was established. But it was suspected by the admitting cardiologist in view of the infant's other syndrome. Dr Turley's point about waiting until the infant goes into congestive failure is a good one, and 1 concur. I d o not see any real need to operate on these infants immediately if they are not in congestive heart failure; rather, I would wait until congestive heart failure begins to be seen, which in our experience does not generally take very long. 13ut if we see an infant who is 1 or 2 days of age and is not in congestive failure, we would probably wait a week or so. Most infants seen with severe failure at that age, however, d o have important truncal valve regurgitation. In fact, we have seen an inordinately high percentage of that in our patients, and it really complicates the surgical management.