Results of a policy of primary repair of truncus arteriosus in the neonate

Results of a policy of primary repair of truncus arteriosus in the neonate

Results of a policy of primary repair of truncus arteriosus in the neonate Although the early mortality for repair of truncus arteriosus has decreased...

1MB Sizes 0 Downloads 27 Views

Results of a policy of primary repair of truncus arteriosus in the neonate Although the early mortality for repair of truncus arteriosus has decreased in the modem era, routine correction in the neonate has not been widely adopted. To assess the results of our protocol of early repair, we reviewed 46 neonates and infants undergoing repair of truncus arteriosus at the University of Michigan Medical Center from January 1986 to January 1992. Their ages ranged from 1 day to 7 months (median 13 days) and weights from 1.8 kg to 5.4 kg (mean 3.1 kg). Repair was performed beyond the first month of life in only 8 patients, because of late referral in 7 and severe noncardiac problems in 1. Associated cardiac anomalies were frequently encountered, the most common being interrupted aortic arch (n = 5), nonconfluent pulmonary arteries (n = 4), hypoplastic pulmonary arteries (n = 4), and major coronary artery anomalies (n = 3). Truncal valve replacement was performed in 5 patients with severe regurgitation, 3 of whom also had truncal valve systolic pressure gradients of 30 mm Hg or more. The truncal valve was replaced with a mechanical prosthesis in 2 patients and with a cryopreserved homograft in 3 patients. Right ventricle-pulmonary artery continuity was established with a homograft in 41 patients (range 8 mm to 15 mm), a valved heterograft conduit in 4 (range 12 mm to 14 mm), and a nonvalved polytetrafluoroethylene tube in the remaining patient (8 mm). There were 5 hospital deaths (11 %, 70 % confidence limits 7 % to 17 %). Multivariate and univariate analyses failed to demonstrate a relationship between hospital mortality and age, weight, or associated cardiac anomalies. Only 1 death occurred among 9 patients with interrupted aortic arch or nonconfluent pulmonary arteries. Hospital survivors were followed-up from 3 months to 6.3 years (mean 3 ± 0.4 years). Late noncardiac deaths occurred in 3 patients, all within 4 months after the operation. Actuarial survival was 81 % ± 6% at 90 days and beyond. Despite the prevalence of major associated conditions, early repair has resulted in excellent survival. We continue to recommend repair promptly after presentation, optimally within the first month of life. (J THORAC CARDIOVASC SURG 1993;105:1057-66)

Edward L. Bove, MD,a Flavian M. Lupinetti, MDa (by invitation), Ara K. Pridjian, MDa (by invitation), Robert H. Beekman III, MD b (by invitation), Louise B. Callow, RN" (by invitation), A. Rebecca Snider, MD b (by invitation), and Amnon Rosenthal, MD b Ann Arbor, Mich.

Luncus arteriosus is a complex congenital anomaly that imposes a severe hemodynamic burden early in life. Congestiveheart failure often occurs within days or weeks after birth as pulmonary vascular resistance falls and From the Department of Surgery, Section of Thoracic Surgery," and the Department of Pediatrics, Division of Pediatric Cardiology," The University of Michigan School of Medicine, Ann Arbor, Mich. Read at the Seventy-second Annual Meeting of The American Association for Thoracic Surgery, Los Angeles, Calif., April 26-29, 1992. Address for reprints: Edward L. Bove, MD, 2120 Taubman Center, Box 0344, 1500 East Medical Center Dr., Ann Arbor, MI 48109. Copyright

1993 by Mosby-Year Book, Inc.

0022-5223/93 $1.00

+ .10 12/6/41632

pulmonary blood flow dramatically increases. When the condition is further complicated by associated lesions such as truncal valve regurgitation, interrupted aortic arch, or coronary and pulmonary artery anomalies, symptoms are more severe and surgical intervention may be urgently required. Although primary repair is preferable to palliation in the treatment of infants with truncus arteriosus, doubt has been raised by some that this approach should be electively adopted in the neonate who may have minimal or even no symptoms.': 2 It is common, however, that this very group of patients will demand early intervention because of severe congestive heart failure, particularly when associated conditions are present. 1057

10 5 8

The Journal of Thoracic and Cardiovascular Surgery June 1993

Bove et al.

Table I. Age and weight distribution in 46 patients with truncus arteriosus No.

Age (days)

<7 7-14

15-21 22-28 >28

Weight (kg) <2

12 13 6

7 8

3

2-4

40

>4

3

Furthermore, even in infants with less pronounced heart failure, medical management often adds little because growth may be poor and additional complications, such as infection or early pulmonary vascular disease, may intervene. After the pioneering work of McGoon, Rastelli, and Ongley.' a number of groups, 1,2,4,5 including our own," have reported improved results with early primary repair for the treatment of infants with truncus arteriosus. The majority of these reports, however, have focused on patients beyond the neonatal period.': 2, 4, 5 Some have continued to recommend palliation for those patients requiring intervention within the first month of life when medical management fails to control congestive heart failure." We, therefore, reviewed and updated our entire experience with surgical intervention for truncus arteriosus since our earlier report in 1989.6 Our policy of early primary repair, ideally within the first month of life, was followed whenever possible. The results of this approach on early and late survival in a large group of neonates and infants (the majority less than I month of age) form the basis of this report. Patients and methods The records of aliSO patients who presented to the C. S. Mott Children's Hospital, University of Michigan Medical Center, between January 1986 and January 1992 with the diagnosis of truncus arteriosus were reviewed. Three patients died without operative intervention and I had received prior pulmonary artery banding at another institution. These 4 patients were excluded from the analysis. None of the remaining 46 patients received any prior palliative procedure and all underwent primary complete repair. Some of these patients were included in a previous report from this institution." There were 26 boys and 20 girls. Their ages ranged from I day to 7 months (median 13 days). Repair was accomplished beyond 28 days of age in only 8 patients because of late referral in 7 and significant noncardiac problems in I (Table 0. Weights ranged from 1.8 kg to 5.4 kg (mean 3.1 kg). Before repair, all patients underwent complete two-dimensional and Doppler echocardiography. Car-

Table II. Associated cardiac diagnoses Diagnosis

Interruptedaortic arch Nonconf'luent pulmonary arteries Hypoplastic pulmonary arteries InfracardiacTAPVR Restrictive VSD Major coronary anomaly Slitlikeostia Multipleintramural RCA branches LCA from LPA

No.

5 4

4 I I

3

TAPVR, Total anomalous pulmonary venous return; VSD. ventricular septal defect; RCA, right coronary artery; LCA, left coronary artery; LPA. left pulmonary artery.

diac catheterization was also performed in 9 patients. Catheterization was generally used to quantify pulmonary vascular resistance in older patients, to image nonconfluent pulmonary arteries when they were not well seen by echocardiography, and to further quantify the degree of truncal valve regurgitation in selected cases. Associated cardiac lesions are listed in Table II. Sixteen patients had one or more associated lesion. The most common cardiac conditions consisted of interrupted aortic arch in 5 patients, nonconfluent pulmonary arteries in 4, hypoplastic pulmonary arteries in 4 and major coronary artery anomalies in 3 patients. The cross-sectional area of the pulmonary arteries in the 4 patients with hypoplastic pulmonary arteries was determined by the method of Nakata and associates.f In I patient with nonconfluent origins of the right upper and right lower pulmonary arteries, the Nakata index was obtained by summing the cross-sectional area of these two vessels separa tely to that of the left pulmonary artery. Hypoplasia was defined as a Nakata index of 250 mm 2/m2 or less. The Nakata indexes in these 4 patients ranged from 87 mm 2/m2 to 224 mm 2/m2 (mean ISS mrn- 1m2). One patient with hypoplastic pulmonary arteries also had obstructed infracardiac total anomalous pulmonary venous return. Truncal valve regurgitation was absent or trivial in 25 patients, mild to moderate in 16, and severe in the remaining 5 patients. Peak instantaneous truncal valve systolic gradients were present by Doppler examination in 7 patients, ranging from 16 to 71 mm Hg (mean 38 mm Hg), Severe truncal valve regurgitation was also present in 3 of these patients (gradients: 30,38, and 71 mm Hg). Either the complete or partial form of DiGeorge's syndrome was present in 7 patients ( 16%). Among the 42 patients with confluent pulmonary arteries, type I anatomy according to the classification of Collett and Edwards? was present in 17 patients, type 2 in 10, and the remaining IS patients had a mixture between those two anatomic types. Among the 4 patients with nonconfluent pulmonary arteries, in 2 the right pulmonary artery arose from the left side of the ascending aorta and the left pulmonary artery arose from the ductus arteriosus (Fig. I). In I patient, the right pulmonary artery arose from the innominate artery and the left pulmonary artery from the ductus arteriosus. In the remaining patient, the right upper lobe pulmonary artery took origin from the ductus arteriosus and the right lower lobe from a right-sided descending thoracic aorta. The left pulmonary artery arose directly from the proximal left coronary sinus of Valsalva and the left coronary artery branched from the pulmonary artery.

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 6

Bove et ai.

10 5 9

A

Fig. 1. Pulmonary artery anatomy in the four patients with nonconfluent pulmonary vessels. In two patients (A~ the right pulmonary artery arose from the usual position on the ascending aorta and the left from the ductus arteriosus. In one patient (B), the right pulmonary artery arose from the innominate artery and the left from the ductus arteriosus. In the final patient (0. the right upper lobe pulmonary artery arose from an occluded ductus arteriosus and the right lower lobe from a right-sided descending thoracic aorta. The left pulmonary artery arose directly from the proximal left coronary sinus of Valsalva and the left coronary artery branched from the pulmonary artery.

Surgical techniques. For the operative repair we used standard techniques of deep hypothermia with either low-flowcontinuous cardiopulmonary bypass or intermittent periods of circulatory arrest, as we6 have previously reported. Venous return was generally provided by a single right-angle DLP cannula (I2F, DLP Corporation, Grand Rapids, Mich.). A single dose of cold dilute blood cardioplegic solution (30 ml/kg) was used in all patients, generally repeated when aortic crossclamp times exceeded 45 to 60 minutes. Aortic crossclamp times ranged from 23 to 103 minutes (mean 65 minutes). Circulatory arrest was used in 15 patients and ranged from 7 to 78 minutes (mean 39 minutes). The pulmonary arteries were excised from the ascending aorta with a generous cuff of arterial wall whenever possible. Often, a brief period of circulatory arrest was required for preciseexposure. When the truncal valveor left coronary artery was close to the pulmonary artery orifice, removal of additional tissue along the inferior edge was not possible. In these situations, removal of the pulmonary arteries was tailored to leave sufficient arterial wall by scalloping into the pulmonary artery confluence to prevent injury to adjacent structures. The defect remaining in the ascending aorta was then repaired with a patch to prevent distortion of the truncal valve or coronary artery. Cardiopulmonary bypass was resumed at this point and the distal anastomosis of the conduit was performed. Right ventriclepulmonary artery continuity (Table Ill) was established with either an aortic or pulmonary cryopreserved homograft in 4 I patients (range 8 mm to 15 mrn), a valved heterograft conduit in 4 patients (range, 12 mm to 14 mm), and a nonvalved polytetrafluoroethylene (PTFE) tube* in I patient (8 mm). A vertical right ventriculotomy was then made and the ventricular septal defect (VSD) was closed with a patch. In I instance, enlargement of the defect was necessary. An additional period of circulatory arrest was often required to properly expose the VSD for accurate closure. Large atrial septal defects were closed through a right atriotomy, but when only a patent fora*Gore-Tex tube, registered trademarkofW. L. Gore & Associates, Inc., Naperville. Ill.

Table III. Conduit type and size Number Type Aortic homograft Pulmonary homograft Heterograft Nonvalved PTFE Size (mm) 8-10 11-15

27 14 4 I 14

32

men ovale was present it was left open. The proximal anastomosis of the conduit was then constructed during systemic rewarming. When truncal valve regurgitation was present, the crossclamp was left in position until sufficient rewarming had occurred to allow for immediate cardiac contraction to prevent left ventricular distention. Early in our experience we extended the homograft proximally with a Dacron conduit, trimming the posterior wall of the Dacron fabric to within 2 or 3 rnm of the homograft to fashion a proximal hood for the ventricular anastomosis. More recently we have avoided all Dacron material, suturing the homograft directly to the ventriculotomy and augmenting the anastomosis with a patch of excess homograft tissue or PTFE. Cardiopulmonary bypass was discontinued when rectal temperature exceeded 36° C, most often with low-dose inotropic support consisting of either amrinone or dobutamine. Mild to moderate degrees of respiratory alkalosis were used and an infusion of fentanyl was begun and continued in the intensive care unit if the pulmonary vascular resistance was elevated or the patient was older than 2 weeks of age. Special circumstances Truncal valve replacement. Five patients (I 1%) underwent replacement of the truncal valve for severe valvular insufficiency (age range, I to IS days). In 2 patients, a 17 mm Omniscience mitral prosthesis (Medical Inc., Inver Grove Heights, Minn.) was implanted in an inverted position. Division of the truncal

I060

The Journal of Thoracic and Cardiovascular Surgery June 1993

Bove et al.

---- -----------TJ---..._-

----------------

~~----!-- Actual

coronary ostium

"""":----~~ Intramural coronary course

;;-;'ir'------!-- Level of incision

~-~-"Apparent" A'J€I.~--"';:""'-..l...-

coronary ostium

Usual coronary ostium

Fig. 2. The distribution of the right coronary artery (larger vessel) and two major right ventricular branches is demonstrated in the patient with an intramural course. Although the external appearance indicated that the apparent ostia of all three vesselswere displaced superiorly, the actual coronary ostia proved to be even considerably higher because the vessels took an intramural course (inset).

A Patch

~ ~ I

I

\\~~I.III

) I

I

~ It

I

I

,

Fig. 3. The technique used in the repair of the patient depicted in Fig. 2 is shown. Because the coronary arteries were transected across their intramural course (A), the internal wall of each vessel was opened inferiorly to provide an adequate lumen. A patch was then sutured to the outer wall to reconstruct the defect in the ascending aorta (B). valve anulus and further enlargement of the VSD (aortoventriculoplasty) was required in both patients to accommodate the prosthesis.!" The remaining 3 patients underwent truncal root replacement with a cylinder of cryopreserved homograft ranging in size from 12 mm to 15 mm (aortic in 2 and pulmonary in I). In each of these 3 patients the coronary arteries were excised with a button of arterial wall and reimplanted into the homograft. Interrupted aortic arch. Interrupted aortic arch was repaired in five patients (I 1%). Their ages ranged from 3 to 71 days, although only one was older than 8 days. The arch was interrupted between the left carotid and left subclavian arteries in each (type B). The aortic cannula was placed low in the non-

coronary sinus of Valsalva and the branch pulmonary arteries were occluded at the onset of cardiopulmonary bypass. Wide mobilization of both the ascending and descending aorta was accomplished during systemic cooling, after which a primary anastomosis was performed with continuous absorbable suture material. Care was taken not to anastomose the descending aorta too far proximally on the ascending aorta to minimize the risk of compression of the pulmonary arteries or the left mainstem bronchus. Nonconfiuent pulmonary arteries. In 4 patients (9%) whose ages ranged from 5 to 16 days, one or more pulmonary arteries were discontinuous. In 2 patients, the right pulmonary artery arose from the usual position on the ascending aorta and the left

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 6

Bove et of.

I06 I

100 80

Survival (%) +/- S.E.M.

60 40 20 30 26

23

12

7

Number of Patients

2

o+---....,..-----,,....----r-----r20 40 o 60 80

Followed Through Interval

Months After Operation

Fig. 4. Actuarial survival for the 46 neonates and young infants undergoing repair of truncus arteriosus. Survival was 87% ± 5% at I month and 81% ± 6% at 3 months and beyond. S.E.M., Standard error of the mean. pulmonary artery arose from a normally positioned ductus arteriosus (Fig. I). In the first patient, repair was accomplished by suturing the distal end of an aortic homograft directly to the left pulmonary artery in the hilum.P: II The right pulmonary artery was then connected to the side of the homograft. A primary anastomosis between the divided ends of the mobilized right and left pulmonary arteries was performed in 2 patients; the homograft from the right ventricle was attached to the newly fashioned and spatulated pulmonary artery anastomosis, In the final patient, the complexity of the pulmonary artery anatomy required a preliminary right thoracotomy before the complete repair.'? In this patient, the right upper lobe pulmonary artery arose from the undersurface of a right aortic arch in the position of the ductus arteriosus and the right lower lobe vessel from the mid-descending right-sided thoracic aorta. The left pulmonary artery arose from the left coronary sinus of Valsalva with the left coronary artery arising from the left pulmonary artery, a few millimeters from its origin, At the first-stage procedure, the right upper and lower lobe vessels were mobilized, divided from their aortic origins and sutured to the ascending aorta. Then, through a midline approach, the left pulmonary artery was transected distal to the origin of the coronary artery and sutured to the ascending aorta, opposite the right-sided vessels. A cylinder of aorta containing all the pulmonary artery branches was then removed, replaced with a PTFE interposition graft, and connected to a cryopreserved homograft from the right ventricle. Coronary artery anomalies. Important coronary artery anomalies were present in three patients (7%). In one patient, noted earlier, the left coronary artery arose from the left pulmonary artery. In another patient, both coronary ostia were slitlike and arose from within the commissures of the truncal valve. In this patient, the valve was severely regurgitant and the leaflets were thickened and myxomatous, Truncal root replacement was required as described earlier and the valve tissue was excised in an effort to provide unobstructed flow to the coronary arteries during reimplantation. The final patient had three separate ostia for the right coronary artery and its major ventricular branches, all taking an intramural course and opening within the pulmonary artery (Figs. 2 and 3). The posterior wall

of each of the coronary vesselswas opened to enlarge their ostia and additional pulmonary artery wall was left on the aorta by scalloping the pulmonary artery during its removal. Statistical analysis. Follow-up was obtained by direct patient contact or discussion with the referring cardiologist and is complete. Survivorship was analyzed according to standard actuarial methods, 13 The following risk factors were evaluated: age, weight, the presence of DiGeorge's syndrome, interrupted aortic arch, major coronary anomalies, nonconfluent or hypoplastic pulmonaryarteries, and severe truncal valve regurgitation necessitating truncal valve replacement or the use of a heterograft valved conduit. Their influence on survival was assessed by univariate (Student's t test or Fisher's exact test) techniques, A logistic regression analysis involving both categoric and noncategoric risk factors was used to construct a model to predict the probability of being alive after repair, Full and reduced models and a stepwise approach were performed to assess any possible influence of each risk factor individually and, when determined simultaneously, in combinations on the model. Mean values are reported with their standard errors,

Results

Survival. There were five hospital deaths (11%, 70% CL * 7% to 17%), In one patient, intractable ventricular fibrillation occurred soon after the induction of general anesthesia. Although bypass was rapidly established and the repair completed, satisfactory cardiac function was never regained. This patient also had a restrictive VSD that had to be enlarged, In two patients, both of whom had sudden circulatory collapse within hours of repair, coronary artery occlusion was the probable cause. In one, compression of the left coronary artery by the conduit was suspected, The other patient underwent truncal root replacement with reimplantation of the coronary arteries. Both coronary ostia were slitlike and partially obstructed *CL

=

Confidence limits.

The Journal of Thoracic and Cardiovascular Surgery June 1993

1 0 6 2 Bove et al.

Table IV. Univariate risk factor analysis Mortality Associated lesion

n

None Interrupted aortic arch Nonconfluent PAs Hypoplastic PAs Truncal valve replacement Coronary artery anomaly Patient tota1*

30 5 4 4 5 3 16

No.

%

70% CL (%)

p Value

2 I 0

7

3-13 8-43 0-21 10-51 8-43 13-62 11-31

0.46 1.00 0.39 0.46 0.30 0.32

It

I:j:

':\:

3

20 0 25 20 33 19

PAs, Pulmonary arteries; CL, confidence limits. "The 21 associated lesions were present in 16 patients: one patient with interrupted aortic arch also had hypoplastic PAs; one patient with hypoplastic PAs also had a truncal valve replacement; and one patient each with nonconftuent PAs also had truncal valve replacement, hypoplastic PAs, and a coronary artery anomaly. t Also had obstructed infracardiac total anomalous pulmonary venous return. :l:5ame patient.

by the leaflets of the thickened, myxomatous truncal valve. An additional patient who died 17 days after repair also had obstructed total anomalous pulmonary venous return below the diaphragm and severe pulmonary artery hypoplasia (Nakata index 87 mm2/m2). Fenestration of the VSD patch was performed because of persistent suprasystemic pulmonary artery pressure but was unsuccessful. The final patient was older than 2 months old and also had interrupted aortic arch. A pulmonary hypertensive crisis led to his death 24 hours after repair. Hospital survivors have been followed-up from 3 months to 6.3 years (mean 3 ± 0.4 years). There were three additional late deaths, none of which was directly cardiac related. In one patient with DiGeorge's syndrome, death resulted from sepsis. Another patient, also with DiGeorge's syndrome, died of inadvertent perforation of a subclavian vein with a central venous catheter during a subsequent hospital admission for noncardiac problems. The final patient died of airway complications related, in part, to an anomalous right upper lobe bronchus. Actuarial survival (Fig. 4) was 87% ± 5% at 1 month and 81% ± 6% at 3 months and beyond. There was no additional mortality after the fourth postoperative month. Risk factor analysis. The 41 hospital survivorsdid not differ from the 5 patients who died early with respect to age, weight, the presence of DiGeorge's syndrome, interrupted aortic arch, nonconfluent or hypoplastic pulmonary arteries, coronary anomalies, or the need for truncal valve replacement (Table IV). Even when the 16 patients with at least one major associated lesion were considered together as a group (because anyone lesion occurred infrequently), no statistically significant increase in early mortality was evident in this series. Among those 16 patients with one or more associated cardiac lesions, there were 3 early deaths (19%, 70% CL 11% to 31%). In contrast, there were 2 early deaths among the remaining 30

patients without a major associated cardiac lesion (7%, 70% CL 3% to 13%) (p = 0.33, Fisher's exact test). Although the presence of DiGeorge's syndrome had no statistically significant effect on either early or late mortality in this series, 2 of the 3 late deaths occurred in patients with this syndrome, 1 of which resulted from sepsis. The risk factors provided a poor model to predict mortality, and none was found to be predictive in the multivariate logistic analysis. Furthermore, stepwise analysis revealed that none of the variables met the 0.05 significance level for entry into the model. Reintervention. Conduit replacement has been performed in fivepatients from 12 months to 52 months after repair (mean 35 ± 8 months). One of these patients had a heterograft placed initially and underwent reoperation 52 months later. The remaining four patients had homografts at the original procedure and were reoperated on at a mean of 31 months (range 12 to 50 months). In the patient undergoing reoperation at the shortest time interval from the original procedure (12 months), patch enlargement of diffusely small central pulmonary arteries was the major indication for operation. The smallest conduit in this series, an 8 mm aortic homograft, was electively replaced 41 months after implantation. The mean homograft size at the initial repair for all five patients was 11 ± mm and was increased to 19 ± 1 mm. Four additional patients underwent reoperation. In one, obstruction at the proximal conduit anastomosis to the right ventricle resulting from peel formation along the Dacron extension was successfully revised 13 months after operation. Another patient underwent successful truncal valve replacement 6 months after repair. The remaining two patients underwent early reoperation to close a residual VSD in one and to fenestrate the VSD patch in the patient with suprasystemic right ventricular

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 6

pressure, hypoplastic pulmonary arteries, and obstructed total anomalous pulmonary venous return. Two patients underwent percutaneous balloon dilation of residual conduit to branch pulmonary artery stenoses at 10 and 20 months after repair.

Discussion Although primary correction of a number of complex congenital heart defects in the neonate and young infant can now be readily accomplished, doubt still persists regarding the optimal timing for repair in patients with truncus arteriosus. I. 2. 4. 5. 14 In an earlier report from this institution, we advocated prompt repair, preferably within the first month oflife, even for those neonates that were free of symptoms at the time of presentation.f The prevalence of major associated anomalies, such as truncal valve regurgitation, nonconfluent pulmonary arteries, and interrupted aortic arch, mandated early operative intervention in many of our patients, often as gravely ill neonates. The improved survival achieved in these patients, in addition to the fact that the seemingly stable condition of infants with uncomplicated truncus arteriosus often deteriorated rapidly, led us to recommend that this anomaly always be repaired as soon in life as possible. However, primary elective repair in the neonate has still not been widely accepted and palliation by internal or external banding of the pulmonary artery, particularly for the neonate, is preferred in some centers.I: 14 The results achieved in this current larger experience, however, continue to favor the policy of elective neonatal repair. Although the surgical technique used for repair has changed very little over the past few years/" 6 there are a number of operative details that deserve emphasis. Removal ofthe pulmonary arteries from the truncus must be done carefully because the distance from the pulmonary artery origin to the ostium of the left coronary artery and to the attachments of the truncal valve may be quite short. In patients with type 2 anatomy, a small counterincision on the anterior wall of the truncus is often helpful in determining the precise location for the incision used to remove the pulmonary confluence. It is occasionally necessary to tailor this incision around the coronary artery or the commissure of the truncal valve to avoid injuring these structures. The wide variety in origin and distribution of the coronary arteries must also be emphasized.P: 16 A single coronary artery has been observed in as many as 18% of patients with truncus arteriosus, and the anterior surface of the right ventricle is crossed by a coronary artery in 13%.17 Because of the considerable impact such anomalies may have on the details of surgical repair, both preoperative and intraoperative assessments of coronary artery anatomy are important.

Bove et al.

I063

The decision to replace the truncal valve is often influenced by the desire not to further complicate an already complicated procedure. However, truncal valve replacement can be accomplished with a low risk and should be performed when severe regurgitation is present. We prefer to use a cryopreserved homograft as reported by Elkins and associates. IS An appropriately sized homograft can often be inserted without aortic root enlargement and the coronary arteries can be reimplanted by means of standard techniques. Although the number of patients is small and the follow-up is short, it appears from both this experience and that gained from other infants in whom we have performed aortic valve replacement that aortic and pulmonary homografts function equally well. The presence of nonconfluent pulmonary arteries can add significantly to the operative procedure and can necessitate individual technical modifications depending on the specific anatomy. It is perhaps even more important in this group of patients to operate early in life because delay may increase the risk of development of pulmonary vascular disease. 19, 20 A direct anastomosis between the branches of the pulmonary arteries can generally be constructed, even when one or both intrapericardial branches are absent. I I , 12,21 Direct anastomosis is preferable to using additional conduit material on the peripheral pulmonary arteries to reduce the risk of stenosis and simplify subsequent reoperation. When this is not possible, using a cylinder of excised ascending aorta to unifocalize the pulmonary arteries to autologous tissue allows the repair to be completed, leaving all anastomoses and prosthetic conduits readily accessible from the midline. An anterior approach to repair of truncus arteriosus and interrupted aortic arch has been advocated by others. 19 , 22 In their report, Fujiwara and colleagues'? advocated placement of the conduit to the right of the ascending aorta. In contrast, we have found it preferable to direct the conduit to the left of the aorta to avoid compression under the sternum. In addition, we and others have found that it is also important to avoid suturing the descending aorta too low on the ascending aorta because obstruction of the left pulmonary artery or left main-stem bronchus may result. 22 Risk factor analysis failed to demonstrate that age, weight, the presence of DiGeorge's syndrome, repair of interrupted aortic arch, reconstruction of nonconfluent pulmonary arteries, major coronary anomalies, or replacement of the truncal valve were associated with an increased operative mortality. Despite the absence of a statistically significant difference in the operative mortality, however, it remains our impression that these risk factors do complicate the perioperative course. Further-

10 6 4

The Journal of Thoracic and Cardiovascular Surgery June 1993

Bove et at.

more, although the overlapping confidence limits indicate that the increased risk of death when the entire group of patients with at least one risk factor was considered as a whole could have been due to chance alone, there is a possibility that the risk is, in fact, increased (Table IV). The low risk among the patients without risk factors (7%, 70% CL 3% to 13%) lends further support to the policy of elective neonatal repair, because this is the group in whom operative timing may be flexible. Additionally, it is important to note that only one death occurred among the nine patients with interrupted aortic arch or nonconfluent pulmonary arteries, two conditions that would generally dictate urgent repair. The influence of pulmonary artery hypoplasia is difficult to assess in this series because only four of our patients met our arbitrarily defined criteria of a Nakata index of 250 mm- I m 2 or less. It is clear, however, that hypoplasia significantly influenced the course in two children. The patient with the smallest index (87 mm2/m2) also had obstructed total anomalous pulmonary venous return and died with suprasystemic pulmonary artery pressure. Another patient with a Nakata index of 103 mm 2/m2had non confluent pulmonary arteries and continues to have near-systemic right ventricular pressure. This patient is scheduled to undergo balloon dilation and possible stenting. The remaining two patients (Nakata indexes of 206 and 224 mm 2/m 2) have both had satisfactory hemodynamics, although one required augmentation of the central pulmonary arteries and conduit replacement at reoperation 12 months after repair. Thus, although this study is not adequate to define the lowest pulmonary artery size compatible with successful complete repair, it appears to be below 200 mm2/m2. The insertion of heterograft conduits failed to predict increased risk, but the use of homografts has nonetheless been an important technical improvement, particularly in extremely small patients. The tissue is easier to handle, is more hemostatic, and is available in smaller sizes than the commercially prepared heterografts. It remains to be seen, however, ifhomografts will significantly prolong the time to reoperation. To date, only five of our patients have required conduit replacement. Reoperation was performed at a mean of 31 months after repair in the four patients with initial homografts (range 12 to 50 months) and at 52 months after repair in the one patient with an initial heterograft. The smallest homograft in our series (8 mm) was inserted in a 1.8 kg neonate and was electively replaced more than 3 years later. In conclusion, survival in this series was 93% among patients without associated conditions (28/30). Although there was an increased risk when the patients with associated lesions were considered as a group, this may have

been due to chance alone. Furthermore, there was no effect of individual defects on the risk of repair. In the presence of severe truncal valve regurgitation, prompt repair with truncal valve replacement still offers excellent survival. Palliation with pulmonary artery banding has no role in the treatment of patients with this lesion, even in the presence of significant associated cardiac anomalies. Because of the consequences of early severe congestive heart failure and pulmonary vascular disease, we currently advocate early primary repair of truncus arteriosus in the neonate. We acknowledgeM. Anthony Schork, PhD, for hisassistance with the statistical evaluation of the datao

1.

2.

3.

4.

REFERENCES Di Donato RM, Fyfe DA, Puga FJ, et al. Fifteen-year experience with surgical repair of truncus arteriosus. J THoRAc CARDIOVASC SURG 1985;89:414-22. Sharma AK, Brawn WJ, Mee RBB. Truncus arteriosus. J THORAC CARDIOVASC SURG 1985;90:45-9. McGoon DC, Rastelli GC, Ongley P A. An operation for the correction of truncus arteriosus. JAM A 1968;205:6973. Pearl JM, Laks H, Drinkwater DC Jr, et al. Repair of truncus arteriosus in infancy. Ann Thorac Surg 1991;

52:780-6. 5. 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 Thorac Surg 1984;200:451-6.

6. Bove EL, Beekman RH, Snider AR, et al. Repair of trun-

cus arteriosus in the neonate and young infant. Ann Thorae Surg 1989;47:499-506. 7. Young JN, Pancastelli MC, Harrell JE Jr, Hardy C, Ahearn EN, Ecker RR. Internal banding for palliation of truncus arteriosus in the neonate. Ann Thorac Surg 1989; 47:620-2. 8. Nakata S, Imai Y, Takanashi Y, et al. A new method for

9.

10.

II.

12.

the quantitative standardization of cross-sectional areas of the pulmonary arteries in congenital heart diseases with decreased pulmonary blood flow. J THoRAc CARDIOVASC SURG 1984;88:610-9. Collett RW, Edwards JE. Persistent truncus arteriosus: a classification according to anatomic types.Surg Clin North Am 1949;29: 1245-70. Konno S, Imai Y, Ilida Y, et al. A new method for prosthetic valve replacement in congenital aortic stenosis associated with hypoplastia of the aortic valvering. J THORAC CARDIOVASC SURG 1975;70:909-17. Ziemer G, Luhmer I, Siclari F, KallfelzHC Truncus arteriosus type A3: complex repair with cryopreserved pulmonary homograft. Eur J Cardiothorac Surg 1987; I:II 0-5. Shanley CJ, Lupinetti FM, Shah NL, Beekman RH. Crowley DC, Bove EL. Primary unifocalization of absent

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 6

intrapericardial pulmonary arteries in the neonate. J THORAC CARDIOVASC SURG [In press]. 13. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457-81. 14. Singh AK, de Leval MR, Pincott JR, Stark J. Pulmonary artery banding for truncus arteriosus in the first year oflife. Circulation I976;54(Suppl):III 17-19. 15. Suzuki A, Ho SY, Anderson RH, Deanfield JE. Coronary arterial and sinusal anatomy in hearts with a common arterial trunk. Ann Thorac Surg 1989;48:792-7. 16. Anderson KR, McGoon DC, Lie JT. Surgical significance of the coronary arterial anatomy in truncus arteriosus communis. Am J Cardiol 1978;41:76-81. 17. de la Cruz MV, Cayre R, Angelini P, Noriega-Ramos N, Sadowinski S. Coronary arteries in truncus arteriosus. Am J Cardiol 1990;66:1482-6. 18. Elkins RC, Steinberg JB, Razook JD, Ward KE, Overholt ED, Thompson WM Jr. Correction of truncus arteriosus with truncal valvar stenosis or insufficiency using two homografts. Ann Thorac Surg 1990;50:728-33. 19. Fujiwara K, Yokota Y, Okamoto F, et al. Successful surgical repair of truncus arteriosus with interrupted aortic arch in infancy by an anterior approach. Ann Thorac Surg 1988;45:441-4. 20. Fyfe DA, Driscoll OJ, Di Donato RM, et al. Truncus arteriosus with single pulmonary artery: influence of pulmonary vascular obstructive disease on early and late operative results. J Am Coli Cardiol 1985;5:1168-72. 21. Rossiter SJ, Silverman JF, Shumway NE. Patterns of pulmonary arterial supply in patients with truncus arteriosus. J THORAC CARDIOVASC SURG 1978;75:73-9. 22. Sano S, Brawn W J, Mee RBB. Repair of truncus arteriosus and interrupted aortic arch. J CardiacSurg 1990;5:15762. Discussion (Papers by Hanley [page 1047l. Bove [page 1057l- and their associates) Dr. Miguel Barbero-Marcial (Sao Paulo. Brazil). Since 1987 at the Heart Institute in Sao Paulo we have used a different technique to repair truncus in 20 infants; that is, we repaired the truncus without an extracardiac conduit. Eighteen patients had type I or 1.5 and two had type 2 truncus arteriosus. Fourteen were in preoperative functional class IV, four in class Ill, and two had required ventilatory support for more than I week. The early mortality rate was 25% and the late, 5%, with only one patient dying of acute gastroenteritis. All 14 survivors are in functional class I, 10 without medication, for a follow-up period up to 63 months. The late postoperative status in 10 patients between 24 and 63 months showed normal left ventricular ejection fraction and gradients between the right ventricle and pulmonary arteries that were less than 30 mm Hg in all but three: two with 40 mm Hg and the other 45 mm Hg. The amount of monocusp valve regurgitation was mild to moderate. We believe that these good late results are explained by the fact that, in this technique, the posterior wall anastomosis is constructed with the patient's own tissue, maintaining the growth of the neo-right ventricular outflow tract.

Bove et al.

I065

Dr. Castaneda, which option do you in those cases prefer in patients needing reoperations-enlargement of the primary tube or its replacement with another homograft? Dr, Gerhard Ziemer (Hannover, Germany). It is obvious that outstanding results can be achieved with valved conduits in truncus repair. As Dr. Barbero-Marcial mentioned in describing his technique of using a nonvalved connection between the right ventricle and central pulmonary arteries, I would like to stress the point that a valved conduit may not be necessary, especially in neonatal repair of truncus arteriosus. In my assessment, this has actually been shown by Dr. Castaneda in his presentation: The postoperative pulmonary artery pressure dropped significantly. Thus my question to Drs. Bove and Castaneda is this: Could you agree to the notion that in neonates there may be no difference to tetralogy repair to which a truncus anatomy could be made similar? One could extensively mobilize the central pulmonary arteries and partially anastomose them to the right ventricle, completing the anastomosis with a right ventricular outflow tract patch anteriorly as in tetralogy. Doing this would avoid the wait for a suitable homograft, and the need for reoperation could be reduced by not having a circular prosthesis. Dr. Joseph J. Amato (New Hyde Park. N.Y.). I agree that the truncus arteriosus should be repaired as soon as possible. I would like to present an unusual complication that recently occurred in our institution. An II-day-old child underwent repair of truncus arteriosus had an uneventful postoperative course. She was sent home in 10 days but returned 3 months after the operation with x-ray evidence of a large mass in the left upper chest cavity. This was diagnosed as pseudoaneurysm of the right ventricular outflow tract. By means of literature research and also personal communication with three other groups, we learned that from a total of 604 cases of homograft insertions, the incidence of pseudoaneurysms was 1.8%. In our case, the pseudoaneurysm occurred between the hood of the pericardium and the homograft. In our discussions of pseudoaneurysms with Cryolife, Inc., Marietta, Georgia, we observed that microscopic sections of aortic homografts prepared by Cryolife showed very little available adventitia. Cryolife stated that almost all of the adventitia was removed in the preparation of the aortic homografts. After our conversation in February 1992, a note was sent out to all implanting surgeons that Cryolife had made the decision to leave the adventitia on the aortic conduit. We believe this will make the suture line stronger. It is obviously equally important to use appropriate suture technique. I would like to ask the authors whether they have any knowledge regarding the occurrence of pseudoaneurysms. Dr. Castaneda. Dr. Bove, it is indeed amazing that our results are so very similar. The data are essentially superimposable except for the complex forms of truncus arteriosus. We have reached the same conclusions and we also agree with you now that a significantly insufficient truncal valve should be replaced at the time of the primary operation. Concerning Dr. Barbero-Marcial's question as to the use of a nonvalved connection between the right ventricle and the pulmonary artery, we believe that a valved conduit is useful particularly during the early postoperative course because of the labile pulmonary vasculature in the young infant. As we have demonstrated, pulmonary hypertensive episodes are frequent, more so beyond the first month of life. However, at later reoperation,

I066

Bove et al.

usually between 4 to 5 years after the initial operation, we tend to do a patch augmentation repair of the previously positioned valved conduit without placing a valve within the right ventricular outflow tract. Dr. Bove. I completely agree with Dr. Castaneda regarding the use of homografts. We have preferred to use them because we believe that the homograft valve confers an important hemodynamic advantage, and, of equal importance, there is no disadvantage. I have not attempted to create a direct connection between the pulmonary arteries and the right ventricular outflow tract. Although this would be a significant advantage if reoperation were avoided, I would be concerned about potential stenosis at the right ventricular anastomosis. It may be difficult to adequately mobilize the peripheral pulmonary arteries without some tension at the anastomosis or possibly compression of the pulmonary artery. Our institution has reported an earlier experience with nonvalved PTFE conduits for truncus arteriosus, but the availability of small-sized homografts offers an important hemodynamic advantage because the pulmonary resistance can still be labile

The Journal of Thoracic and Cardiovascular Surgery June 1993

even in very young patients. The smallest homograft conduit was 8 mm, and we actually have a fair number of homografts in our series between 8 and II mm, smaller than the commercially available heterograft valves. These smaller grafts are a nice technical advantage, too. In response to Dr. Amato's question, we have seen pseudoaneurysms, although not in this truncus series. Pseudoaneurysm has occurred because of a disruption along the anastomosis between the Dacron extension and the inlet, or muscular, portion of the homograft. We have changed our technique to some extent, although not because of that complication. In the past, we extended all of our homografts proximally with a Dacron tube, cutting it obliquely so that there is only a I or 2 mm rim posteriorly where it is anastomosed to the right ventricular outflow tract. Because of a few cases of early obstruction resulting from thickening and retraction along that short piece of Dacron fabric, we now prefer to attach the homograft directly to the upper margin of the ventriculotomy and then augment it anteriorly with pericardium, extra homograft tissue, or PTFE. This has resulted in a marked decrease in technical difficulties with the homograft itself.

Bound volumes available to subscribers Boundvolumesof THEJOURNAL OF THORACIC AND CARDIOVASCULAR SURGERY are availableto subscribers(only) for the 1993 issues from the Publisher, at a cost of $75.00 for domestic, $98.25 for Canadian, and $93.00 for international subscribersfor Vol. 105 (JanuaryJune) and Vol. 106 (July-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertisingis removed.Copiesare shipped within 60 days after publicationof the last issueof the volume. The binding is durable buckram with the JOURNAL name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby, SubscriptionServices, 11830 WestlineIndustrial Drive,St. Louis,Missouri63146-3318, USA; phone I (800) 453-4351 or (314) 453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.